diff --git a/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp b/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
--- a/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
@@ -1520,6 +1520,9 @@
   if (parseOptionalRegister(YamlMFI.VGPRForAGPRCopy, MFI->VGPRForAGPRCopy))
     return true;
 
+  if (parseOptionalRegister(YamlMFI.SGPRForEXECCopy, MFI->SGPRForEXECCopy))
+    return true;
+
   auto diagnoseRegisterClass = [&](const yaml::StringValue &RegName) {
     // Create a diagnostic for a the register string literal.
     const MemoryBuffer &Buffer =
diff --git a/llvm/lib/Target/AMDGPU/SIDefines.h b/llvm/lib/Target/AMDGPU/SIDefines.h
--- a/llvm/lib/Target/AMDGPU/SIDefines.h
+++ b/llvm/lib/Target/AMDGPU/SIDefines.h
@@ -905,6 +905,13 @@
 };
 
 } // namespace ImplicitArg
+
+namespace VirtRegFlag {
+// Virtual Register Flags.
+enum Register_Flag : uint8_t { WWM_REG = 0 };
+
+} // namespace VirtRegFlag
+
 } // namespace AMDGPU
 
 #define R_00B028_SPI_SHADER_PGM_RSRC1_PS                                0x00B028
diff --git a/llvm/lib/Target/AMDGPU/SIFrameLowering.h b/llvm/lib/Target/AMDGPU/SIFrameLowering.h
--- a/llvm/lib/Target/AMDGPU/SIFrameLowering.h
+++ b/llvm/lib/Target/AMDGPU/SIFrameLowering.h
@@ -34,8 +34,8 @@
                             RegScavenger *RS = nullptr) const override;
   void determineCalleeSavesSGPR(MachineFunction &MF, BitVector &SavedRegs,
                                 RegScavenger *RS = nullptr) const;
-  void determinePrologEpilogSGPRSaves(MachineFunction &MF,
-                                      BitVector &SavedRegs) const;
+  void determinePrologEpilogSGPRSaves(MachineFunction &MF, BitVector &SavedRegs,
+                                      bool NeedExecCopyReservedReg) const;
   void emitCSRSpillStores(MachineFunction &MF, MachineBasicBlock &MBB,
                           MachineBasicBlock::iterator MBBI, DebugLoc &DL,
                           LivePhysRegs &LiveRegs, Register FrameReg,
diff --git a/llvm/lib/Target/AMDGPU/SIFrameLowering.cpp b/llvm/lib/Target/AMDGPU/SIFrameLowering.cpp
--- a/llvm/lib/Target/AMDGPU/SIFrameLowering.cpp
+++ b/llvm/lib/Target/AMDGPU/SIFrameLowering.cpp
@@ -66,7 +66,8 @@
 
 static void getVGPRSpillLaneOrTempRegister(
     MachineFunction &MF, LivePhysRegs &LiveRegs, Register SGPR,
-    const TargetRegisterClass &RC = AMDGPU::SReg_32_XM0_XEXECRegClass) {
+    const TargetRegisterClass &RC = AMDGPU::SReg_32_XM0_XEXECRegClass,
+    bool IncludeScratchCopy = true) {
   SIMachineFunctionInfo *MFI = MF.getInfo<SIMachineFunctionInfo>();
   MachineFrameInfo &FrameInfo = MF.getFrameInfo();
 
@@ -77,9 +78,12 @@
 
   // We need to save and restore the given SGPR.
 
+  Register ScratchSGPR;
   // 1: Try to save the given register into an unused scratch SGPR. The LiveRegs
-  // should have all the callee saved registers marked as used.
-  Register ScratchSGPR = findUnusedRegister(MF.getRegInfo(), LiveRegs, RC);
+  // should have all the callee saved registers marked as used.  For certain
+  // cases we skip copy to scratch SGPR.
+  if (IncludeScratchCopy)
+    ScratchSGPR = findUnusedRegister(MF.getRegInfo(), LiveRegs, RC);
 
   if (!ScratchSGPR) {
     int FI = FrameInfo.CreateStackObject(Size, Alignment, true, nullptr,
@@ -1348,8 +1352,8 @@
             TII->getNamedOperand(MI, AMDGPU::OpName::vdata)->getReg();
           if (FuncInfo->allocateVGPRSpillToAGPR(MF, FI,
                                                 TRI->isAGPR(MRI, VReg))) {
-            // FIXME: change to enterBasicBlockEnd()
-            RS->enterBasicBlock(MBB);
+            RS->enterBasicBlockEnd(MBB);
+            RS->backward(MI);
             TRI->eliminateFrameIndex(MI, 0, FIOp, RS);
             SpillFIs.set(FI);
             continue;
@@ -1446,8 +1450,10 @@
 // The special SGPR spills like the one needed for FP, BP or any reserved
 // registers delayed until frame lowering.
 void SIFrameLowering::determinePrologEpilogSGPRSaves(
-    MachineFunction &MF, BitVector &SavedVGPRs) const {
+    MachineFunction &MF, BitVector &SavedVGPRs,
+    bool NeedExecCopyReservedReg) const {
   MachineFrameInfo &FrameInfo = MF.getFrameInfo();
+  MachineRegisterInfo &MRI = MF.getRegInfo();
   SIMachineFunctionInfo *MFI = MF.getInfo<SIMachineFunctionInfo>();
   const GCNSubtarget &ST = MF.getSubtarget<GCNSubtarget>();
   const SIRegisterInfo *TRI = ST.getRegisterInfo();
@@ -1459,6 +1465,27 @@
   for (unsigned I = 0; CSRegs[I]; ++I)
     LiveRegs.addReg(CSRegs[I]);
 
+  if (NeedExecCopyReservedReg) {
+    Register ReservedReg = MFI->getSGPRForEXECCopy();
+    assert(ReservedReg && "Should have reserved an SGPR for EXEC copy.");
+    const TargetRegisterClass &RC = ST.isWave32()
+                                        ? AMDGPU::SReg_32_XM0_XEXECRegClass
+                                        : AMDGPU::SGPR_64RegClass;
+    Register UnusedScratchReg = findUnusedRegister(MRI, LiveRegs, RC);
+    if (UnusedScratchReg) {
+      // If found any unused scratch SGPR, reserve the register itself for Exec
+      // copy and there is no need for any spills in that case.
+      MFI->setSGPRForEXECCopy(UnusedScratchReg);
+      LiveRegs.addReg(UnusedScratchReg);
+    } else {
+      // Needs spill.
+      assert(!MFI->hasPrologEpilogSGPRSpillEntry(ReservedReg) &&
+             "Re-reserving spill slot for EXEC copy register");
+      getVGPRSpillLaneOrTempRegister(MF, LiveRegs, ReservedReg, RC,
+                                     /* IncludeScratchCopy */ false);
+    }
+  }
+
   // hasFP only knows about stack objects that already exist. We're now
   // determining the stack slots that will be created, so we have to predict
   // them. Stack objects force FP usage with calls.
@@ -1497,6 +1524,8 @@
 
   const GCNSubtarget &ST = MF.getSubtarget<GCNSubtarget>();
   const SIRegisterInfo *TRI = ST.getRegisterInfo();
+  const SIInstrInfo *TII = ST.getInstrInfo();
+  bool NeedExecCopyReservedReg = false;
 
   for (MachineBasicBlock &MBB : MF) {
     for (MachineInstr &MI : MBB) {
@@ -1514,6 +1543,8 @@
         MFI->allocateWWMSpill(MF, MI.getOperand(0).getReg());
       else if (MI.getOpcode() == AMDGPU::V_READLANE_B32)
         MFI->allocateWWMSpill(MF, MI.getOperand(1).getReg());
+      else if (TII->isWWMRegSpillOpcode(MI.getOpcode()))
+        NeedExecCopyReservedReg = true;
     }
   }
 
@@ -1526,7 +1557,7 @@
   if (!ST.hasGFX90AInsts())
     SavedVGPRs.clearBitsInMask(TRI->getAllAGPRRegMask());
 
-  determinePrologEpilogSGPRSaves(MF, SavedVGPRs);
+  determinePrologEpilogSGPRSaves(MF, SavedVGPRs, NeedExecCopyReservedReg);
 
   // The Whole-Wave VGPRs need to be specially inserted in the prolog, so don't
   // allow the default insertion to handle them.
diff --git a/llvm/lib/Target/AMDGPU/SIISelLowering.cpp b/llvm/lib/Target/AMDGPU/SIISelLowering.cpp
--- a/llvm/lib/Target/AMDGPU/SIISelLowering.cpp
+++ b/llvm/lib/Target/AMDGPU/SIISelLowering.cpp
@@ -12510,6 +12510,14 @@
     }
   }
 
+  // Reserve the SGPR(s) to save/restore EXEC for WWM spill/copy handling.
+  unsigned MaxNumSGPRs = ST.getMaxNumSGPRs(MF);
+  Register SReg =
+      ST.isWave32()
+          ? AMDGPU::SGPR_32RegClass.getRegister(MaxNumSGPRs - 1)
+          : AMDGPU::SGPR_64RegClass.getRegister((MaxNumSGPRs / 2) - 1);
+  Info->setSGPRForEXECCopy(SReg);
+
   TargetLoweringBase::finalizeLowering(MF);
 }
 
diff --git a/llvm/lib/Target/AMDGPU/SIInstrInfo.h b/llvm/lib/Target/AMDGPU/SIInstrInfo.h
--- a/llvm/lib/Target/AMDGPU/SIInstrInfo.h
+++ b/llvm/lib/Target/AMDGPU/SIInstrInfo.h
@@ -625,6 +625,11 @@
     return get(Opcode).TSFlags & SIInstrFlags::SGPRSpill;
   }
 
+  static bool isWWMRegSpillOpcode(uint16_t Opcode) {
+    return Opcode == AMDGPU::SI_SPILL_WWM_V32_SAVE ||
+           Opcode == AMDGPU::SI_SPILL_WWM_V32_RESTORE;
+  }
+
   static bool isDPP(const MachineInstr &MI) {
     return MI.getDesc().TSFlags & SIInstrFlags::DPP;
   }
diff --git a/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp b/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
--- a/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
+++ b/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
@@ -1511,6 +1511,29 @@
   }
 }
 
+static unsigned getWWMRegSpillSaveOpcode(unsigned Size) {
+  // Currently, there is only 32-bit WWM register spills needed.
+  if (Size != 4)
+    llvm_unreachable("unknown wwm register spill size");
+
+  return AMDGPU::SI_SPILL_WWM_V32_SAVE;
+}
+
+static unsigned getVectorRegSpillSaveOpcode(Register Reg,
+                                            const TargetRegisterClass *RC,
+                                            unsigned Size,
+                                            const SIRegisterInfo &TRI,
+                                            const SIMachineFunctionInfo &MFI,
+                                            const MachineRegisterInfo &MRI) {
+  // Choose the right opcode if spilling a WWM register.
+  if (MFI.checkFlag(MRI, Reg, AMDGPU::VirtRegFlag::WWM_REG))
+    return getWWMRegSpillSaveOpcode(Size);
+
+  return TRI.isVectorSuperClass(RC) ? getAVSpillSaveOpcode(Size)
+         : TRI.isAGPRClass(RC)      ? getAGPRSpillSaveOpcode(Size)
+                                    : getVGPRSpillSaveOpcode(Size);
+}
+
 void SIInstrInfo::storeRegToStackSlot(MachineBasicBlock &MBB,
                                       MachineBasicBlock::iterator MI,
                                       Register SrcReg, bool isKill,
@@ -1557,9 +1580,8 @@
     return;
   }
 
-  unsigned Opcode = RI.isVectorSuperClass(RC) ? getAVSpillSaveOpcode(SpillSize)
-                    : RI.isAGPRClass(RC) ? getAGPRSpillSaveOpcode(SpillSize)
-                                         : getVGPRSpillSaveOpcode(SpillSize);
+  unsigned Opcode =
+      getVectorRegSpillSaveOpcode(SrcReg, RC, SpillSize, RI, *MFI, MRI);
   MFI->setHasSpilledVGPRs();
 
   BuildMI(MBB, MI, DL, get(Opcode))
@@ -1678,6 +1700,29 @@
   }
 }
 
+static unsigned getWWMRegSpillRestoreOpcode(unsigned Size) {
+  // Currently, there is only 32-bit WWM register spills needed.
+  if (Size != 4)
+    llvm_unreachable("unknown wwm register spill size");
+
+  return AMDGPU::SI_SPILL_WWM_V32_RESTORE;
+}
+
+static unsigned getVectorRegSpillRestoreOpcode(Register Reg,
+                                               const TargetRegisterClass *RC,
+                                               unsigned Size,
+                                               const SIRegisterInfo &TRI,
+                                               const SIMachineFunctionInfo &MFI,
+                                               const MachineRegisterInfo &MRI) {
+  // Choose the right opcode if restoring a WWM register.
+  if (MFI.checkFlag(MRI, Reg, AMDGPU::VirtRegFlag::WWM_REG))
+    return getWWMRegSpillRestoreOpcode(Size);
+
+  return TRI.isVectorSuperClass(RC) ? getAVSpillRestoreOpcode(Size)
+         : TRI.isAGPRClass(RC)      ? getAGPRSpillRestoreOpcode(Size)
+                                    : getVGPRSpillRestoreOpcode(Size);
+}
+
 void SIInstrInfo::loadRegFromStackSlot(MachineBasicBlock &MBB,
                                        MachineBasicBlock::iterator MI,
                                        Register DestReg, int FrameIndex,
@@ -1686,6 +1731,7 @@
   MachineFunction *MF = MBB.getParent();
   SIMachineFunctionInfo *MFI = MF->getInfo<SIMachineFunctionInfo>();
   MachineFrameInfo &FrameInfo = MF->getFrameInfo();
+  MachineRegisterInfo &MRI = MF->getRegInfo();
   const DebugLoc &DL = MBB.findDebugLoc(MI);
   unsigned SpillSize = TRI->getSpillSize(*RC);
 
@@ -1706,7 +1752,6 @@
     // lowered to non-memory instructions.
     const MCInstrDesc &OpDesc = get(getSGPRSpillRestoreOpcode(SpillSize));
     if (DestReg.isVirtual() && SpillSize == 4) {
-      MachineRegisterInfo &MRI = MF->getRegInfo();
       MRI.constrainRegClass(DestReg, &AMDGPU::SReg_32_XM0_XEXECRegClass);
     }
 
@@ -1720,10 +1765,9 @@
     return;
   }
 
-  unsigned Opcode = RI.isVectorSuperClass(RC)
-                        ? getAVSpillRestoreOpcode(SpillSize)
-                    : RI.isAGPRClass(RC) ? getAGPRSpillRestoreOpcode(SpillSize)
-                                         : getVGPRSpillRestoreOpcode(SpillSize);
+  unsigned Opcode =
+      getVectorRegSpillRestoreOpcode(DestReg, RC, SpillSize, RI, *MFI, MRI);
+
   BuildMI(MBB, MI, DL, get(Opcode), DestReg)
       .addFrameIndex(FrameIndex)           // vaddr
       .addReg(MFI->getStackPtrOffsetReg()) // scratch_offset
diff --git a/llvm/lib/Target/AMDGPU/SIInstructions.td b/llvm/lib/Target/AMDGPU/SIInstructions.td
--- a/llvm/lib/Target/AMDGPU/SIInstructions.td
+++ b/llvm/lib/Target/AMDGPU/SIInstructions.td
@@ -853,6 +853,8 @@
 defm SI_SPILL_AV512 : SI_SPILL_VGPR <AV_512, 1>;
 defm SI_SPILL_AV1024 : SI_SPILL_VGPR <AV_1024, 1>;
 
+defm SI_SPILL_WWM_V32 : SI_SPILL_VGPR <VGPR_32>;
+
 def SI_PC_ADD_REL_OFFSET : SPseudoInstSI <
   (outs SReg_64:$dst),
   (ins si_ga:$ptr_lo, si_ga:$ptr_hi),
diff --git a/llvm/lib/Target/AMDGPU/SILowerSGPRSpills.cpp b/llvm/lib/Target/AMDGPU/SILowerSGPRSpills.cpp
--- a/llvm/lib/Target/AMDGPU/SILowerSGPRSpills.cpp
+++ b/llvm/lib/Target/AMDGPU/SILowerSGPRSpills.cpp
@@ -20,6 +20,7 @@
 #include "MCTargetDesc/AMDGPUMCTargetDesc.h"
 #include "SIMachineFunctionInfo.h"
 #include "llvm/CodeGen/LiveIntervals.h"
+#include "llvm/CodeGen/MachineDominators.h"
 #include "llvm/CodeGen/MachineFrameInfo.h"
 #include "llvm/CodeGen/RegisterScavenging.h"
 #include "llvm/InitializePasses.h"
@@ -38,6 +39,7 @@
   const SIInstrInfo *TII = nullptr;
   LiveIntervals *LIS = nullptr;
   SlotIndexes *Indexes = nullptr;
+  MachineDominatorTree *MDT = nullptr;
 
   // Save and Restore blocks of the current function. Typically there is a
   // single save block, unless Windows EH funclets are involved.
@@ -51,13 +53,23 @@
 
   void calculateSaveRestoreBlocks(MachineFunction &MF);
   bool spillCalleeSavedRegs(MachineFunction &MF);
+  void updateLaneVGPRDomInstr(
+      int FI, MachineBasicBlock *MBB, MachineBasicBlock::iterator InsertPt,
+      DenseMap<Register, MachineBasicBlock::iterator> &LaneVGPRDomInstr);
 
   bool runOnMachineFunction(MachineFunction &MF) override;
 
   void getAnalysisUsage(AnalysisUsage &AU) const override {
+    AU.addRequired<MachineDominatorTree>();
     AU.setPreservesAll();
     MachineFunctionPass::getAnalysisUsage(AU);
   }
+
+  MachineFunctionProperties getClearedProperties() const override {
+    return MachineFunctionProperties()
+        .set(MachineFunctionProperties::Property::IsSSA)
+        .set(MachineFunctionProperties::Property::NoVRegs);
+  }
 };
 
 } // end anonymous namespace
@@ -68,6 +80,7 @@
                       "SI lower SGPR spill instructions", false, false)
 INITIALIZE_PASS_DEPENDENCY(LiveIntervals)
 INITIALIZE_PASS_DEPENDENCY(VirtRegMap)
+INITIALIZE_PASS_DEPENDENCY(MachineDominatorTree)
 INITIALIZE_PASS_END(SILowerSGPRSpills, DEBUG_TYPE,
                     "SI lower SGPR spill instructions", false, false)
 
@@ -247,6 +260,55 @@
   return false;
 }
 
+void SILowerSGPRSpills::updateLaneVGPRDomInstr(
+    int FI, MachineBasicBlock *MBB, MachineBasicBlock::iterator InsertPt,
+    DenseMap<Register, MachineBasicBlock::iterator> &LaneVGPRDomInstr) {
+  // For the Def of a virtual LaneVPGR to dominate all its uses, we should
+  // insert an IMPLICIT_DEF before the dominating spill. Switching to a
+  // depth first order doesn't really help since the machine function can be in
+  // the unstructured control flow post-SSA. For each virtual register, hence
+  // finding the common dominator to get either the dominating spill or a block
+  // dominating all spills. Is there a better way to handle it?
+  SIMachineFunctionInfo *FuncInfo =
+      MBB->getParent()->getInfo<SIMachineFunctionInfo>();
+  ArrayRef<SIRegisterInfo::SpilledReg> VGPRSpills =
+      FuncInfo->getSGPRSpillToVGPRLanes(FI);
+  Register PrevLaneVGPR;
+  for (auto &Spill : VGPRSpills) {
+    if (PrevLaneVGPR == Spill.VGPR)
+      continue;
+
+    PrevLaneVGPR = Spill.VGPR;
+    auto I = LaneVGPRDomInstr.find(Spill.VGPR);
+    if (Spill.Lane == 0 && I == LaneVGPRDomInstr.end()) {
+      // Initially add the spill instruction itself for Insertion point.
+      LaneVGPRDomInstr[Spill.VGPR] = InsertPt;
+    } else {
+      assert(I != LaneVGPRDomInstr.end());
+      auto PrevInsertPt = I->second;
+      MachineBasicBlock *DomMBB = PrevInsertPt->getParent();
+      if (DomMBB == MBB) {
+        // The insertion point earlier selected in a predecessor block whose
+        // spills are currently being lowered. The earlier InsertPt would be
+        // the one just before the block terminator and it should be changed
+        // if we insert any new spill in it.
+        if (MDT->dominates(&*InsertPt, &*PrevInsertPt))
+          I->second = InsertPt;
+
+        continue;
+      }
+
+      // Find the common dominator block between PrevInsertPt and the
+      // current spill.
+      DomMBB = MDT->findNearestCommonDominator(DomMBB, MBB);
+      if (DomMBB == MBB)
+        I->second = InsertPt;
+      else if (DomMBB != PrevInsertPt->getParent())
+        I->second = &(*DomMBB->getFirstTerminator());
+    }
+  }
+}
+
 bool SILowerSGPRSpills::runOnMachineFunction(MachineFunction &MF) {
   const GCNSubtarget &ST = MF.getSubtarget<GCNSubtarget>();
   TII = ST.getInstrInfo();
@@ -254,6 +316,7 @@
 
   LIS = getAnalysisIfAvailable<LiveIntervals>();
   Indexes = getAnalysisIfAvailable<SlotIndexes>();
+  MDT = &getAnalysis<MachineDominatorTree>();
 
   assert(SaveBlocks.empty() && RestoreBlocks.empty());
 
@@ -263,7 +326,6 @@
   bool HasCSRs = spillCalleeSavedRegs(MF);
 
   MachineFrameInfo &MFI = MF.getFrameInfo();
-  MachineRegisterInfo &MRI = MF.getRegInfo();
   SIMachineFunctionInfo *FuncInfo = MF.getInfo<SIMachineFunctionInfo>();
 
   if (!MFI.hasStackObjects() && !HasCSRs) {
@@ -273,7 +335,6 @@
   }
 
   bool MadeChange = false;
-  bool NewReservedRegs = false;
 
   // TODO: CSR VGPRs will never be spilled to AGPRs. These can probably be
   // handled as SpilledToReg in regular PrologEpilogInserter.
@@ -289,6 +350,9 @@
     // To track the spill frame indices handled in this pass.
     BitVector SpillFIs(MFI.getObjectIndexEnd(), false);
 
+    // To track the IMPLICIT_DEF insertion point for the lane vgprs.
+    DenseMap<Register, MachineBasicBlock::iterator> LaneVGPRDomInstr;
+
     for (MachineBasicBlock &MBB : MF) {
       for (MachineInstr &MI : llvm::make_early_inc_range(MBB)) {
         if (!TII->isSGPRSpill(MI))
@@ -296,23 +360,32 @@
 
         int FI = TII->getNamedOperand(MI, AMDGPU::OpName::addr)->getIndex();
         assert(MFI.getStackID(FI) == TargetStackID::SGPRSpill);
+        MachineInstrSpan MIS(&MI, &MBB);
         if (FuncInfo->allocateSGPRSpillToVGPRLane(MF, FI)) {
-          NewReservedRegs = true;
           bool Spilled = TRI->eliminateSGPRToVGPRSpillFrameIndex(
               MI, FI, nullptr, Indexes, LIS);
           (void)Spilled;
           assert(Spilled && "failed to spill SGPR to VGPR when allocated");
           SpillFIs.set(FI);
+          updateLaneVGPRDomInstr(FI, &MBB, MIS.begin(), LaneVGPRDomInstr);
         }
       }
     }
 
-    // FIXME: Adding to live-ins redundant with reserving registers.
-    for (MachineBasicBlock &MBB : MF) {
-      for (auto Reg : FuncInfo->getSGPRSpillVGPRs())
-        MBB.addLiveIn(Reg);
-      MBB.sortUniqueLiveIns();
+    for (auto Reg : FuncInfo->getSGPRSpillVGPRs()) {
+      auto InsertPt = LaneVGPRDomInstr[Reg];
+      // Insert the IMPLICIT_DEF at the identified points.
+      auto MIB =
+          BuildMI(*InsertPt->getParent(), *InsertPt, InsertPt->getDebugLoc(),
+                  TII->get(AMDGPU::IMPLICIT_DEF), Reg);
+      FuncInfo->setFlag(Reg, AMDGPU::VirtRegFlag::WWM_REG);
+      if (LIS) {
+        LIS->InsertMachineInstrInMaps(*MIB);
+        LIS->createAndComputeVirtRegInterval(Reg);
+      }
+    }
 
+    for (MachineBasicBlock &MBB : MF) {
       // FIXME: The dead frame indices are replaced with a null register from
       // the debug value instructions. We should instead, update it with the
       // correct register value. But not sure the register value alone is
@@ -333,15 +406,26 @@
     // lane".
     FuncInfo->removeDeadFrameIndices(MFI, /*ResetSGPRSpillStackIDs*/ false);
 
+    MachineRegisterInfo &MRI = MF.getRegInfo();
+    const TargetRegisterClass *RC =
+        ST.isWave32() ? &AMDGPU::SGPR_32RegClass : &AMDGPU::SGPR_64RegClass;
+    // Shift back the reserved SGPR for EXEC copy into the lowest range.
+    // This SGPR is reserved to handle the whole-wave spill/copy operations
+    // that might get inserted during vgpr regalloc.
+    Register UnusedLowSGPR = TRI->findUnusedRegister(MRI, RC, MF);
+    if (UnusedLowSGPR && TRI->getHWRegIndex(UnusedLowSGPR) <
+                             TRI->getHWRegIndex(FuncInfo->getSGPRForEXECCopy()))
+      FuncInfo->setSGPRForEXECCopy(UnusedLowSGPR);
+
     MadeChange = true;
+  } else {
+    // No SGPR spills and hence there won't be any WWM spills/copies. Reset the
+    // SGPR reserved for EXEC copy.
+    FuncInfo->setSGPRForEXECCopy(AMDGPU::NoRegister);
   }
 
   SaveBlocks.clear();
   RestoreBlocks.clear();
 
-  // Updated the reserved registers with any VGPRs added for SGPR spills.
-  if (NewReservedRegs)
-    MRI.freezeReservedRegs(MF);
-
   return MadeChange;
 }
diff --git a/llvm/lib/Target/AMDGPU/SIMachineFunctionInfo.h b/llvm/lib/Target/AMDGPU/SIMachineFunctionInfo.h
--- a/llvm/lib/Target/AMDGPU/SIMachineFunctionInfo.h
+++ b/llvm/lib/Target/AMDGPU/SIMachineFunctionInfo.h
@@ -300,6 +300,7 @@
   SIMode Mode;
   Optional<FrameIndex> ScavengeFI;
   StringValue VGPRForAGPRCopy;
+  StringValue SGPRForEXECCopy;
 
   SIMachineFunctionInfo() = default;
   SIMachineFunctionInfo(const llvm::SIMachineFunctionInfo &,
@@ -341,6 +342,8 @@
     YamlIO.mapOptional("scavengeFI", MFI.ScavengeFI);
     YamlIO.mapOptional("vgprForAGPRCopy", MFI.VGPRForAGPRCopy,
                        StringValue()); // Don't print out when it's empty.
+    YamlIO.mapOptional("sgprForEXECCopy", MFI.SGPRForEXECCopy,
+                       StringValue()); // Don't print out when it's empty.
   }
 };
 
@@ -377,7 +380,8 @@
 
 /// This class keeps track of the SPI_SP_INPUT_ADDR config register, which
 /// tells the hardware which interpolation parameters to load.
-class SIMachineFunctionInfo final : public AMDGPUMachineFunction {
+class SIMachineFunctionInfo final : public AMDGPUMachineFunction,
+                                    private MachineRegisterInfo::Delegate {
   friend class GCNTargetMachine;
 
   // State of MODE register, assumed FP mode.
@@ -477,6 +481,9 @@
 
   unsigned HighBitsOf32BitAddress;
 
+  // Flags associated with the virtual registers.
+  IndexedMap<uint8_t, VirtReg2IndexFunctor> VRegFlags;
+
   // Current recorded maximum possible occupancy.
   unsigned Occupancy;
 
@@ -486,6 +493,10 @@
 
   MCPhysReg getNextSystemSGPR() const;
 
+  // MachineRegisterInfo callback functions to notify events.
+  void MRI_NoteNewVirtualRegister(Register Reg) override;
+  void MRI_NotecloneVirtualRegister(Register NewReg, Register SrcReg) override;
+
 public:
   struct VGPRSpillToAGPR {
     SmallVector<MCPhysReg, 32> Lanes;
@@ -494,11 +505,11 @@
   };
 
 private:
-  // To track VGPR + lane index for each subregister of the SGPR spilled to
-  // frameindex key during SILowerSGPRSpills pass.
+  // To track virtual VGPR + lane index for each subregister of the SGPR spilled
+  // to frameindex key during SILowerSGPRSpills pass.
   DenseMap<int, std::vector<SIRegisterInfo::SpilledReg>> SGPRSpillToVGPRLanes;
-  // To track VGPR + lane index for spilling special SGPRs like Frame Pointer
-  // identified during PrologEpilogInserter.
+  // To track physical VGPR + lane index for spilling special SGPRs like Frame
+  // Pointer identified during PrologEpilogInserter.
   DenseMap<int, std::vector<SIRegisterInfo::SpilledReg>>
       PrologEpilogSGPRSpillToVGPRLanes;
   unsigned NumVGPRSpillLanes = 0;
@@ -528,6 +539,9 @@
   // PrologEpilogInserter.
   PrologEpilogSGPRSpillsMap PrologEpilogSGPRSpills;
 
