From e5165490b7ba24bb2f3043399e0d60e7f3bcf8a5 Mon Sep 17 00:00:00 2001
From: Jim Grosbach <grosbach@apple.com>
Date: Mon, 9 Nov 2009 00:11:35 +0000
Subject: [PATCH] Use Unified Assembly Syntax for the ARM backend.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@86494 91177308-0d34-0410-b5e6-96231b3b80d8
---
 lib/Target/ARM/ARMAddressingModes.h          |   4 +-
 lib/Target/ARM/ARMBaseInstrInfo.cpp          |  42 +--
 lib/Target/ARM/ARMBaseRegisterInfo.cpp       |   8 +-
 lib/Target/ARM/ARMConstantIslandPass.cpp     |   4 +-
 lib/Target/ARM/ARMISelDAGToDAG.cpp           |  12 +-
 lib/Target/ARM/ARMISelLowering.cpp           |  38 +--
 lib/Target/ARM/ARMISelLowering.h             |   4 +-
 lib/Target/ARM/ARMInstrInfo.td               |  83 +++---
 lib/Target/ARM/ARMInstrNEON.td               |   2 +-
 lib/Target/ARM/ARMInstrVFP.td                | 261 ++++++++++---------
 lib/Target/ARM/ARMLoadStoreOptimizer.cpp     |  92 +++----
 lib/Target/ARM/AsmPrinter/ARMAsmPrinter.cpp  |  12 +-
 lib/Target/ARM/AsmPrinter/ARMInstPrinter.cpp |   7 +-
 lib/Target/ARM/NEONMoveFix.cpp               |   8 +-
 lib/Target/ARM/README-Thumb.txt              |   8 +-
 lib/Target/ARM/README.txt                    |   2 +-
 lib/Target/ARM/Thumb1RegisterInfo.cpp        |   2 +-
 test/CodeGen/ARM/2009-07-18-RewriterBug.ll   |   2 +-
 test/CodeGen/ARM/2009-09-09-fpcmp-ole.ll     |   2 +-
 test/CodeGen/ARM/2009-09-24-spill-align.ll   |   2 +-
 test/CodeGen/ARM/2009-11-01-NeonMoves.ll     |   5 +-
 test/CodeGen/ARM/arguments_f64_backfill.ll   |   2 +-
 test/CodeGen/ARM/compare-call.ll             |   2 +-
 test/CodeGen/ARM/fabss.ll                    |   6 +-
 test/CodeGen/ARM/fadds.ll                    |   6 +-
 test/CodeGen/ARM/fcopysign.ll                |   2 +-
 test/CodeGen/ARM/fdivs.ll                    |  10 +-
 test/CodeGen/ARM/fixunsdfdi.ll               |   2 +-
 test/CodeGen/ARM/fmacs.ll                    |   6 +-
 test/CodeGen/ARM/fmscs.ll                    |  10 +-
 test/CodeGen/ARM/fmuls.ll                    |   6 +-
 test/CodeGen/ARM/fnegs.ll                    |   6 +-
 test/CodeGen/ARM/fnmacs.ll                   |   8 +-
 test/CodeGen/ARM/fnmscs.ll                   |   4 +-
 test/CodeGen/ARM/fnmul.ll                    |   4 +-
 test/CodeGen/ARM/fp.ll                       |  32 +--
 test/CodeGen/ARM/fp_convert.ll               |   8 +-
 test/CodeGen/ARM/fparith.ll                  |  20 +-
 test/CodeGen/ARM/fpcmp.ll                    |  14 +-
 test/CodeGen/ARM/fpconv.ll                   |  20 +-
 test/CodeGen/ARM/fpmem.ll                    |   4 +-
 test/CodeGen/ARM/fptoint.ll                  |   2 +-
 test/CodeGen/ARM/fsubs.ll                    |   4 +-
 test/CodeGen/ARM/ifcvt5.ll                   |   2 +-
 test/CodeGen/ARM/ifcvt6.ll                   |   2 +-
 test/CodeGen/ARM/ifcvt7.ll                   |   2 +-
 test/CodeGen/ARM/ifcvt8.ll                   |   2 +-
 test/CodeGen/ARM/neon_ld1.ll                 |   6 +-
 test/CodeGen/ARM/neon_ld2.ll                 |   2 +-
 test/CodeGen/ARM/select.ll                   |   2 +-
 test/CodeGen/ARM/vfp.ll                      |  24 +-
 test/CodeGen/ARM/vget_lane.ll                |   4 +-
 test/CodeGen/Thumb2/cross-rc-coalescing-2.ll |   2 +-
 53 files changed, 414 insertions(+), 412 deletions(-)

diff --git a/lib/Target/ARM/ARMAddressingModes.h b/lib/Target/ARM/ARMAddressingModes.h
index 6ee1b68efaa..ddeb1b994e8 100644
--- a/lib/Target/ARM/ARMAddressingModes.h
+++ b/lib/Target/ARM/ARMAddressingModes.h
@@ -520,8 +520,8 @@ namespace ARM_AM {
     return ((AM5Opc >> 8) & 1) ? sub : add;
   }
 
-  /// getAM5Opc - This function encodes the addrmode5 opc field for FLDM and
-  /// FSTM instructions.
+  /// getAM5Opc - This function encodes the addrmode5 opc field for VLDM and
+  /// VSTM instructions.
   static inline unsigned getAM5Opc(AMSubMode SubMode, bool WB,
                                    unsigned char Offset) {
     assert((SubMode == ia || SubMode == db) &&
diff --git a/lib/Target/ARM/ARMBaseInstrInfo.cpp b/lib/Target/ARM/ARMBaseInstrInfo.cpp
index 96b1c9ce041..868d31dbe0e 100644
--- a/lib/Target/ARM/ARMBaseInstrInfo.cpp
+++ b/lib/Target/ARM/ARMBaseInstrInfo.cpp
@@ -509,9 +509,9 @@ ARMBaseInstrInfo::isMoveInstr(const MachineInstr &MI,
 
   switch (MI.getOpcode()) {
   default: break;
-  case ARM::FCPYS:
-  case ARM::FCPYD:
+  case ARM::VMOVS:
   case ARM::VMOVD:
+  case ARM::VMOVDneon:
   case ARM::VMOVQ: {
     SrcReg = MI.getOperand(1).getReg();
     DstReg = MI.getOperand(0).getReg();
@@ -561,8 +561,8 @@ ARMBaseInstrInfo::isLoadFromStackSlot(const MachineInstr *MI,
       return MI->getOperand(0).getReg();
     }
     break;
-  case ARM::FLDD:
-  case ARM::FLDS:
+  case ARM::VLDRD:
+  case ARM::VLDRS:
     if (MI->getOperand(1).isFI() &&
         MI->getOperand(2).isImm() &&
         MI->getOperand(2).getImm() == 0) {
@@ -600,8 +600,8 @@ ARMBaseInstrInfo::isStoreToStackSlot(const MachineInstr *MI,
       return MI->getOperand(0).getReg();
     }
     break;
-  case ARM::FSTD:
-  case ARM::FSTS:
+  case ARM::VSTRD:
+  case ARM::VSTRS:
     if (MI->getOperand(1).isFI() &&
         MI->getOperand(2).isImm() &&
         MI->getOperand(2).getImm() == 0) {
@@ -637,17 +637,17 @@ ARMBaseInstrInfo::copyRegToReg(MachineBasicBlock &MBB,
     AddDefaultCC(AddDefaultPred(BuildMI(MBB, I, DL, get(ARM::MOVr),
                                         DestReg).addReg(SrcReg)));
   } else if (DestRC == ARM::SPRRegisterClass) {
-    AddDefaultPred(BuildMI(MBB, I, DL, get(ARM::FCPYS), DestReg)
+    AddDefaultPred(BuildMI(MBB, I, DL, get(ARM::VMOVS), DestReg)
                    .addReg(SrcReg));
   } else if (DestRC == ARM::DPRRegisterClass) {
-    AddDefaultPred(BuildMI(MBB, I, DL, get(ARM::FCPYD), DestReg)
+    AddDefaultPred(BuildMI(MBB, I, DL, get(ARM::VMOVD), DestReg)
                    .addReg(SrcReg));
   } else if (DestRC == ARM::DPR_VFP2RegisterClass ||
              DestRC == ARM::DPR_8RegisterClass ||
              SrcRC == ARM::DPR_VFP2RegisterClass ||
              SrcRC == ARM::DPR_8RegisterClass) {
     // Always use neon reg-reg move if source or dest is NEON-only regclass.
-    BuildMI(MBB, I, DL, get(ARM::VMOVD), DestReg).addReg(SrcReg);
+    BuildMI(MBB, I, DL, get(ARM::VMOVDneon), DestReg).addReg(SrcReg);
   } else if (DestRC == ARM::QPRRegisterClass ||
              DestRC == ARM::QPR_VFP2RegisterClass ||
              DestRC == ARM::QPR_8RegisterClass) {
@@ -682,11 +682,11 @@ storeRegToStackSlot(MachineBasicBlock &MBB, MachineBasicBlock::iterator I,
   } else if (RC == ARM::DPRRegisterClass ||
              RC == ARM::DPR_VFP2RegisterClass ||
              RC == ARM::DPR_8RegisterClass) {
-    AddDefaultPred(BuildMI(MBB, I, DL, get(ARM::FSTD))
+    AddDefaultPred(BuildMI(MBB, I, DL, get(ARM::VSTRD))
                    .addReg(SrcReg, getKillRegState(isKill))
                    .addFrameIndex(FI).addImm(0).addMemOperand(MMO));
   } else if (RC == ARM::SPRRegisterClass) {
-    AddDefaultPred(BuildMI(MBB, I, DL, get(ARM::FSTS))
+    AddDefaultPred(BuildMI(MBB, I, DL, get(ARM::VSTRS))
                    .addReg(SrcReg, getKillRegState(isKill))
                    .addFrameIndex(FI).addImm(0).addMemOperand(MMO));
   } else {
@@ -728,10 +728,10 @@ loadRegFromStackSlot(MachineBasicBlock &MBB, MachineBasicBlock::iterator I,
   } else if (RC == ARM::DPRRegisterClass ||
              RC == ARM::DPR_VFP2RegisterClass ||
              RC == ARM::DPR_8RegisterClass) {
-    AddDefaultPred(BuildMI(MBB, I, DL, get(ARM::FLDD), DestReg)
+    AddDefaultPred(BuildMI(MBB, I, DL, get(ARM::VLDRD), DestReg)
                    .addFrameIndex(FI).addImm(0).addMemOperand(MMO));
   } else if (RC == ARM::SPRRegisterClass) {
-    AddDefaultPred(BuildMI(MBB, I, DL, get(ARM::FLDS), DestReg)
+    AddDefaultPred(BuildMI(MBB, I, DL, get(ARM::VLDRS), DestReg)
                    .addFrameIndex(FI).addImm(0).addMemOperand(MMO));
   } else {
     assert((RC == ARM::QPRRegisterClass ||
@@ -827,7 +827,7 @@ foldMemoryOperandImpl(MachineFunction &MF, MachineInstr *MI,
                 DstSubReg)
         .addFrameIndex(FI).addImm(0).addImm(ARMCC::AL).addReg(0);
     }
-  } else if (Opc == ARM::FCPYS) {
+  } else if (Opc == ARM::VMOVS) {
     unsigned Pred = MI->getOperand(2).getImm();
     unsigned PredReg = MI->getOperand(3).getReg();
     if (OpNum == 0) { // move -> store
@@ -835,7 +835,7 @@ foldMemoryOperandImpl(MachineFunction &MF, MachineInstr *MI,
       unsigned SrcSubReg = MI->getOperand(1).getSubReg();
       bool isKill = MI->getOperand(1).isKill();
       bool isUndef = MI->getOperand(1).isUndef();
-      NewMI = BuildMI(MF, MI->getDebugLoc(), get(ARM::FSTS))
+      NewMI = BuildMI(MF, MI->getDebugLoc(), get(ARM::VSTRS))
         .addReg(SrcReg, getKillRegState(isKill) | getUndefRegState(isUndef),
                 SrcSubReg)
         .addFrameIndex(FI)
@@ -845,7 +845,7 @@ foldMemoryOperandImpl(MachineFunction &MF, MachineInstr *MI,
       unsigned DstSubReg = MI->getOperand(0).getSubReg();
       bool isDead = MI->getOperand(0).isDead();
       bool isUndef = MI->getOperand(0).isUndef();
-      NewMI = BuildMI(MF, MI->getDebugLoc(), get(ARM::FLDS))
+      NewMI = BuildMI(MF, MI->getDebugLoc(), get(ARM::VLDRS))
         .addReg(DstReg,
                 RegState::Define |
                 getDeadRegState(isDead) |
@@ -854,7 +854,7 @@ foldMemoryOperandImpl(MachineFunction &MF, MachineInstr *MI,
         .addFrameIndex(FI).addImm(0).addImm(Pred).addReg(PredReg);
     }
   }
-  else if (Opc == ARM::FCPYD) {
+  else if (Opc == ARM::VMOVD) {
     unsigned Pred = MI->getOperand(2).getImm();
     unsigned PredReg = MI->getOperand(3).getReg();
     if (OpNum == 0) { // move -> store
@@ -862,7 +862,7 @@ foldMemoryOperandImpl(MachineFunction &MF, MachineInstr *MI,
       unsigned SrcSubReg = MI->getOperand(1).getSubReg();
       bool isKill = MI->getOperand(1).isKill();
       bool isUndef = MI->getOperand(1).isUndef();
-      NewMI = BuildMI(MF, MI->getDebugLoc(), get(ARM::FSTD))
+      NewMI = BuildMI(MF, MI->getDebugLoc(), get(ARM::VSTRD))
         .addReg(SrcReg,
                 getKillRegState(isKill) | getUndefRegState(isUndef),
                 SrcSubReg)
@@ -872,7 +872,7 @@ foldMemoryOperandImpl(MachineFunction &MF, MachineInstr *MI,
       unsigned DstSubReg = MI->getOperand(0).getSubReg();
       bool isDead = MI->getOperand(0).isDead();
       bool isUndef = MI->getOperand(0).isUndef();
-      NewMI = BuildMI(MF, MI->getDebugLoc(), get(ARM::FLDD))
+      NewMI = BuildMI(MF, MI->getDebugLoc(), get(ARM::VLDRD))
         .addReg(DstReg,
                 RegState::Define |
                 getDeadRegState(isDead) |
@@ -908,9 +908,9 @@ ARMBaseInstrInfo::canFoldMemoryOperand(const MachineInstr *MI,
              Opc == ARM::tMOVtgpr2gpr ||
              Opc == ARM::tMOVgpr2tgpr) {
     return true;
-  } else if (Opc == ARM::FCPYS || Opc == ARM::FCPYD) {
+  } else if (Opc == ARM::VMOVS || Opc == ARM::VMOVD) {
     return true;
-  } else if (Opc == ARM::VMOVD || Opc == ARM::VMOVQ) {
+  } else if (Opc == ARM::VMOVDneon || Opc == ARM::VMOVQ) {
     return false; // FIXME
   }
 
diff --git a/lib/Target/ARM/ARMBaseRegisterInfo.cpp b/lib/Target/ARM/ARMBaseRegisterInfo.cpp
index ad5dfc436bc..3197ab08440 100644
--- a/lib/Target/ARM/ARMBaseRegisterInfo.cpp
+++ b/lib/Target/ARM/ARMBaseRegisterInfo.cpp
@@ -1346,7 +1346,7 @@ emitPrologue(MachineFunction &MF) const {
   AFI->setGPRCalleeSavedArea2Offset(GPRCS2Offset);
   AFI->setDPRCalleeSavedAreaOffset(DPRCSOffset);
 
-  movePastCSLoadStoreOps(MBB, MBBI, ARM::FSTD, 0, 3, STI);
+  movePastCSLoadStoreOps(MBB, MBBI, ARM::VSTRD, 0, 3, STI);
   NumBytes = DPRCSOffset;
   if (NumBytes) {
     // Adjust SP after all the callee-save spills.
@@ -1385,7 +1385,7 @@ static bool isCalleeSavedRegister(unsigned Reg, const unsigned *CSRegs) {
 static bool isCSRestore(MachineInstr *MI,
                         const ARMBaseInstrInfo &TII,
                         const unsigned *CSRegs) {
-  return ((MI->getOpcode() == (int)ARM::FLDD ||
+  return ((MI->getOpcode() == (int)ARM::VLDRD ||
            MI->getOpcode() == (int)ARM::LDR ||
            MI->getOpcode() == (int)ARM::t2LDRi12) &&
           MI->getOperand(1).isFI() &&
@@ -1411,7 +1411,7 @@ emitEpilogue(MachineFunction &MF, MachineBasicBlock &MBB) const {
     if (NumBytes != 0)
       emitSPUpdate(isARM, MBB, MBBI, dl, TII, NumBytes);
   } else {
-    // Unwind MBBI to point to first LDR / FLDD.
+    // Unwind MBBI to point to first LDR / VLDRD.
     const unsigned *CSRegs = getCalleeSavedRegs();
     if (MBBI != MBB.begin()) {
       do
@@ -1459,7 +1459,7 @@ emitEpilogue(MachineFunction &MF, MachineBasicBlock &MBB) const {
       emitSPUpdate(isARM, MBB, MBBI, dl, TII, NumBytes);
 
     // Move SP to start of integer callee save spill area 2.
-    movePastCSLoadStoreOps(MBB, MBBI, ARM::FLDD, 0, 3, STI);
+    movePastCSLoadStoreOps(MBB, MBBI, ARM::VLDRD, 0, 3, STI);
     emitSPUpdate(isARM, MBB, MBBI, dl, TII, AFI->getDPRCalleeSavedAreaSize());
 
     // Move SP to start of integer callee save spill area 1.
diff --git a/lib/Target/ARM/ARMConstantIslandPass.cpp b/lib/Target/ARM/ARMConstantIslandPass.cpp
index 981962522db..3235c3f746a 100644
--- a/lib/Target/ARM/ARMConstantIslandPass.cpp
+++ b/lib/Target/ARM/ARMConstantIslandPass.cpp
@@ -541,8 +541,8 @@ void ARMConstantIslands::InitialFunctionScan(MachineFunction &MF,
             Scale = 4;  // +(offset_8*4)
             break;
 
