When dead code elimination removes all but one use, try to fold the single def into the remaining use.

Rematerialization can leave single-use loads behind that we might as well fold whenever possible.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@128918 91177308-0d34-0410-b5e6-96231b3b80d8

3 files changed, 56 insertions, 0 deletions

diff --git a/lib/CodeGen/LiveRangeEdit.cpp b/lib/CodeGen/LiveRangeEdit.cpp
index 4c7b9e6be9..b96575eea0 100644
--- a/lib/CodeGen/LiveRangeEdit.cpp
+++ b/lib/CodeGen/LiveRangeEdit.cpp
@@ -149,6 +149,54 @@ void LiveRangeEdit::eraseVirtReg(unsigned Reg, LiveIntervals &LIS) {
     LIS.removeInterval(Reg);
 }
 
+bool LiveRangeEdit::foldAsLoad(LiveInterval *LI,
+                               SmallVectorImpl<MachineInstr*> &Dead,
+                               MachineRegisterInfo &MRI,
+                               LiveIntervals &LIS,
+                               const TargetInstrInfo &TII) {
+  MachineInstr *DefMI = 0, *UseMI = 0;
+
+  // Check that there is a single def and a single use.
+  for (MachineRegisterInfo::reg_nodbg_iterator I = MRI.reg_nodbg_begin(LI->reg),
+       E = MRI.reg_nodbg_end(); I != E; ++I) {
+    MachineOperand &MO = I.getOperand();
+    MachineInstr *MI = MO.getParent();
+    if (MO.isDef()) {
+      if (DefMI && DefMI != MI)
+        return false;
+      if (!MI->getDesc().canFoldAsLoad())
+        return false;
+      DefMI = MI;
+    } else if (!MO.isUndef()) {
+      if (UseMI && UseMI != MI)
+        return false;
+      // FIXME: Targets don't know how to fold subreg uses.
+      if (MO.getSubReg())
+        return false;
+      UseMI = MI;
+    }
+  }
+  if (!DefMI || !UseMI)
+    return false;
+
+  DEBUG(dbgs() << "Try to fold single def: " << *DefMI
+               << "       into single use: " << *UseMI);
+
+  SmallVector<unsigned, 8> Ops;
+  if (UseMI->readsWritesVirtualRegister(LI->reg, &Ops).second)
+    return false;
+
+  MachineInstr *FoldMI = TII.foldMemoryOperand(UseMI, Ops, DefMI);
+  if (!FoldMI)
+    return false;
+  DEBUG(dbgs() << "                folded: " << *FoldMI);
+  LIS.ReplaceMachineInstrInMaps(UseMI, FoldMI);
+  UseMI->eraseFromParent();
+  DefMI->addRegisterDead(LI->reg, 0);
+  Dead.push_back(DefMI);
+  return true;
+}
+
 void LiveRangeEdit::eliminateDeadDefs(SmallVectorImpl<MachineInstr*> &Dead,
                                       LiveIntervals &LIS, VirtRegMap &VRM,
                                       const TargetInstrInfo &TII) {
@@ -218,6 +266,8 @@ void LiveRangeEdit::eliminateDeadDefs(SmallVectorImpl<MachineInstr*> &Dead,
     // Shrink just one live interval. Then delete new dead defs.
     LiveInterval *LI = ToShrink.back();
     ToShrink.pop_back();
+    if (foldAsLoad(LI, Dead, VRM.getRegInfo(), LIS, TII))
+      continue;
     if (delegate_)
       delegate_->LRE_WillShrinkVirtReg(LI->reg);
     if (!LIS.shrinkToUses(LI, &Dead))
diff --git a/lib/CodeGen/LiveRangeEdit.h b/lib/CodeGen/LiveRangeEdit.h
index c96dfd6998..54a2c83fd6 100644
--- a/lib/CodeGen/LiveRangeEdit.h
+++ b/lib/CodeGen/LiveRangeEdit.h
@@ -80,6 +80,11 @@ private:
   bool allUsesAvailableAt(const MachineInstr *OrigMI, SlotIndex OrigIdx,
                           SlotIndex UseIdx, LiveIntervals &lis);
 
+  /// foldAsLoad - If LI has a single use and a single def that can be folded as
+  /// a load, eliminate the register by folding the def into the use.
+  bool foldAsLoad(LiveInterval *LI, SmallVectorImpl<MachineInstr*> &Dead,
+                  MachineRegisterInfo&, LiveIntervals&, const TargetInstrInfo&);
+
 public:
   /// Create a LiveRangeEdit for breaking down parent into smaller pieces.
   /// @param parent The register being spilled or split.
diff --git a/test/CodeGen/X86/constant-pool-remat-0.ll b/test/CodeGen/X86/constant-pool-remat-0.ll
index f1b061f65a..4be14d2128 100644
--- a/test/CodeGen/X86/constant-pool-remat-0.ll
+++ b/test/CodeGen/X86/constant-pool-remat-0.ll
@@ -1,4 +1,5 @@
 ; RUN: llc < %s -mtriple=x86_64-linux   | FileCheck %s
+; RUN: llc < %s -mtriple=x86_64-linux -regalloc=greedy | FileCheck %s
 ; RUN: llc < %s -march=x86 -mattr=+sse2 | FileCheck %s
 ; CHECK:     LCPI
 ; CHECK:     LCPI