+  // To save/restore EXEC MASK around WWM spills and copies.
+  Register SGPRForEXECCopy;
+
   DenseMap<int, VGPRSpillToAGPR> VGPRToAGPRSpills;
 
   // AGPRs used for VGPR spills.
@@ -651,6 +665,27 @@
                : makeArrayRef(I->second);
   }
 
+  void setFlag(Register Reg, uint8_t Flag) {
+    assert(Reg.isVirtual());
+    if (VRegFlags.inBounds(Reg))
+      VRegFlags[Reg] |= (uint8_t)1 << Flag;
+  }
+
+  bool checkFlag(const MachineRegisterInfo &MRI, Register Reg,
+                 uint8_t Flag) const {
+    if (!Reg.isVirtual()) {
+      // See if a virtReg is available for the physReg. If found, check the
+      // flags of the virtual register.
+      Register VirtReg = MRI.getPhysToCurrentVirtReg();
+      if (!VirtReg)
+        return false;
+
+      Reg = VirtReg;
+    }
+
+    return VRegFlags.inBounds(Reg) && VRegFlags[Reg] & ((uint8_t)1 << Flag);
+  }
+
   void allocateWWMSpill(MachineFunction &MF, Register VGPR, uint64_t Size = 4,
                         Align Alignment = Align(4));
 
@@ -663,6 +698,10 @@
     return SpillAGPR;
   }
 
+  Register getSGPRForEXECCopy() const { return SGPRForEXECCopy; }
+
+  void setSGPRForEXECCopy(Register Reg) { SGPRForEXECCopy = Reg; }
+
   ArrayRef<MCPhysReg> getVGPRSpillAGPRs() const {
     return SpillVGPR;
   }
diff --git a/llvm/lib/Target/AMDGPU/SIMachineFunctionInfo.cpp b/llvm/lib/Target/AMDGPU/SIMachineFunctionInfo.cpp
--- a/llvm/lib/Target/AMDGPU/SIMachineFunctionInfo.cpp
+++ b/llvm/lib/Target/AMDGPU/SIMachineFunctionInfo.cpp
@@ -62,6 +62,9 @@
   Occupancy = ST.computeOccupancy(F, getLDSSize());
   CallingConv::ID CC = F.getCallingConv();
 
+  const_cast<MachineFunction &>(MF).getRegInfo().addDelegate(this);
+  VRegFlags.reserve(256);
+
   // FIXME: Should have analysis or something rather than attribute to detect
   // calls.
   const bool HasCalls = F.hasFnAttribute("amdgpu-calls");
@@ -309,24 +312,11 @@
 bool SIMachineFunctionInfo::allocateVGPRForSGPRSpills(MachineFunction &MF,
                                                       int FI,
                                                       unsigned LaneIndex) {
-  const GCNSubtarget &ST = MF.getSubtarget<GCNSubtarget>();
-  const SIRegisterInfo *TRI = ST.getRegisterInfo();
   MachineRegisterInfo &MRI = MF.getRegInfo();
   Register LaneVGPR;
   if (!LaneIndex) {
-    LaneVGPR = TRI->findUnusedRegister(MRI, &AMDGPU::VGPR_32RegClass, MF);
-    if (LaneVGPR == AMDGPU::NoRegister) {
-      // We have no VGPRs left for spilling SGPRs. Reset because we will not
-      // partially spill the SGPR to VGPRs.
-      SGPRSpillToVGPRLanes.erase(FI);
-      return false;
-    }
-
+    LaneVGPR = MRI.createVirtualRegister(&AMDGPU::VGPR_32RegClass);
     SpillVGPRs.push_back(LaneVGPR);
-    // Add this register as live-in to all blocks to avoid machine verifier
-    // complaining about use of an undefined physical register.
-    for (MachineBasicBlock &BB : MF)
-      BB.addLiveIn(LaneVGPR);
   } else {
     LaneVGPR = SpillVGPRs.back();
   }
@@ -533,6 +523,16 @@
   return AMDGPU::SGPR0 + NumUserSGPRs + NumSystemSGPRs;
 }
 
+void SIMachineFunctionInfo::MRI_NoteNewVirtualRegister(Register Reg) {
+  VRegFlags.grow(Reg);
+}
+
+void SIMachineFunctionInfo::MRI_NotecloneVirtualRegister(Register NewReg,
+                                                         Register SrcReg) {
+  VRegFlags.grow(NewReg);
+  VRegFlags[NewReg] = VRegFlags[SrcReg];
+}
+
 Register
 SIMachineFunctionInfo::getGITPtrLoReg(const MachineFunction &MF) const {
   const GCNSubtarget &ST = MF.getSubtarget<GCNSubtarget>();
@@ -640,6 +640,10 @@
 
   if (MFI.getVGPRForAGPRCopy())
     VGPRForAGPRCopy = regToString(MFI.getVGPRForAGPRCopy(), TRI);
+
+  if (MFI.getSGPRForEXECCopy())
+    SGPRForEXECCopy = regToString(MFI.getSGPRForEXECCopy(), TRI);
+
   auto SFI = MFI.getOptionalScavengeFI();
   if (SFI)
     ScavengeFI = yaml::FrameIndex(*SFI, MF.getFrameInfo());
diff --git a/llvm/lib/Target/AMDGPU/SIRegisterInfo.cpp b/llvm/lib/Target/AMDGPU/SIRegisterInfo.cpp
--- a/llvm/lib/Target/AMDGPU/SIRegisterInfo.cpp
+++ b/llvm/lib/Target/AMDGPU/SIRegisterInfo.cpp
@@ -646,6 +646,11 @@
     assert(!isSubRegister(ScratchRSrcReg, BasePtrReg));
   }
 
+  // SGPR used to preserve EXEC MASK around WWM spill/copy instructions.
+  Register ExecCopyReg = MFI->getSGPRForEXECCopy();
+  if (ExecCopyReg)
+    reserveRegisterTuples(Reserved, ExecCopyReg);
+
   // Reserve VGPRs/AGPRs.
   //
   unsigned MaxNumVGPRs = ST.getMaxNumVGPRs(MF);
@@ -711,9 +716,6 @@
   for (MCPhysReg Reg : MFI->getVGPRSpillAGPRs())
     reserveRegisterTuples(Reserved, Reg);
 
-  for (auto Reg : MFI->getSGPRSpillVGPRs())
-    reserveRegisterTuples(Reserved, Reg);
-
   return Reserved;
 }
 
@@ -1029,6 +1031,8 @@
   case AMDGPU::SI_SPILL_A32_RESTORE:
   case AMDGPU::SI_SPILL_AV32_SAVE:
   case AMDGPU::SI_SPILL_AV32_RESTORE:
+  case AMDGPU::SI_SPILL_WWM_V32_SAVE:
+  case AMDGPU::SI_SPILL_WWM_V32_RESTORE:
     return 1;
   default: llvm_unreachable("Invalid spill opcode");
   }
@@ -1961,6 +1965,40 @@
   }
 }
 
+static void insertScratchExecCopy(MachineFunction &MF, MachineBasicBlock &MBB,
+                                  MachineBasicBlock::iterator MBBI,
+                                  const DebugLoc &DL, Register Reg,
+                                  RegScavenger *RS) {
+  const GCNSubtarget &ST = MF.getSubtarget<GCNSubtarget>();
+  const SIInstrInfo *TII = ST.getInstrInfo();
+  bool IsWave32 = ST.isWave32();
+  if (RS->isRegUsed(AMDGPU::SCC)) {
+    // Insert two move instructions, one to save the original value of EXEC and
+    // the other to turn on all bits in EXEC. This is required as we can't use
+    // the single instruction S_OR_SAVEEXEC that clobbers SCC.
+    unsigned MovOpc = IsWave32 ? AMDGPU::S_MOV_B32 : AMDGPU::S_MOV_B64;
+    MCRegister Exec = IsWave32 ? AMDGPU::EXEC_LO : AMDGPU::EXEC;
+    BuildMI(MBB, MBBI, DL, TII->get(MovOpc), Reg).addReg(Exec, RegState::Kill);
+    BuildMI(MBB, MBBI, DL, TII->get(MovOpc), Exec).addImm(-1);
+  } else {
+    const unsigned OrSaveExec =
+        IsWave32 ? AMDGPU::S_OR_SAVEEXEC_B32 : AMDGPU::S_OR_SAVEEXEC_B64;
+    auto SaveExec =
+        BuildMI(MBB, MBBI, DL, TII->get(OrSaveExec), Reg).addImm(-1);
+    SaveExec->getOperand(3).setIsDead(); // Mark SCC as dead.
+  }
+}
+
+static void restoreExec(MachineFunction &MF, MachineBasicBlock &MBB,
+                        MachineBasicBlock::iterator MBBI, const DebugLoc &DL,
+                        Register Reg) {
+  const GCNSubtarget &ST = MF.getSubtarget<GCNSubtarget>();
+  const SIInstrInfo *TII = ST.getInstrInfo();
+  unsigned ExecMov = ST.isWave32() ? AMDGPU::S_MOV_B32 : AMDGPU::S_MOV_B64;
+  MCRegister Exec = ST.isWave32() ? AMDGPU::EXEC_LO : AMDGPU::EXEC;
+  BuildMI(MBB, MBBI, DL, TII->get(ExecMov), Exec).addReg(Reg, RegState::Kill);
+}
+
 bool SIRegisterInfo::eliminateFrameIndex(MachineBasicBlock::iterator MI,
                                         int SPAdj, unsigned FIOperandNum,
                                         RegScavenger *RS) const {
@@ -2039,7 +2077,8 @@
     case AMDGPU::SI_SPILL_AV128_SAVE:
     case AMDGPU::SI_SPILL_AV96_SAVE:
     case AMDGPU::SI_SPILL_AV64_SAVE:
-    case AMDGPU::SI_SPILL_AV32_SAVE: {
+    case AMDGPU::SI_SPILL_AV32_SAVE:
+    case AMDGPU::SI_SPILL_WWM_V32_SAVE: {
       const MachineOperand *VData = TII->getNamedOperand(*MI,
                                                          AMDGPU::OpName::vdata);
       assert(TII->getNamedOperand(*MI, AMDGPU::OpName::soffset)->getReg() ==
@@ -2048,11 +2087,18 @@
       unsigned Opc = ST.enableFlatScratch() ? AMDGPU::SCRATCH_STORE_DWORD_SADDR
                                             : AMDGPU::BUFFER_STORE_DWORD_OFFSET;
       auto *MBB = MI->getParent();
+      bool IsWWMRegSpill = TII->isWWMRegSpillOpcode(MI->getOpcode());
+      if (IsWWMRegSpill)
+        insertScratchExecCopy(*MF, *MBB, MI, DL, MFI->getSGPRForEXECCopy(), RS);
+
       buildSpillLoadStore(
           *MBB, MI, DL, Opc, Index, VData->getReg(), VData->isKill(), FrameReg,
           TII->getNamedOperand(*MI, AMDGPU::OpName::offset)->getImm(),
           *MI->memoperands_begin(), RS);
       MFI->addToSpilledVGPRs(getNumSubRegsForSpillOp(MI->getOpcode()));
+      if (IsWWMRegSpill)
+        restoreExec(*MF, *MBB, MI, DL, MFI->getSGPRForEXECCopy());
+
       MI->eraseFromParent();
       return true;
     }
@@ -2085,7 +2131,8 @@
     case AMDGPU::SI_SPILL_AV224_RESTORE:
     case AMDGPU::SI_SPILL_AV256_RESTORE:
     case AMDGPU::SI_SPILL_AV512_RESTORE:
-    case AMDGPU::SI_SPILL_AV1024_RESTORE: {
+    case AMDGPU::SI_SPILL_AV1024_RESTORE:
+    case AMDGPU::SI_SPILL_WWM_V32_RESTORE: {
       const MachineOperand *VData = TII->getNamedOperand(*MI,
                                                          AMDGPU::OpName::vdata);
       assert(TII->getNamedOperand(*MI, AMDGPU::OpName::soffset)->getReg() ==
@@ -2094,10 +2141,17 @@
       unsigned Opc = ST.enableFlatScratch() ? AMDGPU::SCRATCH_LOAD_DWORD_SADDR
                                             : AMDGPU::BUFFER_LOAD_DWORD_OFFSET;
       auto *MBB = MI->getParent();
+      bool IsWWMRegSpill = TII->isWWMRegSpillOpcode(MI->getOpcode());
+      if (IsWWMRegSpill)
+        insertScratchExecCopy(*MF, *MBB, MI, DL, MFI->getSGPRForEXECCopy(), RS);
+
       buildSpillLoadStore(
           *MBB, MI, DL, Opc, Index, VData->getReg(), VData->isKill(), FrameReg,
           TII->getNamedOperand(*MI, AMDGPU::OpName::offset)->getImm(),
           *MI->memoperands_begin(), RS);
+      if (IsWWMRegSpill)
+        restoreExec(*MF, *MBB, MI, DL, MFI->getSGPRForEXECCopy());
+
       MI->eraseFromParent();
       return true;
     }
diff --git a/llvm/test/CodeGen/AMDGPU/GlobalISel/assert-align.ll b/llvm/test/CodeGen/AMDGPU/GlobalISel/assert-align.ll
--- a/llvm/test/CodeGen/AMDGPU/GlobalISel/assert-align.ll
+++ b/llvm/test/CodeGen/AMDGPU/GlobalISel/assert-align.ll
@@ -13,6 +13,7 @@
 ; CHECK-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; CHECK-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; CHECK-NEXT:    s_mov_b64 exec, s[18:19]
+; CHECK-NEXT:    ; implicit-def: $vgpr40
 ; CHECK-NEXT:    s_addk_i32 s32, 0x400
 ; CHECK-NEXT:    v_writelane_b32 v40, s30, 0
 ; CHECK-NEXT:    v_mov_b32_e32 v0, 0
diff --git a/llvm/test/CodeGen/AMDGPU/GlobalISel/call-outgoing-stack-args.ll b/llvm/test/CodeGen/AMDGPU/GlobalISel/call-outgoing-stack-args.ll
--- a/llvm/test/CodeGen/AMDGPU/GlobalISel/call-outgoing-stack-args.ll
+++ b/llvm/test/CodeGen/AMDGPU/GlobalISel/call-outgoing-stack-args.ll
@@ -246,6 +246,7 @@
 ; MUBUF-NEXT:    v_mov_b32_e32 v0, 9
 ; MUBUF-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:4
 ; MUBUF-NEXT:    v_mov_b32_e32 v0, 10
+; MUBUF-NEXT:    ; implicit-def: $vgpr40
 ; MUBUF-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:8
 ; MUBUF-NEXT:    v_mov_b32_e32 v0, 11
 ; MUBUF-NEXT:    v_writelane_b32 v40, s30, 0
@@ -283,6 +284,7 @@
 ; FLATSCR-NEXT:    v_mov_b32_e32 v0, 9
 ; FLATSCR-NEXT:    scratch_store_dword off, v0, s32 offset:4
 ; FLATSCR-NEXT:    v_mov_b32_e32 v0, 10
+; FLATSCR-NEXT:    ; implicit-def: $vgpr40
 ; FLATSCR-NEXT:    scratch_store_dword off, v0, s32 offset:8
 ; FLATSCR-NEXT:    v_mov_b32_e32 v0, 11
 ; FLATSCR-NEXT:    v_writelane_b32 v40, s30, 0
@@ -323,8 +325,9 @@
 ; MUBUF-NEXT:    buffer_load_dword v1, v0, s[0:3], 0 offen
 ; MUBUF-NEXT:    buffer_load_dword v2, v0, s[0:3], 0 offen offset:4
 ; MUBUF-NEXT:    s_addk_i32 s32, 0x400
-; MUBUF-NEXT:    v_writelane_b32 v40, s30, 0
+; MUBUF-NEXT:    ; implicit-def: $vgpr40
 ; MUBUF-NEXT:    v_writelane_b32 v41, s4, 0
+; MUBUF-NEXT:    v_writelane_b32 v40, s30, 0
 ; MUBUF-NEXT:    v_writelane_b32 v40, s31, 1
 ; MUBUF-NEXT:    s_getpc_b64 s[4:5]
 ; MUBUF-NEXT:    s_add_u32 s4, s4, external_void_func_byval@rel32@lo+4
@@ -406,8 +409,9 @@
 ; FLATSCR-NEXT:    s_mov_b64 exec, s[2:3]
 ; FLATSCR-NEXT:    scratch_load_dwordx2 v[1:2], v0, off
 ; FLATSCR-NEXT:    s_add_i32 s32, s32, 16
-; FLATSCR-NEXT:    v_writelane_b32 v40, s30, 0
+; FLATSCR-NEXT:    ; implicit-def: $vgpr40
 ; FLATSCR-NEXT:    v_writelane_b32 v41, s0, 0
+; FLATSCR-NEXT:    v_writelane_b32 v40, s30, 0
 ; FLATSCR-NEXT:    v_writelane_b32 v40, s31, 1
 ; FLATSCR-NEXT:    s_getpc_b64 s[0:1]
 ; FLATSCR-NEXT:    s_add_u32 s0, s0, external_void_func_byval@rel32@lo+4
diff --git a/llvm/test/CodeGen/AMDGPU/GlobalISel/image-waterfall-loop-O0.ll b/llvm/test/CodeGen/AMDGPU/GlobalISel/image-waterfall-loop-O0.ll
--- a/llvm/test/CodeGen/AMDGPU/GlobalISel/image-waterfall-loop-O0.ll
+++ b/llvm/test/CodeGen/AMDGPU/GlobalISel/image-waterfall-loop-O0.ll
@@ -8,75 +8,83 @@
 ; CHECK-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; CHECK-NEXT:    s_waitcnt_vscnt null, 0x0
 ; CHECK-NEXT:    s_xor_saveexec_b32 s4, -1
-; CHECK-NEXT:    buffer_store_dword v8, off, s[0:3], s32 offset:44 ; 4-byte Folded Spill
+; CHECK-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:48 ; 4-byte Folded Spill
+; CHECK-NEXT:    buffer_store_dword v2, off, s[0:3], s32 offset:52 ; 4-byte Folded Spill
 ; CHECK-NEXT:    s_mov_b32 exec_lo, s4
-; CHECK-NEXT:    v_mov_b32_e32 v15, v1
-; CHECK-NEXT:    v_mov_b32_e32 v14, v2
-; CHECK-NEXT:    v_mov_b32_e32 v13, v3
-; CHECK-NEXT:    v_mov_b32_e32 v12, v4
-; CHECK-NEXT:    v_mov_b32_e32 v11, v5
-; CHECK-NEXT:    v_mov_b32_e32 v10, v6
-; CHECK-NEXT:    v_mov_b32_e32 v9, v7
+; CHECK-NEXT:    v_mov_b32_e32 v14, v1
+; CHECK-NEXT:    v_mov_b32_e32 v13, v2
+; CHECK-NEXT:    v_mov_b32_e32 v12, v3
+; CHECK-NEXT:    v_mov_b32_e32 v11, v4
+; CHECK-NEXT:    v_mov_b32_e32 v10, v5
+; CHECK-NEXT:    v_mov_b32_e32 v9, v6
+; CHECK-NEXT:    v_mov_b32_e32 v8, v7
 ; CHECK-NEXT:    ; kill: def $vgpr0 killed $vgpr0 def $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7 killed $exec
-; CHECK-NEXT:    v_mov_b32_e32 v1, v15
-; CHECK-NEXT:    v_mov_b32_e32 v2, v14
-; CHECK-NEXT:    v_mov_b32_e32 v3, v13
-; CHECK-NEXT:    v_mov_b32_e32 v4, v12
-; CHECK-NEXT:    v_mov_b32_e32 v5, v11
-; CHECK-NEXT:    v_mov_b32_e32 v6, v10
-; CHECK-NEXT:    v_mov_b32_e32 v7, v9
-; CHECK-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
-; CHECK-NEXT:    buffer_store_dword v1, off, s[0:3], s32 offset:12 ; 4-byte Folded Spill
-; CHECK-NEXT:    buffer_store_dword v2, off, s[0:3], s32 offset:16 ; 4-byte Folded Spill
-; CHECK-NEXT:    buffer_store_dword v3, off, s[0:3], s32 offset:20 ; 4-byte Folded Spill
-; CHECK-NEXT:    buffer_store_dword v4, off, s[0:3], s32 offset:24 ; 4-byte Folded Spill
-; CHECK-NEXT:    buffer_store_dword v5, off, s[0:3], s32 offset:28 ; 4-byte Folded Spill
-; CHECK-NEXT:    buffer_store_dword v6, off, s[0:3], s32 offset:32 ; 4-byte Folded Spill
-; CHECK-NEXT:    buffer_store_dword v7, off, s[0:3], s32 offset:36 ; 4-byte Folded Spill
+; CHECK-NEXT:    v_mov_b32_e32 v1, v14
+; CHECK-NEXT:    v_mov_b32_e32 v2, v13
+; CHECK-NEXT:    v_mov_b32_e32 v3, v12
+; CHECK-NEXT:    v_mov_b32_e32 v4, v11
+; CHECK-NEXT:    v_mov_b32_e32 v5, v10
+; CHECK-NEXT:    v_mov_b32_e32 v6, v9
+; CHECK-NEXT:    v_mov_b32_e32 v7, v8
+; CHECK-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:12 ; 4-byte Folded Spill
+; CHECK-NEXT:    buffer_store_dword v1, off, s[0:3], s32 offset:16 ; 4-byte Folded Spill
+; CHECK-NEXT:    buffer_store_dword v2, off, s[0:3], s32 offset:20 ; 4-byte Folded Spill
+; CHECK-NEXT:    buffer_store_dword v3, off, s[0:3], s32 offset:24 ; 4-byte Folded Spill
+; CHECK-NEXT:    buffer_store_dword v4, off, s[0:3], s32 offset:28 ; 4-byte Folded Spill
+; CHECK-NEXT:    buffer_store_dword v5, off, s[0:3], s32 offset:32 ; 4-byte Folded Spill
+; CHECK-NEXT:    buffer_store_dword v6, off, s[0:3], s32 offset:36 ; 4-byte Folded Spill
+; CHECK-NEXT:    buffer_store_dword v7, off, s[0:3], s32 offset:40 ; 4-byte Folded Spill
 ; CHECK-NEXT:    s_mov_b32 s8, 0
 ; CHECK-NEXT:    s_mov_b32 s4, s8
 ; CHECK-NEXT:    s_mov_b32 s5, s8
 ; CHECK-NEXT:    s_mov_b32 s6, s8
 ; CHECK-NEXT:    s_mov_b32 s7, s8
-; CHECK-NEXT:    v_writelane_b32 v8, s4, 0
-; CHECK-NEXT:    v_writelane_b32 v8, s5, 1
-; CHECK-NEXT:    v_writelane_b32 v8, s6, 2
-; CHECK-NEXT:    v_writelane_b32 v8, s7, 3
+; CHECK-NEXT:    ; implicit-def: $vgpr0
+; CHECK-NEXT:    v_writelane_b32 v0, s4, 0
+; CHECK-NEXT:    v_writelane_b32 v0, s5, 1
+; CHECK-NEXT:    v_writelane_b32 v0, s6, 2
+; CHECK-NEXT:    v_writelane_b32 v0, s7, 3
 ; CHECK-NEXT:    s_mov_b32 s6, 0
 ; CHECK-NEXT:    s_mov_b32 s4, s6
 ; CHECK-NEXT:    s_mov_b32 s5, s6
-; CHECK-NEXT:    v_mov_b32_e32 v0, s4
-; CHECK-NEXT:    v_mov_b32_e32 v1, s5
-; CHECK-NEXT:    buffer_store_dword v0, off, s[0:3], s32 ; 4-byte Folded Spill
+; CHECK-NEXT:    v_mov_b32_e32 v1, s4
+; CHECK-NEXT:    v_mov_b32_e32 v2, s5
 ; CHECK-NEXT:    buffer_store_dword v1, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
+; CHECK-NEXT:    buffer_store_dword v2, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
 ; CHECK-NEXT:    s_mov_b32 s4, exec_lo
-; CHECK-NEXT:    v_writelane_b32 v8, s4, 4
+; CHECK-NEXT:    v_writelane_b32 v0, s4, 4
+; CHECK-NEXT:    s_or_saveexec_b32 s21, -1
+; CHECK-NEXT:    buffer_store_dword v0, off, s[0:3], s32 ; 4-byte Folded Spill
+; CHECK-NEXT:    s_mov_b32 exec_lo, s21
 ; CHECK-NEXT:  .LBB0_1: ; =>This Inner Loop Header: Depth=1
-; CHECK-NEXT:    buffer_load_dword v9, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
-; CHECK-NEXT:    buffer_load_dword v10, off, s[0:3], s32 offset:12 ; 4-byte Folded Reload
-; CHECK-NEXT:    buffer_load_dword v11, off, s[0:3], s32 offset:16 ; 4-byte Folded Reload
-; CHECK-NEXT:    buffer_load_dword v12, off, s[0:3], s32 offset:20 ; 4-byte Folded Reload
-; CHECK-NEXT:    buffer_load_dword v13, off, s[0:3], s32 offset:24 ; 4-byte Folded Reload
-; CHECK-NEXT:    buffer_load_dword v14, off, s[0:3], s32 offset:28 ; 4-byte Folded Reload
-; CHECK-NEXT:    buffer_load_dword v15, off, s[0:3], s32 offset:32 ; 4-byte Folded Reload
-; CHECK-NEXT:    buffer_load_dword v16, off, s[0:3], s32 offset:36 ; 4-byte Folded Reload
+; CHECK-NEXT:    s_or_saveexec_b32 s21, -1
+; CHECK-NEXT:    buffer_load_dword v0, off, s[0:3], s32 ; 4-byte Folded Reload
+; CHECK-NEXT:    s_mov_b32 exec_lo, s21
+; CHECK-NEXT:    buffer_load_dword v9, off, s[0:3], s32 offset:12 ; 4-byte Folded Reload
+; CHECK-NEXT:    buffer_load_dword v10, off, s[0:3], s32 offset:16 ; 4-byte Folded Reload
+; CHECK-NEXT:    buffer_load_dword v11, off, s[0:3], s32 offset:20 ; 4-byte Folded Reload
+; CHECK-NEXT:    buffer_load_dword v12, off, s[0:3], s32 offset:24 ; 4-byte Folded Reload
+; CHECK-NEXT:    buffer_load_dword v13, off, s[0:3], s32 offset:28 ; 4-byte Folded Reload
+; CHECK-NEXT:    buffer_load_dword v14, off, s[0:3], s32 offset:32 ; 4-byte Folded Reload
+; CHECK-NEXT:    buffer_load_dword v15, off, s[0:3], s32 offset:36 ; 4-byte Folded Reload
+; CHECK-NEXT:    buffer_load_dword v16, off, s[0:3], s32 offset:40 ; 4-byte Folded Reload
 ; CHECK-NEXT:    s_waitcnt vmcnt(0)
-; CHECK-NEXT:    v_mov_b32_e32 v7, v9
-; CHECK-NEXT:    v_mov_b32_e32 v6, v10
-; CHECK-NEXT:    v_mov_b32_e32 v5, v11
-; CHECK-NEXT:    v_mov_b32_e32 v4, v12
-; CHECK-NEXT:    v_mov_b32_e32 v3, v13
-; CHECK-NEXT:    v_mov_b32_e32 v2, v14
-; CHECK-NEXT:    v_mov_b32_e32 v1, v15
-; CHECK-NEXT:    v_mov_b32_e32 v0, v16
-; CHECK-NEXT:    v_readfirstlane_b32 s12, v7
-; CHECK-NEXT:    v_readfirstlane_b32 s10, v6
-; CHECK-NEXT:    v_readfirstlane_b32 s9, v5
-; CHECK-NEXT:    v_readfirstlane_b32 s8, v4
-; CHECK-NEXT:    v_readfirstlane_b32 s7, v3
-; CHECK-NEXT:    v_readfirstlane_b32 s6, v2
-; CHECK-NEXT:    v_readfirstlane_b32 s5, v1
-; CHECK-NEXT:    v_readfirstlane_b32 s4, v0
+; CHECK-NEXT:    v_mov_b32_e32 v8, v9
+; CHECK-NEXT:    v_mov_b32_e32 v7, v10
+; CHECK-NEXT:    v_mov_b32_e32 v6, v11
+; CHECK-NEXT:    v_mov_b32_e32 v5, v12
+; CHECK-NEXT:    v_mov_b32_e32 v4, v13
+; CHECK-NEXT:    v_mov_b32_e32 v3, v14
+; CHECK-NEXT:    v_mov_b32_e32 v2, v15
+; CHECK-NEXT:    v_mov_b32_e32 v1, v16
+; CHECK-NEXT:    v_readfirstlane_b32 s12, v8
+; CHECK-NEXT:    v_readfirstlane_b32 s10, v7
+; CHECK-NEXT:    v_readfirstlane_b32 s9, v6
+; CHECK-NEXT:    v_readfirstlane_b32 s8, v5
+; CHECK-NEXT:    v_readfirstlane_b32 s7, v4
+; CHECK-NEXT:    v_readfirstlane_b32 s6, v3
+; CHECK-NEXT:    v_readfirstlane_b32 s5, v2
+; CHECK-NEXT:    v_readfirstlane_b32 s4, v1
 ; CHECK-NEXT:    ; kill: def $sgpr12 killed $sgpr12 def $sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19
 ; CHECK-NEXT:    s_mov_b32 s13, s10
 ; CHECK-NEXT:    s_mov_b32 s14, s9
@@ -85,68 +93,79 @@
 ; CHECK-NEXT:    s_mov_b32 s17, s6
 ; CHECK-NEXT:    s_mov_b32 s18, s5
 ; CHECK-NEXT:    s_mov_b32 s19, s4
-; CHECK-NEXT:    v_writelane_b32 v8, s12, 5
-; CHECK-NEXT:    v_writelane_b32 v8, s13, 6
-; CHECK-NEXT:    v_writelane_b32 v8, s14, 7
-; CHECK-NEXT:    v_writelane_b32 v8, s15, 8
-; CHECK-NEXT:    v_writelane_b32 v8, s16, 9
-; CHECK-NEXT:    v_writelane_b32 v8, s17, 10
-; CHECK-NEXT:    v_writelane_b32 v8, s18, 11
-; CHECK-NEXT:    v_writelane_b32 v8, s19, 12
-; CHECK-NEXT:    v_mov_b32_e32 v6, v9
-; CHECK-NEXT:    v_mov_b32_e32 v7, v10
-; CHECK-NEXT:    v_mov_b32_e32 v4, v11
-; CHECK-NEXT:    v_mov_b32_e32 v5, v12
-; CHECK-NEXT:    v_mov_b32_e32 v2, v13
-; CHECK-NEXT:    v_mov_b32_e32 v3, v14
-; CHECK-NEXT:    v_mov_b32_e32 v0, v15
-; CHECK-NEXT:    v_mov_b32_e32 v1, v16
+; CHECK-NEXT:    v_writelane_b32 v0, s12, 5
+; CHECK-NEXT:    v_writelane_b32 v0, s13, 6
+; CHECK-NEXT:    v_writelane_b32 v0, s14, 7
+; CHECK-NEXT:    v_writelane_b32 v0, s15, 8
+; CHECK-NEXT:    v_writelane_b32 v0, s16, 9
+; CHECK-NEXT:    v_writelane_b32 v0, s17, 10
+; CHECK-NEXT:    v_writelane_b32 v0, s18, 11
+; CHECK-NEXT:    v_writelane_b32 v0, s19, 12
+; CHECK-NEXT:    v_mov_b32_e32 v7, v9
+; CHECK-NEXT:    v_mov_b32_e32 v8, v10
+; CHECK-NEXT:    v_mov_b32_e32 v5, v11
+; CHECK-NEXT:    v_mov_b32_e32 v6, v12
+; CHECK-NEXT:    v_mov_b32_e32 v3, v13
+; CHECK-NEXT:    v_mov_b32_e32 v4, v14
+; CHECK-NEXT:    v_mov_b32_e32 v1, v15
+; CHECK-NEXT:    v_mov_b32_e32 v2, v16
 ; CHECK-NEXT:    s_mov_b64 s[4:5], s[12:13]
 ; CHECK-NEXT:    s_mov_b64 s[10:11], s[14:15]
 ; CHECK-NEXT:    s_mov_b64 s[8:9], s[16:17]
 ; CHECK-NEXT:    s_mov_b64 s[6:7], s[18:19]
-; CHECK-NEXT:    v_cmp_eq_u64_e64 s4, s[4:5], v[6:7]
-; CHECK-NEXT:    v_cmp_eq_u64_e64 s5, s[10:11], v[4:5]
+; CHECK-NEXT:    v_cmp_eq_u64_e64 s4, s[4:5], v[7:8]
+; CHECK-NEXT:    v_cmp_eq_u64_e64 s5, s[10:11], v[5:6]
 ; CHECK-NEXT:    s_and_b32 s4, s4, s5
-; CHECK-NEXT:    v_cmp_eq_u64_e64 s5, s[8:9], v[2:3]
+; CHECK-NEXT:    v_cmp_eq_u64_e64 s5, s[8:9], v[3:4]
 ; CHECK-NEXT:    s_and_b32 s4, s4, s5
-; CHECK-NEXT:    v_cmp_eq_u64_e64 s5, s[6:7], v[0:1]
+; CHECK-NEXT:    v_cmp_eq_u64_e64 s5, s[6:7], v[1:2]
 ; CHECK-NEXT:    s_and_b32 s4, s4, s5
 ; CHECK-NEXT:    s_and_saveexec_b32 s4, s4
-; CHECK-NEXT:    v_writelane_b32 v8, s4, 13
+; CHECK-NEXT:    v_writelane_b32 v0, s4, 13
+; CHECK-NEXT:    s_or_saveexec_b32 s21, -1
+; CHECK-NEXT:    buffer_store_dword v0, off, s[0:3], s32 ; 4-byte Folded Spill
+; CHECK-NEXT:    s_mov_b32 exec_lo, s21
 ; CHECK-NEXT:  ; %bb.2: ; in Loop: Header=BB0_1 Depth=1
-; CHECK-NEXT:    buffer_load_dword v0, off, s[0:3], s32 ; 4-byte Folded Reload
-; CHECK-NEXT:    buffer_load_dword v1, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
-; CHECK-NEXT:    v_readlane_b32 s4, v8, 13
-; CHECK-NEXT:    v_readlane_b32 s8, v8, 5
-; CHECK-NEXT:    v_readlane_b32 s9, v8, 6
-; CHECK-NEXT:    v_readlane_b32 s10, v8, 7
-; CHECK-NEXT:    v_readlane_b32 s11, v8, 8
-; CHECK-NEXT:    v_readlane_b32 s12, v8, 9
-; CHECK-NEXT:    v_readlane_b32 s13, v8, 10
-; CHECK-NEXT:    v_readlane_b32 s14, v8, 11
-; CHECK-NEXT:    v_readlane_b32 s15, v8, 12
-; CHECK-NEXT:    v_readlane_b32 s16, v8, 0
-; CHECK-NEXT:    v_readlane_b32 s17, v8, 1
-; CHECK-NEXT:    v_readlane_b32 s18, v8, 2
-; CHECK-NEXT:    v_readlane_b32 s19, v8, 3
+; CHECK-NEXT:    buffer_load_dword v0, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
+; CHECK-NEXT:    buffer_load_dword v1, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
+; CHECK-NEXT:    s_or_saveexec_b32 s21, -1
+; CHECK-NEXT:    buffer_load_dword v2, off, s[0:3], s32 ; 4-byte Folded Reload
+; CHECK-NEXT:    s_mov_b32 exec_lo, s21
 ; CHECK-NEXT:    s_waitcnt vmcnt(0)
+; CHECK-NEXT:    v_readlane_b32 s4, v2, 13
+; CHECK-NEXT:    v_readlane_b32 s8, v2, 5
+; CHECK-NEXT:    v_readlane_b32 s9, v2, 6
+; CHECK-NEXT:    v_readlane_b32 s10, v2, 7
+; CHECK-NEXT:    v_readlane_b32 s11, v2, 8
+; CHECK-NEXT:    v_readlane_b32 s12, v2, 9
+; CHECK-NEXT:    v_readlane_b32 s13, v2, 10
+; CHECK-NEXT:    v_readlane_b32 s14, v2, 11
+; CHECK-NEXT:    v_readlane_b32 s15, v2, 12
+; CHECK-NEXT:    v_readlane_b32 s16, v2, 0
+; CHECK-NEXT:    v_readlane_b32 s17, v2, 1
+; CHECK-NEXT:    v_readlane_b32 s18, v2, 2
+; CHECK-NEXT:    v_readlane_b32 s19, v2, 3
 ; CHECK-NEXT:    image_sample v0, v[0:1], s[8:15], s[16:19] dmask:0x1 dim:SQ_RSRC_IMG_2D
 ; CHECK-NEXT:    s_waitcnt vmcnt(0)
-; CHECK-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:40 ; 4-byte Folded Spill
+; CHECK-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:44 ; 4-byte Folded Spill
 ; CHECK-NEXT:    s_xor_b32 exec_lo, exec_lo, s4
 ; CHECK-NEXT:    s_cbranch_execnz .LBB0_1
 ; CHECK-NEXT:  ; %bb.3:
-; CHECK-NEXT:    v_readlane_b32 s4, v8, 4
+; CHECK-NEXT:    s_or_saveexec_b32 s21, -1
+; CHECK-NEXT:    buffer_load_dword v0, off, s[0:3], s32 ; 4-byte Folded Reload
+; CHECK-NEXT:    s_mov_b32 exec_lo, s21
+; CHECK-NEXT:    s_waitcnt vmcnt(0)
+; CHECK-NEXT:    v_readlane_b32 s4, v0, 4
 ; CHECK-NEXT:    s_mov_b32 exec_lo, s4
 ; CHECK-NEXT:  ; %bb.4:
-; CHECK-NEXT:    buffer_load_dword v0, off, s[0:3], s32 offset:40 ; 4-byte Folded Reload
+; CHECK-NEXT:    buffer_load_dword v0, off, s[0:3], s32 offset:44 ; 4-byte Folded Reload
 ; CHECK-NEXT:    ; implicit-def: $sgpr4
 ; CHECK-NEXT:    v_mov_b32_e32 v1, s4
 ; CHECK-NEXT:    v_mov_b32_e32 v2, s4
 ; CHECK-NEXT:    v_mov_b32_e32 v3, s4
 ; CHECK-NEXT:    s_xor_saveexec_b32 s4, -1
-; CHECK-NEXT:    buffer_load_dword v8, off, s[0:3], s32 offset:44 ; 4-byte Folded Reload
+; CHECK-NEXT:    buffer_load_dword v0, off, s[0:3], s32 offset:48 ; 4-byte Folded Reload
+; CHECK-NEXT:    buffer_load_dword v2, off, s[0:3], s32 offset:52 ; 4-byte Folded Reload
 ; CHECK-NEXT:    s_mov_b32 exec_lo, s4
 ; CHECK-NEXT:    s_waitcnt vmcnt(0)
 ; CHECK-NEXT:    s_waitcnt_vscnt null, 0x0
diff --git a/llvm/test/CodeGen/AMDGPU/GlobalISel/localizer.ll b/llvm/test/CodeGen/AMDGPU/GlobalISel/localizer.ll
--- a/llvm/test/CodeGen/AMDGPU/GlobalISel/localizer.ll
+++ b/llvm/test/CodeGen/AMDGPU/GlobalISel/localizer.ll
@@ -242,8 +242,9 @@
 ; GFX9-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX9-NEXT:    global_load_dword v0, v[0:1], off glc
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s16, 0
+; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX9-NEXT:    s_swappc_b64 s[30:31], 0
diff --git a/llvm/test/CodeGen/AMDGPU/abi-attribute-hints-undefined-behavior.ll b/llvm/test/CodeGen/AMDGPU/abi-attribute-hints-undefined-behavior.ll
--- a/llvm/test/CodeGen/AMDGPU/abi-attribute-hints-undefined-behavior.ll
+++ b/llvm/test/CodeGen/AMDGPU/abi-attribute-hints-undefined-behavior.ll
@@ -23,6 +23,7 @@
 ; FIXEDABI-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; FIXEDABI-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; FIXEDABI-NEXT:    s_mov_b64 exec, s[18:19]
+; FIXEDABI-NEXT:    ; implicit-def: $vgpr40
 ; FIXEDABI-NEXT:    s_addk_i32 s32, 0x400
 ; FIXEDABI-NEXT:    v_writelane_b32 v40, s30, 0
 ; FIXEDABI-NEXT:    v_writelane_b32 v41, s16, 0
diff --git a/llvm/test/CodeGen/AMDGPU/branch-relax-spill.ll b/llvm/test/CodeGen/AMDGPU/branch-relax-spill.ll
--- a/llvm/test/CodeGen/AMDGPU/branch-relax-spill.ll
+++ b/llvm/test/CodeGen/AMDGPU/branch-relax-spill.ll
@@ -902,6 +902,11 @@
 ; CHECK-NEXT:    buffer_store_dword v0, off, s[0:3], s32 ; 4-byte Folded Spill
 ; CHECK-NEXT:    buffer_store_dword v1, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; CHECK-NEXT:    s_mov_b64 exec, s[4:5]
+; CHECK-NEXT:    ; implicit-def: $vgpr0
+; CHECK-NEXT:    ; implicit-def: $vgpr1
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    s_mov_b32 s0, 0
+; CHECK-NEXT:    ;;#ASMEND
 ; CHECK-NEXT:    s_waitcnt expcnt(1)
 ; CHECK-NEXT:    v_writelane_b32 v0, s30, 0
 ; CHECK-NEXT:    v_writelane_b32 v0, s31, 1
@@ -978,9 +983,6 @@
 ; CHECK-NEXT:    v_writelane_b32 v1, s101, 6
 ; CHECK-NEXT:    s_cmp_eq_u32 s31, 0
 ; CHECK-NEXT:    ;;#ASMSTART
-; CHECK-NEXT:    s_mov_b32 s0, 0
-; CHECK-NEXT:    ;;#ASMEND
-; CHECK-NEXT:    ;;#ASMSTART
 ; CHECK-NEXT:    s_mov_b32 s1, 0
 ; CHECK-NEXT:    ;;#ASMEND
 ; CHECK-NEXT:    ;;#ASMSTART
diff --git a/llvm/test/CodeGen/AMDGPU/call-alias-register-usage-agpr.ll b/llvm/test/CodeGen/AMDGPU/call-alias-register-usage-agpr.ll
--- a/llvm/test/CodeGen/AMDGPU/call-alias-register-usage-agpr.ll
+++ b/llvm/test/CodeGen/AMDGPU/call-alias-register-usage-agpr.ll
@@ -8,12 +8,12 @@
 @alias = hidden alias void (), void ()* @aliasee_default
 