-          case ARM::FLDD:
-          case ARM::FLDS:
+          case ARM::VLDRD:
+          case ARM::VLDRS:
             Bits = 8;
             Scale = 4;  // +-(offset_8*4)
             NegOk = true;
diff --git a/lib/Target/ARM/ARMISelDAGToDAG.cpp b/lib/Target/ARM/ARMISelDAGToDAG.cpp
index 0eb1eb6dc1b..9be74540e23 100644
--- a/lib/Target/ARM/ARMISelDAGToDAG.cpp
+++ b/lib/Target/ARM/ARMISelDAGToDAG.cpp
@@ -1466,8 +1466,8 @@ SDNode *ARMDAGToDAGISel::Select(SDValue Op) {
     }
     break;
   }
-  case ARMISD::FMRRD:
-    return CurDAG->getMachineNode(ARM::FMRRD, dl, MVT::i32, MVT::i32,
+  case ARMISD::VMOVRRD:
+    return CurDAG->getMachineNode(ARM::VMOVRRD, dl, MVT::i32, MVT::i32,
                                   Op.getOperand(0), getAL(CurDAG),
                                   CurDAG->getRegister(0, MVT::i32));
   case ISD::UMUL_LOHI: {
@@ -1656,10 +1656,10 @@ SDNode *ARMDAGToDAGISel::Select(SDValue Op) {
         : ARM::MOVCCr;
       break;
     case MVT::f32:
-      Opc = ARM::FCPYScc;
+      Opc = ARM::VMOVScc;
       break;
     case MVT::f64:
-      Opc = ARM::FCPYDcc;
+      Opc = ARM::VMOVDcc;
       break;
     }
     return CurDAG->SelectNodeTo(Op.getNode(), Opc, VT, Ops, 5);
@@ -1683,10 +1683,10 @@ SDNode *ARMDAGToDAGISel::Select(SDValue Op) {
     default: assert(false && "Illegal conditional move type!");
       break;
     case MVT::f32:
-      Opc = ARM::FNEGScc;
+      Opc = ARM::VNEGScc;
       break;
     case MVT::f64:
-      Opc = ARM::FNEGDcc;
+      Opc = ARM::VNEGDcc;
       break;
     }
     return CurDAG->SelectNodeTo(Op.getNode(), Opc, VT, Ops, 5);
diff --git a/lib/Target/ARM/ARMISelLowering.cpp b/lib/Target/ARM/ARMISelLowering.cpp
index c468b906e52..4945fdfc1d2 100644
--- a/lib/Target/ARM/ARMISelLowering.cpp
+++ b/lib/Target/ARM/ARMISelLowering.cpp
@@ -389,7 +389,7 @@ ARMTargetLowering::ARMTargetLowering(TargetMachine &TM)
   setOperationAction(ISD::SIGN_EXTEND_INREG, MVT::i1, Expand);
 
   if (!UseSoftFloat && Subtarget->hasVFP2() && !Subtarget->isThumb1Only())
-    // Turn f64->i64 into FMRRD, i64 -> f64 to FMDRR iff target supports vfp2.
+    // Turn f64->i64 into VMOVRRD, i64 -> f64 to VMOVDRR iff target supports vfp2.
     setOperationAction(ISD::BIT_CONVERT, MVT::i64, Custom);
 
   // We want to custom lower some of our intrinsics.
@@ -434,7 +434,7 @@ ARMTargetLowering::ARMTargetLowering(TargetMachine &TM)
   }
 
   // We have target-specific dag combine patterns for the following nodes:
-  // ARMISD::FMRRD  - No need to call setTargetDAGCombine
+  // ARMISD::VMOVRRD  - No need to call setTargetDAGCombine
   setTargetDAGCombine(ISD::ADD);
   setTargetDAGCombine(ISD::SUB);
 
@@ -493,8 +493,8 @@ const char *ARMTargetLowering::getTargetNodeName(unsigned Opcode) const {
   case ARMISD::SRA_FLAG:      return "ARMISD::SRA_FLAG";
   case ARMISD::RRX:           return "ARMISD::RRX";
 
-  case ARMISD::FMRRD:         return "ARMISD::FMRRD";
-  case ARMISD::FMDRR:         return "ARMISD::FMDRR";
+  case ARMISD::VMOVRRD:         return "ARMISD::VMOVRRD";
+  case ARMISD::VMOVDRR:         return "ARMISD::VMOVDRR";
 
   case ARMISD::EH_SJLJ_SETJMP: return "ARMISD::EH_SJLJ_SETJMP";
   case ARMISD::EH_SJLJ_LONGJMP:return "ARMISD::EH_SJLJ_LONGJMP";
@@ -790,7 +790,7 @@ ARMTargetLowering::LowerCallResult(SDValue Chain, SDValue InFlag,
                                       InFlag);
       Chain = Hi.getValue(1);
       InFlag = Hi.getValue(2);
-      Val = DAG.getNode(ARMISD::FMDRR, dl, MVT::f64, Lo, Hi);
+      Val = DAG.getNode(ARMISD::VMOVDRR, dl, MVT::f64, Lo, Hi);
 
       if (VA.getLocVT() == MVT::v2f64) {
         SDValue Vec = DAG.getNode(ISD::UNDEF, dl, MVT::v2f64);
@@ -805,7 +805,7 @@ ARMTargetLowering::LowerCallResult(SDValue Chain, SDValue InFlag,
         Hi = DAG.getCopyFromReg(Chain, dl, VA.getLocReg(), MVT::i32, InFlag);
         Chain = Hi.getValue(1);
         InFlag = Hi.getValue(2);
-        Val = DAG.getNode(ARMISD::FMDRR, dl, MVT::f64, Lo, Hi);
+        Val = DAG.getNode(ARMISD::VMOVDRR, dl, MVT::f64, Lo, Hi);
         Val = DAG.getNode(ISD::INSERT_VECTOR_ELT, dl, MVT::v2f64, Vec, Val,
                           DAG.getConstant(1, MVT::i32));
       }
@@ -870,7 +870,7 @@ void ARMTargetLowering::PassF64ArgInRegs(DebugLoc dl, SelectionDAG &DAG,
                                          SmallVector<SDValue, 8> &MemOpChains,
                                          ISD::ArgFlagsTy Flags) {
 
-  SDValue fmrrd = DAG.getNode(ARMISD::FMRRD, dl,
+  SDValue fmrrd = DAG.getNode(ARMISD::VMOVRRD, dl,
                               DAG.getVTList(MVT::i32, MVT::i32), Arg);
   RegsToPass.push_back(std::make_pair(VA.getLocReg(), fmrrd));
 
@@ -1149,7 +1149,7 @@ ARMTargetLowering::LowerReturn(SDValue Chain,
         // Extract the first half and return it in two registers.
         SDValue Half = DAG.getNode(ISD::EXTRACT_VECTOR_ELT, dl, MVT::f64, Arg,
                                    DAG.getConstant(0, MVT::i32));
-        SDValue HalfGPRs = DAG.getNode(ARMISD::FMRRD, dl,
+        SDValue HalfGPRs = DAG.getNode(ARMISD::VMOVRRD, dl,
                                        DAG.getVTList(MVT::i32, MVT::i32), Half);
 
         Chain = DAG.getCopyToReg(Chain, dl, VA.getLocReg(), HalfGPRs, Flag);
@@ -1166,7 +1166,7 @@ ARMTargetLowering::LowerReturn(SDValue Chain,
       }
       // Legalize ret f64 -> ret 2 x i32.  We always have fmrrd if f64 is
       // available.
-      SDValue fmrrd = DAG.getNode(ARMISD::FMRRD, dl,
+      SDValue fmrrd = DAG.getNode(ARMISD::VMOVRRD, dl,
                                   DAG.getVTList(MVT::i32, MVT::i32), &Arg, 1);
       Chain = DAG.getCopyToReg(Chain, dl, VA.getLocReg(), fmrrd, Flag);
       Flag = Chain.getValue(1);
@@ -1556,7 +1556,7 @@ ARMTargetLowering::GetF64FormalArgument(CCValAssign &VA, CCValAssign &NextVA,
     ArgValue2 = DAG.getCopyFromReg(Root, dl, Reg, MVT::i32);
   }
 
-  return DAG.getNode(ARMISD::FMDRR, dl, MVT::f64, ArgValue, ArgValue2);
+  return DAG.getNode(ARMISD::VMOVDRR, dl, MVT::f64, ArgValue, ArgValue2);
 }
 
 SDValue
@@ -2072,16 +2072,16 @@ static SDValue ExpandBIT_CONVERT(SDNode *N, SelectionDAG &DAG) {
   SDValue Op = N->getOperand(0);
   DebugLoc dl = N->getDebugLoc();
   if (N->getValueType(0) == MVT::f64) {
-    // Turn i64->f64 into FMDRR.
+    // Turn i64->f64 into VMOVDRR.
     SDValue Lo = DAG.getNode(ISD::EXTRACT_ELEMENT, dl, MVT::i32, Op,
                              DAG.getConstant(0, MVT::i32));
     SDValue Hi = DAG.getNode(ISD::EXTRACT_ELEMENT, dl, MVT::i32, Op,
                              DAG.getConstant(1, MVT::i32));
-    return DAG.getNode(ARMISD::FMDRR, dl, MVT::f64, Lo, Hi);
+    return DAG.getNode(ARMISD::VMOVDRR, dl, MVT::f64, Lo, Hi);
   }
 
-  // Turn f64->i64 into FMRRD.
-  SDValue Cvt = DAG.getNode(ARMISD::FMRRD, dl,
+  // Turn f64->i64 into VMOVRRD.
+  SDValue Cvt = DAG.getNode(ARMISD::VMOVRRD, dl,
                             DAG.getVTList(MVT::i32, MVT::i32), &Op, 1);
 
   // Merge the pieces into a single i64 value.
@@ -3178,12 +3178,12 @@ static SDValue PerformSUBCombine(SDNode *N,
   return SDValue();
 }
 
-/// PerformFMRRDCombine - Target-specific dag combine xforms for ARMISD::FMRRD.
-static SDValue PerformFMRRDCombine(SDNode *N,
+/// PerformVMOVRRDCombine - Target-specific dag combine xforms for ARMISD::VMOVRRD.
+static SDValue PerformVMOVRRDCombine(SDNode *N,
                                    TargetLowering::DAGCombinerInfo &DCI) {
   // fmrrd(fmdrr x, y) -> x,y
   SDValue InDouble = N->getOperand(0);
-  if (InDouble.getOpcode() == ARMISD::FMDRR)
+  if (InDouble.getOpcode() == ARMISD::VMOVDRR)
     return DCI.CombineTo(N, InDouble.getOperand(0), InDouble.getOperand(1));
   return SDValue();
 }
@@ -3478,7 +3478,7 @@ SDValue ARMTargetLowering::PerformDAGCombine(SDNode *N,
   default: break;
   case ISD::ADD:      return PerformADDCombine(N, DCI);
   case ISD::SUB:      return PerformSUBCombine(N, DCI);
-  case ARMISD::FMRRD: return PerformFMRRDCombine(N, DCI);
+  case ARMISD::VMOVRRD: return PerformVMOVRRDCombine(N, DCI);
   case ISD::INTRINSIC_WO_CHAIN:
     return PerformIntrinsicCombine(N, DCI.DAG);
   case ISD::SHL:
@@ -3760,7 +3760,7 @@ static bool getARMIndexedAddressParts(SDNode *Ptr, EVT VT,
     return true;
   }
 
-  // FIXME: Use FLDM / FSTM to emulate indexed FP load / store.
+  // FIXME: Use VLDM / VSTM to emulate indexed FP load / store.
   return false;
 }
 
diff --git a/lib/Target/ARM/ARMISelLowering.h b/lib/Target/ARM/ARMISelLowering.h
index 9c7a91d68d2..df69ab1cc2c 100644
--- a/lib/Target/ARM/ARMISelLowering.h
+++ b/lib/Target/ARM/ARMISelLowering.h
@@ -62,8 +62,8 @@ namespace llvm {
       SRA_FLAG,     // V,Flag = sra_flag X -> sra X, 1 + save carry out.
       RRX,          // V = RRX X, Flag     -> srl X, 1 + shift in carry flag.
 
-      FMRRD,        // double to two gprs.
-      FMDRR,        // Two gprs to double.
+      VMOVRRD,      // double to two gprs.
+      VMOVDRR,      // Two gprs to double.
 
       EH_SJLJ_SETJMP,    // SjLj exception handling setjmp.
       EH_SJLJ_LONGJMP,   // SjLj exception handling longjmp.
diff --git a/lib/Target/ARM/ARMInstrInfo.td b/lib/Target/ARM/ARMInstrInfo.td
index dfa2bdd071a..893694bfdee 100644
--- a/lib/Target/ARM/ARMInstrInfo.td
+++ b/lib/Target/ARM/ARMInstrInfo.td
@@ -394,13 +394,13 @@ let Defs = [CPSR] in {
 multiclass AI1_bin_s_irs<bits<4> opcod, string opc, PatFrag opnode,
                          bit Commutable = 0> {
   def ri : AI1<opcod, (outs GPR:$dst), (ins GPR:$a, so_imm:$b), DPFrm,
-               IIC_iALUi, opc, "s\t$dst, $a, $b",
+               IIC_iALUi, opc, "\t$dst, $a, $b",
                [(set GPR:$dst, (opnode GPR:$a, so_imm:$b))]> {
     let Inst{20} = 1;
     let Inst{25} = 1;
   }
   def rr : AI1<opcod, (outs GPR:$dst), (ins GPR:$a, GPR:$b), DPFrm,
-               IIC_iALUr, opc, "s\t$dst, $a, $b",
+               IIC_iALUr, opc, "\t$dst, $a, $b",
                [(set GPR:$dst, (opnode GPR:$a, GPR:$b))]> {
     let isCommutable = Commutable;
     let Inst{11-4} = 0b00000000;
@@ -408,7 +408,7 @@ multiclass AI1_bin_s_irs<bits<4> opcod, string opc, PatFrag opnode,
     let Inst{25} = 0;
   }
   def rs : AI1<opcod, (outs GPR:$dst), (ins GPR:$a, so_reg:$b), DPSoRegFrm,
-               IIC_iALUsr, opc, "s\t$dst, $a, $b",
+               IIC_iALUsr, opc, "\t$dst, $a, $b",
                [(set GPR:$dst, (opnode GPR:$a, so_reg:$b))]> {
     let Inst{20} = 1;
     let Inst{25} = 0;
@@ -504,9 +504,13 @@ multiclass AI1_adde_sube_irs<bits<4> opcod, string opc, PatFrag opnode,
                Requires<[IsARM, CarryDefIsUnused]> {
     let Inst{25} = 0;
   }
-  // Carry setting variants
+}
+// Carry setting variants
+let Defs = [CPSR] in {
+multiclass AI1_adde_sube_s_irs<bits<4> opcod, string opc, PatFrag opnode,
+                             bit Commutable = 0> {
   def Sri : AXI1<opcod, (outs GPR:$dst), (ins GPR:$a, so_imm:$b),
-                DPFrm, IIC_iALUi, !strconcat(opc, "s\t$dst, $a, $b"),
+                DPFrm, IIC_iALUi, !strconcat(opc, "\t$dst, $a, $b"),
                [(set GPR:$dst, (opnode GPR:$a, so_imm:$b))]>,
                Requires<[IsARM, CarryDefIsUsed]> {
     let Defs = [CPSR];
@@ -514,7 +518,7 @@ multiclass AI1_adde_sube_irs<bits<4> opcod, string opc, PatFrag opnode,
     let Inst{25} = 1;
   }
   def Srr : AXI1<opcod, (outs GPR:$dst), (ins GPR:$a, GPR:$b),
-                DPFrm, IIC_iALUr, !strconcat(opc, "s\t$dst, $a, $b"),
+                DPFrm, IIC_iALUr, !strconcat(opc, "\t$dst, $a, $b"),
                [(set GPR:$dst, (opnode GPR:$a, GPR:$b))]>,
                Requires<[IsARM, CarryDefIsUsed]> {
     let Defs = [CPSR];
@@ -523,7 +527,7 @@ multiclass AI1_adde_sube_irs<bits<4> opcod, string opc, PatFrag opnode,
     let Inst{25} = 0;
   }
   def Srs : AXI1<opcod, (outs GPR:$dst), (ins GPR:$a, so_reg:$b),
-                DPSoRegFrm, IIC_iALUsr, !strconcat(opc, "s\t$dst, $a, $b"),
+                DPSoRegFrm, IIC_iALUsr, !strconcat(opc, "\t$dst, $a, $b"),
                [(set GPR:$dst, (opnode GPR:$a, so_reg:$b))]>,
                Requires<[IsARM, CarryDefIsUsed]> {
     let Defs = [CPSR];
@@ -532,6 +536,7 @@ multiclass AI1_adde_sube_irs<bits<4> opcod, string opc, PatFrag opnode,
   }
 }
 }
+}
 
 //===----------------------------------------------------------------------===//
 // Instructions
@@ -663,7 +668,7 @@ let isReturn = 1, isTerminator = 1, isBarrier = 1, mayLoad = 1,
     hasExtraDefRegAllocReq = 1 in
   def LDM_RET : AXI4ld<(outs),
                     (ins addrmode4:$addr, pred:$p, reglist:$wb, variable_ops),
-                    LdStMulFrm, IIC_Br, "ldm${p}${addr:submode}\t$addr, $wb",
+                    LdStMulFrm, IIC_Br, "ldm${addr:submode}${p}\t$addr, $wb",
                     []>;
 
 // On non-Darwin platforms R9 is callee-saved.
@@ -803,26 +808,26 @@ def LDRcp : AI2ldw<(outs GPR:$dst), (ins addrmode2:$addr), LdFrm, IIC_iLoadr,
 
 // Loads with zero extension
 def LDRH  : AI3ldh<(outs GPR:$dst), (ins addrmode3:$addr), LdMiscFrm,
-                  IIC_iLoadr, "ldr", "h\t$dst, $addr",
+                  IIC_iLoadr, "ldrh", "\t$dst, $addr",
                   [(set GPR:$dst, (zextloadi16 addrmode3:$addr))]>;
 
 def LDRB  : AI2ldb<(outs GPR:$dst), (ins addrmode2:$addr), LdFrm, 
-                  IIC_iLoadr, "ldr", "b\t$dst, $addr",
+                  IIC_iLoadr, "ldrb", "\t$dst, $addr",
                   [(set GPR:$dst, (zextloadi8 addrmode2:$addr))]>;
 
 // Loads with sign extension
 def LDRSH : AI3ldsh<(outs GPR:$dst), (ins addrmode3:$addr), LdMiscFrm,
-                   IIC_iLoadr, "ldr", "sh\t$dst, $addr",
+                   IIC_iLoadr, "ldrsh", "\t$dst, $addr",
                    [(set GPR:$dst, (sextloadi16 addrmode3:$addr))]>;
 
 def LDRSB : AI3ldsb<(outs GPR:$dst), (ins addrmode3:$addr), LdMiscFrm,
-                   IIC_iLoadr, "ldr", "sb\t$dst, $addr",
+                   IIC_iLoadr, "ldrsb", "\t$dst, $addr",
                    [(set GPR:$dst, (sextloadi8 addrmode3:$addr))]>;
 
 let mayLoad = 1, hasExtraDefRegAllocReq = 1 in {
 // Load doubleword
 def LDRD : AI3ldd<(outs GPR:$dst1, GPR:$dst2), (ins addrmode3:$addr), LdMiscFrm,
-                 IIC_iLoadr, "ldr", "d\t$dst1, $addr",
+                 IIC_iLoadr, "ldrd", "\t$dst1, $addr",
                  []>, Requires<[IsARM, HasV5TE]>;
 