 ; ALL-LABEL: {{^}}kernel:
-; GFX908: .amdhsa_next_free_vgpr 41
+; GFX908: .amdhsa_next_free_vgpr 32
 ; GFX908-NEXT: .amdhsa_next_free_sgpr 33
 
-; GFX90A: .amdhsa_next_free_vgpr 71
+; GFX90A: .amdhsa_next_free_vgpr 59
 ; GFX90A-NEXT: .amdhsa_next_free_sgpr 33
-; GFX90A-NEXT: .amdhsa_accum_offset 44
+; GFX90A-NEXT: .amdhsa_accum_offset 32
 define amdgpu_kernel void @kernel() #0 {
 bb:
   call void @alias() #2
diff --git a/llvm/test/CodeGen/AMDGPU/call-alias-register-usage1.ll b/llvm/test/CodeGen/AMDGPU/call-alias-register-usage1.ll
--- a/llvm/test/CodeGen/AMDGPU/call-alias-register-usage1.ll
+++ b/llvm/test/CodeGen/AMDGPU/call-alias-register-usage1.ll
@@ -9,7 +9,7 @@
 ; The parent kernel has a higher VGPR usage than the possible callees.
 
 ; CHECK-LABEL: {{^}}kernel1:
-; CHECK: .amdhsa_next_free_vgpr 42
+; CHECK: .amdhsa_next_free_vgpr 41
 ; CHECK-NEXT: .amdhsa_next_free_sgpr 33
 define amdgpu_kernel void @kernel1() #0 {
 bb:
diff --git a/llvm/test/CodeGen/AMDGPU/callee-frame-setup.ll b/llvm/test/CodeGen/AMDGPU/callee-frame-setup.ll
--- a/llvm/test/CodeGen/AMDGPU/callee-frame-setup.ll
+++ b/llvm/test/CodeGen/AMDGPU/callee-frame-setup.ll
@@ -178,7 +178,7 @@
 ; restored. No FP is required.
 ;
 ; GCN-LABEL: {{^}}callee_func_sgpr_spill_no_calls:
-; GCN: s_or_saveexec_b64 [[COPY_EXEC0:s\[[0-9]+:[0-9]+\]]], -1{{$}}
+; GCN: s_xor_saveexec_b64 [[COPY_EXEC0:s\[[0-9]+:[0-9]+\]]], -1{{$}}
 ; MUBUF-NEXT:   buffer_store_dword [[CSR_VGPR:v[0-9]+]], off, s[0:3], s32 ; 4-byte Folded Spill
 ; FLATSCR-NEXT: scratch_store_dword off, [[CSR_VGPR:v[0-9]+]], s32 ; 4-byte Folded Spill
 ; GCN-NEXT: s_mov_b64 exec, [[COPY_EXEC0]]
@@ -189,7 +189,7 @@
 ; GCN: v_readlane_b32 s{{[0-9]+}}, [[CSR_VGPR]]
 ; GCN: v_readlane_b32 s{{[0-9]+}}, [[CSR_VGPR]]
 
-; GCN: s_or_saveexec_b64 [[COPY_EXEC1:s\[[0-9]+:[0-9]+\]]], -1{{$}}
+; GCN: s_xor_saveexec_b64 [[COPY_EXEC1:s\[[0-9]+:[0-9]+\]]], -1{{$}}
 ; MUBUF-NEXT:   buffer_load_dword [[CSR_VGPR]], off, s[0:3], s32 ; 4-byte Folded Reload
 ; FLATSCR-NEXT: scratch_load_dword [[CSR_VGPR]], off, s32 ; 4-byte Folded Reload
 ; GCN-NEXT: s_mov_b64 exec, [[COPY_EXEC1]]
@@ -227,6 +227,7 @@
 ; MUBUF-NEXT: buffer_store_dword v0, off, s[0:3], s32 ; 4-byte Folded Spill
 ; FLATSCR-NEXT: scratch_store_dword off, v0, s32 ; 4-byte Folded Spill
 ; GCN-NEXT: s_mov_b64 exec,
+; GCN-NEXT: ; implicit-def: $vgpr0
 ; GCN-NEXT: v_writelane_b32 v0, s42, 0
 ; GCN-NEXT: ;;#ASMSTART
 ; GCN-NEXT: ; clobber s42
@@ -400,8 +401,9 @@
 ; MUBUF-NEXT:   buffer_store_dword [[CSR_VGPR:v[0-9]+]], off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; FLATSCR-NEXT: scratch_store_dword off, [[CSR_VGPR:v[0-9]+]], s33 offset:4 ; 4-byte Folded Spill
 ; GCN-NEXT: s_mov_b64 exec, [[COPY_EXEC0]]
-; GCN: v_writelane_b32 [[CSR_VGPR]], s30, 0
+; GCN-NEXT: ; implicit-def: $vgpr0
 ; GCN: v_mov_b32_e32 [[ZERO:v[0-9]+]], 0
+; GCN: v_writelane_b32 [[CSR_VGPR]], s30, 0
 ; GCN: v_writelane_b32 [[CSR_VGPR]], s31, 1
 ; MUBUF:   buffer_store_dword [[ZERO]], off, s[0:3], s33{{$}}
 ; FLATSCR: scratch_store_dword off, [[ZERO]], s33{{$}}
@@ -439,10 +441,11 @@
 ; GCN: s_waitcnt
 ; GCN-NEXT: s_mov_b32 vcc_lo, s33
 ; GCN-NEXT: s_mov_b32 s33, s32
-; GCN-NEXT: s_or_saveexec_b64 [[COPY_EXEC0:s\[[0-9]+:[0-9]+\]]], -1{{$}}
+; GCN-NEXT: s_xor_saveexec_b64 [[COPY_EXEC0:s\[[0-9]+:[0-9]+\]]], -1{{$}}
 ; MUBUF-NEXT:   buffer_store_dword [[CSR_VGPR:v[0-9]+]], off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; FLATSCR-NEXT: scratch_store_dword off, [[CSR_VGPR:v[0-9]+]], s33 offset:4 ; 4-byte Folded Spill
 ; GCN-NEXT: s_mov_b64 exec, [[COPY_EXEC0]]
+; GCN-NEXT: ; implicit-def: $vgpr48
 
 ; MUBUF-DAG:   buffer_store_dword
 ; FLATSCR-DAG: scratch_store_dword
@@ -450,7 +453,7 @@
 ; FLATSCR:     s_add_i32 s32, s32, 12{{$}}
 
 ; GCN: ;;#ASMSTART
-; GCN: s_or_saveexec_b64 [[COPY_EXEC1:s\[[0-9]+:[0-9]+\]]], -1{{$}}
+; GCN: s_xor_saveexec_b64 [[COPY_EXEC1:s\[[0-9]+:[0-9]+\]]], -1{{$}}
 ; MUBUF-NEXT:   buffer_load_dword [[CSR_VGPR]], off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
 ; FLATSCR-NEXT: scratch_load_dword [[CSR_VGPR]], off, s33 offset:4 ; 4-byte Folded Reload
 ; GCN-NEXT: s_mov_b64 exec, [[COPY_EXEC1]]
@@ -485,7 +488,7 @@
 ; GCN: s_waitcnt
 ; GCN-NEXT: s_mov_b32 vcc_lo, s33
 ; GCN-DAG:  s_mov_b32 s33, s32
-; GCN-NEXT: s_or_saveexec_b64 [[COPY_EXEC0:s\[[0-9]+:[0-9]+\]]], -1{{$}}
+; GCN-NEXT: s_xor_saveexec_b64 [[COPY_EXEC0:s\[[0-9]+:[0-9]+\]]], -1{{$}}
 ; MUBUF-NEXT: s_add_i32 [[SCRATCH_SGPR:s[0-9]+]], s33, 0x40100
 ; FLATSCR-NEXT: s_add_i32 [[SCRATCH_SGPR:s[0-9]+]], s33, 0x1004
 ; MUBUF-NEXT: buffer_store_dword [[CSR_VGPR:v[0-9]+]], off, s[0:3], [[SCRATCH_SGPR]] ; 4-byte Folded Spill
@@ -497,7 +500,7 @@
 ; FLATSCR-DAG: scratch_store_dword
 