 // Indexed loads
@@ -836,35 +841,35 @@ def LDR_POST : AI2ldwpo<(outs GPR:$dst, GPR:$base_wb),
 
 def LDRH_PRE  : AI3ldhpr<(outs GPR:$dst, GPR:$base_wb),
                      (ins addrmode3:$addr), LdMiscFrm, IIC_iLoadru,
-                     "ldr", "h\t$dst, $addr!", "$addr.base = $base_wb", []>;
+                     "ldrh", "\t$dst, $addr!", "$addr.base = $base_wb", []>;
 
 def LDRH_POST : AI3ldhpo<(outs GPR:$dst, GPR:$base_wb),
                      (ins GPR:$base,am3offset:$offset), LdMiscFrm, IIC_iLoadru,
-                     "ldr", "h\t$dst, [$base], $offset", "$base = $base_wb", []>;
+                    "ldrh", "\t$dst, [$base], $offset", "$base = $base_wb", []>;
 
 def LDRB_PRE  : AI2ldbpr<(outs GPR:$dst, GPR:$base_wb),
                      (ins addrmode2:$addr), LdFrm, IIC_iLoadru,
-                     "ldr", "b\t$dst, $addr!", "$addr.base = $base_wb", []>;
+                     "ldrb", "\t$dst, $addr!", "$addr.base = $base_wb", []>;
 
 def LDRB_POST : AI2ldbpo<(outs GPR:$dst, GPR:$base_wb),
                      (ins GPR:$base,am2offset:$offset), LdFrm, IIC_iLoadru,
-                     "ldr", "b\t$dst, [$base], $offset", "$base = $base_wb", []>;
+                    "ldrb", "\t$dst, [$base], $offset", "$base = $base_wb", []>;
 
 def LDRSH_PRE : AI3ldshpr<(outs GPR:$dst, GPR:$base_wb),
                       (ins addrmode3:$addr), LdMiscFrm, IIC_iLoadru,
-                      "ldr", "sh\t$dst, $addr!", "$addr.base = $base_wb", []>;
+                      "ldrsh", "\t$dst, $addr!", "$addr.base = $base_wb", []>;
 
 def LDRSH_POST: AI3ldshpo<(outs GPR:$dst, GPR:$base_wb),
                       (ins GPR:$base,am3offset:$offset), LdMiscFrm, IIC_iLoadru,
-                    "ldr", "sh\t$dst, [$base], $offset", "$base = $base_wb", []>;
+                   "ldrsh", "\t$dst, [$base], $offset", "$base = $base_wb", []>;
 
 def LDRSB_PRE : AI3ldsbpr<(outs GPR:$dst, GPR:$base_wb),
                       (ins addrmode3:$addr), LdMiscFrm, IIC_iLoadru,
-                      "ldr", "sb\t$dst, $addr!", "$addr.base = $base_wb", []>;
+                      "ldrsb", "\t$dst, $addr!", "$addr.base = $base_wb", []>;
 
 def LDRSB_POST: AI3ldsbpo<(outs GPR:$dst, GPR:$base_wb),
                       (ins GPR:$base,am3offset:$offset), LdMiscFrm, IIC_iLoadru,
-                    "ldr", "sb\t$dst, [$base], $offset", "$base = $base_wb", []>;
+                   "ldrsb", "\t$dst, [$base], $offset", "$base = $base_wb", []>;
 }
 
 // Store
@@ -874,18 +879,18 @@ def STR  : AI2stw<(outs), (ins GPR:$src, addrmode2:$addr), StFrm, IIC_iStorer,
 
 // Stores with truncate
 def STRH : AI3sth<(outs), (ins GPR:$src, addrmode3:$addr), StMiscFrm, IIC_iStorer,
-               "str", "h\t$src, $addr",
+               "strh", "\t$src, $addr",
                [(truncstorei16 GPR:$src, addrmode3:$addr)]>;
 
 def STRB : AI2stb<(outs), (ins GPR:$src, addrmode2:$addr), StFrm, IIC_iStorer,
-               "str", "b\t$src, $addr",
+               "strb", "\t$src, $addr",
                [(truncstorei8 GPR:$src, addrmode2:$addr)]>;
 
 // Store doubleword
 let mayStore = 1, hasExtraSrcRegAllocReq = 1 in
 def STRD : AI3std<(outs), (ins GPR:$src1, GPR:$src2, addrmode3:$addr),
                StMiscFrm, IIC_iStorer,
-               "str", "d\t$src1, $addr", []>, Requires<[IsARM, HasV5TE]>;
+               "strd", "\t$src1, $addr", []>, Requires<[IsARM, HasV5TE]>;
 
 // Indexed stores
 def STR_PRE  : AI2stwpr<(outs GPR:$base_wb),
@@ -905,28 +910,28 @@ def STR_POST : AI2stwpo<(outs GPR:$base_wb),
 def STRH_PRE : AI3sthpr<(outs GPR:$base_wb),
                      (ins GPR:$src, GPR:$base,am3offset:$offset), 
                      StMiscFrm, IIC_iStoreru,
-                     "str", "h\t$src, [$base, $offset]!", "$base = $base_wb",
+                     "strh", "\t$src, [$base, $offset]!", "$base = $base_wb",
                     [(set GPR:$base_wb,
                       (pre_truncsti16 GPR:$src, GPR:$base,am3offset:$offset))]>;
 
 def STRH_POST: AI3sthpo<(outs GPR:$base_wb),
                      (ins GPR:$src, GPR:$base,am3offset:$offset), 
                      StMiscFrm, IIC_iStoreru,
-                     "str", "h\t$src, [$base], $offset", "$base = $base_wb",
+                     "strh", "\t$src, [$base], $offset", "$base = $base_wb",
                     [(set GPR:$base_wb, (post_truncsti16 GPR:$src,
                                          GPR:$base, am3offset:$offset))]>;
 
 def STRB_PRE : AI2stbpr<(outs GPR:$base_wb),
                      (ins GPR:$src, GPR:$base,am2offset:$offset), 
                      StFrm, IIC_iStoreru,
-                     "str", "b\t$src, [$base, $offset]!", "$base = $base_wb",
+                     "strb", "\t$src, [$base, $offset]!", "$base = $base_wb",
                     [(set GPR:$base_wb, (pre_truncsti8 GPR:$src,
                                          GPR:$base, am2offset:$offset))]>;
 
 def STRB_POST: AI2stbpo<(outs GPR:$base_wb),
                      (ins GPR:$src, GPR:$base,am2offset:$offset), 
                      StFrm, IIC_iStoreru,
-                     "str", "b\t$src, [$base], $offset", "$base = $base_wb",
+                     "strb", "\t$src, [$base], $offset", "$base = $base_wb",
                     [(set GPR:$base_wb, (post_truncsti8 GPR:$src,
                                          GPR:$base, am2offset:$offset))]>;
 
@@ -937,13 +942,13 @@ def STRB_POST: AI2stbpo<(outs GPR:$base_wb),
 let mayLoad = 1, hasExtraDefRegAllocReq = 1 in
 def LDM : AXI4ld<(outs),
                (ins addrmode4:$addr, pred:$p, reglist:$wb, variable_ops),
-               LdStMulFrm, IIC_iLoadm, "ldm${p}${addr:submode}\t$addr, $wb",
+               LdStMulFrm, IIC_iLoadm, "ldm${addr:submode}${p}\t$addr, $wb",
                []>;
 
 let mayStore = 1, hasExtraSrcRegAllocReq = 1 in
 def STM : AXI4st<(outs),
                (ins addrmode4:$addr, pred:$p, reglist:$wb, variable_ops),
-               LdStMulFrm, IIC_iStorem, "stm${p}${addr:submode}\t$addr, $wb",
+               LdStMulFrm, IIC_iStorem, "stm${addr:submode}${p}\t$addr, $wb",
                []>;
 
 //===----------------------------------------------------------------------===//
@@ -1004,10 +1009,10 @@ def MOVrx : AsI1<0b1101, (outs GPR:$dst), (ins GPR:$src), Pseudo, IIC_iMOVsi,
 
 let Defs = [CPSR] in {
 def MOVsrl_flag : AI1<0b1101, (outs GPR:$dst), (ins GPR:$src), Pseudo, 
-                      IIC_iMOVsi, "mov", "s\t$dst, $src, lsr #1",
+                      IIC_iMOVsi, "movs", "\t$dst, $src, lsr #1",
                       [(set GPR:$dst, (ARMsrl_flag GPR:$src))]>, UnaryDP;
 def MOVsra_flag : AI1<0b1101, (outs GPR:$dst), (ins GPR:$src), Pseudo,
-                      IIC_iMOVsi, "mov", "s\t$dst, $src, asr #1",
+                      IIC_iMOVsi, "movs", "\t$dst, $src, asr #1",
                       [(set GPR:$dst, (ARMsra_flag GPR:$src))]>, UnaryDP;
 }
 
@@ -1083,15 +1088,19 @@ defm SUB  : AsI1_bin_irs<0b0010, "sub",
                          BinOpFrag<(sub  node:$LHS, node:$RHS)>>;
 
 // ADD and SUB with 's' bit set.
-defm ADDS : AI1_bin_s_irs<0b0100, "add",
-                          BinOpFrag<(addc node:$LHS, node:$RHS)>>;
-defm SUBS : AI1_bin_s_irs<0b0010, "sub",
+defm ADDS : AI1_bin_s_irs<0b0100, "adds",
+                          BinOpFrag<(addc node:$LHS, node:$RHS)>, 1>;
+defm SUBS : AI1_bin_s_irs<0b0010, "subs",
                           BinOpFrag<(subc node:$LHS, node:$RHS)>>;
 
 defm ADC : AI1_adde_sube_irs<0b0101, "adc",
                              BinOpFrag<(adde node:$LHS, node:$RHS)>, 1>;
 defm SBC : AI1_adde_sube_irs<0b0110, "sbc",
                              BinOpFrag<(sube node:$LHS, node:$RHS)>>;
+defm ADCS : AI1_adde_sube_s_irs<0b0101, "adcs",
+                             BinOpFrag<(adde node:$LHS, node:$RHS)>, 1>;
+defm SBCS : AI1_adde_sube_s_irs<0b0110, "sbcs",
+                             BinOpFrag<(sube node:$LHS, node:$RHS)>>;
 
 // These don't define reg/reg forms, because they are handled above.
 def RSBri : AsI1<0b0011, (outs GPR:$dst), (ins GPR:$a, so_imm:$b), DPFrm,
@@ -1109,13 +1118,13 @@ def RSBrs : AsI1<0b0011, (outs GPR:$dst), (ins GPR:$a, so_reg:$b), DPSoRegFrm,
 // RSB with 's' bit set.
 let Defs = [CPSR] in {
 def RSBSri : AI1<0b0011, (outs GPR:$dst), (ins GPR:$a, so_imm:$b), DPFrm,
-                 IIC_iALUi, "rsb", "s\t$dst, $a, $b",
+                 IIC_iALUi, "rsbs", "\t$dst, $a, $b",
                  [(set GPR:$dst, (subc so_imm:$b, GPR:$a))]> {
     let Inst{20} = 1;
     let Inst{25} = 1;
 }
 def RSBSrs : AI1<0b0011, (outs GPR:$dst), (ins GPR:$a, so_reg:$b), DPSoRegFrm,
-                 IIC_iALUsr, "rsb", "s\t$dst, $a, $b",
+                 IIC_iALUsr, "rsbs", "\t$dst, $a, $b",
                  [(set GPR:$dst, (subc so_reg:$b, GPR:$a))]> {
     let Inst{20} = 1;
     let Inst{25} = 0;
diff --git a/lib/Target/ARM/ARMInstrNEON.td b/lib/Target/ARM/ARMInstrNEON.td
index 4c34035efa8..05d2918c3d3 100644
--- a/lib/Target/ARM/ARMInstrNEON.td
+++ b/lib/Target/ARM/ARMInstrNEON.td
@@ -2295,7 +2295,7 @@ def  VCNTq    : N2VQInt<0b11, 0b11, 0b00, 0b00, 0b01010, 0,
 
 //   VMOV     : Vector Move (Register)
 
-def  VMOVD    : N3V<0, 0, 0b10, 0b0001, 0, 1, (outs DPR:$dst), (ins DPR:$src),
+def  VMOVDneon: N3V<0, 0, 0b10, 0b0001, 0, 1, (outs DPR:$dst), (ins DPR:$src),
                     IIC_VMOVD, "vmov\t$dst, $src", "", []>;
 def  VMOVQ    : N3V<0, 0, 0b10, 0b0001, 1, 1, (outs QPR:$dst), (ins QPR:$src),
                     IIC_VMOVD, "vmov\t$dst, $src", "", []>;
diff --git a/lib/Target/ARM/ARMInstrVFP.td b/lib/Target/ARM/ARMInstrVFP.td
index 455c33b7959..a1d581d0e4e 100644
--- a/lib/Target/ARM/ARMInstrVFP.td
+++ b/lib/Target/ARM/ARMInstrVFP.td
@@ -17,7 +17,7 @@ def SDT_ITOF :
 SDTypeProfile<1, 1, [SDTCisFP<0>, SDTCisVT<1, f32>]>;
 def SDT_CMPFP0 :
 SDTypeProfile<0, 1, [SDTCisFP<0>]>;
-def SDT_FMDRR :
+def SDT_VMOVDRR :
 SDTypeProfile<1, 2, [SDTCisVT<0, f64>, SDTCisVT<1, i32>,
                      SDTCisSameAs<1, 2>]>;
 
@@ -28,7 +28,7 @@ def arm_uitof  : SDNode<"ARMISD::UITOF",  SDT_ITOF>;
 def arm_fmstat : SDNode<"ARMISD::FMSTAT", SDTNone, [SDNPInFlag,SDNPOutFlag]>;
 def arm_cmpfp  : SDNode<"ARMISD::CMPFP",  SDT_ARMCmp, [SDNPOutFlag]>;
 def arm_cmpfp0 : SDNode<"ARMISD::CMPFPw0",SDT_CMPFP0, [SDNPOutFlag]>;
-def arm_fmdrr  : SDNode<"ARMISD::FMDRR",  SDT_FMDRR>;
+def arm_fmdrr  : SDNode<"ARMISD::VMOVDRR",  SDT_VMOVDRR>;
 
 //===----------------------------------------------------------------------===//
 // Operand Definitions.
@@ -55,21 +55,21 @@ def vfp_f64imm : Operand<f64>,
 //
 
 let canFoldAsLoad = 1 in {
-def FLDD  : ADI5<0b1101, 0b01, (outs DPR:$dst), (ins addrmode5:$addr),
-                 IIC_fpLoad64, "fldd", "\t$dst, $addr",
+def VLDRD : ADI5<0b1101, 0b01, (outs DPR:$dst), (ins addrmode5:$addr),
+                 IIC_fpLoad64, "vldr", ".64\t$dst, $addr",
                  [(set DPR:$dst, (load addrmode5:$addr))]>;
 
-def FLDS  : ASI5<0b1101, 0b01, (outs SPR:$dst), (ins addrmode5:$addr),
-                 IIC_fpLoad32, "flds", "\t$dst, $addr",
+def VLDRS : ASI5<0b1101, 0b01, (outs SPR:$dst), (ins addrmode5:$addr),
+                 IIC_fpLoad32, "vldr", ".32\t$dst, $addr",
                  [(set SPR:$dst, (load addrmode5:$addr))]>;
 } // canFoldAsLoad
 
-def FSTD  : ADI5<0b1101, 0b00, (outs), (ins DPR:$src, addrmode5:$addr),
-                 IIC_fpStore64, "fstd", "\t$src, $addr",
+def VSTRD  : ADI5<0b1101, 0b00, (outs), (ins DPR:$src, addrmode5:$addr),
+                 IIC_fpStore64, "vstr", ".64\t$src, $addr",
                  [(store DPR:$src, addrmode5:$addr)]>;
 
-def FSTS  : ASI5<0b1101, 0b00, (outs), (ins SPR:$src, addrmode5:$addr),
-                 IIC_fpStore32, "fsts", "\t$src, $addr",
+def VSTRS  : ASI5<0b1101, 0b00, (outs), (ins SPR:$src, addrmode5:$addr),
+                 IIC_fpStore32, "vstr", ".32\t$src, $addr",
                  [(store SPR:$src, addrmode5:$addr)]>;
 
 //===----------------------------------------------------------------------===//
@@ -77,32 +77,32 @@ def FSTS  : ASI5<0b1101, 0b00, (outs), (ins SPR:$src, addrmode5:$addr),
 //
 
 let mayLoad = 1, hasExtraDefRegAllocReq = 1 in {
-def FLDMD : AXDI5<(outs), (ins addrmode5:$addr, pred:$p, reglist:$wb,
+def VLDMD : AXDI5<(outs), (ins addrmode5:$addr, pred:$p, reglist:$wb,
                            variable_ops), IIC_fpLoadm,
-                  "fldm${addr:submode}d${p}\t${addr:base}, $wb",
+                  "vldm${addr:submode}${p}\t${addr:base}, $wb",
                   []> {
   let Inst{20} = 1;
 }
 
-def FLDMS : AXSI5<(outs), (ins addrmode5:$addr, pred:$p, reglist:$wb,
+def VLDMS : AXSI5<(outs), (ins addrmode5:$addr, pred:$p, reglist:$wb,
                            variable_ops), IIC_fpLoadm, 
-                  "fldm${addr:submode}s${p}\t${addr:base}, $wb",
+                  "vldm${addr:submode}${p}\t${addr:base}, $wb",
                   []> {
   let Inst{20} = 1;
 }
 } // mayLoad, hasExtraDefRegAllocReq
 
 let mayStore = 1, hasExtraSrcRegAllocReq = 1 in {
-def FSTMD : AXDI5<(outs), (ins addrmode5:$addr, pred:$p, reglist:$wb,
+def VSTMD : AXDI5<(outs), (ins addrmode5:$addr, pred:$p, reglist:$wb,
                            variable_ops), IIC_fpStorem,
-                 "fstm${addr:submode}d${p}\t${addr:base}, $wb",
+                 "vstm${addr:submode}${p}\t${addr:base}, $wb",
                  []> {
   let Inst{20} = 0;
 }
 
-def FSTMS : AXSI5<(outs), (ins addrmode5:$addr, pred:$p, reglist:$wb,
+def VSTMS : AXSI5<(outs), (ins addrmode5:$addr, pred:$p, reglist:$wb,
                            variable_ops), IIC_fpStorem,
-                 "fstm${addr:submode}s${p}\t${addr:base}, $wb",
+                 "vstm${addr:submode}${p}\t${addr:base}, $wb",
                  []> {
   let Inst{20} = 0;
 }
@@ -114,68 +114,68 @@ def FSTMS : AXSI5<(outs), (ins addrmode5:$addr, pred:$p, reglist:$wb,
 // FP Binary Operations.
 //
 
-def FADDD  : ADbI<0b11100011, (outs DPR:$dst), (ins DPR:$a, DPR:$b),
-                 IIC_fpALU64, "faddd", "\t$dst, $a, $b",
+def VADDD  : ADbI<0b11100011, (outs DPR:$dst), (ins DPR:$a, DPR:$b),
+                 IIC_fpALU64, "vadd", ".f64\t$dst, $a, $b",
                  [(set DPR:$dst, (fadd DPR:$a, DPR:$b))]>;
 
-def FADDS  : ASbIn<0b11100011, (outs SPR:$dst), (ins SPR:$a, SPR:$b),
-                  IIC_fpALU32, "fadds", "\t$dst, $a, $b",
+def VADDS  : ASbIn<0b11100011, (outs SPR:$dst), (ins SPR:$a, SPR:$b),
+                  IIC_fpALU32, "vadd", ".f32\t$dst, $a, $b",
                   [(set SPR:$dst, (fadd SPR:$a, SPR:$b))]>;
 
 // These are encoded as unary instructions.
 let Defs = [FPSCR] in {
-def FCMPED : ADuI<0b11101011, 0b0100, 0b1100, (outs), (ins DPR:$a, DPR:$b),
-                 IIC_fpCMP64, "fcmped", "\t$a, $b",
+def VCMPED : ADuI<0b11101011, 0b0100, 0b1100, (outs), (ins DPR:$a, DPR:$b),
+                 IIC_fpCMP64, "vcmpe", ".f64\t$a, $b",
                  [(arm_cmpfp DPR:$a, DPR:$b)]>;
 
-def FCMPES : ASuI<0b11101011, 0b0100, 0b1100, (outs), (ins SPR:$a, SPR:$b),
-                 IIC_fpCMP32, "fcmpes", "\t$a, $b",
+def VCMPES : ASuI<0b11101011, 0b0100, 0b1100, (outs), (ins SPR:$a, SPR:$b),
+                 IIC_fpCMP32, "vcmpe", ".f32\t$a, $b",
                  [(arm_cmpfp SPR:$a, SPR:$b)]>;
 }
 