 ; GCN: ;;#ASMSTART
-; GCN: s_or_saveexec_b64 [[COPY_EXEC1:s\[[0-9]+:[0-9]+\]]], -1{{$}}
+; GCN: s_xor_saveexec_b64 [[COPY_EXEC1:s\[[0-9]+:[0-9]+\]]], -1{{$}}
 ; MUBUF-NEXT: s_add_i32 [[SCRATCH_SGPR:s[0-9]+]], s33, 0x40100
 ; MUBUF-NEXT: buffer_load_dword [[CSR_VGPR]], off, s[0:3], [[SCRATCH_SGPR]] ; 4-byte Folded Reload
 ; FLATSCR-NEXT: s_add_i32 [[SCRATCH_SGPR:s[0-9]+]], s33, 0x1004
@@ -591,7 +594,7 @@
 ; VGPR.
 ; GCN-LABEL: {{^}}callee_need_to_spill_fp_to_memory_full_reserved_vgpr:
 ; MUBUF:   s_mov_b32 [[FP_SCRATCH_COPY:s[0-9]+]], s33
-; FLATSCR: s_mov_b32 s33, s0
+; FLATSCR: s_mov_b32 s33, s2
 ; MUBUF:   s_mov_b32 s33, s32
 ; MUBUF:   s_xor_saveexec_b64 [[COPY_EXEC1:s\[[0-9]+:[0-9]+\]]], -1{{$}}
 ; MUBUF:   s_mov_b64 exec, [[COPY_EXEC1]]
@@ -632,14 +635,14 @@
 ; Make sure that the FP save happens after restoring exec from the same
 ; register.
 ; GCN-LABEL: {{^}}callee_need_to_spill_fp_to_reg:
-; FLATSCR: s_mov_b32 s0, s33
+; FLATSCR: s_mov_b32 [[FP_SCRATCH_COPY:s[0-9]+]], s33
 ; FLATSCR: s_mov_b32 s33, s32
 ; GCN-NOT: v_writelane_b32 v40, s33
-; FLATSCR: s_or_saveexec_b64 s[2:3], -1
-; FLATSCR: s_mov_b64 exec, s[2:3]
-; FLATSCR: s_or_saveexec_b64 s[2:3], -1
+; FLATSCR: s_xor_saveexec_b64 [[COPY_EXEC0:s\[[0-9]+:[0-9]+\]]], -1{{$}}
+; FLATSCR: s_mov_b64 exec, [[COPY_EXEC0]]
+; FLATSCR: s_xor_saveexec_b64 [[COPY_EXEC1:s\[[0-9]+:[0-9]+\]]], -1{{$}}
 ; GCN-NOT: v_readlane_b32 s33, v40
-; FLATSCR: s_mov_b32 s33, s0
+; FLATSCR: s_mov_b32 s33, [[FP_SCRATCH_COPY]]
 ; GCN:     s_setpc_b64
 define void @callee_need_to_spill_fp_to_reg() #1 {
   call void asm sideeffect "; clobber nonpreserved SGPRs and 64 CSRs",
@@ -672,7 +675,7 @@
 ; MUBUF-NEXT: buffer_store_dword v39, off, s[0:3], [[SCRATCH_SGPR]] ; 4-byte Folded Spill
 ; MUBUF: v_mov_b32_e32 v0, [[FP_SCRATCH_COPY]]
 ; GCN-NOT: v_mov_b32_e32 v0, 0x100c
-; MUBUF-NEXT: s_add_i32 [[SCRATCH_SGPR:s[0-9]+]], s33, 0x40200
+; MUBUF: s_add_i32 [[SCRATCH_SGPR:s[0-9]+]], s33, 0x40200
 ; MUBUF: buffer_store_dword v0, off, s[0:3], [[SCRATCH_SGPR]] ; 4-byte Folded Spill
 ; FLATSCR: v_mov_b32_e32 v0, 0
 ; FLATSCR: s_add_i32 [[SOFF:s[0-9]+]], s33, 0x1000
diff --git a/llvm/test/CodeGen/AMDGPU/cf-loop-on-constant.ll b/llvm/test/CodeGen/AMDGPU/cf-loop-on-constant.ll
--- a/llvm/test/CodeGen/AMDGPU/cf-loop-on-constant.ll
+++ b/llvm/test/CodeGen/AMDGPU/cf-loop-on-constant.ll
@@ -30,7 +30,14 @@
 ;
 ; GCN_DBG-LABEL: test_loop:
 ; GCN_DBG:       ; %bb.0: ; %entry
+; GCN_DBG-NEXT:    s_mov_b32 s8, SCRATCH_RSRC_DWORD0
+; GCN_DBG-NEXT:    s_mov_b32 s9, SCRATCH_RSRC_DWORD1
+; GCN_DBG-NEXT:    s_mov_b32 s10, -1
+; GCN_DBG-NEXT:    s_mov_b32 s11, 0xe8f000
+; GCN_DBG-NEXT:    s_add_u32 s8, s8, s3
+; GCN_DBG-NEXT:    s_addc_u32 s9, s9, 0
 ; GCN_DBG-NEXT:    s_load_dword s2, s[0:1], 0x9
+; GCN_DBG-NEXT:    ; implicit-def: $vgpr0
 ; GCN_DBG-NEXT:    s_waitcnt lgkmcnt(0)
 ; GCN_DBG-NEXT:    v_writelane_b32 v0, s2, 0
 ; GCN_DBG-NEXT:    s_load_dword s1, s[0:1], 0xa
@@ -39,11 +46,20 @@
 ; GCN_DBG-NEXT:    s_waitcnt lgkmcnt(0)
 ; GCN_DBG-NEXT:    s_cmp_lg_u32 s1, s2
 ; GCN_DBG-NEXT:    v_writelane_b32 v0, s0, 1
+; GCN_DBG-NEXT:    s_mov_b64 s[4:5], exec
+; GCN_DBG-NEXT:    s_mov_b64 exec, -1
+; GCN_DBG-NEXT:    buffer_store_dword v0, off, s[8:11], 0 offset:4 ; 4-byte Folded Spill
+; GCN_DBG-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN_DBG-NEXT:    s_cbranch_scc1 .LBB0_2
 ; GCN_DBG-NEXT:  ; %bb.1: ; %for.exit
 ; GCN_DBG-NEXT:    s_endpgm
 ; GCN_DBG-NEXT:  .LBB0_2: ; %for.body
 ; GCN_DBG-NEXT:    ; =>This Inner Loop Header: Depth=1
+; GCN_DBG-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GCN_DBG-NEXT:    s_waitcnt expcnt(0)
+; GCN_DBG-NEXT:    buffer_load_dword v0, off, s[8:11], 0 offset:4 ; 4-byte Folded Reload
+; GCN_DBG-NEXT:    s_mov_b64 exec, s[4:5]
+; GCN_DBG-NEXT:    s_waitcnt vmcnt(0)
 ; GCN_DBG-NEXT:    v_readlane_b32 s0, v0, 1
 ; GCN_DBG-NEXT:    v_readlane_b32 s2, v0, 0
 ; GCN_DBG-NEXT:    s_mov_b32 s1, 2
@@ -65,6 +81,9 @@
 ; GCN_DBG-NEXT:    s_mov_b64 s[2:3], -1
 ; GCN_DBG-NEXT:    s_and_b64 vcc, exec, s[2:3]
 ; GCN_DBG-NEXT:    v_writelane_b32 v0, s0, 1
+; GCN_DBG-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GCN_DBG-NEXT:    buffer_store_dword v0, off, s[8:11], 0 offset:4 ; 4-byte Folded Spill
+; GCN_DBG-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN_DBG-NEXT:    s_cbranch_vccnz .LBB0_2
 ; GCN_DBG-NEXT:  ; %bb.3: ; %DummyReturnBlock
 ; GCN_DBG-NEXT:    s_endpgm
@@ -105,16 +124,31 @@
 ;
 ; GCN_DBG-LABEL: loop_const_true:
 ; GCN_DBG:       ; %bb.0: ; %entry
+; GCN_DBG-NEXT:    s_mov_b32 s8, SCRATCH_RSRC_DWORD0
+; GCN_DBG-NEXT:    s_mov_b32 s9, SCRATCH_RSRC_DWORD1
+; GCN_DBG-NEXT:    s_mov_b32 s10, -1
+; GCN_DBG-NEXT:    s_mov_b32 s11, 0xe8f000
+; GCN_DBG-NEXT:    s_add_u32 s8, s8, s3
+; GCN_DBG-NEXT:    s_addc_u32 s9, s9, 0
 ; GCN_DBG-NEXT:    s_load_dword s0, s[0:1], 0x9
+; GCN_DBG-NEXT:    ; implicit-def: $vgpr0
 ; GCN_DBG-NEXT:    s_waitcnt lgkmcnt(0)
 ; GCN_DBG-NEXT:    v_writelane_b32 v0, s0, 0
 ; GCN_DBG-NEXT:    s_mov_b32 s0, 0
 ; GCN_DBG-NEXT:    v_writelane_b32 v0, s0, 1
+; GCN_DBG-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GCN_DBG-NEXT:    buffer_store_dword v0, off, s[8:11], 0 offset:4 ; 4-byte Folded Spill
+; GCN_DBG-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN_DBG-NEXT:    s_branch .LBB1_2
 ; GCN_DBG-NEXT:  .LBB1_1: ; %for.exit
 ; GCN_DBG-NEXT:    s_endpgm
 ; GCN_DBG-NEXT:  .LBB1_2: ; %for.body
 ; GCN_DBG-NEXT:    ; =>This Inner Loop Header: Depth=1
+; GCN_DBG-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GCN_DBG-NEXT:    s_waitcnt expcnt(0)
+; GCN_DBG-NEXT:    buffer_load_dword v0, off, s[8:11], 0 offset:4 ; 4-byte Folded Reload
+; GCN_DBG-NEXT:    s_mov_b64 exec, s[4:5]
+; GCN_DBG-NEXT:    s_waitcnt vmcnt(0)
 ; GCN_DBG-NEXT:    v_readlane_b32 s0, v0, 1
 ; GCN_DBG-NEXT:    v_readlane_b32 s2, v0, 0
 ; GCN_DBG-NEXT:    s_mov_b32 s1, 2
@@ -136,6 +170,9 @@
 ; GCN_DBG-NEXT:    s_mov_b64 s[2:3], 0
 ; GCN_DBG-NEXT:    s_and_b64 vcc, exec, s[2:3]
 ; GCN_DBG-NEXT:    v_writelane_b32 v0, s0, 1
+; GCN_DBG-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GCN_DBG-NEXT:    buffer_store_dword v0, off, s[8:11], 0 offset:4 ; 4-byte Folded Spill
+; GCN_DBG-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN_DBG-NEXT:    s_cbranch_vccnz .LBB1_1
 ; GCN_DBG-NEXT:    s_branch .LBB1_2
 entry:
@@ -170,16 +207,31 @@
 ;
 ; GCN_DBG-LABEL: loop_const_false:
 ; GCN_DBG:       ; %bb.0: ; %entry
+; GCN_DBG-NEXT:    s_mov_b32 s8, SCRATCH_RSRC_DWORD0
+; GCN_DBG-NEXT:    s_mov_b32 s9, SCRATCH_RSRC_DWORD1
+; GCN_DBG-NEXT:    s_mov_b32 s10, -1
+; GCN_DBG-NEXT:    s_mov_b32 s11, 0xe8f000
+; GCN_DBG-NEXT:    s_add_u32 s8, s8, s3
+; GCN_DBG-NEXT:    s_addc_u32 s9, s9, 0
 ; GCN_DBG-NEXT:    s_load_dword s0, s[0:1], 0x9
+; GCN_DBG-NEXT:    ; implicit-def: $vgpr0
 ; GCN_DBG-NEXT:    s_waitcnt lgkmcnt(0)
 ; GCN_DBG-NEXT:    v_writelane_b32 v0, s0, 0
 ; GCN_DBG-NEXT:    s_mov_b32 s0, 0
 ; GCN_DBG-NEXT:    v_writelane_b32 v0, s0, 1
+; GCN_DBG-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GCN_DBG-NEXT:    buffer_store_dword v0, off, s[8:11], 0 offset:4 ; 4-byte Folded Spill
+; GCN_DBG-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN_DBG-NEXT:    s_branch .LBB2_2
 ; GCN_DBG-NEXT:  .LBB2_1: ; %for.exit
 ; GCN_DBG-NEXT:    s_endpgm
 ; GCN_DBG-NEXT:  .LBB2_2: ; %for.body
 ; GCN_DBG-NEXT:    ; =>This Inner Loop Header: Depth=1
+; GCN_DBG-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GCN_DBG-NEXT:    s_waitcnt expcnt(0)
+; GCN_DBG-NEXT:    buffer_load_dword v0, off, s[8:11], 0 offset:4 ; 4-byte Folded Reload
+; GCN_DBG-NEXT:    s_mov_b64 exec, s[4:5]
+; GCN_DBG-NEXT:    s_waitcnt vmcnt(0)
 ; GCN_DBG-NEXT:    v_readlane_b32 s0, v0, 1
 ; GCN_DBG-NEXT:    v_readlane_b32 s2, v0, 0
 ; GCN_DBG-NEXT:    s_mov_b32 s1, 2
@@ -201,6 +253,9 @@
 ; GCN_DBG-NEXT:    s_mov_b64 s[2:3], -1
 ; GCN_DBG-NEXT:    s_and_b64 vcc, exec, s[2:3]
 ; GCN_DBG-NEXT:    v_writelane_b32 v0, s0, 1
+; GCN_DBG-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GCN_DBG-NEXT:    buffer_store_dword v0, off, s[8:11], 0 offset:4 ; 4-byte Folded Spill
+; GCN_DBG-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN_DBG-NEXT:    s_cbranch_vccnz .LBB2_1
 ; GCN_DBG-NEXT:    s_branch .LBB2_2
 entry:
@@ -236,16 +291,31 @@
 ;
 ; GCN_DBG-LABEL: loop_const_undef:
 ; GCN_DBG:       ; %bb.0: ; %entry
+; GCN_DBG-NEXT:    s_mov_b32 s8, SCRATCH_RSRC_DWORD0
+; GCN_DBG-NEXT:    s_mov_b32 s9, SCRATCH_RSRC_DWORD1
+; GCN_DBG-NEXT:    s_mov_b32 s10, -1
+; GCN_DBG-NEXT:    s_mov_b32 s11, 0xe8f000
+; GCN_DBG-NEXT:    s_add_u32 s8, s8, s3
+; GCN_DBG-NEXT:    s_addc_u32 s9, s9, 0
 ; GCN_DBG-NEXT:    s_load_dword s0, s[0:1], 0x9
+; GCN_DBG-NEXT:    ; implicit-def: $vgpr0
 ; GCN_DBG-NEXT:    s_waitcnt lgkmcnt(0)
 ; GCN_DBG-NEXT:    v_writelane_b32 v0, s0, 0
 ; GCN_DBG-NEXT:    s_mov_b32 s0, 0
 ; GCN_DBG-NEXT:    v_writelane_b32 v0, s0, 1
+; GCN_DBG-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GCN_DBG-NEXT:    buffer_store_dword v0, off, s[8:11], 0 offset:4 ; 4-byte Folded Spill
+; GCN_DBG-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN_DBG-NEXT:    s_branch .LBB3_2
 ; GCN_DBG-NEXT:  .LBB3_1: ; %for.exit
 ; GCN_DBG-NEXT:    s_endpgm
 ; GCN_DBG-NEXT:  .LBB3_2: ; %for.body
 ; GCN_DBG-NEXT:    ; =>This Inner Loop Header: Depth=1
+; GCN_DBG-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GCN_DBG-NEXT:    s_waitcnt expcnt(0)
+; GCN_DBG-NEXT:    buffer_load_dword v0, off, s[8:11], 0 offset:4 ; 4-byte Folded Reload
+; GCN_DBG-NEXT:    s_mov_b64 exec, s[4:5]
+; GCN_DBG-NEXT:    s_waitcnt vmcnt(0)
 ; GCN_DBG-NEXT:    v_readlane_b32 s0, v0, 1
 ; GCN_DBG-NEXT:    v_readlane_b32 s2, v0, 0
 ; GCN_DBG-NEXT:    s_mov_b32 s1, 2
@@ -265,6 +335,9 @@
 ; GCN_DBG-NEXT:    s_mov_b32 s1, 1
 ; GCN_DBG-NEXT:    s_add_i32 s0, s0, s1
 ; GCN_DBG-NEXT:    v_writelane_b32 v0, s0, 1
+; GCN_DBG-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GCN_DBG-NEXT:    buffer_store_dword v0, off, s[8:11], 0 offset:4 ; 4-byte Folded Spill
+; GCN_DBG-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN_DBG-NEXT:    s_cbranch_scc1 .LBB3_1
 ; GCN_DBG-NEXT:    s_branch .LBB3_2
 entry:
@@ -314,7 +387,14 @@
 ;
 ; GCN_DBG-LABEL: loop_arg_0:
 ; GCN_DBG:       ; %bb.0: ; %entry
+; GCN_DBG-NEXT:    s_mov_b32 s8, SCRATCH_RSRC_DWORD0
+; GCN_DBG-NEXT:    s_mov_b32 s9, SCRATCH_RSRC_DWORD1
+; GCN_DBG-NEXT:    s_mov_b32 s10, -1
+; GCN_DBG-NEXT:    s_mov_b32 s11, 0xe8f000
+; GCN_DBG-NEXT:    s_add_u32 s8, s8, s3
+; GCN_DBG-NEXT:    s_addc_u32 s9, s9, 0
 ; GCN_DBG-NEXT:    s_load_dword s0, s[0:1], 0x9
+; GCN_DBG-NEXT:    ; implicit-def: $vgpr0
 ; GCN_DBG-NEXT:    s_waitcnt lgkmcnt(0)
 ; GCN_DBG-NEXT:    v_writelane_b32 v0, s0, 0
 ; GCN_DBG-NEXT:    v_mov_b32_e32 v1, 0
@@ -331,11 +411,19 @@
 ; GCN_DBG-NEXT:    v_writelane_b32 v0, s1, 2
 ; GCN_DBG-NEXT:    s_mov_b32 s0, 0
 ; GCN_DBG-NEXT:    v_writelane_b32 v0, s0, 3
+; GCN_DBG-NEXT:    s_or_saveexec_b64 s[6:7], -1
+; GCN_DBG-NEXT:    buffer_store_dword v0, off, s[8:11], 0 offset:4 ; 4-byte Folded Spill
+; GCN_DBG-NEXT:    s_mov_b64 exec, s[6:7]
 ; GCN_DBG-NEXT:    s_branch .LBB4_2
 ; GCN_DBG-NEXT:  .LBB4_1: ; %for.exit
 ; GCN_DBG-NEXT:    s_endpgm
 ; GCN_DBG-NEXT:  .LBB4_2: ; %for.body
 ; GCN_DBG-NEXT:    ; =>This Inner Loop Header: Depth=1
+; GCN_DBG-NEXT:    s_or_saveexec_b64 s[6:7], -1
+; GCN_DBG-NEXT:    s_waitcnt expcnt(0)
+; GCN_DBG-NEXT:    buffer_load_dword v0, off, s[8:11], 0 offset:4 ; 4-byte Folded Reload
+; GCN_DBG-NEXT:    s_mov_b64 exec, s[6:7]
+; GCN_DBG-NEXT:    s_waitcnt vmcnt(0)
 ; GCN_DBG-NEXT:    v_readlane_b32 s0, v0, 3
 ; GCN_DBG-NEXT:    v_readlane_b32 s2, v0, 1
 ; GCN_DBG-NEXT:    v_readlane_b32 s3, v0, 2
@@ -358,6 +446,9 @@
 ; GCN_DBG-NEXT:    s_add_i32 s0, s0, s1
 ; GCN_DBG-NEXT:    s_and_b64 vcc, exec, s[2:3]
 ; GCN_DBG-NEXT:    v_writelane_b32 v0, s0, 3
+; GCN_DBG-NEXT:    s_or_saveexec_b64 s[6:7], -1
+; GCN_DBG-NEXT:    buffer_store_dword v0, off, s[8:11], 0 offset:4 ; 4-byte Folded Spill
+; GCN_DBG-NEXT:    s_mov_b64 exec, s[6:7]
 ; GCN_DBG-NEXT:    s_cbranch_vccnz .LBB4_1
 ; GCN_DBG-NEXT:    s_branch .LBB4_2
 entry:
diff --git a/llvm/test/CodeGen/AMDGPU/collapse-endcf.ll b/llvm/test/CodeGen/AMDGPU/collapse-endcf.ll
--- a/llvm/test/CodeGen/AMDGPU/collapse-endcf.ll
+++ b/llvm/test/CodeGen/AMDGPU/collapse-endcf.ll
@@ -19,14 +19,14 @@
 ; GCN-O0:      s_mov_b64 s[{{[0-9:]+}}], exec
 ; GCN-O0-DAG:  v_writelane_b32 [[VGPR:v[0-9]+]], s{{[0-9]+}}, [[OUTER_SPILL_LANE_0:[0-9]+]]
 ; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[OUTER_SPILL_LANE_1:[0-9]+]]
-; GCN-O0-NEXT: s_and_b64 s[{{[0-9:]+}}], s[{{[0-9:]+}}], s[{{[0-9:]+}}]
+; GCN-O0-DAG: s_and_b64 s[{{[0-9:]+}}], s[{{[0-9:]+}}], s[{{[0-9:]+}}]
 ; GCN-O0-NEXT: s_mov_b64 exec, s[{{[0-9:]+}}]
 ; GCN-O0-NEXT: s_cbranch_execz [[ENDIF_OUTER:.LBB[0-9_]+]]
 ; GCN-O0-NEXT: ; %bb.{{[0-9]+}}:
 ; GCN-O0:      s_mov_b64 s[{{[0-9:]+}}], exec
 ; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[INNER_SPILL_LANE_0:[0-9]+]]
 ; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[INNER_SPILL_LANE_1:[0-9]+]]
-; GCN-O0-NEXT: s_and_b64 s[{{[0-9:]+}}], s[{{[0-9:]+}}], s[{{[0-9:]+}}]
+; GCN-O0-DAG: s_and_b64 s[{{[0-9:]+}}], s[{{[0-9:]+}}], s[{{[0-9:]+}}]
 ; GCN-O0-NEXT: s_mov_b64 exec, s[{{[0-9:]+}}]
 ; GCN-O0-NEXT: s_cbranch_execz [[ENDIF_INNER:.LBB[0-9_]+]]
 ; GCN-O0-NEXT: ; %bb.{{[0-9]+}}:
@@ -84,14 +84,14 @@
 ; GCN-O0:      s_mov_b64 s[{{[0-9:]+}}], exec
 ; GCN-O0-DAG:  v_writelane_b32 [[VGPR:v[0-9]+]], s{{[0-9]+}}, [[OUTER_SPILL_LANE_0:[0-9]+]]
 ; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[OUTER_SPILL_LANE_1:[0-9]+]]
-; GCN-O0-NEXT: s_and_b64 s[{{[0-9:]+}}]
+; GCN-O0-DAG: s_and_b64 s[{{[0-9:]+}}]
 ; GCN-O0-NEXT: s_mov_b64 exec, s[{{[0-9:]+}}]
 ; GCN-O0-NEXT: s_cbranch_execz [[ENDIF_OUTER:.LBB[0-9_]+]]
 ; GCN-O0-NEXT: ; %bb.{{[0-9]+}}:
 ; GCN-O0:      s_mov_b64 s[{{[0-9:]+}}], exec
 ; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[INNER_SPILL_LANE_0:[0-9]+]]
 ; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[INNER_SPILL_LANE_1:[0-9]+]]
-; GCN-O0-NEXT: s_and_b64 s[{{[0-9:]+}}]
+; GCN-O0-DAG: s_and_b64 s[{{[0-9:]+}}]
 ; GCN-O0-NEXT: s_mov_b64 exec, s[{{[0-9:]+}}]
 ; GCN-O0-NEXT: s_cbranch_execz [[ENDIF_INNER:.LBB[0-9_]+]]
 ; GCN-O0-NEXT: ; %bb.{{[0-9]+}}:
@@ -161,7 +161,7 @@
 ; GCN-O0:      s_mov_b64 s[{{[0-9:]+}}], exec
 ; GCN-O0-DAG:  v_writelane_b32 [[VGPR:v[0-9]+]], s{{[0-9]+}}, [[OUTER_SPILL_LANE_0:[0-9]+]]
 ; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[OUTER_SPILL_LANE_1:[0-9]+]]
-; GCN-O0-NEXT: s_and_b64 s[{{[0-9:]+}}]
+; GCN-O0-DAG: s_and_b64 s[{{[0-9:]+}}]
 ; GCN-O0-NEXT: s_mov_b64 exec, s[{{[0-9:]+}}]
 ; GCN-O0-NEXT: s_cbranch_execz [[ENDIF_OUTER:.LBB[0-9_]+]]
 ; GCN-O0-NEXT: ; %bb.{{[0-9]+}}:
@@ -170,6 +170,9 @@
 ; GCN-O0-NEXT: s_xor_b64 s[{{[0-9:]+}}], s[{{[0-9:]+}}], s[{{[0-9:]+}}]
 ; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[THEN_SPILL_LANE_0:[0-9]+]]
 ; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[THEN_SPILL_LANE_1:[0-9]+]]
+; GCN-O0-NEXT: s_or_saveexec_b64 [[EXEC_COPY:s\[[0-9]+:[0-9]+\]]], -1
+; GCN-O0-NEXT: buffer_store_dword
+; GCN-O0-NEXT: s_mov_b64 exec, [[EXEC_COPY]]
 ; GCN-O0-NEXT: s_mov_b64 exec, s[{{[0-9:]+}}]
 ; GCN-O0-NEXT: s_cbranch_execz [[THEN_INNER:.LBB[0-9_]+]]
 ; GCN-O0-NEXT: s_branch [[TEMP_BB:.LBB[0-9_]+]]
@@ -180,7 +183,7 @@
 ; GCN-O0-NEXT: s_and_b64 s[{{[0-9:]+}}], exec, s[{{[0-9:]+}}]
 ; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[INNER_SPILL_LANE_0:[0-9]+]]
 ; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[INNER_SPILL_LANE_1:[0-9]+]]
-; GCN-O0-NEXT: s_xor_b64 exec, exec, s[{{[0-9:]+}}]
+; GCN-O0-DAG: s_xor_b64 exec, exec, s[{{[0-9:]+}}]
 ; GCN-O0-NEXT: s_cbranch_execz [[ENDIF_INNER:.LBB[0-9_]+]]
 ; GCN-O0-NEXT: ; %bb.{{[0-9]+}}:
 ; GCN-O0:      store_dword
@@ -260,6 +263,9 @@
 ; GCN-O0-NEXT: s_xor_b64 s[{{[0-9:]+}}], s[{{[0-9:]+}}], s[{{[0-9:]+}}]
 ; GCN-O0-DAG:  v_writelane_b32 [[VGPR:v[0-9]+]], s{{[0-9]+}}, [[OUTER_SPILL_LANE_0:[0-9]+]]
 ; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[OUTER_SPILL_LANE_1:[0-9]+]]
+; GCN-O0-NEXT: s_or_saveexec_b64 [[EXEC_COPY:s\[[0-9]+:[0-9]+\]]], -1
+; GCN-O0-NEXT: buffer_store_dword [[VGPR]]
+; GCN-O0-NEXT: s_mov_b64 exec, [[EXEC_COPY]]
 ; GCN-O0-NEXT: s_mov_b64 exec, s[{{[0-9:]+}}]
 ; GCN-O0-NEXT: s_cbranch_execz [[THEN_OUTER:.LBB[0-9_]+]]
 ; GCN-O0-NEXT: s_branch [[INNER_IF_OUTER_ELSE:.LBB[0-9_]+]]
@@ -270,14 +276,14 @@
 ; GCN-O0-NEXT: s_and_b64 s[{{[0-9:]+}}], exec, s[{{[0-9:]+}}]
 ; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[OUTER_2_SPILL_LANE_0:[0-9]+]]
 ; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[OUTER_2_SPILL_LANE_1:[0-9]+]]
-; GCN-O0-NEXT: s_xor_b64 exec, exec, s[{{[0-9:]+}}]
+; GCN-O0-DAG: s_xor_b64 exec, exec, s[{{[0-9:]+}}]
 ; GCN-O0-NEXT: s_cbranch_execz [[ENDIF_OUTER:.LBB[0-9_]+]]
 ; GCN-O0-NEXT: ; %bb.{{[0-9]+}}:
 ; GCN-O0:      store_dword
 ; GCN-O0:      s_mov_b64 s[{{[0-9:]+}}], exec
 ; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[ELSE_SPILL_LANE_0:[0-9]+]]
 ; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[ELSE_SPILL_LANE_1:[0-9]+]]
-; GCN-O0-NEXT: s_and_b64 s[{{[0-9:]+}}], s[{{[0-9:]+}}], s[{{[0-9:]+}}]
+; GCN-O0-DAG: s_and_b64 s[{{[0-9:]+}}], s[{{[0-9:]+}}], s[{{[0-9:]+}}]
 ; GCN-O0-NEXT: s_mov_b64 exec, s[{{[0-9:]+}}]
 ; GCN-O0-NEXT: s_cbranch_execz [[FLOW1:.LBB[0-9_]+]]
 ; GCN-O0-NEXT: ; %bb.{{[0-9]+}}:
@@ -287,7 +293,7 @@
 ; GCN-O0:      s_mov_b64 s[{{[0-9:]+}}], exec
 ; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[INNER_IF_OUTER_ELSE_SPILL_LANE_0:[0-9]+]]
 ; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[INNER_IF_OUTER_ELSE_SPILL_LANE_1:[0-9]+]]
-; GCN-O0-NEXT: s_and_b64 s[{{[0-9:]+}}], s[{{[0-9:]+}}], s[{{[0-9:]+}}]
+; GCN-O0-DAG: s_and_b64 s[{{[0-9:]+}}], s[{{[0-9:]+}}], s[{{[0-9:]+}}]
 ; GCN-O0-NEXT: s_mov_b64 exec, s[{{[0-9:]+}}]
 ; GCN-O0-NEXT: s_cbranch_execz [[THEN_OUTER_FLOW:.LBB[0-9_]+]]
 ; GCN-O0-NEXT: ; %bb.{{[0-9]+}}:
@@ -357,7 +363,7 @@
 ; GCN-O0:      s_mov_b64 s[{{[0-9:]+}}], exec
 ; GCN-O0-DAG:  v_writelane_b32 [[VGPR:v[0-9]+]], s{{[0-9]+}}, [[SPILL_LANE_0:[0-9]+]]
 ; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[SPILL_LANE_1:[0-9]+]]
-; GCN-O0-NEXT: s_and_b64 s[{{[0-9:]+}}]
+; GCN-O0-DAG: s_and_b64 s[{{[0-9:]+}}]
 ; GCN-O0-NEXT: s_mov_b64 exec, s[{{[0-9:]+}}]
 ; GCN-O0-NEXT: s_cbranch_execz [[ENDIF:.LBB[0-9_]+]]
 ; GCN-O0-NEXT: ; %bb.{{[0-9]+}}:
@@ -419,61 +425,81 @@
 ; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[INNER_LOOP_IN_EXEC_SPILL_LANE_1:[0-9]+]]
 ; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[INNER_LOOP_BACK_EDGE_EXEC_SPILL_LANE_0:[0-9]+]]
 ; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[INNER_LOOP_BACK_EDGE_EXEC_SPILL_LANE_1:[0-9]+]]
+; GCN-O0-NEXT: s_or_saveexec_b64 [[EXEC_COPY:s\[[0-9]+:[0-9]+\]]], -1
+; GCN-O0-NEXT: buffer_store_dword [[VGPR]], off, s{{\[[0-9]+:[0-9]+\]}}, s32 ; 4-byte Folded Spill
+; GCN-O0-NEXT: s_mov_b64 exec, [[EXEC_COPY]]
 ; GCN-O0: [[INNER_LOOP:.LBB[0-9]+_[0-9]+]]:
-; GCN-O0: buffer_load_dword
-; GCN-O0-DAG:  v_readlane_b32 s{{[0-9]+}}, [[VGPR]], [[INNER_LOOP_BACK_EDGE_EXEC_SPILL_LANE_0]]
-; GCN-O0-DAG:  v_readlane_b32 s{{[0-9]+}}, [[VGPR]], [[INNER_LOOP_BACK_EDGE_EXEC_SPILL_LANE_1]]
-; GCN-O0-DAG:  v_readlane_b32 s{{[0-9]+}}, [[VGPR]], [[INNER_LOOP_IN_EXEC_SPILL_LANE_0]]
-; GCN-O0-DAG:  v_readlane_b32 s{{[0-9]+}}, [[VGPR]], [[INNER_LOOP_IN_EXEC_SPILL_LANE_1]]
-; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[OUTER_LOOP_EXEC_SPILL_LANE_0:[0-9]+]]
-; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[OUTER_LOOP_EXEC_SPILL_LANE_1:[0-9]+]]
+; GCN-O0:      buffer_load_dword [[RESTORED_VGPR:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, s32 ; 4-byte Folded Reload
+; GCN-O0-DAG:  v_readlane_b32 s{{[0-9]+}}, [[RESTORED_VGPR]], [[INNER_LOOP_BACK_EDGE_EXEC_SPILL_LANE_0]]
+; GCN-O0-DAG:  v_readlane_b32 s{{[0-9]+}}, [[RESTORED_VGPR]], [[INNER_LOOP_BACK_EDGE_EXEC_SPILL_LANE_1]]
+; GCN-O0-DAG:  v_readlane_b32 s{{[0-9]+}}, [[RESTORED_VGPR]], [[INNER_LOOP_IN_EXEC_SPILL_LANE_0]]
+; GCN-O0-DAG:  v_readlane_b32 s{{[0-9]+}}, [[RESTORED_VGPR]], [[INNER_LOOP_IN_EXEC_SPILL_LANE_1]]
+; GCN-O0-DAG:  v_writelane_b32 [[RESTORED_VGPR]], s{{[0-9]+}}, [[OUTER_LOOP_EXEC_SPILL_LANE_0:[0-9]+]]
+; GCN-O0-DAG:  v_writelane_b32 [[RESTORED_VGPR]], s{{[0-9]+}}, [[OUTER_LOOP_EXEC_SPILL_LANE_1:[0-9]+]]
 ; GCN-O0:      s_or_b64 s[{{[0-9:]+}}], s[{{[0-9:]+}}], s[{{[0-9:]+}}]
-; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[INNER_LOOP_OUT_EXEC_SPILL_LANE_0:[0-9]+]]
-; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[INNER_LOOP_OUT_EXEC_SPILL_LANE_1:[0-9]+]]
-; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[INNER_LOOP_IN_EXEC_SPILL_LANE_0]]
-; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[INNER_LOOP_IN_EXEC_SPILL_LANE_1]]
+; GCN-O0-DAG:  v_writelane_b32 [[RESTORED_VGPR]], s{{[0-9]+}}, [[INNER_LOOP_OUT_EXEC_SPILL_LANE_0:[0-9]+]]
+; GCN-O0-DAG:  v_writelane_b32 [[RESTORED_VGPR]], s{{[0-9]+}}, [[INNER_LOOP_OUT_EXEC_SPILL_LANE_1:[0-9]+]]
+; GCN-O0-DAG:  v_writelane_b32 [[RESTORED_VGPR]], s{{[0-9]+}}, [[INNER_LOOP_IN_EXEC_SPILL_LANE_0]]
+; GCN-O0-DAG:  v_writelane_b32 [[RESTORED_VGPR]], s{{[0-9]+}}, [[INNER_LOOP_IN_EXEC_SPILL_LANE_1]]
 ; GCN-O0-NEXT: s_mov_b64 s[{{[0-9:]+}}], s[{{[0-9:]+}}]
-; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[INNER_LOOP_BACK_EDGE_EXEC_SPILL_LANE_0]]
-; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[INNER_LOOP_BACK_EDGE_EXEC_SPILL_LANE_1]]
+; GCN-O0-DAG:  v_writelane_b32 [[RESTORED_VGPR]], s{{[0-9]+}}, [[INNER_LOOP_BACK_EDGE_EXEC_SPILL_LANE_0]]
+; GCN-O0-DAG:  v_writelane_b32 [[RESTORED_VGPR]], s{{[0-9]+}}, [[INNER_LOOP_BACK_EDGE_EXEC_SPILL_LANE_1]]
+; GCN-O0-NEXT: s_or_saveexec_b64 [[EXEC_COPY:s\[[0-9]+:[0-9]+\]]], -1
+; GCN-O0-NEXT: buffer_store_dword [[RESTORED_VGPR]], off, s{{\[[0-9]+:[0-9]+\]}}, s32 ; 4-byte Folded Spill
+; GCN-O0-NEXT: s_mov_b64 exec, [[EXEC_COPY]]
 ; GCN-O0-NEXT: s_andn2_b64 exec, exec, s[{{[0-9:]+}}]
 ; GCN-O0-NEXT: s_cbranch_execnz [[INNER_LOOP]]
 ; GCN-O0-NEXT: ; %bb.{{[0-9]+}}:
-; GCN-O0-DAG:  v_readlane_b32 s{{[0-9]+}}, [[VGPR]], [[INNER_LOOP_OUT_EXEC_SPILL_LANE_0]]
-; GCN-O0-DAG:  v_readlane_b32 s{{[0-9]+}}, [[VGPR]], [[INNER_LOOP_OUT_EXEC_SPILL_LANE_1]]
+; GCN-O0:      buffer_load_dword [[RESTORED_1_VGPR:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, s32  ; 4-byte Folded Reload
+; GCN-O0-DAG:  v_readlane_b32 s{{[0-9]+}}, [[RESTORED_1_VGPR]], [[INNER_LOOP_OUT_EXEC_SPILL_LANE_0]]
+; GCN-O0-DAG:  v_readlane_b32 s{{[0-9]+}}, [[RESTORED_1_VGPR]], [[INNER_LOOP_OUT_EXEC_SPILL_LANE_1]]
 ; GCN-O0-NEXT: s_or_b64 exec, exec, s[{{[0-9:]+}}]
 ; GCN-O0:      s_mov_b64 s[{{[0-9:]+}}], exec
-; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[FLOW2_IN_EXEC_SPILL_LANE_0:[0-9]+]]
-; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[FLOW2_IN_EXEC_SPILL_LANE_1:[0-9]+]]
+; GCN-O0-DAG:  v_writelane_b32 [[RESTORED_1_VGPR]], s{{[0-9]+}}, [[FLOW2_IN_EXEC_SPILL_LANE_0:[0-9]+]]
+; GCN-O0-DAG:  v_writelane_b32 [[RESTORED_1_VGPR]], s{{[0-9]+}}, [[FLOW2_IN_EXEC_SPILL_LANE_1:[0-9]+]]
+; GCN-O0-NEXT: s_or_saveexec_b64 [[EXEC_COPY:s\[[0-9]+:[0-9]+\]]], -1
+; GCN-O0-NEXT: buffer_store_dword [[RESTORED_1_VGPR]], off, s{{\[[0-9]+:[0-9]+\]}}, s32 ; 4-byte Folded Spill
+; GCN-O0-NEXT: s_mov_b64 exec, [[EXEC_COPY]]
 ; GCN-O0-NEXT: s_and_b64 s[{{[0-9:]+}}], s[{{[0-9:]+}}], s[{{[0-9:]+}}]
 ; GCN-O0-NEXT: s_mov_b64 exec, s[{{[0-9:]+}}]
 ; GCN-O0-NEXT: s_cbranch_execz [[FLOW2:.LBB[0-9_]+]]
 ; GCN-O0: {{^}}[[FLOW2]]:
-; GCN-O0-DAG:  v_readlane_b32 s{{[0-9]+}}, [[VGPR]], [[FLOW2_IN_EXEC_SPILL_LANE_0]]
-; GCN-O0-DAG:  v_readlane_b32 s{{[0-9]+}}, [[VGPR]], [[FLOW2_IN_EXEC_SPILL_LANE_1]]
+; GCN-O0:      buffer_load_dword [[RESTORED_2_VGPR:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, s32 ; 4-byte Folded Reload
+; GCN-O0-DAG:  v_readlane_b32 s{{[0-9]+}}, [[RESTORED_2_VGPR]], [[FLOW2_IN_EXEC_SPILL_LANE_0]]
+; GCN-O0-DAG:  v_readlane_b32 s{{[0-9]+}}, [[RESTORED_2_VGPR]], [[FLOW2_IN_EXEC_SPILL_LANE_1]]
 ; GCN-O0:      s_branch [[FLOW:.LBB[0-9_]+]]
 ; GCN-O0: {{^}}[[FLOW]]:
+; GCN-O0:      buffer_load_dword [[RESTORED_3_VGPR:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, s32 ; 4-byte Folded Reload
 ; GCN-O0:      s_mov_b64 s[{{[0-9:]+}}], exec
-; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[FLOW3_IN_EXEC_SPILL_LANE_0:[0-9]+]]
-; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[FLOW3_IN_EXEC_SPILL_LANE_1:[0-9]+]]
+; GCN-O0-DAG:  v_writelane_b32 [[RESTORED_3_VGPR]], s{{[0-9]+}}, [[FLOW3_IN_EXEC_SPILL_LANE_0:[0-9]+]]
+; GCN-O0-DAG:  v_writelane_b32 [[RESTORED_3_VGPR]], s{{[0-9]+}}, [[FLOW3_IN_EXEC_SPILL_LANE_1:[0-9]+]]
+; GCN-O0-NEXT: s_or_saveexec_b64 [[EXEC_COPY:s\[[0-9]+:[0-9]+\]]], -1
+; GCN-O0-NEXT: buffer_store_dword [[RESTORED_3_VGPR]], off, s{{\[[0-9]+:[0-9]+\]}}, s32 ; 4-byte Folded Spill
+; GCN-O0-NEXT: s_mov_b64 exec, [[EXEC_COPY]]
 ; GCN-O0-NEXT: s_and_b64 s[{{[0-9:]+}}], s[{{[0-9:]+}}], s[{{[0-9:]+}}]
 ; GCN-O0-NEXT: s_mov_b64 exec, s[{{[0-9:]+}}]
 ; GCN-O0-NEXT: s_cbranch_execz [[FLOW3:.LBB[0-9_]+]]
 ; GCN-O0:      ; %bb.{{[0-9]+}}:
-; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[FLOW1_OUT_EXEC_SPILL_LANE_0:[0-9]+]]
-; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[FLOW1_OUT_EXEC_SPILL_LANE_1:[0-9]+]]
+; GCN-O0:      buffer_load_dword [[RESTORED_4_VGPR:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, s32 ; 4-byte Folded Reload
+; GCN-O0-DAG:  v_writelane_b32 [[RESTORED_4_VGPR]], s{{[0-9]+}}, [[FLOW1_OUT_EXEC_SPILL_LANE_0:[0-9]+]]
+; GCN-O0-DAG:  v_writelane_b32 [[RESTORED_4_VGPR]], s{{[0-9]+}}, [[FLOW1_OUT_EXEC_SPILL_LANE_1:[0-9]+]]
+; GCN-O0-NEXT: s_or_saveexec_b64 [[EXEC_COPY:s\[[0-9]+:[0-9]+\]]], -1
+; GCN-O0-NEXT: buffer_store_dword [[RESTORED_4_VGPR]], off, s{{\[[0-9]+:[0-9]+\]}}, s32 ; 4-byte Folded Spill
+; GCN-O0-NEXT: s_mov_b64 exec, [[EXEC_COPY]]
 ; GCN-O0: {{^}}[[FLOW3]]:
 ; GCN-O0-COUNT-4: buffer_load_dword
-; GCN-O0-DAG:  v_readlane_b32 s{{[0-9]+}}, [[VGPR]], [[OUTER_LOOP_EXEC_SPILL_LANE_0]]
-; GCN-O0-DAG:  v_readlane_b32 s{{[0-9]+}}, [[VGPR]], [[OUTER_LOOP_EXEC_SPILL_LANE_1]]
-; GCN-O0-DAG:  v_readlane_b32 s{{[0-9]+}}, [[VGPR]], [[FLOW1_OUT_EXEC_SPILL_LANE_0]]
-; GCN-O0-DAG:  v_readlane_b32 s{{[0-9]+}}, [[VGPR]], [[FLOW1_OUT_EXEC_SPILL_LANE_1]]
+; GCN-O0:      buffer_load_dword [[RESTORED_5_VGPR:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, s32 ; 4-byte Folded Reload
+; GCN-O0-DAG:  v_readlane_b32 s{{[0-9]+}}, [[RESTORED_5_VGPR]], [[OUTER_LOOP_EXEC_SPILL_LANE_0]]
+; GCN-O0-DAG:  v_readlane_b32 s{{[0-9]+}}, [[RESTORED_5_VGPR]], [[OUTER_LOOP_EXEC_SPILL_LANE_1]]
+; GCN-O0-DAG:  v_readlane_b32 s{{[0-9]+}}, [[RESTORED_5_VGPR]], [[FLOW1_OUT_EXEC_SPILL_LANE_0]]
+; GCN-O0-DAG:  v_readlane_b32 s{{[0-9]+}}, [[RESTORED_5_VGPR]], [[FLOW1_OUT_EXEC_SPILL_LANE_1]]
 ; GCN-O0:      s_and_b64 s[{{[0-9:]+}}], exec, s[{{[0-9:]+}}]
 ; GCN-O0-NEXT: s_or_b64 s[{{[0-9:]+}}], s[{{[0-9:]+}}], s[{{[0-9:]+}}]
 ; GCN-O0-COUNT-2: s_mov_b64
-; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[INNER_LOOP_IN_EXEC_SPILL_LANE_0]]
-; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[INNER_LOOP_IN_EXEC_SPILL_LANE_1]]
-; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[INNER_LOOP_BACK_EDGE_EXEC_SPILL_LANE_0]]
-; GCN-O0-DAG:  v_writelane_b32 [[VGPR]], s{{[0-9]+}}, [[INNER_LOOP_BACK_EDGE_EXEC_SPILL_LANE_1]]
+; GCN-O0-DAG:  v_writelane_b32 [[RESTORED_5_VGPR]], s{{[0-9]+}}, [[INNER_LOOP_IN_EXEC_SPILL_LANE_0]]
+; GCN-O0-DAG:  v_writelane_b32 [[RESTORED_5_VGPR]], s{{[0-9]+}}, [[INNER_LOOP_IN_EXEC_SPILL_LANE_1]]
+; GCN-O0-DAG:  v_writelane_b32 [[RESTORED_5_VGPR]], s{{[0-9]+}}, [[INNER_LOOP_BACK_EDGE_EXEC_SPILL_LANE_0]]
+; GCN-O0-DAG:  v_writelane_b32 [[RESTORED_5_VGPR]], s{{[0-9]+}}, [[INNER_LOOP_BACK_EDGE_EXEC_SPILL_LANE_1]]
 ; GCN-O0-COUNT-4: buffer_store_dword
 ; GCN-O0:      s_andn2_b64 exec, exec, s[{{[0-9:]+}}]
 ; GCN-O0-NEXT: s_cbranch_execnz [[INNER_LOOP]]
diff --git a/llvm/test/CodeGen/AMDGPU/control-flow-fastregalloc.ll b/llvm/test/CodeGen/AMDGPU/control-flow-fastregalloc.ll
--- a/llvm/test/CodeGen/AMDGPU/control-flow-fastregalloc.ll
+++ b/llvm/test/CodeGen/AMDGPU/control-flow-fastregalloc.ll
@@ -10,7 +10,7 @@
 
 
 ; GCN-LABEL: {{^}}divergent_if_endif:
-; VGPR: workitem_private_segment_byte_size = 12{{$}}
+; VGPR: workitem_private_segment_byte_size = 16{{$}}
 
 
 ; GCN: {{^}}; %bb.0:
@@ -19,7 +19,7 @@
 
 ; Spill load
 ; GCN: buffer_store_dword [[LOAD0]], off, s[0:3], 0 offset:[[LOAD0_OFFSET:[0-9]+]] ; 4-byte Folded Spill
-; GCN: v_cmp_eq_u32_e64 [[CMP0:s\[[0-9]+:[0-9]\]]], v0, s{{[0-9]+}}
+; GCN: v_cmp_eq_u32_e64 [[CMP0:s\[[0-9]+:[0-9]\]]], v{{[0-9]+}}, s{{[0-9]+}}
 
 ; Spill saved exec
 ; GCN: s_mov_b64 s[[[SAVEEXEC_LO:[0-9]+]]:[[SAVEEXEC_HI:[0-9]+]]], exec
@@ -82,13 +82,13 @@
 }
 
 ; GCN-LABEL: {{^}}divergent_loop:
-; VGPR: workitem_private_segment_byte_size = 16{{$}}
+; VGPR: workitem_private_segment_byte_size = 20{{$}}
 
 ; GCN: {{^}}; %bb.0:
 ; GCN-DAG: s_mov_b32 m0, -1
 ; GCN-DAG: v_mov_b32_e32 [[PTR0:v[0-9]+]], 0{{$}}
 ; GCN: ds_read_b32 [[LOAD0:v[0-9]+]], [[PTR0]]
-; GCN: v_cmp_eq_u32_e64 [[CMP0:s\[[0-9]+:[0-9]+\]]], v0, s{{[0-9]+}}
+; GCN: v_cmp_eq_u32_e64 [[CMP0:s\[[0-9]+:[0-9]+\]]], v{{[0-9]+}}, s{{[0-9]+}}
 
 ; Spill load
 ; GCN: buffer_store_dword [[LOAD0]], off, s[0:3], 0 offset:[[LOAD0_OFFSET:[0-9]+]] ; 4-byte Folded Spill
@@ -166,7 +166,7 @@
 ; GCN: buffer_store_dword [[LOAD0]], off, s[0:3], 0 offset:[[LOAD0_OFFSET:[0-9]+]] ; 4-byte Folded Spill
 
 ; GCN: s_mov_b32 [[ZERO:s[0-9]+]], 0
-; GCN: v_cmp_ne_u32_e64 [[CMP0:s\[[0-9]+:[0-9]\]]], v0, [[ZERO]]
+; GCN: v_cmp_ne_u32_e64 [[CMP0:s\[[0-9]+:[0-9]\]]], v{{[0-9]+}}, [[ZERO]]
 
 ; GCN: s_mov_b64 s[[[SAVEEXEC_LO:[0-9]+]]:[[SAVEEXEC_HI:[0-9]+]]], exec
 ; GCN: s_and_b64 s[[[ANDEXEC_LO:[0-9]+]]:[[ANDEXEC_HI:[0-9]+]]], s[[[SAVEEXEC_LO:[0-9]+]]:[[SAVEEXEC_HI:[0-9]+]]], [[CMP0]]
@@ -187,6 +187,7 @@
 ; GCN-NEXT: s_branch [[ELSE:.LBB[0-9]+_[0-9]+]]
 