-def FDIVD  : ADbI<0b11101000, (outs DPR:$dst), (ins DPR:$a, DPR:$b),
-                 IIC_fpDIV64, "fdivd", "\t$dst, $a, $b",
+def VDIVD  : ADbI<0b11101000, (outs DPR:$dst), (ins DPR:$a, DPR:$b),
+                 IIC_fpDIV64, "vdiv", ".f64\t$dst, $a, $b",
                  [(set DPR:$dst, (fdiv DPR:$a, DPR:$b))]>;
 
-def FDIVS  : ASbI<0b11101000, (outs SPR:$dst), (ins SPR:$a, SPR:$b),
-                 IIC_fpDIV32, "fdivs", "\t$dst, $a, $b",
+def VDIVS  : ASbI<0b11101000, (outs SPR:$dst), (ins SPR:$a, SPR:$b),
+                 IIC_fpDIV32, "vdiv", ".f32\t$dst, $a, $b",
                  [(set SPR:$dst, (fdiv SPR:$a, SPR:$b))]>;
 
-def FMULD  : ADbI<0b11100010, (outs DPR:$dst), (ins DPR:$a, DPR:$b),
-                 IIC_fpMUL64, "fmuld", "\t$dst, $a, $b",
+def VMULD  : ADbI<0b11100010, (outs DPR:$dst), (ins DPR:$a, DPR:$b),
+                 IIC_fpMUL64, "vmul", ".f64\t$dst, $a, $b",
                  [(set DPR:$dst, (fmul DPR:$a, DPR:$b))]>;
 
-def FMULS  : ASbIn<0b11100010, (outs SPR:$dst), (ins SPR:$a, SPR:$b),
-                  IIC_fpMUL32, "fmuls", "\t$dst, $a, $b",
+def VMULS  : ASbIn<0b11100010, (outs SPR:$dst), (ins SPR:$a, SPR:$b),
+                  IIC_fpMUL32, "vmul", ".f32\t$dst, $a, $b",
                   [(set SPR:$dst, (fmul SPR:$a, SPR:$b))]>;
-                 
-def FNMULD  : ADbI<0b11100010, (outs DPR:$dst), (ins DPR:$a, DPR:$b),
-                  IIC_fpMUL64, "fnmuld", "\t$dst, $a, $b",
+
+def VNMULD  : ADbI<0b11100010, (outs DPR:$dst), (ins DPR:$a, DPR:$b),
+                  IIC_fpMUL64, "vnmul", ".f64\t$dst, $a, $b",
                   [(set DPR:$dst, (fneg (fmul DPR:$a, DPR:$b)))]> {
   let Inst{6} = 1;
 }
 
-def FNMULS  : ASbI<0b11100010, (outs SPR:$dst), (ins SPR:$a, SPR:$b),
-                  IIC_fpMUL32, "fnmuls", "\t$dst, $a, $b",
+def VNMULS  : ASbI<0b11100010, (outs SPR:$dst), (ins SPR:$a, SPR:$b),
+                  IIC_fpMUL32, "vnmul", ".f32\t$dst, $a, $b",
                   [(set SPR:$dst, (fneg (fmul SPR:$a, SPR:$b)))]> {
   let Inst{6} = 1;
 }
 
 // Match reassociated forms only if not sign dependent rounding.
 def : Pat<(fmul (fneg DPR:$a), DPR:$b),
-          (FNMULD DPR:$a, DPR:$b)>, Requires<[NoHonorSignDependentRounding]>;
+          (VNMULD DPR:$a, DPR:$b)>, Requires<[NoHonorSignDependentRounding]>;
 def : Pat<(fmul (fneg SPR:$a), SPR:$b),
-          (FNMULS SPR:$a, SPR:$b)>, Requires<[NoHonorSignDependentRounding]>;
+          (VNMULS SPR:$a, SPR:$b)>, Requires<[NoHonorSignDependentRounding]>;
 
 
-def FSUBD  : ADbI<0b11100011, (outs DPR:$dst), (ins DPR:$a, DPR:$b),
-                 IIC_fpALU64, "fsubd", "\t$dst, $a, $b",
+def VSUBD  : ADbI<0b11100011, (outs DPR:$dst), (ins DPR:$a, DPR:$b),
+                 IIC_fpALU64, "vsub", ".f64\t$dst, $a, $b",
                  [(set DPR:$dst, (fsub DPR:$a, DPR:$b))]> {
   let Inst{6} = 1;
 }
 
-def FSUBS  : ASbIn<0b11100011, (outs SPR:$dst), (ins SPR:$a, SPR:$b),
-                  IIC_fpALU32, "fsubs", "\t$dst, $a, $b",
+def VSUBS  : ASbIn<0b11100011, (outs SPR:$dst), (ins SPR:$a, SPR:$b),
+                  IIC_fpALU32, "vsub", ".f32\t$dst, $a, $b",
                   [(set SPR:$dst, (fsub SPR:$a, SPR:$b))]> {
   let Inst{6} = 1;
 }
@@ -184,31 +184,31 @@ def FSUBS  : ASbIn<0b11100011, (outs SPR:$dst), (ins SPR:$a, SPR:$b),
 // FP Unary Operations.
 //
 
-def FABSD  : ADuI<0b11101011, 0b0000, 0b1100, (outs DPR:$dst), (ins DPR:$a),
-                 IIC_fpUNA64, "fabsd", "\t$dst, $a",
+def VABSD  : ADuI<0b11101011, 0b0000, 0b1100, (outs DPR:$dst), (ins DPR:$a),
+                 IIC_fpUNA64, "vabs", ".f64\t$dst, $a",
                  [(set DPR:$dst, (fabs DPR:$a))]>;
 
-def FABSS  : ASuIn<0b11101011, 0b0000, 0b1100, (outs SPR:$dst), (ins SPR:$a),
-                  IIC_fpUNA32, "fabss", "\t$dst, $a",
+def VABSS  : ASuIn<0b11101011, 0b0000, 0b1100, (outs SPR:$dst), (ins SPR:$a),
+                  IIC_fpUNA32, "vabs", ".f32\t$dst, $a",
                   [(set SPR:$dst, (fabs SPR:$a))]>;
 
 let Defs = [FPSCR] in {
-def FCMPEZD : ADuI<0b11101011, 0b0101, 0b1100, (outs), (ins DPR:$a),
-                  IIC_fpCMP64, "fcmpezd", "\t$a",
+def VCMPEZD : ADuI<0b11101011, 0b0101, 0b1100, (outs), (ins DPR:$a),
+                  IIC_fpCMP64, "vcmpe", ".f64\t$a, #0.0",
                   [(arm_cmpfp0 DPR:$a)]>;
 
-def FCMPEZS : ASuI<0b11101011, 0b0101, 0b1100, (outs), (ins SPR:$a),
-                  IIC_fpCMP32, "fcmpezs", "\t$a",
+def VCMPEZS : ASuI<0b11101011, 0b0101, 0b1100, (outs), (ins SPR:$a),
+                  IIC_fpCMP32, "vcmpe", ".f32\t$a, #0.0",
                   [(arm_cmpfp0 SPR:$a)]>;
 }
 
-def FCVTDS : ASuI<0b11101011, 0b0111, 0b1100, (outs DPR:$dst), (ins SPR:$a),
-                 IIC_fpCVTDS, "fcvtds", "\t$dst, $a",
+def VCVTDS : ASuI<0b11101011, 0b0111, 0b1100, (outs DPR:$dst), (ins SPR:$a),
+                 IIC_fpCVTDS, "vcvt", ".f64.f32\t$dst, $a",
                  [(set DPR:$dst, (fextend SPR:$a))]>;
 
 // Special case encoding: bits 11-8 is 0b1011.
-def FCVTSD : VFPAI<(outs SPR:$dst), (ins DPR:$a), VFPUnaryFrm,
-                   IIC_fpCVTSD, "fcvtsd", "\t$dst, $a",
+def VCVTSD : VFPAI<(outs SPR:$dst), (ins DPR:$a), VFPUnaryFrm,
+                   IIC_fpCVTSD, "vcvt", ".f32.f64\t$dst, $a",
                    [(set SPR:$dst, (fround DPR:$a))]> {
   let Inst{27-23} = 0b11101;
   let Inst{21-16} = 0b110111;
@@ -217,52 +217,52 @@ def FCVTSD : VFPAI<(outs SPR:$dst), (ins DPR:$a), VFPUnaryFrm,
 }
 
 let neverHasSideEffects = 1 in {
-def FCPYD  : ADuI<0b11101011, 0b0000, 0b0100, (outs DPR:$dst), (ins DPR:$a),
-                 IIC_fpUNA64, "fcpyd", "\t$dst, $a", []>;
+def VMOVD: ADuI<0b11101011, 0b0000, 0b0100, (outs DPR:$dst), (ins DPR:$a),
+                 IIC_fpUNA64, "vmov", ".f64\t$dst, $a", []>;
 
-def FCPYS  : ASuI<0b11101011, 0b0000, 0b0100, (outs SPR:$dst), (ins SPR:$a),
-                 IIC_fpUNA32, "fcpys", "\t$dst, $a", []>;
+def VMOVS: ASuI<0b11101011, 0b0000, 0b0100, (outs SPR:$dst), (ins SPR:$a),
+                 IIC_fpUNA32, "vmov", ".f32\t$dst, $a", []>;
 } // neverHasSideEffects
 
-def FNEGD  : ADuI<0b11101011, 0b0001, 0b0100, (outs DPR:$dst), (ins DPR:$a),
-                 IIC_fpUNA64, "fnegd", "\t$dst, $a",
+def VNEGD  : ADuI<0b11101011, 0b0001, 0b0100, (outs DPR:$dst), (ins DPR:$a),
+                 IIC_fpUNA64, "vneg", ".f64\t$dst, $a",
                  [(set DPR:$dst, (fneg DPR:$a))]>;
 
-def FNEGS  : ASuIn<0b11101011, 0b0001, 0b0100, (outs SPR:$dst), (ins SPR:$a),
-                  IIC_fpUNA32, "fnegs", "\t$dst, $a",
+def VNEGS  : ASuIn<0b11101011, 0b0001, 0b0100, (outs SPR:$dst), (ins SPR:$a),
+                  IIC_fpUNA32, "vneg", ".f32\t$dst, $a",
                   [(set SPR:$dst, (fneg SPR:$a))]>;
 
-def FSQRTD  : ADuI<0b11101011, 0b0001, 0b1100, (outs DPR:$dst), (ins DPR:$a),
-                 IIC_fpSQRT64, "fsqrtd", "\t$dst, $a",
+def VSQRTD  : ADuI<0b11101011, 0b0001, 0b1100, (outs DPR:$dst), (ins DPR:$a),
+                 IIC_fpSQRT64, "vsqrt", ".f64\t$dst, $a",
                  [(set DPR:$dst, (fsqrt DPR:$a))]>;
 
-def FSQRTS  : ASuI<0b11101011, 0b0001, 0b1100, (outs SPR:$dst), (ins SPR:$a),
-                 IIC_fpSQRT32, "fsqrts", "\t$dst, $a",
+def VSQRTS  : ASuI<0b11101011, 0b0001, 0b1100, (outs SPR:$dst), (ins SPR:$a),
+                 IIC_fpSQRT32, "vsqrt", ".f32\t$dst, $a",
                  [(set SPR:$dst, (fsqrt SPR:$a))]>;
 
 //===----------------------------------------------------------------------===//
 // FP <-> GPR Copies.  Int <-> FP Conversions.
 //
 
-def FMRS   : AVConv2I<0b11100001, 0b1010, (outs GPR:$dst), (ins SPR:$src),
-                 IIC_VMOVSI, "fmrs", "\t$dst, $src",
+def VMOVRS : AVConv2I<0b11100001, 0b1010, (outs GPR:$dst), (ins SPR:$src),
+                 IIC_VMOVSI, "vmov", "\t$dst, $src",
                  [(set GPR:$dst, (bitconvert SPR:$src))]>;
 
-def FMSR   : AVConv4I<0b11100000, 0b1010, (outs SPR:$dst), (ins GPR:$src),
-                 IIC_VMOVIS, "fmsr", "\t$dst, $src",
+def VMOVSR : AVConv4I<0b11100000, 0b1010, (outs SPR:$dst), (ins GPR:$src),
+                 IIC_VMOVIS, "vmov", "\t$dst, $src",
                  [(set SPR:$dst, (bitconvert GPR:$src))]>;
 
-def FMRRD  : AVConv3I<0b11000101, 0b1011,
+def VMOVRRD  : AVConv3I<0b11000101, 0b1011,
                       (outs GPR:$wb, GPR:$dst2), (ins DPR:$src),
-                 IIC_VMOVDI, "fmrrd", "\t$wb, $dst2, $src",
+                 IIC_VMOVDI, "vmov", "\t$wb, $dst2, $src",
                  [/* FIXME: Can't write pattern for multiple result instr*/]>;
 
 // FMDHR: GPR -> SPR
 // FMDLR: GPR -> SPR
 
-def FMDRR : AVConv5I<0b11000100, 0b1011,
+def VMOVDRR : AVConv5I<0b11000100, 0b1011,
                      (outs DPR:$dst), (ins GPR:$src1, GPR:$src2),
-                IIC_VMOVID, "fmdrr", "\t$dst, $src1, $src2",
+                IIC_VMOVID, "vmov", "\t$dst, $src1, $src2",
                 [(set DPR:$dst, (arm_fmdrr GPR:$src1, GPR:$src2))]>;
 
 // FMRDH: SPR -> GPR
@@ -277,53 +277,53 @@ def FMDRR : AVConv5I<0b11000100, 0b1011,
 
 // Int to FP:
 
-def FSITOD : AVConv1I<0b11101011, 0b1000, 0b1011, (outs DPR:$dst), (ins SPR:$a),
-                 IIC_fpCVTID, "fsitod", "\t$dst, $a",
+def VSITOD : AVConv1I<0b11101011, 0b1000, 0b1011, (outs DPR:$dst), (ins SPR:$a),
+                 IIC_fpCVTID, "vcvt", ".f64.s32\t$dst, $a",
                  [(set DPR:$dst, (arm_sitof SPR:$a))]> {
   let Inst{7} = 1;
 }
 
-def FSITOS : AVConv1In<0b11101011, 0b1000, 0b1010, (outs SPR:$dst),(ins SPR:$a),
-                 IIC_fpCVTIS, "fsitos", "\t$dst, $a",
+def VSITOS : AVConv1In<0b11101011, 0b1000, 0b1010, (outs SPR:$dst),(ins SPR:$a),
+                 IIC_fpCVTIS, "vcvt", ".f32.s32\t$dst, $a",
                  [(set SPR:$dst, (arm_sitof SPR:$a))]> {
   let Inst{7} = 1;
 }
 
-def FUITOD : AVConv1I<0b11101011, 0b1000, 0b1011, (outs DPR:$dst), (ins SPR:$a),
-                 IIC_fpCVTID, "fuitod", "\t$dst, $a",
+def VUITOD : AVConv1I<0b11101011, 0b1000, 0b1011, (outs DPR:$dst), (ins SPR:$a),
+                 IIC_fpCVTID, "vcvt", ".f64.u32\t$dst, $a",
                  [(set DPR:$dst, (arm_uitof SPR:$a))]>;
 
-def FUITOS : AVConv1In<0b11101011, 0b1000, 0b1010, (outs SPR:$dst),(ins SPR:$a),
-                 IIC_fpCVTIS, "fuitos", "\t$dst, $a",
+def VUITOS : AVConv1In<0b11101011, 0b1000, 0b1010, (outs SPR:$dst),(ins SPR:$a),
+                 IIC_fpCVTIS, "vcvt", ".f32.u32\t$dst, $a",
                  [(set SPR:$dst, (arm_uitof SPR:$a))]>;
 
 // FP to Int:
 // Always set Z bit in the instruction, i.e. "round towards zero" variants.
 
-def FTOSIZD : AVConv1I<0b11101011, 0b1101, 0b1011,
+def VTOSIZD : AVConv1I<0b11101011, 0b1101, 0b1011,
                        (outs SPR:$dst), (ins DPR:$a),
-                 IIC_fpCVTDI, "ftosizd", "\t$dst, $a",
+                 IIC_fpCVTDI, "vcvt", ".s32.f64\t$dst, $a",
                  [(set SPR:$dst, (arm_ftosi DPR:$a))]> {
   let Inst{7} = 1; // Z bit
 }
 
-def FTOSIZS : AVConv1In<0b11101011, 0b1101, 0b1010,
+def VTOSIZS : AVConv1In<0b11101011, 0b1101, 0b1010,
                         (outs SPR:$dst), (ins SPR:$a),
-                 IIC_fpCVTSI, "ftosizs", "\t$dst, $a",
+                 IIC_fpCVTSI, "vcvt", ".s32.f32\t$dst, $a",
                  [(set SPR:$dst, (arm_ftosi SPR:$a))]> {
   let Inst{7} = 1; // Z bit
 }
 
-def FTOUIZD : AVConv1I<0b11101011, 0b1100, 0b1011,
+def VTOUIZD : AVConv1I<0b11101011, 0b1100, 0b1011,
                        (outs SPR:$dst), (ins DPR:$a),
-                 IIC_fpCVTDI, "ftouizd", "\t$dst, $a",
+                 IIC_fpCVTDI, "vcvt", ".u32.f64\t$dst, $a",
                  [(set SPR:$dst, (arm_ftoui DPR:$a))]> {
   let Inst{7} = 1; // Z bit
 }
 
-def FTOUIZS : AVConv1In<0b11101011, 0b1100, 0b1010,
+def VTOUIZS : AVConv1In<0b11101011, 0b1100, 0b1010,
                         (outs SPR:$dst), (ins SPR:$a),
-                 IIC_fpCVTSI, "ftouizs", "\t$dst, $a",
+                 IIC_fpCVTSI, "vcvt", ".u32.f32\t$dst, $a",
                  [(set SPR:$dst, (arm_ftoui SPR:$a))]> {
   let Inst{7} = 1; // Z bit
 }
@@ -332,54 +332,54 @@ def FTOUIZS : AVConv1In<0b11101011, 0b1100, 0b1010,
 // FP FMA Operations.
 //
 
-def FMACD : ADbI<0b11100000, (outs DPR:$dst), (ins DPR:$dstin, DPR:$a, DPR:$b),
-                IIC_fpMAC64, "fmacd", "\t$dst, $a, $b",
+def VMLAD : ADbI<0b11100000, (outs DPR:$dst), (ins DPR:$dstin, DPR:$a, DPR:$b),
+                IIC_fpMAC64, "vmla", ".f64\t$dst, $a, $b",
                 [(set DPR:$dst, (fadd (fmul DPR:$a, DPR:$b), DPR:$dstin))]>,
                 RegConstraint<"$dstin = $dst">;
 
-def FMACS : ASbIn<0b11100000, (outs SPR:$dst), (ins SPR:$dstin, SPR:$a, SPR:$b),
-                 IIC_fpMAC32, "fmacs", "\t$dst, $a, $b",
+def VMLAS : ASbIn<0b11100000, (outs SPR:$dst), (ins SPR:$dstin, SPR:$a, SPR:$b),
+                 IIC_fpMAC32, "vmla", ".f32\t$dst, $a, $b",
                  [(set SPR:$dst, (fadd (fmul SPR:$a, SPR:$b), SPR:$dstin))]>,
                  RegConstraint<"$dstin = $dst">;
 
-def FMSCD : ADbI<0b11100001, (outs DPR:$dst), (ins DPR:$dstin, DPR:$a, DPR:$b),
-                IIC_fpMAC64, "fmscd", "\t$dst, $a, $b",
+def VNMLSD : ADbI<0b11100001, (outs DPR:$dst), (ins DPR:$dstin, DPR:$a, DPR:$b),
+                IIC_fpMAC64, "vnmls", ".f64\t$dst, $a, $b",
                 [(set DPR:$dst, (fsub (fmul DPR:$a, DPR:$b), DPR:$dstin))]>,
                 RegConstraint<"$dstin = $dst">;
 
-def FMSCS : ASbI<0b11100001, (outs SPR:$dst), (ins SPR:$dstin, SPR:$a, SPR:$b),
-                IIC_fpMAC32, "fmscs", "\t$dst, $a, $b",
+def VNMLSS : ASbI<0b11100001, (outs SPR:$dst), (ins SPR:$dstin, SPR:$a, SPR:$b),
+                IIC_fpMAC32, "vnmls", ".f32\t$dst, $a, $b",
                 [(set SPR:$dst, (fsub (fmul SPR:$a, SPR:$b), SPR:$dstin))]>,
                 RegConstraint<"$dstin = $dst">;
 
-def FNMACD : ADbI<0b11100000, (outs DPR:$dst), (ins DPR:$dstin, DPR:$a, DPR:$b),
-                 IIC_fpMAC64, "fnmacd", "\t$dst, $a, $b",
+def VMLSD : ADbI<0b11100000, (outs DPR:$dst), (ins DPR:$dstin, DPR:$a, DPR:$b),
+                 IIC_fpMAC64, "vmls", ".f64\t$dst, $a, $b",
              [(set DPR:$dst, (fadd (fneg (fmul DPR:$a, DPR:$b)), DPR:$dstin))]>,
                 RegConstraint<"$dstin = $dst"> {
   let Inst{6} = 1;
 }
 
-def FNMACS : ASbIn<0b11100000, (outs SPR:$dst), (ins SPR:$dstin, SPR:$a, SPR:$b),
-                  IIC_fpMAC32, "fnmacs", "\t$dst, $a, $b",
+def VMLSS : ASbIn<0b11100000, (outs SPR:$dst), (ins SPR:$dstin, SPR:$a, SPR:$b),
+                  IIC_fpMAC32, "vmls", ".f32\t$dst, $a, $b",
              [(set SPR:$dst, (fadd (fneg (fmul SPR:$a, SPR:$b)), SPR:$dstin))]>,
                 RegConstraint<"$dstin = $dst"> {
   let Inst{6} = 1;
 }
 
 def : Pat<(fsub DPR:$dstin, (fmul DPR:$a, DPR:$b)),
-          (FNMACD DPR:$dstin, DPR:$a, DPR:$b)>, Requires<[DontUseNEONForFP]>;
+          (VMLSD DPR:$dstin, DPR:$a, DPR:$b)>, Requires<[DontUseNEONForFP]>;
 def : Pat<(fsub SPR:$dstin, (fmul SPR:$a, SPR:$b)),
-          (FNMACS SPR:$dstin, SPR:$a, SPR:$b)>, Requires<[DontUseNEONForFP]>;
+          (VMLSS SPR:$dstin, SPR:$a, SPR:$b)>, Requires<[DontUseNEONForFP]>;
 
-def FNMSCD : ADbI<0b11100001, (outs DPR:$dst), (ins DPR:$dstin, DPR:$a, DPR:$b),
-                 IIC_fpMAC64, "fnmscd", "\t$dst, $a, $b",
+def VNMLAD : ADbI<0b11100001, (outs DPR:$dst), (ins DPR:$dstin, DPR:$a, DPR:$b),
+                 IIC_fpMAC64, "vnmla", ".f64\t$dst, $a, $b",
              [(set DPR:$dst, (fsub (fneg (fmul DPR:$a, DPR:$b)), DPR:$dstin))]>,
                 RegConstraint<"$dstin = $dst"> {
   let Inst{6} = 1;
 }
 
-def FNMSCS : ASbI<0b11100001, (outs SPR:$dst), (ins SPR:$dstin, SPR:$a, SPR:$b),
-                IIC_fpMAC32, "fnmscs", "\t$dst, $a, $b",
+def VNMLAS : ASbI<0b11100001, (outs SPR:$dst), (ins SPR:$dstin, SPR:$a, SPR:$b),
+                IIC_fpMAC32, "vnmla", ".f32\t$dst, $a, $b",
              [(set SPR:$dst, (fsub (fneg (fmul SPR:$a, SPR:$b)), SPR:$dstin))]>,
                 RegConstraint<"$dstin = $dst"> {
   let Inst{6} = 1;
@@ -389,27 +389,27 @@ def FNMSCS : ASbI<0b11100001, (outs SPR:$dst), (ins SPR:$dstin, SPR:$a, SPR:$b),
 // FP Conditional moves.
 //
 
-def FCPYDcc  : ADuI<0b11101011, 0b0000, 0b0100,
+def VMOVDcc  : ADuI<0b11101011, 0b0000, 0b0100,
                     (outs DPR:$dst), (ins DPR:$false, DPR:$true),
-                    IIC_fpUNA64, "fcpyd", "\t$dst, $true",
+                    IIC_fpUNA64, "vmov", ".f64\t$dst, $true",
                 [/*(set DPR:$dst, (ARMcmov DPR:$false, DPR:$true, imm:$cc))*/]>,
                     RegConstraint<"$false = $dst">;
 
-def FCPYScc  : ASuI<0b11101011, 0b0000, 0b0100,
+def VMOVScc  : ASuI<0b11101011, 0b0000, 0b0100,
                     (outs SPR:$dst), (ins SPR:$false, SPR:$true),
-                    IIC_fpUNA32, "fcpys", "\t$dst, $true",
+                    IIC_fpUNA32, "vmov", ".f32\t$dst, $true",
                 [/*(set SPR:$dst, (ARMcmov SPR:$false, SPR:$true, imm:$cc))*/]>,
                     RegConstraint<"$false = $dst">;
 
-def FNEGDcc  : ADuI<0b11101011, 0b0001, 0b0100,
+def VNEGDcc  : ADuI<0b11101011, 0b0001, 0b0100,
                     (outs DPR:$dst), (ins DPR:$false, DPR:$true),
-                    IIC_fpUNA64, "fnegd", "\t$dst, $true",
+                    IIC_fpUNA64, "vneg", ".f64\t$dst, $true",
                 [/*(set DPR:$dst, (ARMcneg DPR:$false, DPR:$true, imm:$cc))*/]>,
                     RegConstraint<"$false = $dst">;
 
-def FNEGScc  : ASuI<0b11101011, 0b0001, 0b0100,
+def VNEGScc  : ASuI<0b11101011, 0b0001, 0b0100,
                     (outs SPR:$dst), (ins SPR:$false, SPR:$true),
-                    IIC_fpUNA32, "fnegs", "\t$dst, $true",
+                    IIC_fpUNA32, "vneg", ".f32\t$dst, $true",
                 [/*(set SPR:$dst, (ARMcneg SPR:$false, SPR:$true, imm:$cc))*/]>,
                     RegConstraint<"$false = $dst">;
 
@@ -419,7 +419,8 @@ def FNEGScc  : ASuI<0b11101011, 0b0001, 0b0100,
 //
 
 let Defs = [CPSR], Uses = [FPSCR] in
-def FMSTAT : VFPAI<(outs), (ins), VFPMiscFrm, IIC_fpSTAT, "fmstat", "",
+def FMSTAT : VFPAI<(outs), (ins), VFPMiscFrm, IIC_fpSTAT, "vmrs",
+                   "\tAPSR_nzcv, FPSCR",
              [(arm_fmstat)]> {
   let Inst{27-20} = 0b11101111;
   let Inst{19-16} = 0b0001;
@@ -431,26 +432,26 @@ def FMSTAT : VFPAI<(outs), (ins), VFPMiscFrm, IIC_fpSTAT, "fmstat", "",
 
 
 // Materialize FP immediates. VFP3 only.
-let isReMaterializable = 1 in 
-def FCONSTS : VFPAI<(outs SPR:$dst), (ins vfp_f32imm:$imm),
+let isReMaterializable = 1 in {
+def FCONSTD : VFPAI<(outs DPR:$dst), (ins vfp_f64imm:$imm),
                     VFPMiscFrm, IIC_VMOVImm,
-                    "fconsts", "\t$dst, $imm",
-                    [(set SPR:$dst, vfp_f32imm:$imm)]>, Requires<[HasVFP3]> {
+                    "fconstd", "\t$dst, $imm",
+                    [(set DPR:$dst, vfp_f64imm:$imm)]>, Requires<[HasVFP3]> {
   let Inst{27-23} = 0b11101;
   let Inst{21-20} = 0b11;
   let Inst{11-9}  = 0b101;
-  let Inst{8}     = 0;
+  let Inst{8}     = 1;
   let Inst{7-4}   = 0b0000;
 }
 
-let isReMaterializable = 1 in 
-def FCONSTD : VFPAI<(outs DPR:$dst), (ins vfp_f64imm:$imm),
+def FCONSTS : VFPAI<(outs SPR:$dst), (ins vfp_f32imm:$imm),
                     VFPMiscFrm, IIC_VMOVImm,
-                    "fconstd", "\t$dst, $imm",
-                    [(set DPR:$dst, vfp_f64imm:$imm)]>, Requires<[HasVFP3]> {
+                    "fconsts", "\t$dst, $imm",
+                    [(set SPR:$dst, vfp_f32imm:$imm)]>, Requires<[HasVFP3]> {
   let Inst{27-23} = 0b11101;
   let Inst{21-20} = 0b11;
   let Inst{11-9}  = 0b101;
-  let Inst{8}     = 1;
+  let Inst{8}     = 0;
   let Inst{7-4}   = 0b0000;
 }
+}
diff --git a/lib/Target/ARM/ARMLoadStoreOptimizer.cpp b/lib/Target/ARM/ARMLoadStoreOptimizer.cpp
index 7e1783bd44b..873a8d3cfee 100644
--- a/lib/Target/ARM/ARMLoadStoreOptimizer.cpp
+++ b/lib/Target/ARM/ARMLoadStoreOptimizer.cpp
@@ -41,8 +41,8 @@ using namespace llvm;
 
 STATISTIC(NumLDMGened , "Number of ldm instructions generated");
 STATISTIC(NumSTMGened , "Number of stm instructions generated");
-STATISTIC(NumFLDMGened, "Number of fldm instructions generated");
-STATISTIC(NumFSTMGened, "Number of fstm instructions generated");
+STATISTIC(NumVLDMGened, "Number of vldm instructions generated");
+STATISTIC(NumVSTMGened, "Number of vstm instructions generated");
 STATISTIC(NumLdStMoved, "Number of load / store instructions moved");
 STATISTIC(NumLDRDFormed,"Number of ldrd created before allocation");
 STATISTIC(NumSTRDFormed,"Number of strd created before allocation");
@@ -127,18 +127,18 @@ static int getLoadStoreMultipleOpcode(int Opcode) {
   case ARM::t2STRi12:
     NumSTMGened++;
     return ARM::t2STM;
-  case ARM::FLDS:
-    NumFLDMGened++;
-    return ARM::FLDMS;
-  case ARM::FSTS:
-    NumFSTMGened++;
-    return ARM::FSTMS;
-  case ARM::FLDD:
-    NumFLDMGened++;
-    return ARM::FLDMD;
-  case ARM::FSTD:
-    NumFSTMGened++;
-    return ARM::FSTMD;
+  case ARM::VLDRS:
+    NumVLDMGened++;
+    return ARM::VLDMS;
+  case ARM::VSTRS:
+    NumVSTMGened++;
+    return ARM::VSTMS;
+  case ARM::VLDRD:
+    NumVLDMGened++;
+    return ARM::VLDMD;
+  case ARM::VSTRD:
+    NumVSTMGened++;
+    return ARM::VSTMD;
   default: llvm_unreachable("Unhandled opcode!");
   }
   return 0;
@@ -229,8 +229,8 @@ ARMLoadStoreOpt::MergeOps(MachineBasicBlock &MBB,
     BaseKill = true;  // New base is always killed right its use.
   }
 
-  bool isDPR = Opcode == ARM::FLDD || Opcode == ARM::FSTD;
-  bool isDef = isi32Load(Opcode) || Opcode == ARM::FLDS || Opcode == ARM::FLDD;
+  bool isDPR = Opcode == ARM::VLDRD || Opcode == ARM::VSTRD;
+  bool isDef = isi32Load(Opcode) || Opcode == ARM::VLDRS || Opcode == ARM::VLDRD;
   Opcode = getLoadStoreMultipleOpcode(Opcode);
   MachineInstrBuilder MIB = (isAM4)
     ? BuildMI(MBB, MBBI, dl, TII->get(Opcode))
@@ -373,27 +373,27 @@ static inline unsigned getLSMultipleTransferSize(MachineInstr *MI) {
   case ARM::t2LDRi12:
   case ARM::t2STRi8:
   case ARM::t2STRi12:
-  case ARM::FLDS:
-  case ARM::FSTS:
+  case ARM::VLDRS:
+  case ARM::VSTRS:
     return 4;
-  case ARM::FLDD:
-  case ARM::FSTD:
+  case ARM::VLDRD:
+  case ARM::VSTRD:
     return 8;
   case ARM::LDM:
   case ARM::STM:
   case ARM::t2LDM:
   case ARM::t2STM:
     return (MI->getNumOperands() - 5) * 4;
-  case ARM::FLDMS:
-  case ARM::FSTMS:
-  case ARM::FLDMD:
-  case ARM::FSTMD:
+  case ARM::VLDMS:
+  case ARM::VSTMS:
+  case ARM::VLDMD:
+  case ARM::VSTMD:
     return ARM_AM::getAM5Offset(MI->getOperand(1).getImm()) * 4;
   }
 }
 
 /// MergeBaseUpdateLSMultiple - Fold proceeding/trailing inc/dec of base
-/// register into the LDM/STM/FLDM{D|S}/FSTM{D|S} op when possible:
+/// register into the LDM/STM/VLDM{D|S}/VSTM{D|S} op when possible:
 ///
 /// stmia rn, <ra, rb, rc>
 /// rn := rn + 4 * 3;
@@ -475,7 +475,7 @@ bool ARMLoadStoreOpt::MergeBaseUpdateLSMultiple(MachineBasicBlock &MBB,
       }
     }
   } else {
-    // FLDM{D|S}, FSTM{D|S} addressing mode 5 ops.
+    // VLDM{D|S}, VSTM{D|S} addressing mode 5 ops.
     if (ARM_AM::getAM5WBFlag(MI->getOperand(1).getImm()))
       return false;
 
@@ -517,10 +517,10 @@ static unsigned getPreIndexedLoadStoreOpcode(unsigned Opc) {
   switch (Opc) {
   case ARM::LDR: return ARM::LDR_PRE;
   case ARM::STR: return ARM::STR_PRE;
-  case ARM::FLDS: return ARM::FLDMS;
-  case ARM::FLDD: return ARM::FLDMD;
-  case ARM::FSTS: return ARM::FSTMS;
-  case ARM::FSTD: return ARM::FSTMD;
+  case ARM::VLDRS: return ARM::VLDMS;
+  case ARM::VLDRD: return ARM::VLDMD;
+  case ARM::VSTRS: return ARM::VSTMS;
+  case ARM::VSTRD: return ARM::VSTMD;
   case ARM::t2LDRi8:
   case ARM::t2LDRi12:
     return ARM::t2LDR_PRE;
@@ -536,10 +536,10 @@ static unsigned getPostIndexedLoadStoreOpcode(unsigned Opc) {
   switch (Opc) {
   case ARM::LDR: return ARM::LDR_POST;
   case ARM::STR: return ARM::STR_POST;
-  case ARM::FLDS: return ARM::FLDMS;
-  case ARM::FLDD: return ARM::FLDMD;
-  case ARM::FSTS: return ARM::FSTMS;
-  case ARM::FSTD: return ARM::FSTMD;
+  case ARM::VLDRS: return ARM::VLDMS;
+  case ARM::VLDRD: return ARM::VLDMD;
+  case ARM::VSTRS: return ARM::VSTMS;
+  case ARM::VSTRD: return ARM::VSTMD;
   case ARM::t2LDRi8:
   case ARM::t2LDRi12:
     return ARM::t2LDR_POST;
@@ -564,8 +564,8 @@ bool ARMLoadStoreOpt::MergeBaseUpdateLoadStore(MachineBasicBlock &MBB,
   unsigned Bytes = getLSMultipleTransferSize(MI);
   int Opcode = MI->getOpcode();
   DebugLoc dl = MI->getDebugLoc();
-  bool isAM5 = Opcode == ARM::FLDD || Opcode == ARM::FLDS ||
-    Opcode == ARM::FSTD || Opcode == ARM::FSTS;
+  bool isAM5 = Opcode == ARM::VLDRD || Opcode == ARM::VLDRS ||
+    Opcode == ARM::VSTRD || Opcode == ARM::VSTRS;
   bool isAM2 = Opcode == ARM::LDR || Opcode == ARM::STR;
   if (isAM2 && ARM_AM::getAM2Offset(MI->getOperand(3).getImm()) != 0)
     return false;
@@ -575,7 +575,7 @@ bool ARMLoadStoreOpt::MergeBaseUpdateLoadStore(MachineBasicBlock &MBB,
     if (MI->getOperand(2).getImm() != 0)
       return false;
 
-  bool isLd = isi32Load(Opcode) || Opcode == ARM::FLDS || Opcode == ARM::FLDD;
+  bool isLd = isi32Load(Opcode) || Opcode == ARM::VLDRS || Opcode == ARM::VLDRD;
   // Can't do the merge if the destination register is the same as the would-be
   // writeback register.
   if (isLd && MI->getOperand(0).getReg() == Base)
@@ -626,7 +626,7 @@ bool ARMLoadStoreOpt::MergeBaseUpdateLoadStore(MachineBasicBlock &MBB,
   if (!DoMerge)
     return false;
 
-  bool isDPR = NewOpc == ARM::FLDMD || NewOpc == ARM::FSTMD;
+  bool isDPR = NewOpc == ARM::VLDMD || NewOpc == ARM::VSTMD;
   unsigned Offset = 0;
   if (isAM5)
     Offset = ARM_AM::getAM5Opc((AddSub == ARM_AM::sub)
@@ -638,7 +638,7 @@ bool ARMLoadStoreOpt::MergeBaseUpdateLoadStore(MachineBasicBlock &MBB,
     Offset = AddSub == ARM_AM::sub ? -Bytes : Bytes;
   if (isLd) {
     if (isAM5)
-      // FLDMS, FLDMD
+      // VLDMS, VLDMD
       BuildMI(MBB, MBBI, dl, TII->get(NewOpc))
         .addReg(Base, getKillRegState(BaseKill))
         .addImm(Offset).addImm(Pred).addReg(PredReg)
@@ -657,7 +657,7 @@ bool ARMLoadStoreOpt::MergeBaseUpdateLoadStore(MachineBasicBlock &MBB,
   } else {
     MachineOperand &MO = MI->getOperand(0);
     if (isAM5)
-      // FSTMS, FSTMD
+      // VSTMS, VSTMD
       BuildMI(MBB, MBBI, dl, TII->get(NewOpc)).addReg(Base).addImm(Offset)
         .addImm(Pred).addReg(PredReg)
         .addReg(Base, getDefRegState(true)) // WB base register
@@ -687,11 +687,11 @@ static bool isMemoryOp(const MachineInstr *MI) {
   case ARM::LDR:
   case ARM::STR:
     return MI->getOperand(1).isReg() && MI->getOperand(2).getReg() == 0;
-  case ARM::FLDS:
-  case ARM::FSTS:
+  case ARM::VLDRS:
+  case ARM::VSTRS:
     return MI->getOperand(1).isReg();
-  case ARM::FLDD:
-  case ARM::FSTD:
+  case ARM::VLDRD:
+  case ARM::VSTRD:
     return MI->getOperand(1).isReg();
   case ARM::t2LDRi8:
   case ARM::t2LDRi12:
@@ -1214,7 +1214,7 @@ ARMPreAllocLoadStoreOpt::CanFormLdStDWord(MachineInstr *Op0, MachineInstr *Op1,
   if (!STI->hasV5TEOps())
     return false;
 