 ; GCN: [[FLOW]]: ; %Flow
+; VGPR: buffer_load_dword
 ; GCN: buffer_load_dword [[FLOW_VAL:v[0-9]+]], off, s[0:3], 0 offset:[[FLOW_VAL_OFFSET:[0-9]+]] ; 4-byte Folded Reload
 ; VGPR: v_readlane_b32 s[[FLOW_S_RELOAD_SAVEEXEC_LO:[0-9]+]], [[SPILL_VGPR]], [[SAVEEXEC_LO_LANE]]
 ; VGPR: v_readlane_b32 s[[FLOW_S_RELOAD_SAVEEXEC_HI:[0-9]+]], [[SPILL_VGPR]], [[SAVEEXEC_HI_LANE]]
diff --git a/llvm/test/CodeGen/AMDGPU/cross-block-use-is-not-abi-copy.ll b/llvm/test/CodeGen/AMDGPU/cross-block-use-is-not-abi-copy.ll
--- a/llvm/test/CodeGen/AMDGPU/cross-block-use-is-not-abi-copy.ll
+++ b/llvm/test/CodeGen/AMDGPU/cross-block-use-is-not-abi-copy.ll
@@ -33,6 +33,7 @@
 ; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GCN-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[18:19]
+; GCN-NEXT:    ; implicit-def: $vgpr40
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
 ; GCN-NEXT:    v_writelane_b32 v40, s30, 0
 ; GCN-NEXT:    v_writelane_b32 v41, s16, 0
@@ -71,6 +72,7 @@
 ; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GCN-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[18:19]
+; GCN-NEXT:    ; implicit-def: $vgpr40
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
 ; GCN-NEXT:    v_writelane_b32 v40, s30, 0
 ; GCN-NEXT:    v_writelane_b32 v41, s16, 0
@@ -109,6 +111,7 @@
 ; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GCN-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[18:19]
+; GCN-NEXT:    ; implicit-def: $vgpr40
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
 ; GCN-NEXT:    v_writelane_b32 v40, s30, 0
 ; GCN-NEXT:    v_writelane_b32 v41, s16, 0
@@ -147,6 +150,7 @@
 ; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GCN-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[18:19]
+; GCN-NEXT:    ; implicit-def: $vgpr40
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
 ; GCN-NEXT:    v_writelane_b32 v40, s30, 0
 ; GCN-NEXT:    v_writelane_b32 v41, s16, 0
diff --git a/llvm/test/CodeGen/AMDGPU/csr-sgpr-spill-live-ins.mir b/llvm/test/CodeGen/AMDGPU/csr-sgpr-spill-live-ins.mir
--- a/llvm/test/CodeGen/AMDGPU/csr-sgpr-spill-live-ins.mir
+++ b/llvm/test/CodeGen/AMDGPU/csr-sgpr-spill-live-ins.mir
@@ -19,10 +19,11 @@
   ; CHECK-NEXT:   $sgpr4_sgpr5 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
   ; CHECK-NEXT:   BUFFER_STORE_DWORD_OFFSET $vgpr0, $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr32, 0, 0, 0, implicit $exec :: (store (s32) into %stack.4, addrspace 5)
   ; CHECK-NEXT:   $exec = S_MOV_B64 killed $sgpr4_sgpr5
-  ; CHECK-NEXT:   $vgpr0 = V_WRITELANE_B32 $sgpr42, 0, $vgpr0
-  ; CHECK-NEXT:   $vgpr0 = V_WRITELANE_B32 $sgpr43, 1, $vgpr0
-  ; CHECK-NEXT:   $vgpr0 = V_WRITELANE_B32 $sgpr46, 2, $vgpr0
-  ; CHECK-NEXT:   $vgpr0 = V_WRITELANE_B32 $sgpr47, 3, $vgpr0
+  ; CHECK-NEXT:   renamable $vgpr0 = IMPLICIT_DEF
+  ; CHECK-NEXT:   renamable $vgpr0 = V_WRITELANE_B32 $sgpr42, 0, killed $vgpr0
+  ; CHECK-NEXT:   renamable $vgpr0 = V_WRITELANE_B32 $sgpr43, 1, killed $vgpr0
+  ; CHECK-NEXT:   renamable $vgpr0 = V_WRITELANE_B32 $sgpr46, 2, killed $vgpr0
+  ; CHECK-NEXT:   dead renamable $vgpr0 = V_WRITELANE_B32 $sgpr47, 3, killed $vgpr0
   ; CHECK-NEXT:   S_NOP 0
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.1:
diff --git a/llvm/test/CodeGen/AMDGPU/dwarf-multi-register-use-crash.ll b/llvm/test/CodeGen/AMDGPU/dwarf-multi-register-use-crash.ll
--- a/llvm/test/CodeGen/AMDGPU/dwarf-multi-register-use-crash.ll
+++ b/llvm/test/CodeGen/AMDGPU/dwarf-multi-register-use-crash.ll
@@ -19,25 +19,26 @@
 ; CHECK-NEXT:    s_mov_b32 s16, s33
 ; CHECK-NEXT:    s_mov_b32 s33, s32
 ; CHECK-NEXT:    s_or_saveexec_b64 s[18:19], -1
-; CHECK-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
+; CHECK-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; CHECK-NEXT:    buffer_store_dword v42, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
 ; CHECK-NEXT:    s_mov_b64 exec, s[18:19]
-; CHECK-NEXT:    v_writelane_b32 v40, s30, 0
-; CHECK-NEXT:    v_writelane_b32 v40, s31, 1
-; CHECK-NEXT:    v_writelane_b32 v40, s34, 2
-; CHECK-NEXT:    v_writelane_b32 v40, s35, 3
-; CHECK-NEXT:    v_writelane_b32 v40, s36, 4
-; CHECK-NEXT:    v_writelane_b32 v40, s37, 5
-; CHECK-NEXT:    v_writelane_b32 v40, s38, 6
-; CHECK-NEXT:    v_writelane_b32 v40, s39, 7
-; CHECK-NEXT:    v_writelane_b32 v40, s40, 8
-; CHECK-NEXT:    v_writelane_b32 v40, s41, 9
-; CHECK-NEXT:    v_writelane_b32 v40, s42, 10
-; CHECK-NEXT:    v_writelane_b32 v40, s43, 11
-; CHECK-NEXT:    v_writelane_b32 v40, s44, 12
+; CHECK-NEXT:    ; implicit-def: $vgpr41
 ; CHECK-NEXT:    s_addk_i32 s32, 0x400
-; CHECK-NEXT:    v_writelane_b32 v40, s45, 13
-; CHECK-NEXT:    v_writelane_b32 v40, s46, 14
+; CHECK-NEXT:    v_writelane_b32 v41, s30, 0
+; CHECK-NEXT:    v_writelane_b32 v41, s31, 1
+; CHECK-NEXT:    v_writelane_b32 v41, s34, 2
+; CHECK-NEXT:    v_writelane_b32 v41, s35, 3
+; CHECK-NEXT:    v_writelane_b32 v41, s36, 4
+; CHECK-NEXT:    v_writelane_b32 v41, s37, 5
+; CHECK-NEXT:    v_writelane_b32 v41, s38, 6
+; CHECK-NEXT:    v_writelane_b32 v41, s39, 7
+; CHECK-NEXT:    v_writelane_b32 v41, s40, 8
+; CHECK-NEXT:    v_writelane_b32 v41, s41, 9
+; CHECK-NEXT:    v_writelane_b32 v41, s42, 10
+; CHECK-NEXT:    v_writelane_b32 v41, s43, 11
+; CHECK-NEXT:    v_writelane_b32 v41, s44, 12
+; CHECK-NEXT:    v_writelane_b32 v41, s45, 13
+; CHECK-NEXT:    v_writelane_b32 v41, s46, 14
 ; CHECK-NEXT:    s_mov_b64 s[40:41], s[4:5]
 ; CHECK-NEXT:    ;DEBUG_VALUE: dummy:dummy <- undef
 ; CHECK-NEXT:  .Ltmp0:
@@ -45,12 +46,12 @@
 ; CHECK-NEXT:    s_getpc_b64 s[4:5]
 ; CHECK-NEXT:    s_add_u32 s4, s4, __kmpc_alloc_shared@gotpcrel32@lo+4
 ; CHECK-NEXT:    s_addc_u32 s5, s5, __kmpc_alloc_shared@gotpcrel32@hi+12
-; CHECK-NEXT:    v_writelane_b32 v40, s47, 15
+; CHECK-NEXT:    v_writelane_b32 v41, s47, 15
 ; CHECK-NEXT:    s_load_dwordx2 s[46:47], s[4:5], 0x0
 ; CHECK-NEXT:    s_mov_b64 s[4:5], s[40:41]
 ; CHECK-NEXT:    v_writelane_b32 v42, s16, 0
-; CHECK-NEXT:    buffer_store_dword v41, off, s[0:3], s33 ; 4-byte Folded Spill
-; CHECK-NEXT:    v_mov_b32_e32 v41, v31
+; CHECK-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
+; CHECK-NEXT:    v_mov_b32_e32 v40, v31
 ; CHECK-NEXT:    s_mov_b32 s42, s15
 ; CHECK-NEXT:    s_mov_b32 s43, s14
 ; CHECK-NEXT:    s_mov_b32 s44, s13
@@ -68,33 +69,33 @@
 ; CHECK-NEXT:    s_mov_b32 s13, s44
 ; CHECK-NEXT:    s_mov_b32 s14, s43
 ; CHECK-NEXT:    s_mov_b32 s15, s42
-; CHECK-NEXT:    v_mov_b32_e32 v31, v41
+; CHECK-NEXT:    v_mov_b32_e32 v31, v40
 ; CHECK-NEXT:    s_swappc_b64 s[30:31], s[46:47]
 ; CHECK-NEXT:  .Ltmp1:
 ; CHECK-NEXT:    ;DEBUG_VALUE: dummy:dummy <- [$vgpr0_vgpr1+0]
 ; CHECK-NEXT:    .loc 1 0 9 is_stmt 0 ; dummy:0:9
-; CHECK-NEXT:    buffer_load_dword v41, off, s[0:3], s33 ; 4-byte Folded Reload
+; CHECK-NEXT:    buffer_load_dword v40, off, s[0:3], s33 ; 4-byte Folded Reload
 ; CHECK-NEXT:    v_mov_b32_e32 v2, 0
 ; CHECK-NEXT:    flat_store_dword v[0:1], v2
-; CHECK-NEXT:    v_readlane_b32 s47, v40, 15
-; CHECK-NEXT:    v_readlane_b32 s46, v40, 14
-; CHECK-NEXT:    v_readlane_b32 s45, v40, 13
-; CHECK-NEXT:    v_readlane_b32 s44, v40, 12
-; CHECK-NEXT:    v_readlane_b32 s43, v40, 11
-; CHECK-NEXT:    v_readlane_b32 s42, v40, 10
-; CHECK-NEXT:    v_readlane_b32 s41, v40, 9
-; CHECK-NEXT:    v_readlane_b32 s40, v40, 8
-; CHECK-NEXT:    v_readlane_b32 s39, v40, 7
-; CHECK-NEXT:    v_readlane_b32 s38, v40, 6
-; CHECK-NEXT:    v_readlane_b32 s37, v40, 5
-; CHECK-NEXT:    v_readlane_b32 s36, v40, 4
-; CHECK-NEXT:    v_readlane_b32 s35, v40, 3
-; CHECK-NEXT:    v_readlane_b32 s34, v40, 2
-; CHECK-NEXT:    v_readlane_b32 s31, v40, 1
-; CHECK-NEXT:    v_readlane_b32 s30, v40, 0
+; CHECK-NEXT:    v_readlane_b32 s47, v41, 15
+; CHECK-NEXT:    v_readlane_b32 s46, v41, 14
+; CHECK-NEXT:    v_readlane_b32 s45, v41, 13
+; CHECK-NEXT:    v_readlane_b32 s44, v41, 12
+; CHECK-NEXT:    v_readlane_b32 s43, v41, 11
+; CHECK-NEXT:    v_readlane_b32 s42, v41, 10
+; CHECK-NEXT:    v_readlane_b32 s41, v41, 9
+; CHECK-NEXT:    v_readlane_b32 s40, v41, 8
+; CHECK-NEXT:    v_readlane_b32 s39, v41, 7
+; CHECK-NEXT:    v_readlane_b32 s38, v41, 6
+; CHECK-NEXT:    v_readlane_b32 s37, v41, 5
+; CHECK-NEXT:    v_readlane_b32 s36, v41, 4
+; CHECK-NEXT:    v_readlane_b32 s35, v41, 3
+; CHECK-NEXT:    v_readlane_b32 s34, v41, 2
+; CHECK-NEXT:    v_readlane_b32 s31, v41, 1
+; CHECK-NEXT:    v_readlane_b32 s30, v41, 0
 ; CHECK-NEXT:    v_readlane_b32 s4, v42, 0
 ; CHECK-NEXT:    s_or_saveexec_b64 s[6:7], -1
-; CHECK-NEXT:    buffer_load_dword v40, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
+; CHECK-NEXT:    buffer_load_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
 ; CHECK-NEXT:    buffer_load_dword v42, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload
 ; CHECK-NEXT:    s_mov_b64 exec, s[6:7]
 ; CHECK-NEXT:    s_addk_i32 s32, 0xfc00
diff --git a/llvm/test/CodeGen/AMDGPU/fix-frame-reg-in-custom-csr-spills.ll b/llvm/test/CodeGen/AMDGPU/fix-frame-reg-in-custom-csr-spills.ll
--- a/llvm/test/CodeGen/AMDGPU/fix-frame-reg-in-custom-csr-spills.ll
+++ b/llvm/test/CodeGen/AMDGPU/fix-frame-reg-in-custom-csr-spills.ll
@@ -25,6 +25,7 @@
 ; GCN-NEXT:    s_add_u32 s16, s16, extern_func@gotpcrel32@lo+4
 ; GCN-NEXT:    s_addc_u32 s17, s17, extern_func@gotpcrel32@hi+12
 ; GCN-NEXT:    s_load_dwordx2 s[16:17], s[16:17], 0x0
+; GCN-NEXT:    ; implicit-def: $vgpr42
 ; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GCN-NEXT:    buffer_store_dword v41, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GCN-NEXT:    v_writelane_b32 v42, s30, 0
diff --git a/llvm/test/CodeGen/AMDGPU/flat-scratch-init.ll b/llvm/test/CodeGen/AMDGPU/flat-scratch-init.ll
--- a/llvm/test/CodeGen/AMDGPU/flat-scratch-init.ll
+++ b/llvm/test/CodeGen/AMDGPU/flat-scratch-init.ll
@@ -119,21 +119,15 @@
 ; FLAT_SCR_OPT-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s2
 ; FLAT_SCR_OPT-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s3
 ; FLAT_SCR_OPT-NEXT:    s_load_dwordx2 s[2:3], s[0:1], 0x0
-; FLAT_SCR_OPT-NEXT:    s_mov_b32 s104, exec_lo
-; FLAT_SCR_OPT-NEXT:    s_mov_b32 exec_lo, 3
-; FLAT_SCR_OPT-NEXT:    s_mov_b32 s4, 0
-; FLAT_SCR_OPT-NEXT:    scratch_store_dword off, v72, s4
+; FLAT_SCR_OPT-NEXT:    ; implicit-def: $vgpr0
 ; FLAT_SCR_OPT-NEXT:    s_waitcnt lgkmcnt(0)
-; FLAT_SCR_OPT-NEXT:    v_writelane_b32 v72, s2, 0
-; FLAT_SCR_OPT-NEXT:    s_mov_b32 s4, 4
-; FLAT_SCR_OPT-NEXT:    v_writelane_b32 v72, s3, 1
-; FLAT_SCR_OPT-NEXT:    scratch_store_dword off, v72, s4 ; 4-byte Folded Spill
+; FLAT_SCR_OPT-NEXT:    v_writelane_b32 v0, s2, 0
+; FLAT_SCR_OPT-NEXT:    v_writelane_b32 v0, s3, 1
+; FLAT_SCR_OPT-NEXT:    s_or_saveexec_b32 s105, -1
+; FLAT_SCR_OPT-NEXT:    s_mov_b32 s2, 4
+; FLAT_SCR_OPT-NEXT:    scratch_store_dword off, v0, s2 ; 4-byte Folded Spill
 ; FLAT_SCR_OPT-NEXT:    s_waitcnt_depctr 0xffe3
-; FLAT_SCR_OPT-NEXT:    s_mov_b32 s4, 0
-; FLAT_SCR_OPT-NEXT:    scratch_load_dword v72, off, s4
-; FLAT_SCR_OPT-NEXT:    s_waitcnt vmcnt(0)
-; FLAT_SCR_OPT-NEXT:    s_waitcnt_depctr 0xffe3
-; FLAT_SCR_OPT-NEXT:    s_mov_b32 exec_lo, s104
+; FLAT_SCR_OPT-NEXT:    s_mov_b32 exec_lo, s105
 ; FLAT_SCR_OPT-NEXT:    s_load_dword vcc_lo, s[0:1], 0x8
 ; FLAT_SCR_OPT-NEXT:    ; kill: killed $sgpr0_sgpr1
 ; FLAT_SCR_OPT-NEXT:    ;;#ASMSTART
@@ -230,22 +224,14 @@
 ; FLAT_SCR_OPT-NEXT:    ;;#ASMEND
 ; FLAT_SCR_OPT-NEXT:    ;;#ASMSTART
 ; FLAT_SCR_OPT-NEXT:    ;;#ASMEND
-; FLAT_SCR_OPT-NEXT:    s_mov_b32 s2, exec_lo
-; FLAT_SCR_OPT-NEXT:    s_mov_b32 exec_lo, 3
-; FLAT_SCR_OPT-NEXT:    s_mov_b32 s3, 0
-; FLAT_SCR_OPT-NEXT:    scratch_store_dword off, v2, s3
-; FLAT_SCR_OPT-NEXT:    s_waitcnt_depctr 0xffe3
-; FLAT_SCR_OPT-NEXT:    s_mov_b32 s3, 4
-; FLAT_SCR_OPT-NEXT:    scratch_load_dword v2, off, s3 ; 4-byte Folded Reload
+; FLAT_SCR_OPT-NEXT:    s_or_saveexec_b32 s105, -1
+; FLAT_SCR_OPT-NEXT:    s_mov_b32 s0, 4
+; FLAT_SCR_OPT-NEXT:    scratch_load_dword v1, off, s0 ; 4-byte Folded Reload
 ; FLAT_SCR_OPT-NEXT:    s_waitcnt_depctr 0xffe3
-; FLAT_SCR_OPT-NEXT:    s_mov_b32 s3, 0
+; FLAT_SCR_OPT-NEXT:    s_mov_b32 exec_lo, s105
 ; FLAT_SCR_OPT-NEXT:    s_waitcnt vmcnt(0)
-; FLAT_SCR_OPT-NEXT:    v_readlane_b32 s0, v2, 0
-; FLAT_SCR_OPT-NEXT:    v_readlane_b32 s1, v2, 1
-; FLAT_SCR_OPT-NEXT:    scratch_load_dword v2, off, s3
-; FLAT_SCR_OPT-NEXT:    s_waitcnt vmcnt(0)
-; FLAT_SCR_OPT-NEXT:    s_waitcnt_depctr 0xffe3
-; FLAT_SCR_OPT-NEXT:    s_mov_b32 exec_lo, s2
+; FLAT_SCR_OPT-NEXT:    v_readlane_b32 s0, v1, 0
+; FLAT_SCR_OPT-NEXT:    v_readlane_b32 s1, v1, 1
 ; FLAT_SCR_OPT-NEXT:    v_mov_b32_e32 v1, 0
 ; FLAT_SCR_OPT-NEXT:    global_store_dword v1, v0, s[0:1]
 ; FLAT_SCR_OPT-NEXT:    s_endpgm
@@ -253,21 +239,15 @@
 ; FLAT_SCR_ARCH-LABEL: test:
 ; FLAT_SCR_ARCH:       ; %bb.0:
 ; FLAT_SCR_ARCH-NEXT:    s_load_dwordx2 s[2:3], s[0:1], 0x0
-; FLAT_SCR_ARCH-NEXT:    s_mov_b32 s104, exec_lo
-; FLAT_SCR_ARCH-NEXT:    s_mov_b32 exec_lo, 3
-; FLAT_SCR_ARCH-NEXT:    s_mov_b32 s4, 0
-; FLAT_SCR_ARCH-NEXT:    scratch_store_dword off, v72, s4
+; FLAT_SCR_ARCH-NEXT:    ; implicit-def: $vgpr0
 ; FLAT_SCR_ARCH-NEXT:    s_waitcnt lgkmcnt(0)
-; FLAT_SCR_ARCH-NEXT:    v_writelane_b32 v72, s2, 0
-; FLAT_SCR_ARCH-NEXT:    s_mov_b32 s4, 4
-; FLAT_SCR_ARCH-NEXT:    v_writelane_b32 v72, s3, 1
-; FLAT_SCR_ARCH-NEXT:    scratch_store_dword off, v72, s4 ; 4-byte Folded Spill
+; FLAT_SCR_ARCH-NEXT:    v_writelane_b32 v0, s2, 0
+; FLAT_SCR_ARCH-NEXT:    v_writelane_b32 v0, s3, 1
+; FLAT_SCR_ARCH-NEXT:    s_or_saveexec_b32 s105, -1
+; FLAT_SCR_ARCH-NEXT:    s_mov_b32 s2, 4
+; FLAT_SCR_ARCH-NEXT:    scratch_store_dword off, v0, s2 ; 4-byte Folded Spill
 ; FLAT_SCR_ARCH-NEXT:    s_waitcnt_depctr 0xffe3
-; FLAT_SCR_ARCH-NEXT:    s_mov_b32 s4, 0
-; FLAT_SCR_ARCH-NEXT:    scratch_load_dword v72, off, s4
-; FLAT_SCR_ARCH-NEXT:    s_waitcnt vmcnt(0)
-; FLAT_SCR_ARCH-NEXT:    s_waitcnt_depctr 0xffe3
-; FLAT_SCR_ARCH-NEXT:    s_mov_b32 exec_lo, s104
+; FLAT_SCR_ARCH-NEXT:    s_mov_b32 exec_lo, s105
 ; FLAT_SCR_ARCH-NEXT:    s_load_dword vcc_lo, s[0:1], 0x8
 ; FLAT_SCR_ARCH-NEXT:    ; kill: killed $sgpr0_sgpr1
 ; FLAT_SCR_ARCH-NEXT:    ;;#ASMSTART
@@ -364,22 +344,14 @@
 ; FLAT_SCR_ARCH-NEXT:    ;;#ASMEND
 ; FLAT_SCR_ARCH-NEXT:    ;;#ASMSTART
 ; FLAT_SCR_ARCH-NEXT:    ;;#ASMEND
-; FLAT_SCR_ARCH-NEXT:    s_mov_b32 s2, exec_lo
-; FLAT_SCR_ARCH-NEXT:    s_mov_b32 exec_lo, 3
-; FLAT_SCR_ARCH-NEXT:    s_mov_b32 s3, 0
-; FLAT_SCR_ARCH-NEXT:    scratch_store_dword off, v2, s3
-; FLAT_SCR_ARCH-NEXT:    s_waitcnt_depctr 0xffe3
-; FLAT_SCR_ARCH-NEXT:    s_mov_b32 s3, 4
-; FLAT_SCR_ARCH-NEXT:    scratch_load_dword v2, off, s3 ; 4-byte Folded Reload
+; FLAT_SCR_ARCH-NEXT:    s_or_saveexec_b32 s105, -1
+; FLAT_SCR_ARCH-NEXT:    s_mov_b32 s0, 4
+; FLAT_SCR_ARCH-NEXT:    scratch_load_dword v1, off, s0 ; 4-byte Folded Reload
 ; FLAT_SCR_ARCH-NEXT:    s_waitcnt_depctr 0xffe3
-; FLAT_SCR_ARCH-NEXT:    s_mov_b32 s3, 0
+; FLAT_SCR_ARCH-NEXT:    s_mov_b32 exec_lo, s105
 ; FLAT_SCR_ARCH-NEXT:    s_waitcnt vmcnt(0)
-; FLAT_SCR_ARCH-NEXT:    v_readlane_b32 s0, v2, 0
-; FLAT_SCR_ARCH-NEXT:    v_readlane_b32 s1, v2, 1
-; FLAT_SCR_ARCH-NEXT:    scratch_load_dword v2, off, s3
-; FLAT_SCR_ARCH-NEXT:    s_waitcnt vmcnt(0)
-; FLAT_SCR_ARCH-NEXT:    s_waitcnt_depctr 0xffe3
-; FLAT_SCR_ARCH-NEXT:    s_mov_b32 exec_lo, s2
+; FLAT_SCR_ARCH-NEXT:    v_readlane_b32 s0, v1, 0
+; FLAT_SCR_ARCH-NEXT:    v_readlane_b32 s1, v1, 1
 ; FLAT_SCR_ARCH-NEXT:    v_mov_b32_e32 v1, 0
 ; FLAT_SCR_ARCH-NEXT:    global_store_dword v1, v0, s[0:1]
 ; FLAT_SCR_ARCH-NEXT:    s_endpgm
diff --git a/llvm/test/CodeGen/AMDGPU/fold-reload-into-exec.mir b/llvm/test/CodeGen/AMDGPU/fold-reload-into-exec.mir
--- a/llvm/test/CodeGen/AMDGPU/fold-reload-into-exec.mir
+++ b/llvm/test/CodeGen/AMDGPU/fold-reload-into-exec.mir
@@ -12,14 +12,13 @@
 body:             |
   bb.0:
     ; CHECK-LABEL: name: merge_sgpr_spill_into_copy_from_exec_lo
-    ; CHECK: liveins: $vgpr0
-    ; CHECK-NEXT: {{  $}}
-    ; CHECK-NEXT: S_NOP 0, implicit-def $exec_lo
+    ; CHECK: S_NOP 0, implicit-def $exec_lo
     ; CHECK-NEXT: $sgpr0 = S_MOV_B32 $exec_lo
-    ; CHECK-NEXT: $vgpr0 = V_WRITELANE_B32 killed $sgpr0, 0, $vgpr0
+    ; CHECK-NEXT: renamable $vgpr0 = IMPLICIT_DEF
+    ; CHECK-NEXT: renamable $vgpr0 = V_WRITELANE_B32 killed $sgpr0, 0, killed $vgpr0
     ; CHECK-NEXT: $sgpr0 = V_READLANE_B32 $vgpr0, 0
     ; CHECK-NEXT: S_NOP 0, implicit-def dead renamable $sgpr1, implicit-def dead renamable $sgpr0, implicit killed renamable $sgpr0
-    ; CHECK-NEXT: $sgpr0 = V_READLANE_B32 $vgpr0, 0
+    ; CHECK-NEXT: $sgpr0 = V_READLANE_B32 killed $vgpr0, 0
     ; CHECK-NEXT: $exec_lo = S_MOV_B32 killed $sgpr0
     ; CHECK-NEXT: S_SENDMSG 0, implicit $m0, implicit $exec
     S_NOP 0, implicit-def $exec_lo
@@ -38,14 +37,13 @@
 body:             |
   bb.0:
     ; CHECK-LABEL: name: merge_sgpr_spill_into_copy_from_exec_hi
-    ; CHECK: liveins: $vgpr0
-    ; CHECK-NEXT: {{  $}}
-    ; CHECK-NEXT: S_NOP 0, implicit-def $exec_hi
+    ; CHECK: S_NOP 0, implicit-def $exec_hi
     ; CHECK-NEXT: $sgpr0 = S_MOV_B32 $exec_hi
-    ; CHECK-NEXT: $vgpr0 = V_WRITELANE_B32 killed $sgpr0, 0, $vgpr0
+    ; CHECK-NEXT: renamable $vgpr0 = IMPLICIT_DEF
+    ; CHECK-NEXT: renamable $vgpr0 = V_WRITELANE_B32 killed $sgpr0, 0, killed $vgpr0
     ; CHECK-NEXT: $sgpr0 = V_READLANE_B32 $vgpr0, 0
     ; CHECK-NEXT: S_NOP 0, implicit-def dead renamable $sgpr1, implicit-def dead renamable $sgpr0, implicit killed renamable $sgpr0
-    ; CHECK-NEXT: $sgpr0 = V_READLANE_B32 $vgpr0, 0
+    ; CHECK-NEXT: $sgpr0 = V_READLANE_B32 killed $vgpr0, 0
     ; CHECK-NEXT: $exec_hi = S_MOV_B32 killed $sgpr0
     ; CHECK-NEXT: S_SENDMSG 0, implicit $m0, implicit $exec
     S_NOP 0, implicit-def $exec_hi
@@ -64,17 +62,16 @@
 body:             |
   bb.0:
     ; CHECK-LABEL: name: merge_sgpr_spill_into_copy_from_exec
-    ; CHECK: liveins: $vgpr0
-    ; CHECK-NEXT: {{  $}}
-    ; CHECK-NEXT: S_NOP 0, implicit-def $exec
+    ; CHECK: S_NOP 0, implicit-def $exec
     ; CHECK-NEXT: $sgpr0_sgpr1 = S_MOV_B64 $exec
-    ; CHECK-NEXT: $vgpr0 = V_WRITELANE_B32 killed $sgpr0, 0, $vgpr0, implicit-def $sgpr0_sgpr1, implicit $sgpr0_sgpr1
-    ; CHECK-NEXT: $vgpr0 = V_WRITELANE_B32 killed $sgpr1, 1, $vgpr0, implicit $sgpr0_sgpr1
+    ; CHECK-NEXT: renamable $vgpr0 = IMPLICIT_DEF
+    ; CHECK-NEXT: renamable $vgpr0 = V_WRITELANE_B32 killed $sgpr0, 0, killed $vgpr0, implicit-def $sgpr0_sgpr1, implicit $sgpr0_sgpr1
+    ; CHECK-NEXT: renamable $vgpr0 = V_WRITELANE_B32 killed $sgpr1, 1, killed $vgpr0, implicit $sgpr0_sgpr1
     ; CHECK-NEXT: $sgpr0 = V_READLANE_B32 $vgpr0, 0, implicit-def $sgpr0_sgpr1
     ; CHECK-NEXT: $sgpr1 = V_READLANE_B32 $vgpr0, 1
     ; CHECK-NEXT: S_NOP 0, implicit-def dead renamable $sgpr2_sgpr3, implicit-def dead renamable $sgpr0_sgpr1, implicit killed renamable $sgpr0_sgpr1
     ; CHECK-NEXT: $sgpr0 = V_READLANE_B32 $vgpr0, 0, implicit-def $sgpr0_sgpr1
-    ; CHECK-NEXT: $sgpr1 = V_READLANE_B32 $vgpr0, 1
+    ; CHECK-NEXT: $sgpr1 = V_READLANE_B32 killed $vgpr0, 1
     ; CHECK-NEXT: $exec = S_MOV_B64 killed $sgpr0_sgpr1
     ; CHECK-NEXT: S_SENDMSG 0, implicit $m0, implicit $exec
     S_NOP 0, implicit-def $exec
@@ -96,13 +93,12 @@
 body:             |
   bb.0:
     ; CHECK-LABEL: name: reload_sgpr_spill_into_copy_to_exec_lo
-    ; CHECK: liveins: $vgpr0
-    ; CHECK-NEXT: {{  $}}
-    ; CHECK-NEXT: S_NOP 0, implicit-def renamable $sgpr0, implicit-def dead renamable $sgpr1, implicit-def $exec_lo
-    ; CHECK-NEXT: $vgpr0 = V_WRITELANE_B32 killed $sgpr0, 0, $vgpr0
+    ; CHECK: S_NOP 0, implicit-def renamable $sgpr0, implicit-def dead renamable $sgpr1, implicit-def $exec_lo
+    ; CHECK-NEXT: renamable $vgpr0 = IMPLICIT_DEF
+    ; CHECK-NEXT: renamable $vgpr0 = V_WRITELANE_B32 killed $sgpr0, 0, killed $vgpr0
     ; CHECK-NEXT: $sgpr0 = V_READLANE_B32 $vgpr0, 0
     ; CHECK-NEXT: S_NOP 0, implicit killed renamable $sgpr0, implicit-def dead renamable $sgpr1, implicit-def dead renamable $sgpr0
-    ; CHECK-NEXT: $sgpr0 = V_READLANE_B32 $vgpr0, 0
+    ; CHECK-NEXT: $sgpr0 = V_READLANE_B32 killed $vgpr0, 0
     ; CHECK-NEXT: $exec_lo = S_MOV_B32 killed $sgpr0
     ; CHECK-NEXT: S_SENDMSG 0, implicit $m0, implicit $exec
     S_NOP 0, implicit-def %0:sreg_32, implicit-def %1:sreg_32, implicit-def $exec_lo
@@ -120,13 +116,12 @@
 body:             |
   bb.0:
     ; CHECK-LABEL: name: reload_sgpr_spill_into_copy_to_exec_hi
-    ; CHECK: liveins: $vgpr0
-    ; CHECK-NEXT: {{  $}}
-    ; CHECK-NEXT: S_NOP 0, implicit-def renamable $sgpr0, implicit-def dead renamable $sgpr1, implicit-def $exec_hi
-    ; CHECK-NEXT: $vgpr0 = V_WRITELANE_B32 killed $sgpr0, 0, $vgpr0
+    ; CHECK: S_NOP 0, implicit-def renamable $sgpr0, implicit-def dead renamable $sgpr1, implicit-def $exec_hi
+    ; CHECK-NEXT: renamable $vgpr0 = IMPLICIT_DEF
+    ; CHECK-NEXT: renamable $vgpr0 = V_WRITELANE_B32 killed $sgpr0, 0, killed $vgpr0
     ; CHECK-NEXT: $sgpr0 = V_READLANE_B32 $vgpr0, 0
     ; CHECK-NEXT: S_NOP 0, implicit killed renamable $sgpr0, implicit-def dead renamable $sgpr1, implicit-def dead renamable $sgpr0
-    ; CHECK-NEXT: $sgpr0 = V_READLANE_B32 $vgpr0, 0
+    ; CHECK-NEXT: $sgpr0 = V_READLANE_B32 killed $vgpr0, 0
     ; CHECK-NEXT: $exec_hi = S_MOV_B32 killed $sgpr0
     ; CHECK-NEXT: S_SENDMSG 0, implicit $m0, implicit $exec
     S_NOP 0, implicit-def %0:sreg_32, implicit-def %1:sreg_32, implicit-def $exec_hi
@@ -144,16 +139,15 @@
 body:             |
   bb.0:
     ; CHECK-LABEL: name: reload_sgpr_spill_into_copy_to_exec
-    ; CHECK: liveins: $vgpr0
-    ; CHECK-NEXT: {{  $}}
-    ; CHECK-NEXT: S_NOP 0, implicit-def renamable $sgpr0_sgpr1, implicit-def dead renamable $sgpr2_sgpr3, implicit-def $exec
-    ; CHECK-NEXT: $vgpr0 = V_WRITELANE_B32 killed $sgpr0, 0, $vgpr0, implicit-def $sgpr0_sgpr1, implicit $sgpr0_sgpr1
-    ; CHECK-NEXT: $vgpr0 = V_WRITELANE_B32 killed $sgpr1, 1, $vgpr0, implicit $sgpr0_sgpr1
+    ; CHECK: S_NOP 0, implicit-def renamable $sgpr0_sgpr1, implicit-def dead renamable $sgpr2_sgpr3, implicit-def $exec
+    ; CHECK-NEXT: renamable $vgpr0 = IMPLICIT_DEF
+    ; CHECK-NEXT: renamable $vgpr0 = V_WRITELANE_B32 killed $sgpr0, 0, killed $vgpr0, implicit-def $sgpr0_sgpr1, implicit $sgpr0_sgpr1
+    ; CHECK-NEXT: renamable $vgpr0 = V_WRITELANE_B32 killed $sgpr1, 1, killed $vgpr0, implicit $sgpr0_sgpr1
     ; CHECK-NEXT: $sgpr0 = V_READLANE_B32 $vgpr0, 0, implicit-def $sgpr0_sgpr1
     ; CHECK-NEXT: $sgpr1 = V_READLANE_B32 $vgpr0, 1
     ; CHECK-NEXT: S_NOP 0, implicit killed renamable $sgpr0_sgpr1, implicit-def dead renamable $sgpr2_sgpr3, implicit-def dead renamable $sgpr0_sgpr1
     ; CHECK-NEXT: $sgpr0 = V_READLANE_B32 $vgpr0, 0, implicit-def $sgpr0_sgpr1
-    ; CHECK-NEXT: $sgpr1 = V_READLANE_B32 $vgpr0, 1
+    ; CHECK-NEXT: $sgpr1 = V_READLANE_B32 killed $vgpr0, 1
     ; CHECK-NEXT: $exec = S_MOV_B64 killed $sgpr0_sgpr1
     ; CHECK-NEXT: S_SENDMSG 0, implicit $m0, implicit $exec
     S_NOP 0, implicit-def %0:sreg_64, implicit-def %1:sreg_64, implicit-def $exec
diff --git a/llvm/test/CodeGen/AMDGPU/fold-reload-into-m0.mir b/llvm/test/CodeGen/AMDGPU/fold-reload-into-m0.mir
--- a/llvm/test/CodeGen/AMDGPU/fold-reload-into-m0.mir
+++ b/llvm/test/CodeGen/AMDGPU/fold-reload-into-m0.mir
@@ -13,14 +13,13 @@
   bb.0:
 
     ; CHECK-LABEL: name: merge_sgpr_spill_into_copy_from_m0
-    ; CHECK: liveins: $vgpr0
-    ; CHECK-NEXT: {{  $}}
-    ; CHECK-NEXT: S_NOP 0, implicit-def $m0
+    ; CHECK: S_NOP 0, implicit-def $m0
     ; CHECK-NEXT: $sgpr0 = S_MOV_B32 $m0
-    ; CHECK-NEXT: $vgpr0 = V_WRITELANE_B32 killed $sgpr0, 0, $vgpr0
+    ; CHECK-NEXT: renamable $vgpr0 = IMPLICIT_DEF
+    ; CHECK-NEXT: renamable $vgpr0 = V_WRITELANE_B32 killed $sgpr0, 0, killed $vgpr0
     ; CHECK-NEXT: $sgpr0 = V_READLANE_B32 $vgpr0, 0
     ; CHECK-NEXT: S_NOP 0, implicit-def dead renamable $sgpr1, implicit-def dead renamable $sgpr0, implicit killed renamable $sgpr0
-    ; CHECK-NEXT: $sgpr0 = V_READLANE_B32 $vgpr0, 0
+    ; CHECK-NEXT: $sgpr0 = V_READLANE_B32 killed $vgpr0, 0
     ; CHECK-NEXT: $m0 = S_MOV_B32 killed $sgpr0
     ; CHECK-NEXT: S_NOP 0
     ; CHECK-NEXT: S_SENDMSG 0, implicit $m0, implicit $exec
@@ -44,13 +43,12 @@
   bb.0:
 