-  // FIXME: FLDS / FSTS -> FLDD / FSTD
+  // FIXME: VLDRS / VSTRS -> VLDRD / VSTRD
   unsigned Scale = 1;
   unsigned Opcode = Op0->getOpcode();
   if (Opcode == ARM::LDR)
@@ -1456,7 +1456,7 @@ ARMPreAllocLoadStoreOpt::RescheduleLoadStoreInstrs(MachineBasicBlock *MBB) {
         continue;
 
       int Opc = MI->getOpcode();
-      bool isLd = isi32Load(Opc) || Opc == ARM::FLDS || Opc == ARM::FLDD;
+      bool isLd = isi32Load(Opc) || Opc == ARM::VLDRS || Opc == ARM::VLDRD;
       unsigned Base = MI->getOperand(1).getReg();
       int Offset = getMemoryOpOffset(MI);
 
diff --git a/lib/Target/ARM/AsmPrinter/ARMAsmPrinter.cpp b/lib/Target/ARM/AsmPrinter/ARMAsmPrinter.cpp
index 020710bb45d..e9d89d75767 100644
--- a/lib/Target/ARM/AsmPrinter/ARMAsmPrinter.cpp
+++ b/lib/Target/ARM/AsmPrinter/ARMAsmPrinter.cpp
@@ -609,12 +609,7 @@ void ARMAsmPrinter::printAddrMode5Operand(const MachineInstr *MI, int Op,
 
   if (Modifier && strcmp(Modifier, "submode") == 0) {
     ARM_AM::AMSubMode Mode = ARM_AM::getAM5SubMode(MO2.getImm());
-    if (MO1.getReg() == ARM::SP) {
-      bool isFLDM = (MI->getOpcode() == ARM::FLDMD ||
-                     MI->getOpcode() == ARM::FLDMS);
-      O << ARM_AM::getAMSubModeAltStr(Mode, isFLDM);
-    } else
-      O << ARM_AM::getAMSubModeStr(Mode);
+    O << ARM_AM::getAMSubModeStr(Mode);
     return;
   } else if (Modifier && strcmp(Modifier, "base") == 0) {
     // Used for FSTM{D|S} and LSTM{D|S} operations.
@@ -1131,9 +1126,8 @@ void ARMAsmPrinter::EmitStartOfAsmFile(Module &M) {
     }
   }
 
-  // Use unified assembler syntax mode for Thumb.
-  if (Subtarget->isThumb())
-    O << "\t.syntax unified\n";
+  // Use unified assembler syntax.
+  O << "\t.syntax unified\n";
 
   // Emit ARM Build Attributes
   if (Subtarget->isTargetELF()) {
diff --git a/lib/Target/ARM/AsmPrinter/ARMInstPrinter.cpp b/lib/Target/ARM/AsmPrinter/ARMInstPrinter.cpp
index f422798e315..00479257217 100644
--- a/lib/Target/ARM/AsmPrinter/ARMInstPrinter.cpp
+++ b/lib/Target/ARM/AsmPrinter/ARMInstPrinter.cpp
@@ -259,12 +259,7 @@ void ARMInstPrinter::printAddrMode5Operand(const MCInst *MI, unsigned OpNum,
   
   if (Modifier && strcmp(Modifier, "submode") == 0) {
     ARM_AM::AMSubMode Mode = ARM_AM::getAM5SubMode(MO2.getImm());
-    if (MO1.getReg() == ARM::SP) {
-      bool isFLDM = (MI->getOpcode() == ARM::FLDMD ||
-                     MI->getOpcode() == ARM::FLDMS);
-      O << ARM_AM::getAMSubModeAltStr(Mode, isFLDM);
-    } else
-      O << ARM_AM::getAMSubModeStr(Mode);
+    O << ARM_AM::getAMSubModeStr(Mode);
     return;
   } else if (Modifier && strcmp(Modifier, "base") == 0) {
     // Used for FSTM{D|S} and LSTM{D|S} operations.
diff --git a/lib/Target/ARM/NEONMoveFix.cpp b/lib/Target/ARM/NEONMoveFix.cpp
index f307e3b3108..7d767ec5261 100644
--- a/lib/Target/ARM/NEONMoveFix.cpp
+++ b/lib/Target/ARM/NEONMoveFix.cpp
@@ -54,10 +54,10 @@ bool NEONMoveFixPass::InsertMoves(MachineBasicBlock &MBB) {
     NextMII = next(MII);
     MachineInstr *MI = &*MII;
 
-    if (MI->getOpcode() == ARM::FCPYD &&
+    if (MI->getOpcode() == ARM::VMOVD &&
         !TII->isPredicated(MI)) {
       unsigned SrcReg = MI->getOperand(1).getReg();
-      // If we do not found an instruction defining the reg, this means the
+      // If we do not find an instruction defining the reg, this means the
       // register should be live-in for this BB. It's always to better to use
       // NEON reg-reg moves.
       unsigned Domain = ARMII::DomainNEON;
@@ -71,7 +71,7 @@ bool NEONMoveFixPass::InsertMoves(MachineBasicBlock &MBB) {
       }
 
       if (Domain & ARMII::DomainNEON) {
-        // Convert FCPYD to VMOVD.
+        // Convert VMOVD to VMOVDneon
         unsigned DestReg = MI->getOperand(0).getReg();
 
         DEBUG({errs() << "vmov convert: "; MI->dump();});
@@ -82,7 +82,7 @@ bool NEONMoveFixPass::InsertMoves(MachineBasicBlock &MBB) {
         //  - The imp-defs / imp-uses are superregs only, we don't care about
         //    them.
         BuildMI(MBB, *MI, MI->getDebugLoc(),
-                TII->get(ARM::VMOVD), DestReg).addReg(SrcReg);
+                TII->get(ARM::VMOVDneon), DestReg).addReg(SrcReg);
         MBB.erase(MI);
         MachineBasicBlock::iterator I = prior(NextMII);
         MI = &*I;
diff --git a/lib/Target/ARM/README-Thumb.txt b/lib/Target/ARM/README-Thumb.txt
index e7770b2292e..6b605bb0a7c 100644
--- a/lib/Target/ARM/README-Thumb.txt
+++ b/lib/Target/ARM/README-Thumb.txt
@@ -37,7 +37,7 @@ LPCRELL0:
 	mov r1, #PCRELV0
 	add r1, pc
 	ldr r0, [r0, r1]
-	cpy pc, r0 
+	mov pc, r0 
 	.align	2
 LJTI1_0_0:
 	.long	 LBB1_3
@@ -51,7 +51,7 @@ We should be able to generate:
 LPCRELL0:
 	add r1, LJTI1_0_0
 	ldr r0, [r0, r1]
-	cpy pc, r0 
+	mov pc, r0 
 	.align	2
 LJTI1_0_0:
 	.long	 LBB1_3
@@ -206,8 +206,8 @@ LPC0:
 	add r5, pc
 	ldr r6, LCPI1_1
 	ldr r2, LCPI1_2
-	cpy r3, r6
-	cpy lr, pc
+	mov r3, r6
+	mov lr, pc
 	bx r5
 
 //===---------------------------------------------------------------------===//
diff --git a/lib/Target/ARM/README.txt b/lib/Target/ARM/README.txt
index e4c7e6b5a6d..11c48add21b 100644
--- a/lib/Target/ARM/README.txt
+++ b/lib/Target/ARM/README.txt
@@ -321,7 +321,7 @@ time.
 4) Once we added support for multiple result patterns, write indexed loads
    patterns instead of C++ instruction selection code.
 
-5) Use FLDM / FSTM to emulate indexed FP load / store.
+5) Use VLDM / VSTM to emulate indexed FP load / store.
 
 //===---------------------------------------------------------------------===//
 
diff --git a/lib/Target/ARM/Thumb1RegisterInfo.cpp b/lib/Target/ARM/Thumb1RegisterInfo.cpp
index 5ef7801ba8b..37adf37af4c 100644
--- a/lib/Target/ARM/Thumb1RegisterInfo.cpp
+++ b/lib/Target/ARM/Thumb1RegisterInfo.cpp
@@ -795,7 +795,7 @@ void Thumb1RegisterInfo::emitEpilogue(MachineFunction &MF,
     if (NumBytes != 0)
       emitSPUpdate(MBB, MBBI, TII, dl, *this, NumBytes);
   } else {
-    // Unwind MBBI to point to first LDR / FLDD.
+    // Unwind MBBI to point to first LDR / VLDRD.
     const unsigned *CSRegs = getCalleeSavedRegs();
     if (MBBI != MBB.begin()) {
       do
diff --git a/test/CodeGen/ARM/2009-07-18-RewriterBug.ll b/test/CodeGen/ARM/2009-07-18-RewriterBug.ll
index ee93fde998c..2b7ccd86152 100644
--- a/test/CodeGen/ARM/2009-07-18-RewriterBug.ll
+++ b/test/CodeGen/ARM/2009-07-18-RewriterBug.ll
@@ -1,4 +1,4 @@
-; RUN: llc < %s -mtriple=armv6-apple-darwin10 -mattr=+vfp2 | grep fcmpezd | count 13
+; RUN: llc < %s -mtriple=armv6-apple-darwin10 -mattr=+vfp2 | grep vcmpe | count 13
 
 	%struct.EDGE_PAIR = type { %struct.edge_rec*, %struct.edge_rec* }
 	%struct.VEC2 = type { double, double, double }
diff --git a/test/CodeGen/ARM/2009-09-09-fpcmp-ole.ll b/test/CodeGen/ARM/2009-09-09-fpcmp-ole.ll
index 98cab9a9149..3909c6a526e 100644
--- a/test/CodeGen/ARM/2009-09-09-fpcmp-ole.ll
+++ b/test/CodeGen/ARM/2009-09-09-fpcmp-ole.ll
@@ -9,7 +9,7 @@ define void @test(double* %x, double* %y) nounwind {
   br i1 %4, label %bb1, label %bb2
 
 bb1:
-;CHECK: fstdhi
+;CHECK: vstrhi.64
   store double %1, double* %y, align 4
   br label %bb2
 
diff --git a/test/CodeGen/ARM/2009-09-24-spill-align.ll b/test/CodeGen/ARM/2009-09-24-spill-align.ll
index 6281775d061..5476d5f7961 100644
--- a/test/CodeGen/ARM/2009-09-24-spill-align.ll
+++ b/test/CodeGen/ARM/2009-09-24-spill-align.ll
@@ -6,7 +6,7 @@ entry:
   %arg0_poly16x4_t = alloca <4 x i16>             ; <<4 x i16>*> [#uses=1]
   %out_poly16_t = alloca i16                      ; <i16*> [#uses=1]
   %"alloca point" = bitcast i32 0 to i32          ; <i32> [#uses=0]
-; CHECK: fldd
+; CHECK: vldr.64
   %0 = load <4 x i16>* %arg0_poly16x4_t, align 8  ; <<4 x i16>> [#uses=1]
   %1 = extractelement <4 x i16> %0, i32 1         ; <i16> [#uses=1]
   store i16 %1, i16* %out_poly16_t, align 2
diff --git a/test/CodeGen/ARM/2009-11-01-NeonMoves.ll b/test/CodeGen/ARM/2009-11-01-NeonMoves.ll
index c260b973b5a..62f3786e206 100644
--- a/test/CodeGen/ARM/2009-11-01-NeonMoves.ll
+++ b/test/CodeGen/ARM/2009-11-01-NeonMoves.ll
@@ -1,4 +1,4 @@
-; RUN: llc -mcpu=cortex-a8 < %s | grep vmov | count 1
+; RUN: llc -mcpu=cortex-a8 < %s | FileCheck %s
 
 target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64"
 target triple = "armv7-eabi"
@@ -11,12 +11,15 @@ entry:
   %0 = getelementptr inbounds %foo* %quat_addr, i32 0, i32 0 ; <<4 x float>*> [#uses=1]
   store <4 x float> %quat.0, <4 x float>* %0
   %1 = call arm_aapcs_vfpcc  <4 x float> @quux(%foo* %quat_addr) nounwind ; <<4 x float>> [#uses=3]
+;CHECK: vmov.f32
+;CHECK: vmov.f32
   %2 = fmul <4 x float> %1, %1                    ; <<4 x float>> [#uses=2]
   %3 = shufflevector <4 x float> %2, <4 x float> undef, <2 x i32> <i32 0, i32 1> ; <<2 x float>> [#uses=1]
   %4 = shufflevector <4 x float> %2, <4 x float> undef, <2 x i32> <i32 2, i32 3> ; <<2 x float>> [#uses=1]
   %5 = call <2 x float> @llvm.arm.neon.vpadd.v2f32(<2 x float> %3, <2 x float> %4) nounwind ; <<2 x float>> [#uses=2]
   %6 = call <2 x float> @llvm.arm.neon.vpadd.v2f32(<2 x float> %5, <2 x float> %5) nounwind ; <<2 x float>> [#uses=2]
   %7 = shufflevector <2 x float> %6, <2 x float> %6, <4 x i32> <i32 0, i32 1, i32 2, i32 3> ; <<4 x float>> [#uses=2]
+;CHECK: vmov
   %8 = call <4 x float> @llvm.arm.neon.vrsqrte.v4f32(<4 x float> %7) nounwind ; <<4 x float>> [#uses=3]
   %9 = fmul <4 x float> %8, %8                    ; <<4 x float>> [#uses=1]
   %10 = call <4 x float> @llvm.arm.neon.vrsqrts.v4f32(<4 x float> %9, <4 x float> %7) nounwind ; <<4 x float>> [#uses=1]
diff --git a/test/CodeGen/ARM/arguments_f64_backfill.ll b/test/CodeGen/ARM/arguments_f64_backfill.ll
index d8019a07fab..062133e8645 100644
--- a/test/CodeGen/ARM/arguments_f64_backfill.ll
+++ b/test/CodeGen/ARM/arguments_f64_backfill.ll
@@ -1,7 +1,7 @@
 ; RUN: llc < %s -mtriple=arm-linux-gnueabi -mattr=+vfp2 -float-abi=hard | FileCheck %s
 
 define float @f(float %z, double %a, float %b) {
-; CHECK: fcpys s0, s1
+; CHECK: vmov.f32 s0, s1
         %tmp = call float @g(float %b)
         ret float %tmp
 }
diff --git a/test/CodeGen/ARM/compare-call.ll b/test/CodeGen/ARM/compare-call.ll
index 5f3ed1d2743..fac2bc5e432 100644
--- a/test/CodeGen/ARM/compare-call.ll
+++ b/test/CodeGen/ARM/compare-call.ll
@@ -1,5 +1,5 @@
 ; RUN: llc < %s -march=arm -mattr=+v6,+vfp2 | \
-; RUN:   grep fcmpes
+; RUN:   grep vcmpe.f32
 
 define void @test3(float* %glob, i32 %X) {
 entry:
diff --git a/test/CodeGen/ARM/fabss.ll b/test/CodeGen/ARM/fabss.ll
index 5690a01d750..46f136ba1fe 100644
--- a/test/CodeGen/ARM/fabss.ll
+++ b/test/CodeGen/ARM/fabss.ll
@@ -1,8 +1,8 @@
-; RUN: llc < %s -march=arm -mattr=+vfp2 | grep -E {fabss\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
+; RUN: llc < %s -march=arm -mattr=+vfp2 | grep -E {vabs.f32\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
 ; RUN: llc < %s -march=arm -mattr=+neon -arm-use-neon-fp=1 | grep -E {vabs.f32\\W*d\[0-9\]+,\\W*d\[0-9\]+} | count 1
-; RUN: llc < %s -march=arm -mattr=+neon -arm-use-neon-fp=0 | grep -E {fabss\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
+; RUN: llc < %s -march=arm -mattr=+neon -arm-use-neon-fp=0 | grep -E {vabs.f32\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
 ; RUN: llc < %s -march=arm -mcpu=cortex-a8 | grep -E {vabs.f32\\W*d\[0-9\]+,\\W*d\[0-9\]+} | count 1
-; RUN: llc < %s -march=arm -mcpu=cortex-a9 | grep -E {fabss\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
+; RUN: llc < %s -march=arm -mcpu=cortex-a9 | grep -E {vabs.f32\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
 
 define float @test(float %a, float %b) {
 entry:
diff --git a/test/CodeGen/ARM/fadds.ll b/test/CodeGen/ARM/fadds.ll
index a01f868d18b..1426a2dc883 100644
--- a/test/CodeGen/ARM/fadds.ll
+++ b/test/CodeGen/ARM/fadds.ll
@@ -1,8 +1,8 @@
-; RUN: llc < %s -march=arm -mattr=+vfp2 | grep -E {fadds\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
+; RUN: llc < %s -march=arm -mattr=+vfp2 | grep -E {vadd.f32\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
 ; RUN: llc < %s -march=arm -mattr=+neon -arm-use-neon-fp=1 | grep -E {vadd.f32\\W*d\[0-9\]+,\\W*d\[0-9\]+,\\W*d\[0-9\]+} | count 1
-; RUN: llc < %s -march=arm -mattr=+neon -arm-use-neon-fp=0 | grep -E {fadds\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
+; RUN: llc < %s -march=arm -mattr=+neon -arm-use-neon-fp=0 | grep -E {vadd.f32\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
 ; RUN: llc < %s -march=arm -mcpu=cortex-a8 | grep -E {vadd.f32\\W*d\[0-9\]+,\\W*d\[0-9\]+,\\W*d\[0-9\]+} | count 1
-; RUN: llc < %s -march=arm -mcpu=cortex-a9 | grep -E {fadds\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
+; RUN: llc < %s -march=arm -mcpu=cortex-a9 | grep -E {vadd.f32\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
 
 define float @test(float %a, float %b) {
 entry:
diff --git a/test/CodeGen/ARM/fcopysign.ll b/test/CodeGen/ARM/fcopysign.ll
index bf7c305c895..a6d741087a8 100644
--- a/test/CodeGen/ARM/fcopysign.ll
+++ b/test/CodeGen/ARM/fcopysign.ll
@@ -1,6 +1,6 @@
 ; RUN: llc < %s -march=arm | grep bic | count 2
 ; RUN: llc < %s -march=arm -mattr=+v6,+vfp2 | \
-; RUN:   grep fneg | count 2
+; RUN:   grep vneg | count 2
 
 define float @test1(float %x, double %y) {
 	%tmp = fpext float %x to double
diff --git a/test/CodeGen/ARM/fdivs.ll b/test/CodeGen/ARM/fdivs.ll
index 2af250d121d..45803f6d3c3 100644
--- a/test/CodeGen/ARM/fdivs.ll
+++ b/test/CodeGen/ARM/fdivs.ll
@@ -1,8 +1,8 @@
-; RUN: llc < %s -march=arm -mattr=+vfp2 | grep -E {fdivs\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
-; RUN: llc < %s -march=arm -mattr=+neon -arm-use-neon-fp=1 | grep -E {fdivs\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
-; RUN: llc < %s -march=arm -mattr=+neon -arm-use-neon-fp=0 | grep -E {fdivs\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
-; RUN: llc < %s -march=arm -mcpu=cortex-a8 | grep -E {fdivs\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
-; RUN: llc < %s -march=arm -mcpu=cortex-a9 | grep -E {fdivs\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
+; RUN: llc < %s -march=arm -mattr=+vfp2 | grep -E {vdiv.f32\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
+; RUN: llc < %s -march=arm -mattr=+neon -arm-use-neon-fp=1 | grep -E {vdiv.f32\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
+; RUN: llc < %s -march=arm -mattr=+neon -arm-use-neon-fp=0 | grep -E {vdiv.f32\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
+; RUN: llc < %s -march=arm -mcpu=cortex-a8 | grep -E {vdiv.f32\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
+; RUN: llc < %s -march=arm -mcpu=cortex-a9 | grep -E {vdiv.f32\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
 