     ; CHECK-LABEL: name: reload_sgpr_spill_into_copy_to_m0
-    ; CHECK: liveins: $vgpr0
-    ; CHECK-NEXT: {{  $}}
+    ; CHECK: renamable $vgpr0 = IMPLICIT_DEF
     ; CHECK-NEXT: S_NOP 0, implicit-def renamable $sgpr0, implicit-def dead renamable $sgpr1, implicit-def $m0
-    ; CHECK-NEXT: $vgpr0 = V_WRITELANE_B32 killed $sgpr0, 0, $vgpr0
+    ; CHECK-NEXT: renamable $vgpr0 = V_WRITELANE_B32 killed $sgpr0, 0, killed $vgpr0
     ; CHECK-NEXT: $sgpr0 = V_READLANE_B32 $vgpr0, 0
     ; CHECK-NEXT: S_NOP 0, implicit killed renamable $sgpr0, implicit-def dead renamable $sgpr1, implicit-def dead renamable $sgpr0
-    ; CHECK-NEXT: $sgpr0 = V_READLANE_B32 $vgpr0, 0
+    ; CHECK-NEXT: $sgpr0 = V_READLANE_B32 killed $vgpr0, 0
     ; CHECK-NEXT: $m0 = S_MOV_B32 killed $sgpr0
     ; CHECK-NEXT: S_NOP 0
     ; CHECK-NEXT: S_SENDMSG 0, implicit $m0, implicit $exec
diff --git a/llvm/test/CodeGen/AMDGPU/frame-setup-without-sgpr-to-vgpr-spills.ll b/llvm/test/CodeGen/AMDGPU/frame-setup-without-sgpr-to-vgpr-spills.ll
--- a/llvm/test/CodeGen/AMDGPU/frame-setup-without-sgpr-to-vgpr-spills.ll
+++ b/llvm/test/CodeGen/AMDGPU/frame-setup-without-sgpr-to-vgpr-spills.ll
@@ -16,6 +16,7 @@
 ; SPILL-TO-VGPR-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; SPILL-TO-VGPR-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
 ; SPILL-TO-VGPR-NEXT:    s_mov_b64 exec, s[8:9]
+; SPILL-TO-VGPR-NEXT:    ; implicit-def: $vgpr40
 ; SPILL-TO-VGPR-NEXT:    s_addk_i32 s32, 0x400
 ; SPILL-TO-VGPR-NEXT:    v_writelane_b32 v40, s30, 0
 ; SPILL-TO-VGPR-NEXT:    v_mov_b32_e32 v0, 0
diff --git a/llvm/test/CodeGen/AMDGPU/gfx-call-non-gfx-func.ll b/llvm/test/CodeGen/AMDGPU/gfx-call-non-gfx-func.ll
--- a/llvm/test/CodeGen/AMDGPU/gfx-call-non-gfx-func.ll
+++ b/llvm/test/CodeGen/AMDGPU/gfx-call-non-gfx-func.ll
@@ -8,11 +8,13 @@
 ; SDAG-LABEL: gfx_func:
 ; SDAG:       ; %bb.0:
 ; SDAG-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; SDAG-NEXT:    s_mov_b32 s36, s33
+; SDAG-NEXT:    s_mov_b32 s38, s33
 ; SDAG-NEXT:    s_mov_b32 s33, s32
 ; SDAG-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; SDAG-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; SDAG-NEXT:    s_mov_b64 exec, s[34:35]
+; SDAG-NEXT:    ; implicit-def: $vgpr40
+; SDAG-NEXT:    s_addk_i32 s32, 0x400
 ; SDAG-NEXT:    v_writelane_b32 v40, s4, 0
 ; SDAG-NEXT:    v_writelane_b32 v40, s5, 1
 ; SDAG-NEXT:    v_writelane_b32 v40, s6, 2
@@ -33,7 +35,6 @@
 ; SDAG-NEXT:    v_writelane_b32 v40, s21, 17
 ; SDAG-NEXT:    v_writelane_b32 v40, s22, 18
 ; SDAG-NEXT:    v_writelane_b32 v40, s23, 19
-; SDAG-NEXT:    s_addk_i32 s32, 0x400
 ; SDAG-NEXT:    v_writelane_b32 v40, s24, 20
 ; SDAG-NEXT:    v_writelane_b32 v40, s25, 21
 ; SDAG-NEXT:    s_getpc_b64 s[34:35]
@@ -81,18 +82,20 @@
 ; SDAG-NEXT:    buffer_load_dword v40, off, s[0:3], s33 ; 4-byte Folded Reload
 ; SDAG-NEXT:    s_mov_b64 exec, s[34:35]
 ; SDAG-NEXT:    s_addk_i32 s32, 0xfc00
-; SDAG-NEXT:    s_mov_b32 s33, s36
+; SDAG-NEXT:    s_mov_b32 s33, s38
 ; SDAG-NEXT:    s_waitcnt vmcnt(0)
 ; SDAG-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GISEL-LABEL: gfx_func:
 ; GISEL:       ; %bb.0:
 ; GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GISEL-NEXT:    s_mov_b32 s36, s33
+; GISEL-NEXT:    s_mov_b32 s38, s33
 ; GISEL-NEXT:    s_mov_b32 s33, s32
 ; GISEL-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GISEL-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GISEL-NEXT:    s_mov_b64 exec, s[34:35]
+; GISEL-NEXT:    ; implicit-def: $vgpr40
+; GISEL-NEXT:    s_addk_i32 s32, 0x400
 ; GISEL-NEXT:    v_writelane_b32 v40, s4, 0
 ; GISEL-NEXT:    v_writelane_b32 v40, s5, 1
 ; GISEL-NEXT:    v_writelane_b32 v40, s6, 2
@@ -113,7 +116,6 @@
 ; GISEL-NEXT:    v_writelane_b32 v40, s21, 17
 ; GISEL-NEXT:    v_writelane_b32 v40, s22, 18
 ; GISEL-NEXT:    v_writelane_b32 v40, s23, 19
-; GISEL-NEXT:    s_addk_i32 s32, 0x400
 ; GISEL-NEXT:    v_writelane_b32 v40, s24, 20
 ; GISEL-NEXT:    v_writelane_b32 v40, s25, 21
 ; GISEL-NEXT:    s_getpc_b64 s[34:35]
@@ -161,7 +163,7 @@
 ; GISEL-NEXT:    buffer_load_dword v40, off, s[0:3], s33 ; 4-byte Folded Reload
 ; GISEL-NEXT:    s_mov_b64 exec, s[34:35]
 ; GISEL-NEXT:    s_addk_i32 s32, 0xfc00
-; GISEL-NEXT:    s_mov_b32 s33, s36
+; GISEL-NEXT:    s_mov_b32 s33, s38
 ; GISEL-NEXT:    s_waitcnt vmcnt(0)
 ; GISEL-NEXT:    s_setpc_b64 s[30:31]
   call void @extern_c_func()
diff --git a/llvm/test/CodeGen/AMDGPU/gfx-callable-argument-types.ll b/llvm/test/CodeGen/AMDGPU/gfx-callable-argument-types.ll
--- a/llvm/test/CodeGen/AMDGPU/gfx-callable-argument-types.ll
+++ b/llvm/test/CodeGen/AMDGPU/gfx-callable-argument-types.ll
@@ -103,6 +103,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 1
@@ -136,15 +137,16 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 1
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i1@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_i1@rel32@hi+12
-; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-NEXT:    buffer_store_byte v0, off, s[0:3], s32
+; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
@@ -171,16 +173,18 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_mov_b32_e32 v0, 1
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_i1@rel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, external_void_func_i1@rel32@hi+12
-; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX11-NEXT:    scratch_store_b8 off, v0, s32
+; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
+; GFX11-NEXT:    s_delay_alu instid0(VALU_DEP_1)
 ; GFX11-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX11-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX11-NEXT:    v_readlane_b32 s0, v41, 0
@@ -205,15 +209,16 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 1
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_i1@rel32@lo+4
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_i1@rel32@hi+12
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-SCRATCH-NEXT:    scratch_store_byte off, v0, s32
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-SCRATCH-NEXT:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s30, v40, 0
@@ -244,6 +249,7 @@
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
 ; GFX9-NEXT:    global_load_ubyte v0, v[0:1], off glc
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
@@ -251,6 +257,7 @@
 ; GFX9-NEXT:    s_getpc_b64 s[34:35]
 ; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_i1_signext@rel32@lo+4
 ; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_i1_signext@rel32@hi+12
+; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    v_and_b32_e32 v0, 1, v0
 ; GFX9-NEXT:    buffer_store_byte v0, off, s[0:3], s32
 ; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
@@ -279,13 +286,15 @@
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
 ; GFX10-NEXT:    global_load_ubyte v0, v[0:1], off glc dlc
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i1_signext@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_i1_signext@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    v_and_b32_e32 v0, 1, v0
 ; GFX10-NEXT:    buffer_store_byte v0, off, s[0:3], s32
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
@@ -316,13 +325,15 @@
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX11-NEXT:    global_load_u8 v0, v[0:1], off glc dlc
 ; GFX11-NEXT:    s_waitcnt vmcnt(0)
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_i1_signext@rel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, external_void_func_i1_signext@rel32@hi+12
 ; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
 ; GFX11-NEXT:    v_and_b32_e32 v0, 1, v0
 ; GFX11-NEXT:    scratch_store_b8 off, v0, s32
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
@@ -352,13 +363,15 @@
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX10-SCRATCH-NEXT:    global_load_ubyte v0, v[0:1], off glc dlc
 ; GFX10-SCRATCH-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_i1_signext@rel32@lo+4
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_i1_signext@rel32@hi+12
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 1
+; GFX10-SCRATCH-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-SCRATCH-NEXT:    v_and_b32_e32 v0, 1, v0
 ; GFX10-SCRATCH-NEXT:    scratch_store_byte off, v0, s32
 ; GFX10-SCRATCH-NEXT:    s_swappc_b64 s[30:31], s[0:1]
@@ -392,6 +405,7 @@
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
 ; GFX9-NEXT:    global_load_ubyte v0, v[0:1], off glc
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
@@ -399,6 +413,7 @@
 ; GFX9-NEXT:    s_getpc_b64 s[34:35]
 ; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_i1_zeroext@rel32@lo+4
 ; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_i1_zeroext@rel32@hi+12
+; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    v_and_b32_e32 v0, 1, v0
 ; GFX9-NEXT:    buffer_store_byte v0, off, s[0:3], s32
 ; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
@@ -427,13 +442,15 @@
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
 ; GFX10-NEXT:    global_load_ubyte v0, v[0:1], off glc dlc
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i1_zeroext@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_i1_zeroext@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    v_and_b32_e32 v0, 1, v0
 ; GFX10-NEXT:    buffer_store_byte v0, off, s[0:3], s32
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
@@ -464,13 +481,15 @@
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX11-NEXT:    global_load_u8 v0, v[0:1], off glc dlc
 ; GFX11-NEXT:    s_waitcnt vmcnt(0)
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_i1_zeroext@rel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, external_void_func_i1_zeroext@rel32@hi+12
 ; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
 ; GFX11-NEXT:    v_and_b32_e32 v0, 1, v0
 ; GFX11-NEXT:    scratch_store_b8 off, v0, s32
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
@@ -500,13 +519,15 @@
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX10-SCRATCH-NEXT:    global_load_ubyte v0, v[0:1], off glc dlc
 ; GFX10-SCRATCH-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_i1_zeroext@rel32@lo+4
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_i1_zeroext@rel32@hi+12
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 1
+; GFX10-SCRATCH-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-SCRATCH-NEXT:    v_and_b32_e32 v0, 1, v0
 ; GFX10-SCRATCH-NEXT:    scratch_store_byte off, v0, s32
 ; GFX10-SCRATCH-NEXT:    s_swappc_b64 s[30:31], s[0:1]
@@ -538,6 +559,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0x7b
@@ -570,8 +592,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
@@ -604,8 +627,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
@@ -638,8 +662,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
@@ -676,6 +701,7 @@
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
 ; GFX9-NEXT:    global_load_sbyte v0, v[0:1], off glc
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
@@ -709,8 +735,9 @@
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
 ; GFX10-NEXT:    global_load_sbyte v0, v[0:1], off glc dlc
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i8_signext@rel32@lo+4
@@ -744,8 +771,9 @@
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX11-NEXT:    global_load_i8 v0, v[0:1], off glc dlc
 ; GFX11-NEXT:    s_waitcnt vmcnt(0)
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_i8_signext@rel32@lo+4
@@ -779,8 +807,9 @@
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX10-SCRATCH-NEXT:    global_load_sbyte v0, v[0:1], off glc dlc
 ; GFX10-SCRATCH-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_i8_signext@rel32@lo+4
@@ -817,6 +846,7 @@
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
 ; GFX9-NEXT:    global_load_ubyte v0, v[0:1], off glc
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
@@ -850,8 +880,9 @@
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
 ; GFX10-NEXT:    global_load_ubyte v0, v[0:1], off glc dlc
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i8_zeroext@rel32@lo+4
@@ -885,8 +916,9 @@
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX11-NEXT:    global_load_u8 v0, v[0:1], off glc dlc
 ; GFX11-NEXT:    s_waitcnt vmcnt(0)
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_i8_zeroext@rel32@lo+4
@@ -920,8 +952,9 @@
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX10-SCRATCH-NEXT:    global_load_ubyte v0, v[0:1], off glc dlc
 ; GFX10-SCRATCH-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_i8_zeroext@rel32@lo+4
@@ -956,6 +989,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0x7b
@@ -988,8 +1022,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
@@ -1022,8 +1057,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
@@ -1056,8 +1092,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
@@ -1094,6 +1131,7 @@
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
 ; GFX9-NEXT:    global_load_ushort v0, v[0:1], off glc
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
@@ -1127,8 +1165,9 @@
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
 ; GFX10-NEXT:    global_load_ushort v0, v[0:1], off glc dlc
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i16_signext@rel32@lo+4
@@ -1162,8 +1201,9 @@
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX11-NEXT:    global_load_u16 v0, v[0:1], off glc dlc
 ; GFX11-NEXT:    s_waitcnt vmcnt(0)
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_i16_signext@rel32@lo+4
@@ -1197,8 +1237,9 @@
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX10-SCRATCH-NEXT:    global_load_ushort v0, v[0:1], off glc dlc
 ; GFX10-SCRATCH-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_i16_signext@rel32@lo+4
@@ -1235,6 +1276,7 @@
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
 ; GFX9-NEXT:    global_load_ushort v0, v[0:1], off glc
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
@@ -1268,8 +1310,9 @@
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
 ; GFX10-NEXT:    global_load_ushort v0, v[0:1], off glc dlc
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i16_zeroext@rel32@lo+4
@@ -1303,8 +1346,9 @@
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX11-NEXT:    global_load_u16 v0, v[0:1], off glc dlc
 ; GFX11-NEXT:    s_waitcnt vmcnt(0)
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_i16_zeroext@rel32@lo+4
@@ -1338,8 +1382,9 @@
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX10-SCRATCH-NEXT:    global_load_ushort v0, v[0:1], off glc dlc
 ; GFX10-SCRATCH-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_i16_zeroext@rel32@lo+4
@@ -1374,6 +1419,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 42
@@ -1406,8 +1452,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 42
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
@@ -1440,8 +1487,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_mov_b32_e32 v0, 42
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
@@ -1474,8 +1522,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 42
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
@@ -1510,6 +1559,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0x7b
@@ -1543,15 +1593,16 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
-; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i64@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_i64@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
@@ -1578,16 +1629,17 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_dual_mov_b32 v0, 0x7b :: v_dual_mov_b32 v1, 0
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
-; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_i64@rel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, external_void_func_i64@rel32@hi+12
-; GFX11-NEXT:    s_delay_alu instid0(SALU_CYCLE_1)
+; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
+; GFX11-NEXT:    s_delay_alu instid0(VALU_DEP_1)
 ; GFX11-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX11-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX11-NEXT:    v_readlane_b32 s0, v41, 0
@@ -1612,15 +1664,16 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_i64@rel32@lo+4
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_i64@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-SCRATCH-NEXT:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s30, v40, 0
@@ -1652,6 +1705,7 @@
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX9-NEXT:    global_load_dwordx4 v[0:3], v[0:1], off
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
@@ -1685,8 +1739,9 @@
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 0
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2i64@rel32@lo+4
@@ -1721,8 +1776,9 @@
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX11-NEXT:    v_mov_b32_e32 v0, 0
 ; GFX11-NEXT:    v_mov_b32_e32 v1, 0
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v2i64@rel32@lo+4
@@ -1757,8 +1813,9 @@
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v1, 0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v2i64@rel32@lo+4
@@ -1794,6 +1851,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 1
@@ -1829,8 +1887,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 1
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 3
 ; GFX10-NEXT:    v_mov_b32_e32 v3, 4
@@ -1866,8 +1925,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_dual_mov_b32 v0, 1 :: v_dual_mov_b32 v1, 2
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_dual_mov_b32 v2, 3 :: v_dual_mov_b32 v3, 4
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
@@ -1901,8 +1961,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 1
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v1, 2
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v2, 3
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v3, 4
@@ -1943,6 +2004,7 @@
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX9-NEXT:    global_load_dwordx4 v[0:3], v[0:1], off
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v4, 1
@@ -1978,8 +2040,9 @@
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 0
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_mov_b32_e32 v4, 1
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v5, 2
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    global_load_dwordx4 v[0:3], v[0:1], off
@@ -2016,8 +2079,9 @@
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX11-NEXT:    v_dual_mov_b32 v0, 0 :: v_dual_mov_b32 v5, 2
 ; GFX11-NEXT:    v_dual_mov_b32 v1, 0 :: v_dual_mov_b32 v4, 1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    global_load_b128 v[0:3], v[0:1], off
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
@@ -2052,8 +2116,9 @@
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v1, 0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v4, 1
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v5, 2
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    global_load_dwordx4 v[0:3], v[0:1], off
@@ -2096,6 +2161,7 @@
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX9-NEXT:    global_load_dwordx4 v[0:3], v[0:1], off
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v4, 1
@@ -2133,8 +2199,9 @@
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 0
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_mov_b32_e32 v4, 1
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v5, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v6, 3
 ; GFX10-NEXT:    global_load_dwordx4 v[0:3], v[0:1], off
@@ -2173,17 +2240,18 @@
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX11-NEXT:    v_dual_mov_b32 v0, 0 :: v_dual_mov_b32 v5, 2
 ; GFX11-NEXT:    v_dual_mov_b32 v1, 0 :: v_dual_mov_b32 v4, 1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_dual_mov_b32 v6, 3 :: v_dual_mov_b32 v7, 4
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    global_load_b128 v[0:3], v[0:1], off
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
-; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v4i64@rel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, external_void_func_v4i64@rel32@hi+12
-; GFX11-NEXT:    s_delay_alu instid0(SALU_CYCLE_1)
+; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
+; GFX11-NEXT:    s_delay_alu instid0(VALU_DEP_1)
 ; GFX11-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX11-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX11-NEXT:    v_readlane_b32 s0, v41, 0
@@ -2210,8 +2278,9 @@
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v1, 0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v4, 1
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v5, 2
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v6, 3
 ; GFX10-SCRATCH-NEXT:    global_load_dwordx4 v[0:3], v[0:1], off
@@ -2252,6 +2321,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0x4400
@@ -2284,8 +2354,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 0x4400
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
@@ -2318,8 +2389,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_mov_b32_e32 v0, 0x4400
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
@@ -2352,8 +2424,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 0x4400
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
@@ -2388,6 +2461,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 4.0
@@ -2420,8 +2494,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 4.0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
@@ -2454,8 +2529,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_mov_b32_e32 v0, 4.0
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
@@ -2488,8 +2564,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 4.0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
@@ -2524,6 +2601,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 1.0
@@ -2557,15 +2635,16 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 1.0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 2.0
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
-; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2f32@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v2f32@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
@@ -2592,16 +2671,17 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_dual_mov_b32 v0, 1.0 :: v_dual_mov_b32 v1, 2.0
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
-; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v2f32@rel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, external_void_func_v2f32@rel32@hi+12
-; GFX11-NEXT:    s_delay_alu instid0(SALU_CYCLE_1)
+; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
+; GFX11-NEXT:    s_delay_alu instid0(VALU_DEP_1)
 ; GFX11-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX11-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX11-NEXT:    v_readlane_b32 s0, v41, 0
@@ -2626,15 +2706,16 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 1.0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v1, 2.0
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v2f32@rel32@lo+4
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_v2f32@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-SCRATCH-NEXT:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s30, v40, 0
@@ -2663,6 +2744,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 1.0
@@ -2697,8 +2779,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 1.0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 2.0
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 4.0
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
@@ -2733,17 +2816,18 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_dual_mov_b32 v0, 1.0 :: v_dual_mov_b32 v1, 2.0
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_mov_b32_e32 v2, 4.0
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
-; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v3f32@rel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, external_void_func_v3f32@rel32@hi+12
-; GFX11-NEXT:    s_delay_alu instid0(SALU_CYCLE_1)
+; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
+; GFX11-NEXT:    s_delay_alu instid0(VALU_DEP_1)
 ; GFX11-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX11-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX11-NEXT:    v_readlane_b32 s0, v41, 0
@@ -2768,8 +2852,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 1.0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v1, 2.0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v2, 4.0
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
@@ -2806,6 +2891,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 1.0
@@ -2842,8 +2928,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 1.0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 2.0
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 4.0
 ; GFX10-NEXT:    v_mov_b32_e32 v3, -1.0
@@ -2880,8 +2967,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_dual_mov_b32 v0, 1.0 :: v_dual_mov_b32 v1, 2.0
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_dual_mov_b32 v2, 4.0 :: v_dual_mov_b32 v3, -1.0
 ; GFX11-NEXT:    v_mov_b32_e32 v4, 0.5
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
@@ -2916,8 +3004,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 1.0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v1, 2.0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v2, 4.0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v3, -1.0
@@ -2956,6 +3045,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0
@@ -2989,15 +3079,16 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 0x40100000
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
-; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_f64@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_f64@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
@@ -3024,16 +3115,17 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_dual_mov_b32 v0, 0 :: v_dual_mov_b32 v1, 0x40100000
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
-; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_f64@rel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, external_void_func_f64@rel32@hi+12
-; GFX11-NEXT:    s_delay_alu instid0(SALU_CYCLE_1)
+; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
+; GFX11-NEXT:    s_delay_alu instid0(VALU_DEP_1)
 ; GFX11-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX11-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX11-NEXT:    v_readlane_b32 s0, v41, 0
@@ -3058,15 +3150,16 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v1, 0x40100000
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_f64@rel32@lo+4
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_f64@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-SCRATCH-NEXT:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s30, v40, 0
@@ -3095,6 +3188,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0
@@ -3130,8 +3224,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 2.0
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v3, 0x40100000
@@ -3167,8 +3262,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_dual_mov_b32 v0, 0 :: v_dual_mov_b32 v1, 2.0
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_dual_mov_b32 v2, 0 :: v_dual_mov_b32 v3, 0x40100000
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
@@ -3202,8 +3298,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v1, 2.0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v3, 0x40100000
@@ -3241,6 +3338,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0
@@ -3278,8 +3376,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 2.0
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v3, 0x40100000
@@ -3317,8 +3416,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_dual_mov_b32 v0, 0 :: v_dual_mov_b32 v1, 2.0
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_dual_mov_b32 v2, 0 :: v_dual_mov_b32 v3, 0x40100000
 ; GFX11-NEXT:    v_dual_mov_b32 v4, 0 :: v_dual_mov_b32 v5, 0x40200000
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
@@ -3353,8 +3453,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v1, 2.0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v3, 0x40100000
@@ -3395,6 +3496,7 @@
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
 ; GFX9-NEXT:    global_load_dword v0, v[0:1], off
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
@@ -3427,8 +3529,9 @@
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
 ; GFX10-NEXT:    global_load_dword v0, v[0:1], off
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2i16@rel32@lo+4
@@ -3461,8 +3564,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX11-NEXT:    global_load_b32 v0, v[0:1], off
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v2i16@rel32@lo+4
@@ -3495,8 +3599,9 @@
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX10-SCRATCH-NEXT:    global_load_dword v0, v[0:1], off
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v2i16@rel32@lo+4
@@ -3532,6 +3637,7 @@
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
 ; GFX9-NEXT:    global_load_dwordx2 v[0:1], v[0:1], off
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
@@ -3564,8 +3670,9 @@
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
 ; GFX10-NEXT:    global_load_dwordx2 v[0:1], v[0:1], off
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3i16@rel32@lo+4
@@ -3598,8 +3705,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX11-NEXT:    global_load_b64 v[0:1], v[0:1], off
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v3i16@rel32@lo+4
@@ -3632,8 +3740,9 @@
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX10-SCRATCH-NEXT:    global_load_dwordx2 v[0:1], v[0:1], off
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v3i16@rel32@lo+4
@@ -3669,6 +3778,7 @@
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
 ; GFX9-NEXT:    global_load_dwordx2 v[0:1], v[0:1], off
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
@@ -3701,8 +3811,9 @@
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
 ; GFX10-NEXT:    global_load_dwordx2 v[0:1], v[0:1], off
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3f16@rel32@lo+4
@@ -3735,8 +3846,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX11-NEXT:    global_load_b64 v[0:1], v[0:1], off
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v3f16@rel32@lo+4
@@ -3769,8 +3881,9 @@
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX10-SCRATCH-NEXT:    global_load_dwordx2 v[0:1], v[0:1], off
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v3f16@rel32@lo+4
@@ -3805,6 +3918,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0x20001
@@ -3838,15 +3952,16 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 0x20001
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 3
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
-; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3i16@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v3i16@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
@@ -3873,16 +3988,17 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_dual_mov_b32 v0, 0x20001 :: v_dual_mov_b32 v1, 3
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
-; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v3i16@rel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, external_void_func_v3i16@rel32@hi+12
-; GFX11-NEXT:    s_delay_alu instid0(SALU_CYCLE_1)
+; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
+; GFX11-NEXT:    s_delay_alu instid0(VALU_DEP_1)
 ; GFX11-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX11-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX11-NEXT:    v_readlane_b32 s0, v41, 0
@@ -3907,15 +4023,16 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 0x20001
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v1, 3
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v3i16@rel32@lo+4
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_v3i16@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-SCRATCH-NEXT:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s30, v40, 0
@@ -3944,6 +4061,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0x40003c00
@@ -3977,15 +4095,16 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 0x40003c00
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 0x4400
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
-; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3f16@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v3f16@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
@@ -4012,17 +4131,18 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_mov_b32_e32 v0, 0x40003c00
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_mov_b32_e32 v1, 0x4400
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
-; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v3f16@rel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, external_void_func_v3f16@rel32@hi+12
-; GFX11-NEXT:    s_delay_alu instid0(SALU_CYCLE_1)
+; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
+; GFX11-NEXT:    s_delay_alu instid0(VALU_DEP_1)
 ; GFX11-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX11-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX11-NEXT:    v_readlane_b32 s0, v41, 0
@@ -4047,15 +4167,16 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 0x40003c00
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v1, 0x4400
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v3f16@rel32@lo+4
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_v3f16@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-SCRATCH-NEXT:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s30, v40, 0
@@ -4085,6 +4206,7 @@
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
 ; GFX9-NEXT:    global_load_dwordx2 v[0:1], v[0:1], off
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
@@ -4117,8 +4239,9 @@
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
 ; GFX10-NEXT:    global_load_dwordx2 v[0:1], v[0:1], off
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v4i16@rel32@lo+4
@@ -4151,8 +4274,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX11-NEXT:    global_load_b64 v[0:1], v[0:1], off
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v4i16@rel32@lo+4
@@ -4185,8 +4309,9 @@
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX10-SCRATCH-NEXT:    global_load_dwordx2 v[0:1], v[0:1], off
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v4i16@rel32@lo+4
@@ -4221,6 +4346,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0x20001
@@ -4254,15 +4380,16 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 0x20001
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 0x40003
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
-; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v4i16@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v4i16@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
@@ -4289,17 +4416,18 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_mov_b32_e32 v0, 0x20001
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_mov_b32_e32 v1, 0x40003
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
-; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v4i16@rel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, external_void_func_v4i16@rel32@hi+12
-; GFX11-NEXT:    s_delay_alu instid0(SALU_CYCLE_1)
+; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
+; GFX11-NEXT:    s_delay_alu instid0(VALU_DEP_1)
 ; GFX11-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX11-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX11-NEXT:    v_readlane_b32 s0, v41, 0
@@ -4324,15 +4452,16 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 0x20001
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v1, 0x40003
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v4i16@rel32@lo+4
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_v4i16@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-SCRATCH-NEXT:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s30, v40, 0
@@ -4362,6 +4491,7 @@
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
 ; GFX9-NEXT:    global_load_dword v0, v[0:1], off
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
@@ -4394,8 +4524,9 @@
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
 ; GFX10-NEXT:    global_load_dword v0, v[0:1], off
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2f16@rel32@lo+4
@@ -4428,8 +4559,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX11-NEXT:    global_load_b32 v0, v[0:1], off
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v2f16@rel32@lo+4
@@ -4462,8 +4594,9 @@
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX10-SCRATCH-NEXT:    global_load_dword v0, v[0:1], off
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v2f16@rel32@lo+4
@@ -4499,6 +4632,7 @@
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
 ; GFX9-NEXT:    global_load_dwordx2 v[0:1], v[0:1], off
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
@@ -4531,8 +4665,9 @@
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
 ; GFX10-NEXT:    global_load_dwordx2 v[0:1], v[0:1], off
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2i32@rel32@lo+4
@@ -4565,8 +4700,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX11-NEXT:    global_load_b64 v[0:1], v[0:1], off
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v2i32@rel32@lo+4
@@ -4599,8 +4735,9 @@
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX10-SCRATCH-NEXT:    global_load_dwordx2 v[0:1], v[0:1], off
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v2i32@rel32@lo+4
@@ -4635,6 +4772,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 1
@@ -4668,15 +4806,16 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 1
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 2
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
-; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2i32@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v2i32@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
@@ -4703,16 +4842,17 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_dual_mov_b32 v0, 1 :: v_dual_mov_b32 v1, 2
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
-; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v2i32@rel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, external_void_func_v2i32@rel32@hi+12
-; GFX11-NEXT:    s_delay_alu instid0(SALU_CYCLE_1)
+; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
+; GFX11-NEXT:    s_delay_alu instid0(VALU_DEP_1)
 ; GFX11-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX11-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX11-NEXT:    v_readlane_b32 s0, v41, 0
@@ -4737,15 +4877,16 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 1
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v1, 2
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v2i32@rel32@lo+4
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_v2i32@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-SCRATCH-NEXT:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s30, v40, 0
@@ -4774,6 +4915,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 3
@@ -4808,8 +4950,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 3
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 4
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 5
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
@@ -4844,17 +4987,18 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_dual_mov_b32 v0, 3 :: v_dual_mov_b32 v1, 4
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_mov_b32_e32 v2, 5
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
-; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v3i32@rel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, external_void_func_v3i32@rel32@hi+12
-; GFX11-NEXT:    s_delay_alu instid0(SALU_CYCLE_1)
+; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
+; GFX11-NEXT:    s_delay_alu instid0(VALU_DEP_1)
 ; GFX11-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX11-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX11-NEXT:    v_readlane_b32 s0, v41, 0
@@ -4879,8 +5023,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 3
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v1, 4