 define float @test(float %a, float %b) {
 entry:
diff --git a/test/CodeGen/ARM/fixunsdfdi.ll b/test/CodeGen/ARM/fixunsdfdi.ll
index ebf1d84536e..6db2385a63e 100644
--- a/test/CodeGen/ARM/fixunsdfdi.ll
+++ b/test/CodeGen/ARM/fixunsdfdi.ll
@@ -1,5 +1,5 @@
 ; RUN: llc < %s -march=arm -mattr=+vfp2
-; RUN: llc < %s -march=arm -mattr=vfp2 | not grep fstd
+; RUN: llc < %s -march=arm -mattr=vfp2 | not grep vstr.64
 
 define hidden i64 @__fixunsdfdi(double %x) nounwind readnone {
 entry:
diff --git a/test/CodeGen/ARM/fmacs.ll b/test/CodeGen/ARM/fmacs.ll
index 5c31ea641de..57efa826404 100644
--- a/test/CodeGen/ARM/fmacs.ll
+++ b/test/CodeGen/ARM/fmacs.ll
@@ -1,8 +1,8 @@
-; RUN: llc < %s -march=arm -mattr=+vfp2 | grep -E {fmacs\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
+; RUN: llc < %s -march=arm -mattr=+vfp2 | grep -E {vmla.f32\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
 ; RUN: llc < %s -march=arm -mattr=+neon -arm-use-neon-fp=1 | grep -E {vmul.f32\\W*d\[0-9\]+,\\W*d\[0-9\]+,\\W*d\[0-9\]+} | count 1
-; RUN: llc < %s -march=arm -mattr=+neon -arm-use-neon-fp=0 | grep -E {fmacs\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
+; RUN: llc < %s -march=arm -mattr=+neon -arm-use-neon-fp=0 | grep -E {vmla.f32\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
 ; RUN: llc < %s -march=arm -mcpu=cortex-a8 | grep -E {vmul.f32\\W*d\[0-9\]+,\\W*d\[0-9\]+,\\W*d\[0-9\]+} | count 1
-; RUN: llc < %s -march=arm -mcpu=cortex-a9 | grep -E {fmacs\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
+; RUN: llc < %s -march=arm -mcpu=cortex-a9 | grep -E {vmla.f32\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
 
 define float @test(float %acc, float %a, float %b) {
 entry:
diff --git a/test/CodeGen/ARM/fmscs.ll b/test/CodeGen/ARM/fmscs.ll
index c6e6d406040..31b5c52d38d 100644
--- a/test/CodeGen/ARM/fmscs.ll
+++ b/test/CodeGen/ARM/fmscs.ll
@@ -1,8 +1,8 @@
-; RUN: llc < %s -march=arm -mattr=+vfp2 | grep -E {fmscs\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
-; RUN: llc < %s -march=arm -mattr=+neon -arm-use-neon-fp=1 | grep -E {fmscs\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
-; RUN: llc < %s -march=arm -mattr=+neon -arm-use-neon-fp=0 | grep -E {fmscs\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
-; RUN: llc < %s -march=arm -mcpu=cortex-a8 | grep -E {fmscs\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
-; RUN: llc < %s -march=arm -mcpu=cortex-a9 | grep -E {fmscs\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
+; RUN: llc < %s -march=arm -mattr=+vfp2 | grep -E {vnmls.f32\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
+; RUN: llc < %s -march=arm -mattr=+neon -arm-use-neon-fp=1 | grep -E {vnmls.f32\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
+; RUN: llc < %s -march=arm -mattr=+neon -arm-use-neon-fp=0 | grep -E {vnmls.f32\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
+; RUN: llc < %s -march=arm -mcpu=cortex-a8 | grep -E {vnmls.f32\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
+; RUN: llc < %s -march=arm -mcpu=cortex-a9 | grep -E {vnmls.f32\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
 
 define float @test(float %acc, float %a, float %b) {
 entry:
diff --git a/test/CodeGen/ARM/fmuls.ll b/test/CodeGen/ARM/fmuls.ll
index cb5dadeb210..735263c9a31 100644
--- a/test/CodeGen/ARM/fmuls.ll
+++ b/test/CodeGen/ARM/fmuls.ll
@@ -1,8 +1,8 @@
-; RUN: llc < %s -march=arm -mattr=+vfp2 | grep -E {fmuls\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
+; RUN: llc < %s -march=arm -mattr=+vfp2 | grep -E {vmul.f32\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
 ; RUN: llc < %s -march=arm -mattr=+neon -arm-use-neon-fp=1 | grep -E {vmul.f32\\W*d\[0-9\]+,\\W*d\[0-9\]+,\\W*d\[0-9\]+} | count 1
-; RUN: llc < %s -march=arm -mattr=+neon -arm-use-neon-fp=0 | grep -E {fmuls\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
+; RUN: llc < %s -march=arm -mattr=+neon -arm-use-neon-fp=0 | grep -E {vmul.f32\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
 ; RUN: llc < %s -march=arm -mcpu=cortex-a8 | grep -E {vmul.f32\\W*d\[0-9\]+,\\W*d\[0-9\]+,\\W*d\[0-9\]+} | count 1
-; RUN: llc < %s -march=arm -mcpu=cortex-a9 | grep -E {fmuls\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
+; RUN: llc < %s -march=arm -mcpu=cortex-a9 | grep -E {vmul.f32\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
 
 define float @test(float %a, float %b) {
 entry:
diff --git a/test/CodeGen/ARM/fnegs.ll b/test/CodeGen/ARM/fnegs.ll
index 7da443dd93f..bc3d42de75c 100644
--- a/test/CodeGen/ARM/fnegs.ll
+++ b/test/CodeGen/ARM/fnegs.ll
@@ -1,8 +1,8 @@
-; RUN: llc < %s -march=arm -mattr=+vfp2 | grep -E {fnegs\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 2
+; RUN: llc < %s -march=arm -mattr=+vfp2 | grep -E {vneg.f32\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 2
 ; RUN: llc < %s -march=arm -mattr=+neon -arm-use-neon-fp=1 | grep -E {vneg.f32\\W*d\[0-9\]+,\\W*d\[0-9\]+} | count 2
-; RUN: llc < %s -march=arm -mattr=+neon -arm-use-neon-fp=0 | grep -E {fnegs\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 2
+; RUN: llc < %s -march=arm -mattr=+neon -arm-use-neon-fp=0 | grep -E {vneg.f32\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 2
 ; RUN: llc < %s -march=arm -mcpu=cortex-a8 | grep -E {vneg.f32\\W*d\[0-9\]+,\\W*d\[0-9\]+} | count 2
-; RUN: llc < %s -march=arm -mcpu=cortex-a9 | grep -E {fnegs\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 2
+; RUN: llc < %s -march=arm -mcpu=cortex-a9 | grep -E {vneg.f32\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 2
 
 define float @test1(float* %a) {
 entry:
diff --git a/test/CodeGen/ARM/fnmacs.ll b/test/CodeGen/ARM/fnmacs.ll
index 8fc13e78bc3..724947ea04d 100644
--- a/test/CodeGen/ARM/fnmacs.ll
+++ b/test/CodeGen/ARM/fnmacs.ll
@@ -4,14 +4,14 @@
 
 define float @test(float %acc, float %a, float %b) {
 entry:
-; VFP2: fnmacs
-; NEON: fnmacs
+; VFP2: vmls.f32
+; NEON: vmls.f32
 
 ; NEONFP-NOT: vmls
-; NEONFP-NOT: fcpys
+; NEONFP-NOT: vmov.f32
 ; NEONFP:     vmul.f32
 ; NEONFP:     vsub.f32
-; NEONFP:     fmrs
+; NEONFP:     vmov
 
 	%0 = fmul float %a, %b
         %1 = fsub float %acc, %0
diff --git a/test/CodeGen/ARM/fnmscs.ll b/test/CodeGen/ARM/fnmscs.ll
index 3ae437d69db..ad2188218e4 100644
--- a/test/CodeGen/ARM/fnmscs.ll
+++ b/test/CodeGen/ARM/fnmscs.ll
@@ -5,7 +5,7 @@
 ; RUN: llc < %s -march=arm -mcpu=cortex-a9 | FileCheck %s
 
 define float @test1(float %acc, float %a, float %b) nounwind {
-; CHECK: fnmscs s2, s1, s0 
+; CHECK: vnmla.f32 s2, s1, s0
 entry:
 	%0 = fmul float %a, %b
 	%1 = fsub float -0.0, %0
@@ -14,7 +14,7 @@ entry:
 }
 
 define float @test2(float %acc, float %a, float %b) nounwind {
-; CHECK: fnmscs s2, s1, s0 
+; CHECK: vnmla.f32 s2, s1, s0
 entry:
 	%0 = fmul float %a, %b
 	%1 = fmul float -1.0, %0
diff --git a/test/CodeGen/ARM/fnmul.ll b/test/CodeGen/ARM/fnmul.ll
index 613b347cdbf..6d7bc05ffa9 100644
--- a/test/CodeGen/ARM/fnmul.ll
+++ b/test/CodeGen/ARM/fnmul.ll
@@ -1,5 +1,5 @@
-; RUN: llc < %s -march=arm -mattr=+v6,+vfp2 | grep fnmuld
-; RUN: llc < %s -march=arm -mattr=+v6,+vfp2 -enable-sign-dependent-rounding-fp-math | grep fmul
+; RUN: llc < %s -march=arm -mattr=+v6,+vfp2 | grep vnmul.f64
+; RUN: llc < %s -march=arm -mattr=+v6,+vfp2 -enable-sign-dependent-rounding-fp-math | grep vmul.f64
 
 
 define double @t1(double %a, double %b) {
diff --git a/test/CodeGen/ARM/fp.ll b/test/CodeGen/ARM/fp.ll
index 4e4ef722f97..8fbd45b9757 100644
--- a/test/CodeGen/ARM/fp.ll
+++ b/test/CodeGen/ARM/fp.ll
@@ -2,9 +2,9 @@
 
 define float @f(i32 %a) {
 ;CHECK: f:
-;CHECK: fmsr
-;CHECK-NEXT: fsitos
-;CHECK-NEXT: fmrs
+;CHECK: vmov
+;CHECK-NEXT: vcvt.f32.s32
+;CHECK-NEXT: vmov
 entry:
         %tmp = sitofp i32 %a to float           ; <float> [#uses=1]
         ret float %tmp
@@ -12,9 +12,9 @@ entry:
 
 define double @g(i32 %a) {
 ;CHECK: g:
-;CHECK: fmsr
-;CHECK-NEXT: fsitod
-;CHECK-NEXT: fmrrd
+;CHECK: vmov
+;CHECK-NEXT: vcvt.f64.s32
+;CHECK-NEXT: vmov
 entry:
         %tmp = sitofp i32 %a to double          ; <double> [#uses=1]
         ret double %tmp
@@ -22,9 +22,9 @@ entry:
 
 define double @uint_to_double(i32 %a) {
 ;CHECK: uint_to_double:
-;CHECK: fmsr
-;CHECK-NEXT: fuitod
-;CHECK-NEXT: fmrrd
+;CHECK: vmov
+;CHECK-NEXT: vcvt.f64.u32
+;CHECK-NEXT: vmov
 entry:
         %tmp = uitofp i32 %a to double          ; <double> [#uses=1]
         ret double %tmp
@@ -32,9 +32,9 @@ entry:
 
 define float @uint_to_float(i32 %a) {
 ;CHECK: uint_to_float:
-;CHECK: fmsr
-;CHECK-NEXT: fuitos
-;CHECK-NEXT: fmrs
+;CHECK: vmov
+;CHECK-NEXT: vcvt.f32.u32
+;CHECK-NEXT: vmov
 entry:
         %tmp = uitofp i32 %a to float           ; <float> [#uses=1]
         ret float %tmp
@@ -42,8 +42,8 @@ entry:
 
 define double @h(double* %v) {
 ;CHECK: h:
-;CHECK: fldd
-;CHECK-NEXT: fmrrd
+;CHECK: vldr.64 
+;CHECK-NEXT: vmov
 entry:
         %tmp = load double* %v          ; <double> [#uses=1]
         ret double %tmp
@@ -58,13 +58,13 @@ entry:
 
 define double @f2(double %a) {
 ;CHECK: f2:
-;CHECK-NOT: fmdrr
+;CHECK-NOT: vmov
         ret double %a
 }
 
 define void @f3() {
 ;CHECK: f3:
-;CHECK-NOT: fmdrr
+;CHECK-NOT: vmov
 ;CHECK: f4
 entry:
         %tmp = call double @f5( )               ; <double> [#uses=1]
diff --git a/test/CodeGen/ARM/fp_convert.ll b/test/CodeGen/ARM/fp_convert.ll
index 9ce2ac549b5..2adac78cf80 100644
--- a/test/CodeGen/ARM/fp_convert.ll
+++ b/test/CodeGen/ARM/fp_convert.ll
@@ -6,7 +6,7 @@
 
 define i32 @test1(float %a, float %b) {
 ; VFP2: test1:
-; VFP2: ftosizs s0, s0
+; VFP2: vcvt.s32.f32 s0, s0
 ; NEON: test1:
 ; NEON: vcvt.s32.f32 d0, d0
 entry:
@@ -17,7 +17,7 @@ entry:
 
 define i32 @test2(float %a, float %b) {
 ; VFP2: test2:
-; VFP2: ftouizs s0, s0
+; VFP2: vcvt.u32.f32 s0, s0
 ; NEON: test2:
 ; NEON: vcvt.u32.f32 d0, d0
 entry:
@@ -28,7 +28,7 @@ entry:
 
 define float @test3(i32 %a, i32 %b) {
 ; VFP2: test3:
-; VFP2: fuitos s0, s0
+; VFP2: vcvt.f32.u32 s0, s0
 ; NEON: test3:
 ; NEON: vcvt.f32.u32 d0, d0
 entry:
@@ -39,7 +39,7 @@ entry:
 
 define float @test4(i32 %a, i32 %b) {
 ; VFP2: test4:
-; VFP2: fsitos s0, s0
+; VFP2: vcvt.f32.s32 s0, s0
 ; NEON: test4:
 ; NEON: vcvt.f32.s32 d0, d0
 entry:
diff --git a/test/CodeGen/ARM/fparith.ll b/test/CodeGen/ARM/fparith.ll
index ebeeb184121..ce6d6b29e9d 100644
--- a/test/CodeGen/ARM/fparith.ll
+++ b/test/CodeGen/ARM/fparith.ll
@@ -2,7 +2,7 @@
 
 define float @f1(float %a, float %b) {
 ;CHECK: f1:
-;CHECK: fadds
+;CHECK: vadd.f32
 entry:
 	%tmp = fadd float %a, %b		; <float> [#uses=1]
 	ret float %tmp
@@ -10,7 +10,7 @@ entry:
 
 define double @f2(double %a, double %b) {
 ;CHECK: f2:
-;CHECK: faddd
+;CHECK: vadd.f64
 entry:
 	%tmp = fadd double %a, %b		; <double> [#uses=1]
 	ret double %tmp
@@ -18,7 +18,7 @@ entry:
 
 define float @f3(float %a, float %b) {
 ;CHECK: f3:
-;CHECK: fmuls
+;CHECK: vmul.f32
 entry:
 	%tmp = fmul float %a, %b		; <float> [#uses=1]
 	ret float %tmp
@@ -26,7 +26,7 @@ entry:
 
 define double @f4(double %a, double %b) {
 ;CHECK: f4:
-;CHECK: fmuld
+;CHECK: vmul.f64
 entry:
 	%tmp = fmul double %a, %b		; <double> [#uses=1]
 	ret double %tmp
@@ -34,7 +34,7 @@ entry:
 
 define float @f5(float %a, float %b) {
 ;CHECK: f5:
-;CHECK: fsubs
+;CHECK: vsub.f32
 entry:
 	%tmp = fsub float %a, %b		; <float> [#uses=1]
 	ret float %tmp
@@ -42,7 +42,7 @@ entry:
 
 define double @f6(double %a, double %b) {
 ;CHECK: f6:
-;CHECK: fsubd
+;CHECK: vsub.f64
 entry:
 	%tmp = fsub double %a, %b		; <double> [#uses=1]
 	ret double %tmp
@@ -58,7 +58,7 @@ entry:
 
 define double @f8(double %a) {
 ;CHECK: f8:
-;CHECK: fnegd
+;CHECK: vneg.f64
 entry:
 	%tmp1 = fsub double -0.000000e+00, %a		; <double> [#uses=1]
 	ret double %tmp1
@@ -66,7 +66,7 @@ entry:
 
 define float @f9(float %a, float %b) {
 ;CHECK: f9:
-;CHECK: fdivs
+;CHECK: vdiv.f32
 entry:
 	%tmp1 = fdiv float %a, %b		; <float> [#uses=1]
 	ret float %tmp1
@@ -74,7 +74,7 @@ entry:
 
 define double @f10(double %a, double %b) {
 ;CHECK: f10:
-;CHECK: fdivd
+;CHECK: vdiv.f64
 entry:
 	%tmp1 = fdiv double %a, %b		; <double> [#uses=1]
 	ret double %tmp1
@@ -92,7 +92,7 @@ declare float @fabsf(float)
 
 define double @f12(double %a) {
 ;CHECK: f12:
-;CHECK: fabsd
+;CHECK: vabs.f64
 entry:
 	%tmp1 = call double @fabs( double %a )		; <double> [#uses=1]
 	ret double %tmp1
diff --git a/test/CodeGen/ARM/fpcmp.ll b/test/CodeGen/ARM/fpcmp.ll
index 2c9591ca542..260ec49cd86 100644
--- a/test/CodeGen/ARM/fpcmp.ll
+++ b/test/CodeGen/ARM/fpcmp.ll
@@ -2,7 +2,7 @@
 
 define i32 @f1(float %a) {
 ;CHECK: f1:
-;CHECK: fcmpes
+;CHECK: vcmpe.f32
 ;CHECK: movmi
 entry:
         %tmp = fcmp olt float %a, 1.000000e+00          ; <i1> [#uses=1]
@@ -12,7 +12,7 @@ entry:
 
 define i32 @f2(float %a) {
 ;CHECK: f2:
-;CHECK: fcmpes
+;CHECK: vcmpe.f32
 ;CHECK: moveq
 entry:
         %tmp = fcmp oeq float %a, 1.000000e+00          ; <i1> [#uses=1]
@@ -22,7 +22,7 @@ entry:
 
 define i32 @f3(float %a) {
 ;CHECK: f3:
-;CHECK: fcmpes
+;CHECK: vcmpe.f32
 ;CHECK: movgt
 entry:
         %tmp = fcmp ogt float %a, 1.000000e+00          ; <i1> [#uses=1]
@@ -32,7 +32,7 @@ entry:
 
 define i32 @f4(float %a) {
 ;CHECK: f4:
-;CHECK: fcmpes
+;CHECK: vcmpe.f32
 ;CHECK: movge
 entry:
         %tmp = fcmp oge float %a, 1.000000e+00          ; <i1> [#uses=1]
@@ -42,7 +42,7 @@ entry:
 
 define i32 @f5(float %a) {
 ;CHECK: f5:
-;CHECK: fcmpes
+;CHECK: vcmpe.f32
 ;CHECK: movls
 entry:
         %tmp = fcmp ole float %a, 1.000000e+00          ; <i1> [#uses=1]
@@ -52,7 +52,7 @@ entry:
 
 define i32 @f6(float %a) {
 ;CHECK: f6:
-;CHECK: fcmpes
+;CHECK: vcmpe.f32
 ;CHECK: movne
 entry:
         %tmp = fcmp une float %a, 1.000000e+00          ; <i1> [#uses=1]
@@ -62,7 +62,7 @@ entry:
 
 define i32 @g1(double %a) {
 ;CHECK: g1:
-;CHECK: fcmped
+;CHECK: vcmpe.f64
 ;CHECK: movmi
 entry:
         %tmp = fcmp olt double %a, 1.000000e+00         ; <i1> [#uses=1]
diff --git a/test/CodeGen/ARM/fpconv.ll b/test/CodeGen/ARM/fpconv.ll
index ee3c338e3b3..bf197a46cb7 100644
--- a/test/CodeGen/ARM/fpconv.ll
+++ b/test/CodeGen/ARM/fpconv.ll
@@ -3,7 +3,7 @@
 
 define float @f1(double %x) {
 ;CHECK-VFP: f1:
-;CHECK-VFP: fcvtsd
+;CHECK-VFP: vcvt.f32.f64
 ;CHECK: f1:
 ;CHECK: truncdfsf2
 entry:
@@ -13,7 +13,7 @@ entry:
 
 define double @f2(float %x) {
 ;CHECK-VFP: f2:
-;CHECK-VFP: fcvtds
+;CHECK-VFP: vcvt.f64.f32
 ;CHECK: f2:
 ;CHECK: extendsfdf2
 entry:
@@ -23,7 +23,7 @@ entry:
 
 define i32 @f3(float %x) {
 ;CHECK-VFP: f3:
-;CHECK-VFP: ftosizs
+;CHECK-VFP: vcvt.s32.f32
 ;CHECK: f3:
 ;CHECK: fixsfsi
 entry:
@@ -33,7 +33,7 @@ entry:
 
 define i32 @f4(float %x) {
 ;CHECK-VFP: f4:
-;CHECK-VFP: ftouizs
+;CHECK-VFP: vcvt.u32.f32
 ;CHECK: f4:
 ;CHECK: fixunssfsi
 entry:
@@ -43,7 +43,7 @@ entry:
 
 define i32 @f5(double %x) {
 ;CHECK-VFP: f5:
-;CHECK-VFP: ftosizd
+;CHECK-VFP: vcvt.s32.f64
 ;CHECK: f5:
 ;CHECK: fixdfsi
 entry:
@@ -53,7 +53,7 @@ entry:
 
 define i32 @f6(double %x) {
 ;CHECK-VFP: f6:
-;CHECK-VFP: ftouizd
+;CHECK-VFP: vcvt.u32.f64
 ;CHECK: f6:
 ;CHECK: fixunsdfsi
 entry:
@@ -63,7 +63,7 @@ entry:
 
 define float @f7(i32 %a) {
 ;CHECK-VFP: f7:
-;CHECK-VFP: fsitos
+;CHECK-VFP: vcvt.f32.s32
 ;CHECK: f7:
 ;CHECK: floatsisf
 entry:
@@ -73,7 +73,7 @@ entry:
 
 define double @f8(i32 %a) {
 ;CHECK-VFP: f8:
-;CHECK-VFP: fsitod
+;CHECK-VFP: vcvt.f64.s32
 ;CHECK: f8:
 ;CHECK: floatsidf
 entry:
@@ -83,7 +83,7 @@ entry:
 
 define float @f9(i32 %a) {
 ;CHECK-VFP: f9:
-;CHECK-VFP: fuitos
+;CHECK-VFP: vcvt.f32.u32
 ;CHECK: f9:
 ;CHECK: floatunsisf
 entry:
@@ -93,7 +93,7 @@ entry:
 
 define double @f10(i32 %a) {
 ;CHECK-VFP: f10:
-;CHECK-VFP: fuitod
+;CHECK-VFP: vcvt.f64.u32
 ;CHECK: f10:
 ;CHECK: floatunsidf
 entry:
diff --git a/test/CodeGen/ARM/fpmem.ll b/test/CodeGen/ARM/fpmem.ll
index 0822fbff653..c3cff18c959 100644
--- a/test/CodeGen/ARM/fpmem.ll
+++ b/test/CodeGen/ARM/fpmem.ll
@@ -8,7 +8,7 @@ define float @f1(float %a) {
 
 define float @f2(float* %v, float %u) {
 ; CHECK: f2:
-; CHECK: flds{{.*}}[
+; CHECK: vldr.32{{.*}}[
         %tmp = load float* %v           ; <float> [#uses=1]
         %tmp1 = fadd float %tmp, %u              ; <float> [#uses=1]
         ret float %tmp1
@@ -16,7 +16,7 @@ define float @f2(float* %v, float %u) {
 
 define void @f3(float %a, float %b, float* %v) {
 ; CHECK: f3:
-; CHECK: fsts{{.*}}[
+; CHECK: vstr.32{{.*}}[
         %tmp = fadd float %a, %b         ; <float> [#uses=1]
         store float %tmp, float* %v
         ret void
diff --git a/test/CodeGen/ARM/fptoint.ll b/test/CodeGen/ARM/fptoint.ll
index 0d270b0c056..4cacc5de7ee 100644
--- a/test/CodeGen/ARM/fptoint.ll
+++ b/test/CodeGen/ARM/fptoint.ll
@@ -1,4 +1,4 @@
-; RUN: llc < %s -march=arm -mattr=+v6,+vfp2 | grep fmrs | count 1
+; RUN: llc < %s -march=arm -mattr=+v6,+vfp2 | grep -E {vmov\\W*r\[0-9\]+,\\W*s\[0-9\]+} | count 1
 ; RUN: llc < %s -march=arm -mattr=+v6,+vfp2 | not grep fmrrd
 
 @i = weak global i32 0		; <i32*> [#uses=2]
diff --git a/test/CodeGen/ARM/fsubs.ll b/test/CodeGen/ARM/fsubs.ll
index 060dd464f1b..f84ccdd480b 100644
--- a/test/CodeGen/ARM/fsubs.ll
+++ b/test/CodeGen/ARM/fsubs.ll
@@ -1,6 +1,6 @@
-; RUN: llc < %s -march=arm -mattr=+vfp2 | grep -E {fsubs\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
+; RUN: llc < %s -march=arm -mattr=+vfp2 | grep -E {vsub.f32\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
 ; RUN: llc < %s -march=arm -mattr=+neon -arm-use-neon-fp=1 | grep -E {vsub.f32\\W*d\[0-9\]+,\\W*d\[0-9\]+,\\W*d\[0-9\]+} | count 1
-; RUN: llc < %s -march=arm -mattr=+neon -arm-use-neon-fp=0 | grep -E {fsubs\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
+; RUN: llc < %s -march=arm -mattr=+neon -arm-use-neon-fp=0 | grep -E {vsub.f32\\W*s\[0-9\]+,\\W*s\[0-9\]+,\\W*s\[0-9\]+} | count 1
 
 define float @test(float %a, float %b) {
 entry:
diff --git a/test/CodeGen/ARM/ifcvt5.ll b/test/CodeGen/ARM/ifcvt5.ll
index e9145ac36dd..623f2cb1dfa 100644
--- a/test/CodeGen/ARM/ifcvt5.ll
+++ b/test/CodeGen/ARM/ifcvt5.ll
@@ -11,7 +11,7 @@ entry:
 
 define void @t1(i32 %a, i32 %b) {
 ; CHECK: t1:
-; CHECK: ldmltfd sp!, {r7, pc}
+; CHECK: ldmfdlt sp!, {r7, pc}
 entry:
 	%tmp1 = icmp sgt i32 %a, 10		; <i1> [#uses=1]
 	br i1 %tmp1, label %cond_true, label %UnifiedReturnBlock
diff --git a/test/CodeGen/ARM/ifcvt6.ll b/test/CodeGen/ARM/ifcvt6.ll
index 58241157580..d7fcf7d6440 100644
--- a/test/CodeGen/ARM/ifcvt6.ll
+++ b/test/CodeGen/ARM/ifcvt6.ll
@@ -1,7 +1,7 @@
 ; RUN: llc < %s -march=arm -mtriple=arm-apple-darwin | \
 ; RUN:   grep cmpne | count 1
 ; RUN: llc < %s -march=arm -mtriple=arm-apple-darwin | \
-; RUN:   grep ldmhi | count 1
+; RUN:   grep ldmfdhi | count 1
 
 define void @foo(i32 %X, i32 %Y) {
 entry:
diff --git a/test/CodeGen/ARM/ifcvt7.ll b/test/CodeGen/ARM/ifcvt7.ll
index f9cf88f7292..c60ad93699f 100644
--- a/test/CodeGen/ARM/ifcvt7.ll
+++ b/test/CodeGen/ARM/ifcvt7.ll
@@ -3,7 +3,7 @@
 ; RUN: llc < %s -march=arm -mtriple=arm-apple-darwin | \
 ; RUN:   grep moveq | count 1
 ; RUN: llc < %s -march=arm -mtriple=arm-apple-darwin | \
-; RUN:   grep ldmeq | count 1
+; RUN:   grep ldmfdeq | count 1
 ; FIXME: Need post-ifcvt branch folding to get rid of the extra br at end of BB1.
 
 	%struct.quad_struct = type { i32, i32, %struct.quad_struct*, %struct.quad_struct*, %struct.quad_struct*, %struct.quad_struct*, %struct.quad_struct* }
diff --git a/test/CodeGen/ARM/ifcvt8.ll b/test/CodeGen/ARM/ifcvt8.ll
index 6cb8e7bb69f..a7da834f781 100644
--- a/test/CodeGen/ARM/ifcvt8.ll
+++ b/test/CodeGen/ARM/ifcvt8.ll
@@ -1,5 +1,5 @@
 ; RUN: llc < %s -march=arm -mtriple=arm-apple-darwin | \
-; RUN:   grep ldmne | count 1
+; RUN:   grep ldmfdne | count 1
 
 	%struct.SString = type { i8*, i32, i32 }
 
diff --git a/test/CodeGen/ARM/neon_ld1.ll b/test/CodeGen/ARM/neon_ld1.ll
index 2796dec5b97..c78872a4bca 100644
--- a/test/CodeGen/ARM/neon_ld1.ll
+++ b/test/CodeGen/ARM/neon_ld1.ll
@@ -1,6 +1,6 @@
-; RUN: llc < %s -march=arm -mattr=+neon | grep fldd | count 4
-; RUN: llc < %s -march=arm -mattr=+neon | grep fstd
-; RUN: llc < %s -march=arm -mattr=+neon | grep fmrrd
+; RUN: llc < %s -march=arm -mattr=+neon | grep vldr.64 | count 4
+; RUN: llc < %s -march=arm -mattr=+neon | grep vstr.64
+; RUN: llc < %s -march=arm -mattr=+neon | grep vmov
 
 define void @t1(<2 x i32>* %r, <4 x i16>* %a, <4 x i16>* %b) nounwind {
 entry:
diff --git a/test/CodeGen/ARM/neon_ld2.ll b/test/CodeGen/ARM/neon_ld2.ll
index 547bab76356..130277b31c3 100644
--- a/test/CodeGen/ARM/neon_ld2.ll
+++ b/test/CodeGen/ARM/neon_ld2.ll
@@ -1,6 +1,6 @@
 ; RUN: llc < %s -march=arm -mattr=+neon | grep vldmia | count 4
 ; RUN: llc < %s -march=arm -mattr=+neon | grep vstmia | count 1
-; RUN: llc < %s -march=arm -mattr=+neon | grep fmrrd  | count 2
+; RUN: llc < %s -march=arm -mattr=+neon | grep vmov  | count 2
 
 define void @t1(<4 x i32>* %r, <2 x i64>* %a, <2 x i64>* %b) nounwind {
 entry:
diff --git a/test/CodeGen/ARM/select.ll b/test/CodeGen/ARM/select.ll
index 85c8b5b8477..29c55c6bd97 100644
--- a/test/CodeGen/ARM/select.ll
+++ b/test/CodeGen/ARM/select.ll
@@ -60,7 +60,7 @@ define double @f7(double %a, double %b) {
 ;CHECK: movlt
 ;CHECK: movlt
 ;CHECK-VFP: f7:
-;CHECK-VFP: fcpydmi
+;CHECK-VFP: vmovmi
     %tmp = fcmp olt double %a, 1.234e+00
     %tmp1 = select i1 %tmp, double -1.000e+00, double %b
     ret double %tmp1
diff --git a/test/CodeGen/ARM/vfp.ll b/test/CodeGen/ARM/vfp.ll
index 50000e31e11..44a44afe9af 100644
--- a/test/CodeGen/ARM/vfp.ll
+++ b/test/CodeGen/ARM/vfp.ll
@@ -15,11 +15,11 @@ declare double @fabs(double)
 define void @test_abs(float* %P, double* %D) {
 ;CHECK: test_abs:
 	%a = load float* %P		; <float> [#uses=1]
-;CHECK: fabss
+;CHECK: vabs.f32
 	%b = call float @fabsf( float %a )		; <float> [#uses=1]
 	store float %b, float* %P
 	%A = load double* %D		; <double> [#uses=1]
-;CHECK: fabsd
+;CHECK: vabs.f64
 	%B = call double @fabs( double %A )		; <double> [#uses=1]
 	store double %B, double* %D
 	ret void
@@ -39,10 +39,10 @@ define void @test_add(float* %P, double* %D) {
 define void @test_ext_round(float* %P, double* %D) {
 ;CHECK: test_ext_round:
 	%a = load float* %P		; <float> [#uses=1]
-;CHECK: fcvtds
+;CHECK: vcvt.f64.f32
 	%b = fpext float %a to double		; <double> [#uses=1]
 	%A = load double* %D		; <double> [#uses=1]
-;CHECK: fcvtsd
+;CHECK: vcvt.f32.f64
 	%B = fptrunc double %A to float		; <float> [#uses=1]
 	store double %b, double* %D
 	store float %B, float* %P
@@ -54,7 +54,7 @@ define void @test_fma(float* %P1, float* %P2, float* %P3) {
 	%a1 = load float* %P1		; <float> [#uses=1]
 	%a2 = load float* %P2		; <float> [#uses=1]
 	%a3 = load float* %P3		; <float> [#uses=1]
-;CHECK: fmscs
+;CHECK: vnmls.f32
 	%X = fmul float %a1, %a2		; <float> [#uses=1]
 	%Y = fsub float %X, %a3		; <float> [#uses=1]
 	store float %Y, float* %P1
@@ -64,7 +64,7 @@ define void @test_fma(float* %P1, float* %P2, float* %P3) {
 define i32 @test_ftoi(float* %P1) {
 ;CHECK: test_ftoi:
 	%a1 = load float* %P1		; <float> [#uses=1]
-;CHECK: ftosizs
+;CHECK: vcvt.s32.f32
 	%b1 = fptosi float %a1 to i32		; <i32> [#uses=1]
 	ret i32 %b1
 }
@@ -72,7 +72,7 @@ define i32 @test_ftoi(float* %P1) {
 define i32 @test_ftou(float* %P1) {
 ;CHECK: test_ftou:
 	%a1 = load float* %P1		; <float> [#uses=1]
-;CHECK: ftouizs
+;CHECK: vcvt.u32.f32
 	%b1 = fptoui float %a1 to i32		; <i32> [#uses=1]
 	ret i32 %b1
 }
@@ -80,7 +80,7 @@ define i32 @test_ftou(float* %P1) {
 define i32 @test_dtoi(double* %P1) {
 ;CHECK: test_dtoi:
 	%a1 = load double* %P1		; <double> [#uses=1]
-;CHECK: ftosizd
+;CHECK: vcvt.s32.f64
 	%b1 = fptosi double %a1 to i32		; <i32> [#uses=1]
 	ret i32 %b1
 }
@@ -88,14 +88,14 @@ define i32 @test_dtoi(double* %P1) {
 define i32 @test_dtou(double* %P1) {
 ;CHECK: test_dtou:
 	%a1 = load double* %P1		; <double> [#uses=1]
-;CHECK: ftouizd
+;CHECK: vcvt.u32.f64
 	%b1 = fptoui double %a1 to i32		; <i32> [#uses=1]
 	ret i32 %b1
 }
 
 define void @test_utod(double* %P1, i32 %X) {
 ;CHECK: test_utod:
-;CHECK: fuitod
+;CHECK: vcvt.f64.u32
 	%b1 = uitofp i32 %X to double		; <double> [#uses=1]
 	store double %b1, double* %P1
 	ret void
@@ -103,7 +103,7 @@ define void @test_utod(double* %P1, i32 %X) {
 
 define void @test_utod2(double* %P1, i8 %X) {
 ;CHECK: test_utod2:
-;CHECK: fuitod
+;CHECK: vcvt.f64.u32
 	%b1 = uitofp i8 %X to double		; <double> [#uses=1]
 	store double %b1, double* %P1
 	ret void
@@ -141,7 +141,7 @@ define void @test_cmpfp0(float* %glob, i32 %X) {
 ;CHECK: test_cmpfp0:
 entry:
 	%tmp = load float* %glob		; <float> [#uses=1]
-;CHECK: fcmpezs
+;CHECK: vcmpe.f32
 	%tmp.upgrd.3 = fcmp ogt float %tmp, 0.000000e+00		; <i1> [#uses=1]
 	br i1 %tmp.upgrd.3, label %cond_true, label %cond_false
 
diff --git a/test/CodeGen/ARM/vget_lane.ll b/test/CodeGen/ARM/vget_lane.ll
index f0df7982ef4..5dd87d66c15 100644
--- a/test/CodeGen/ARM/vget_lane.ll
+++ b/test/CodeGen/ARM/vget_lane.ll
@@ -204,8 +204,8 @@ define <4 x i32> @vsetQ_lane32(<4 x i32>* %A, i32 %B) nounwind {
 
 define arm_aapcs_vfpcc <2 x float> @test_vset_lanef32(float %arg0_float32_t, <2 x float> %arg1_float32x2_t) nounwind {
 ;CHECK: test_vset_lanef32:
-;CHECK: fcpys
-;CHECK: fcpys
+;CHECK: vmov.f32
+;CHECK: vmov.f32
 entry:
   %0 = insertelement <2 x float> %arg1_float32x2_t, float %arg0_float32_t, i32 1 ; <<2 x float>> [#uses=1]
   ret <2 x float> %0
diff --git a/test/CodeGen/Thumb2/cross-rc-coalescing-2.ll b/test/CodeGen/Thumb2/cross-rc-coalescing-2.ll
index 4320328e9c1..eefbae53e72 100644
--- a/test/CodeGen/Thumb2/cross-rc-coalescing-2.ll
+++ b/test/CodeGen/Thumb2/cross-rc-coalescing-2.ll
@@ -1,4 +1,4 @@
-; RUN: llc < %s -mtriple=thumbv7-apple-darwin9 -mcpu=cortex-a8 | grep fcpys | count 4
+; RUN: llc < %s -mtriple=thumbv7-apple-darwin9 -mcpu=cortex-a8 | grep vmov.f32 | count 4
 
 define arm_apcscc void @fht(float* nocapture %fz, i16 signext %n) nounwind {
 entry:
-- 
2.34.1