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v2, 5
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
@@ -4917,6 +5062,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 3
@@ -4952,8 +5098,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 3
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 4
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 5
 ; GFX10-NEXT:    v_mov_b32_e32 v3, 6
@@ -4989,8 +5136,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_dual_mov_b32 v0, 3 :: v_dual_mov_b32 v1, 4
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_dual_mov_b32 v2, 5 :: v_dual_mov_b32 v3, 6
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
@@ -5024,8 +5172,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 3
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v1, 4
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v2, 5
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v3, 6
@@ -5064,6 +5213,7 @@
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
 ; GFX9-NEXT:    global_load_dwordx4 v[0:3], v[0:1], off
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
@@ -5096,8 +5246,9 @@
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
 ; GFX10-NEXT:    global_load_dwordx4 v[0:3], v[0:1], off
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v4i32@rel32@lo+4
@@ -5130,8 +5281,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX11-NEXT:    global_load_b128 v[0:3], v[0:1], off
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v4i32@rel32@lo+4
@@ -5164,8 +5316,9 @@
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX10-SCRATCH-NEXT:    global_load_dwordx4 v[0:3], v[0:1], off
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v4i32@rel32@lo+4
@@ -5200,6 +5353,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 1
@@ -5235,8 +5389,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 1
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 3
 ; GFX10-NEXT:    v_mov_b32_e32 v3, 4
@@ -5272,8 +5427,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_dual_mov_b32 v0, 1 :: v_dual_mov_b32 v1, 2
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_dual_mov_b32 v2, 3 :: v_dual_mov_b32 v3, 4
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
@@ -5307,8 +5463,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 1
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v1, 2
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v2, 3
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v3, 4
@@ -5346,6 +5503,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 1
@@ -5382,8 +5540,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 1
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 3
 ; GFX10-NEXT:    v_mov_b32_e32 v3, 4
@@ -5420,8 +5579,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_dual_mov_b32 v0, 1 :: v_dual_mov_b32 v1, 2
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_dual_mov_b32 v2, 3 :: v_dual_mov_b32 v3, 4
 ; GFX11-NEXT:    v_mov_b32_e32 v4, 5
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
@@ -5456,8 +5616,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 1
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v1, 2
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v2, 3
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v3, 4
@@ -5499,6 +5660,7 @@
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX9-NEXT:    v_mov_b32_e32 v8, 0
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
@@ -5535,8 +5697,9 @@
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX10-NEXT:    v_mov_b32_e32 v8, 0
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX10-NEXT:    s_clause 0x1
 ; GFX10-NEXT:    global_load_dwordx4 v[0:3], v8, s[34:35]
@@ -5574,8 +5737,9 @@
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_load_b64 s[0:1], s[0:1], 0x0
 ; GFX11-NEXT:    v_mov_b32_e32 v4, 0
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX11-NEXT:    s_clause 0x1
 ; GFX11-NEXT:    global_load_b128 v[0:3], v4, s[0:1]
@@ -5613,8 +5777,9 @@
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v8, 0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX10-SCRATCH-NEXT:    s_clause 0x1
 ; GFX10-SCRATCH-NEXT:    global_load_dwordx4 v[0:3], v8, s[0:1]
@@ -5653,6 +5818,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 1
@@ -5692,8 +5858,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 1
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 3
 ; GFX10-NEXT:    v_mov_b32_e32 v3, 4
@@ -5733,8 +5900,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_dual_mov_b32 v0, 1 :: v_dual_mov_b32 v1, 2
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_dual_mov_b32 v2, 3 :: v_dual_mov_b32 v3, 4
 ; GFX11-NEXT:    v_dual_mov_b32 v4, 5 :: v_dual_mov_b32 v5, 6
 ; GFX11-NEXT:    v_dual_mov_b32 v6, 7 :: v_dual_mov_b32 v7, 8
@@ -5770,8 +5938,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 1
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v1, 2
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v2, 3
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v3, 4
@@ -5816,6 +5985,7 @@
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX9-NEXT:    v_mov_b32_e32 v16, 0
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
@@ -5854,8 +6024,9 @@
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX10-NEXT:    v_mov_b32_e32 v16, 0
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX10-NEXT:    s_clause 0x3
 ; GFX10-NEXT:    global_load_dwordx4 v[0:3], v16, s[34:35]
@@ -5895,8 +6066,9 @@
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_load_b64 s[0:1], s[0:1], 0x0
 ; GFX11-NEXT:    v_mov_b32_e32 v12, 0
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX11-NEXT:    s_clause 0x3
 ; GFX11-NEXT:    global_load_b128 v[0:3], v12, s[0:1]
@@ -5936,8 +6108,9 @@
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v16, 0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX10-SCRATCH-NEXT:    s_clause 0x3
 ; GFX10-SCRATCH-NEXT:    global_load_dwordx4 v[0:3], v16, s[0:1]
@@ -5981,6 +6154,7 @@
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX9-NEXT:    v_mov_b32_e32 v28, 0
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
@@ -6024,8 +6198,9 @@
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX10-NEXT:    v_mov_b32_e32 v32, 0
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX10-NEXT:    s_clause 0x7
 ; GFX10-NEXT:    global_load_dwordx4 v[0:3], v32, s[34:35]
@@ -6069,8 +6244,9 @@
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_load_b64 s[0:1], s[0:1], 0x0
 ; GFX11-NEXT:    v_mov_b32_e32 v28, 0
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX11-NEXT:    s_clause 0x7
 ; GFX11-NEXT:    global_load_b128 v[0:3], v28, s[0:1]
@@ -6114,8 +6290,9 @@
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v32, 0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX10-SCRATCH-NEXT:    s_clause 0x7
 ; GFX10-SCRATCH-NEXT:    global_load_dwordx4 v[0:3], v32, s[0:1]
@@ -6164,6 +6341,7 @@
 ; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX9-NEXT:    v_mov_b32_e32 v28, 0
 ; GFX9-NEXT:    global_load_dword v32, v[0:1], off
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX9-NEXT:    global_load_dwordx4 v[0:3], v28, s[34:35]
@@ -6209,8 +6387,9 @@
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX10-NEXT:    v_mov_b32_e32 v32, 0
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    global_load_dword v33, v[0:1], off
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX10-NEXT:    s_clause 0x7
@@ -6257,8 +6436,9 @@
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_load_b64 s[0:1], s[0:1], 0x0
 ; GFX11-NEXT:    v_mov_b32_e32 v28, 0
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    global_load_b32 v32, v[0:1], off
 ; GFX11-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX11-NEXT:    s_clause 0x7
@@ -6304,8 +6484,9 @@
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v32, 0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    global_load_dword v33, v[0:1], off
 ; GFX10-SCRATCH-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX10-SCRATCH-NEXT:    s_clause 0x7
@@ -6351,31 +6532,32 @@
 ; GFX9-NEXT:    s_mov_b32 s34, s33
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_or_saveexec_b64 s[36:37], -1
-; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
+; GFX9-NEXT:    buffer_store_dword v42, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v43, off, s[0:3], s33 offset:12 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr42
 ; GFX9-NEXT:    s_addk_i32 s32, 0x800
-; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
-; GFX9-NEXT:    buffer_store_dword v42, off, s[0:3], s33 ; 4-byte Folded Spill
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
-; GFX9-NEXT:    v_mov_b32_e32 v41, v0
+; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
+; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 ; 4-byte Folded Spill
+; GFX9-NEXT:    v_writelane_b32 v42, s30, 0
+; GFX9-NEXT:    v_mov_b32_e32 v40, v0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 42
 ; GFX9-NEXT:    v_writelane_b32 v43, s34, 0
-; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    v_mov_b32_e32 v42, v1
+; GFX9-NEXT:    v_writelane_b32 v42, s31, 1
+; GFX9-NEXT:    v_mov_b32_e32 v41, v1
 ; GFX9-NEXT:    s_getpc_b64 s[34:35]
 ; GFX9-NEXT:    s_add_u32 s34, s34, external_i32_func_i32@rel32@lo+4
 ; GFX9-NEXT:    s_addc_u32 s35, s35, external_i32_func_i32@rel32@hi+12
 ; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
-; GFX9-NEXT:    global_store_dword v[41:42], v0, off
+; GFX9-NEXT:    global_store_dword v[40:41], v0, off
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
-; GFX9-NEXT:    buffer_load_dword v42, off, s[0:3], s33 ; 4-byte Folded Reload
-; GFX9-NEXT:    buffer_load_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
-; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
-; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
+; GFX9-NEXT:    buffer_load_dword v41, off, s[0:3], s33 ; 4-byte Folded Reload
+; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
+; GFX9-NEXT:    v_readlane_b32 s31, v42, 1
+; GFX9-NEXT:    v_readlane_b32 s30, v42, 0
 ; GFX9-NEXT:    v_readlane_b32 s34, v43, 0
 ; GFX9-NEXT:    s_or_saveexec_b64 s[36:37], -1
-; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload
+; GFX9-NEXT:    buffer_load_dword v42, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload
 ; GFX9-NEXT:    buffer_load_dword v43, off, s[0:3], s33 offset:12 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
 ; GFX9-NEXT:    s_addk_i32 s32, 0xf800
@@ -6390,34 +6572,35 @@
 ; GFX10-NEXT:    s_mov_b32 s34, s33
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_or_saveexec_b32 s35, -1
-; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
+; GFX10-NEXT:    buffer_store_dword v42, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
 ; GFX10-NEXT:    buffer_store_dword v43, off, s[0:3], s33 offset:12 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
-; GFX10-NEXT:    buffer_store_dword v42, off, s[0:3], s33 ; 4-byte Folded Spill
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
-; GFX10-NEXT:    v_mov_b32_e32 v41, v0
+; GFX10-NEXT:    ; implicit-def: $vgpr42
+; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
+; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 ; 4-byte Folded Spill
+; GFX10-NEXT:    v_writelane_b32 v42, s30, 0
+; GFX10-NEXT:    v_mov_b32_e32 v40, v0
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 42
 ; GFX10-NEXT:    s_addk_i32 s32, 0x400
 ; GFX10-NEXT:    v_writelane_b32 v43, s34, 0
-; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    v_mov_b32_e32 v42, v1
+; GFX10-NEXT:    v_writelane_b32 v42, s31, 1
+; GFX10-NEXT:    v_mov_b32_e32 v41, v1
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_i32_func_i32@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_i32_func_i32@rel32@hi+12
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
-; GFX10-NEXT:    global_store_dword v[41:42], v0, off
+; GFX10-NEXT:    global_store_dword v[40:41], v0, off
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    s_clause 0x1
-; GFX10-NEXT:    buffer_load_dword v42, off, s[0:3], s33
-; GFX10-NEXT:    buffer_load_dword v41, off, s[0:3], s33 offset:4
-; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
-; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
+; GFX10-NEXT:    buffer_load_dword v41, off, s[0:3], s33
+; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s33 offset:4
+; GFX10-NEXT:    v_readlane_b32 s31, v42, 1
+; GFX10-NEXT:    v_readlane_b32 s30, v42, 0
 ; GFX10-NEXT:    v_readlane_b32 s34, v43, 0
 ; GFX10-NEXT:    s_or_saveexec_b32 s35, -1
 ; GFX10-NEXT:    s_clause 0x1
-; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s33 offset:8
+; GFX10-NEXT:    buffer_load_dword v42, off, s[0:3], s33 offset:8
 ; GFX10-NEXT:    buffer_load_dword v43, off, s[0:3], s33 offset:12
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
@@ -6434,34 +6617,35 @@
 ; GFX11-NEXT:    s_mov_b32 s33, s32
 ; GFX11-NEXT:    s_or_saveexec_b32 s1, -1
 ; GFX11-NEXT:    s_clause 0x1
-; GFX11-NEXT:    scratch_store_b32 off, v40, s33 offset:8
+; GFX11-NEXT:    scratch_store_b32 off, v42, s33 offset:8
 ; GFX11-NEXT:    scratch_store_b32 off, v43, s33 offset:12
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
+; GFX11-NEXT:    ; implicit-def: $vgpr42
 ; GFX11-NEXT:    s_clause 0x1
-; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
-; GFX11-NEXT:    scratch_store_b32 off, v42, s33
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
-; GFX11-NEXT:    v_dual_mov_b32 v42, v1 :: v_dual_mov_b32 v41, v0
+; GFX11-NEXT:    scratch_store_b32 off, v40, s33 offset:4
+; GFX11-NEXT:    scratch_store_b32 off, v41, s33
+; GFX11-NEXT:    v_writelane_b32 v42, s30, 0
+; GFX11-NEXT:    v_dual_mov_b32 v41, v1 :: v_dual_mov_b32 v40, v0
 ; GFX11-NEXT:    v_mov_b32_e32 v0, 42
 ; GFX11-NEXT:    s_add_i32 s32, s32, 32
 ; GFX11-NEXT:    v_writelane_b32 v43, s0, 0
-; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
+; GFX11-NEXT:    v_writelane_b32 v42, s31, 1
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_i32_func_i32@rel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, external_i32_func_i32@rel32@hi+12
 ; GFX11-NEXT:    s_delay_alu instid0(SALU_CYCLE_1)
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
-; GFX11-NEXT:    global_store_b32 v[41:42], v0, off dlc
+; GFX11-NEXT:    global_store_b32 v[40:41], v0, off dlc
 ; GFX11-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX11-NEXT:    s_clause 0x1
-; GFX11-NEXT:    scratch_load_b32 v42, off, s33
-; GFX11-NEXT:    scratch_load_b32 v41, off, s33 offset:4
-; GFX11-NEXT:    v_readlane_b32 s31, v40, 1
-; GFX11-NEXT:    v_readlane_b32 s30, v40, 0
+; GFX11-NEXT:    scratch_load_b32 v41, off, s33
+; GFX11-NEXT:    scratch_load_b32 v40, off, s33 offset:4
+; GFX11-NEXT:    v_readlane_b32 s31, v42, 1
+; GFX11-NEXT:    v_readlane_b32 s30, v42, 0
 ; GFX11-NEXT:    v_readlane_b32 s0, v43, 0
 ; GFX11-NEXT:    s_or_saveexec_b32 s1, -1
 ; GFX11-NEXT:    s_clause 0x1
-; GFX11-NEXT:    scratch_load_b32 v40, off, s33 offset:8
+; GFX11-NEXT:    scratch_load_b32 v42, off, s33 offset:8
 ; GFX11-NEXT:    scratch_load_b32 v43, off, s33 offset:12
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX11-NEXT:    s_addk_i32 s32, 0xffe0
@@ -6476,34 +6660,35 @@
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s0, s33
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_or_saveexec_b32 s1, -1
-; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v40, s33 offset:8 ; 4-byte Folded Spill
+; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v42, s33 offset:8 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v43, s33 offset:12 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
-; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v42, s33 ; 4-byte Folded Spill
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
-; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v41, v0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr42
+; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v40, s33 offset:4 ; 4-byte Folded Spill
+; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 ; 4-byte Folded Spill
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v42, s30, 0
+; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v40, v0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 42
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 32
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v43, s0, 0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v42, v1
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v42, s31, 1
+; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v41, v1
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_i32_func_i32@rel32@lo+4
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_i32_func_i32@rel32@hi+12
 ; GFX10-SCRATCH-NEXT:    s_swappc_b64 s[30:31], s[0:1]
-; GFX10-SCRATCH-NEXT:    global_store_dword v[41:42], v0, off
+; GFX10-SCRATCH-NEXT:    global_store_dword v[40:41], v0, off
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-SCRATCH-NEXT:    s_clause 0x1
-; GFX10-SCRATCH-NEXT:    scratch_load_dword v42, off, s33
-; GFX10-SCRATCH-NEXT:    scratch_load_dword v41, off, s33 offset:4
-; GFX10-SCRATCH-NEXT:    v_readlane_b32 s31, v40, 1
-; GFX10-SCRATCH-NEXT:    v_readlane_b32 s30, v40, 0
+; GFX10-SCRATCH-NEXT:    scratch_load_dword v41, off, s33
+; GFX10-SCRATCH-NEXT:    scratch_load_dword v40, off, s33 offset:4
+; GFX10-SCRATCH-NEXT:    v_readlane_b32 s31, v42, 1
+; GFX10-SCRATCH-NEXT:    v_readlane_b32 s30, v42, 0
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s0, v43, 0
 ; GFX10-SCRATCH-NEXT:    s_or_saveexec_b32 s1, -1
 ; GFX10-SCRATCH-NEXT:    s_clause 0x1
-; GFX10-SCRATCH-NEXT:    scratch_load_dword v40, off, s33 offset:8
+; GFX10-SCRATCH-NEXT:    scratch_load_dword v42, off, s33 offset:8
 ; GFX10-SCRATCH-NEXT:    scratch_load_dword v43, off, s33 offset:12
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
@@ -6529,6 +6714,7 @@
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX9-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
@@ -6565,8 +6751,9 @@
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 0
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX10-NEXT:    s_clause 0x1
 ; GFX10-NEXT:    global_load_ubyte v0, v2, s[34:35]
@@ -6604,8 +6791,9 @@
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_load_b64 s[0:1], s[0:1], 0x0
 ; GFX11-NEXT:    v_mov_b32_e32 v1, 0
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX11-NEXT:    s_clause 0x1
 ; GFX11-NEXT:    global_load_u8 v0, v1, s[0:1]
@@ -6643,8 +6831,9 @@
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v2, 0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX10-SCRATCH-NEXT:    s_clause 0x1
 ; GFX10-SCRATCH-NEXT:    global_load_ubyte v0, v2, s[0:1]
@@ -6684,6 +6873,7 @@
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:12 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 3
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    buffer_store_byte v0, off, s[0:3], s33
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 8
 ; GFX9-NEXT:    s_addk_i32 s32, 0x800
@@ -6721,16 +6911,17 @@
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 3
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 8
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x400
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    buffer_store_byte v0, off, s[0:3], s33
 ; GFX10-NEXT:    buffer_store_dword v1, off, s[0:3], s33 offset:4
 ; GFX10-NEXT:    v_lshrrev_b32_e64 v0, 5, s33
-; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_byval_struct_i8_i32@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_byval_struct_i8_i32@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
@@ -6758,19 +6949,20 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:12
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX11-NEXT:    v_dual_mov_b32 v0, 3 :: v_dual_mov_b32 v1, 8
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 32
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_clause 0x1
 ; GFX11-NEXT:    scratch_store_b8 off, v0, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v1, s33 offset:4
 ; GFX11-NEXT:    v_mov_b32_e32 v0, s33
-; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_byval_struct_i8_i32@rel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, external_void_func_byval_struct_i8_i32@rel32@hi+12
-; GFX11-NEXT:    s_delay_alu instid0(SALU_CYCLE_1)
+; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
+; GFX11-NEXT:    s_delay_alu instid0(VALU_DEP_1)
 ; GFX11-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX11-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX11-NEXT:    v_readlane_b32 s0, v41, 0
@@ -6797,16 +6989,17 @@
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 3
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v1, 8
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 32
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    scratch_store_byte off, v0, s33
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v1, s33 offset:4
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, s33
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_byval_struct_i8_i32@rel32@lo+4
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_byval_struct_i8_i32@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-SCRATCH-NEXT:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s30, v40, 0
@@ -6843,6 +7036,7 @@
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 3
 ; GFX9-NEXT:    buffer_store_byte v0, off, s[0:3], s33
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 8
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], s33 offset:4
 ; GFX9-NEXT:    v_lshrrev_b32_e64 v0, 6, s33
 ; GFX9-NEXT:    s_addk_i32 s32, 0x800
@@ -6887,8 +7081,9 @@
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 3
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 8
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x400
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    buffer_store_byte v0, off, s[0:3], s33
 ; GFX10-NEXT:    buffer_store_dword v1, off, s[0:3], s33 offset:4
@@ -6940,10 +7135,11 @@
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_sret_struct_i8_i32_byval_struct_i8_i32@rel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, external_void_func_sret_struct_i8_i32_byval_struct_i8_i32@rel32@hi+12
 ; GFX11-NEXT:    s_add_i32 vcc_lo, s33, 8
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_clause 0x1
 ; GFX11-NEXT:    scratch_store_b8 off, v0, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v1, s33 offset:4
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_dual_mov_b32 v0, vcc_lo :: v_dual_mov_b32 v1, s33
 ; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
@@ -6987,9 +7183,10 @@
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_sret_struct_i8_i32_byval_struct_i8_i32@rel32@lo+4
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_sret_struct_i8_i32_byval_struct_i8_i32@rel32@hi+12
 ; GFX10-SCRATCH-NEXT:    s_add_i32 vcc_lo, s33, 8
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    scratch_store_byte off, v0, s33
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v1, s33 offset:4
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, vcc_lo
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v1, s33
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 1
@@ -7045,6 +7242,7 @@
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
@@ -7099,8 +7297,9 @@
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 0
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX10-NEXT:    global_load_dwordx4 v[0:3], v0, s[34:35]
@@ -7156,8 +7355,9 @@
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_load_b64 s[0:1], s[0:1], 0x0
 ; GFX11-NEXT:    v_mov_b32_e32 v0, 0
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX11-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX11-NEXT:    global_load_b128 v[0:3], v0, s[0:1]
@@ -7209,8 +7409,9 @@
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-SCRATCH-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX10-SCRATCH-NEXT:    global_load_dwordx4 v[0:3], v0, s[0:1]
@@ -7261,7 +7462,7 @@
 ; GFX9-LABEL: tail_call_byval_align16:
 ; GFX9:       ; %bb.0: ; %entry
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX9-NEXT:    s_mov_b32 s6, s33
+; GFX9-NEXT:    s_mov_b32 s8, s33
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:24 ; 4-byte Folded Spill
@@ -7269,6 +7470,8 @@
 ; GFX9-NEXT:    buffer_load_dword v32, off, s[0:3], s33 offset:20
 ; GFX9-NEXT:    buffer_load_dword v33, off, s[0:3], s33 offset:16
 ; GFX9-NEXT:    buffer_load_dword v31, off, s[0:3], s33
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_addk_i32 s32, 0x800
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s34, 2
@@ -7299,7 +7502,6 @@
 ; GFX9-NEXT:    v_writelane_b32 v40, s59, 27
 ; GFX9-NEXT:    v_writelane_b32 v40, s60, 28
 ; GFX9-NEXT:    v_writelane_b32 v40, s61, 29
-; GFX9-NEXT:    s_addk_i32 s32, 0x800
 ; GFX9-NEXT:    v_writelane_b32 v40, s62, 30
 ; GFX9-NEXT:    v_writelane_b32 v40, s63, 31
 ; GFX9-NEXT:    s_getpc_b64 s[4:5]
@@ -7346,7 +7548,7 @@
 ; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s33 offset:24 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX9-NEXT:    s_addk_i32 s32, 0xf800
-; GFX9-NEXT:    s_mov_b32 s33, s6
+; GFX9-NEXT:    s_mov_b32 s33, s8
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
 ;
@@ -7354,7 +7556,7 @@
 ; GFX10:       ; %bb.0: ; %entry
 ; GFX10-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
-; GFX10-NEXT:    s_mov_b32 s6, s33
+; GFX10-NEXT:    s_mov_b32 s7, s33
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_or_saveexec_b32 s4, -1
 ; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:24 ; 4-byte Folded Spill
@@ -7364,8 +7566,9 @@
 ; GFX10-NEXT:    buffer_load_dword v32, off, s[0:3], s33 offset:20
 ; GFX10-NEXT:    buffer_load_dword v33, off, s[0:3], s33 offset:16
 ; GFX10-NEXT:    buffer_load_dword v31, off, s[0:3], s33
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x400
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_getpc_b64 s[4:5]
 ; GFX10-NEXT:    s_add_u32 s4, s4, byval_align16_f64_arg@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s5, s5, byval_align16_f64_arg@rel32@hi+12
@@ -7442,7 +7645,7 @@
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s4
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfc00
-; GFX10-NEXT:    s_mov_b32 s33, s6
+; GFX10-NEXT:    s_mov_b32 s33, s7
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    s_setpc_b64 s[30:31]
 ;
@@ -7450,7 +7653,7 @@
 ; GFX11:       ; %bb.0: ; %entry
 ; GFX11-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX11-NEXT:    s_waitcnt_vscnt null, 0x0
-; GFX11-NEXT:    s_mov_b32 s4, s33
+; GFX11-NEXT:    s_mov_b32 s5, s33
 ; GFX11-NEXT:    s_mov_b32 s33, s32
 ; GFX11-NEXT:    s_or_saveexec_b32 s0, -1
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33 offset:24 ; 4-byte Folded Spill
@@ -7458,8 +7661,9 @@
 ; GFX11-NEXT:    s_clause 0x1
 ; GFX11-NEXT:    scratch_load_b64 v[32:33], off, s33 offset:16
 ; GFX11-NEXT:    scratch_load_b32 v31, off, s33
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 32
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, byval_align16_f64_arg@rel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, byval_align16_f64_arg@rel32@hi+12
@@ -7533,7 +7737,7 @@
 ; GFX11-NEXT:    scratch_load_b32 v40, off, s33 offset:24 ; 4-byte Folded Reload
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s0
 ; GFX11-NEXT:    s_addk_i32 s32, 0xffe0
-; GFX11-NEXT:    s_mov_b32 s33, s4
+; GFX11-NEXT:    s_mov_b32 s33, s5
 ; GFX11-NEXT:    s_waitcnt vmcnt(0)
 ; GFX11-NEXT:    s_setpc_b64 s[30:31]
 ;
@@ -7541,7 +7745,7 @@
 ; GFX10-SCRATCH:       ; %bb.0: ; %entry
 ; GFX10-SCRATCH-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_vscnt null, 0x0
-; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, s33
+; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, s33
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_or_saveexec_b32 s0, -1
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v40, s33 offset:24 ; 4-byte Folded Spill
@@ -7550,8 +7754,9 @@
 ; GFX10-SCRATCH-NEXT:    s_clause 0x1
 ; GFX10-SCRATCH-NEXT:    scratch_load_dwordx2 v[32:33], off, s33 offset:16
 ; GFX10-SCRATCH-NEXT:    scratch_load_dword v31, off, s33
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 32
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, byval_align16_f64_arg@rel32@lo+4
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, byval_align16_f64_arg@rel32@hi+12
@@ -7626,7 +7831,7 @@
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
 ; GFX10-SCRATCH-NEXT:    s_addk_i32 s32, 0xffe0
-; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s4
+; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s5
 ; GFX10-SCRATCH-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-SCRATCH-NEXT:    s_setpc_b64 s[30:31]
 entry:
@@ -7646,6 +7851,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 1
@@ -7679,15 +7885,16 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 1
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i1_inreg@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_i1_inreg@rel32@hi+12
-; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-NEXT:    buffer_store_byte v0, off, s[0:3], s32
+; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
@@ -7714,16 +7921,18 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_mov_b32_e32 v0, 1
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_i1_inreg@rel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, external_void_func_i1_inreg@rel32@hi+12
-; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX11-NEXT:    scratch_store_b8 off, v0, s32
+; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
+; GFX11-NEXT:    s_delay_alu instid0(VALU_DEP_1)
 ; GFX11-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX11-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX11-NEXT:    v_readlane_b32 s0, v41, 0
@@ -7748,15 +7957,16 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 1
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_i1_inreg@rel32@lo+4
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_i1_inreg@rel32@hi+12
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-SCRATCH-NEXT:    scratch_store_byte off, v0, s32
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-SCRATCH-NEXT:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s30, v40, 0
@@ -7785,8 +7995,9 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
-; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
+; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX9-NEXT:    s_movk_i32 s4, 0x7b
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
@@ -7819,9 +8030,10 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_movk_i32 s4, 0x7b
-; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i8_inreg@rel32@lo+4
@@ -7855,9 +8067,10 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
+; GFX11-NEXT:    ; implicit-def: $vgpr40
+; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_movk_i32 s4, 0x7b
-; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_i8_inreg@rel32@lo+4
@@ -7891,9 +8104,10 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_movk_i32 s4, 0x7b
-; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_i8_inreg@rel32@lo+4
@@ -7929,8 +8143,9 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
-; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
+; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX9-NEXT:    s_movk_i32 s4, 0x7b
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
@@ -7963,9 +8178,10 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_movk_i32 s4, 0x7b
-; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i16_inreg@rel32@lo+4
@@ -7999,9 +8215,10 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
+; GFX11-NEXT:    ; implicit-def: $vgpr40
+; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_movk_i32 s4, 0x7b
-; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_i16_inreg@rel32@lo+4
@@ -8035,9 +8252,10 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_movk_i32 s4, 0x7b
-; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_i16_inreg@rel32@lo+4
@@ -8073,8 +8291,9 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
-; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
+; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX9-NEXT:    s_mov_b32 s4, 42
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
@@ -8107,9 +8326,10 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 42
-; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i32_inreg@rel32@lo+4
@@ -8143,9 +8363,10 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
+; GFX11-NEXT:    ; implicit-def: $vgpr40
+; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_mov_b32 s4, 42
-; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_i32_inreg@rel32@lo+4
@@ -8179,9 +8400,10 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 42
-; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_i32_inreg@rel32@lo+4
@@ -8217,9 +8439,10 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX9-NEXT:    s_movk_i32 s4, 0x7b
 ; GFX9-NEXT:    s_mov_b32 s5, 0
@@ -8254,9 +8477,10 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_movk_i32 s4, 0x7b
-; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i64_inreg@rel32@lo+4
@@ -8293,9 +8517,10 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
+; GFX11-NEXT:    ; implicit-def: $vgpr40
+; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_movk_i32 s4, 0x7b
-; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_i64_inreg@rel32@lo+4
@@ -8332,9 +8557,10 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_movk_i32 s4, 0x7b
-; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_i64_inreg@rel32@lo+4
@@ -8373,9 +8599,10 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX9-NEXT:    s_mov_b64 s[34:35], 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s7, 3
@@ -8414,8 +8641,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b64 s[34:35], 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
@@ -8457,8 +8685,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
+; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_mov_b64 s[0:1], 0
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s5, 1
@@ -8500,8 +8729,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b64 s[0:1], 0
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
@@ -8546,11 +8776,12 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX9-NEXT:    v_writelane_b32 v40, s7, 3
-; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 4
 ; GFX9-NEXT:    s_mov_b32 s4, 1
 ; GFX9-NEXT:    s_mov_b32 s5, 2
@@ -8589,9 +8820,10 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 1
-; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2i64_inreg@rel32@lo+4
@@ -8634,9 +8866,10 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
+; GFX11-NEXT:    ; implicit-def: $vgpr40
+; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_mov_b32 s4, 1
-; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v2i64_inreg@rel32@lo+4
@@ -8679,9 +8912,10 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 1
-; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v2i64_inreg@rel32@lo+4
@@ -8726,9 +8960,10 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX9-NEXT:    s_mov_b64 s[34:35], 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s7, 3
@@ -8773,8 +9008,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b64 s[34:35], 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
@@ -8822,8 +9058,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
+; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_mov_b64 s[0:1], 0
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s5, 1
@@ -8871,8 +9108,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b64 s[0:1], 0
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
@@ -8925,10 +9163,11 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s6, 2
-; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s7, 3
 ; GFX9-NEXT:    s_mov_b64 s[34:35], 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s8, 4
@@ -8978,8 +9217,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b64 s[34:35], 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
@@ -9033,8 +9273,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
+; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_mov_b64 s[0:1], 0
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s5, 1
@@ -9088,8 +9329,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b64 s[0:1], 0
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
@@ -9147,8 +9389,9 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
-; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
+; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX9-NEXT:    s_movk_i32 s4, 0x4400
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
@@ -9181,9 +9424,10 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_movk_i32 s4, 0x4400
-; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_f16_inreg@rel32@lo+4
@@ -9217,9 +9461,10 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
+; GFX11-NEXT:    ; implicit-def: $vgpr40
+; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_movk_i32 s4, 0x4400
-; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_f16_inreg@rel32@lo+4
@@ -9253,9 +9498,10 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_movk_i32 s4, 0x4400
-; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_f16_inreg@rel32@lo+4
@@ -9291,8 +9537,9 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
-; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
+; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX9-NEXT:    s_mov_b32 s4, 4.0
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
@@ -9325,9 +9572,10 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 4.0
-; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_f32_inreg@rel32@lo+4
@@ -9361,9 +9609,10 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
+; GFX11-NEXT:    ; implicit-def: $vgpr40
+; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_mov_b32 s4, 4.0
-; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_f32_inreg@rel32@lo+4
@@ -9397,9 +9646,10 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 4.0
-; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_f32_inreg@rel32@lo+4
@@ -9435,9 +9685,10 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX9-NEXT:    s_mov_b32 s4, 1.0
 ; GFX9-NEXT:    s_mov_b32 s5, 2.0
@@ -9472,9 +9723,10 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 1.0
-; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2f32_inreg@rel32@lo+4
@@ -9511,9 +9763,10 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
+; GFX11-NEXT:    ; implicit-def: $vgpr40
+; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_mov_b32 s4, 1.0
-; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v2f32_inreg@rel32@lo+4
@@ -9550,9 +9803,10 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 1.0
-; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v2f32_inreg@rel32@lo+4
@@ -9591,10 +9845,11 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s6, 2
-; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 3
 ; GFX9-NEXT:    s_mov_b32 s4, 1.0
 ; GFX9-NEXT:    s_mov_b32 s5, 2.0
@@ -9631,9 +9886,10 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 1.0
-; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3f32_inreg@rel32@lo+4
@@ -9673,9 +9929,10 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
+; GFX11-NEXT:    ; implicit-def: $vgpr40
+; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_mov_b32 s4, 1.0
-; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v3f32_inreg@rel32@lo+4
@@ -9715,9 +9972,10 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 1.0
-; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v3f32_inreg@rel32@lo+4
@@ -9759,12 +10017,13 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX9-NEXT:    v_writelane_b32 v40, s7, 3
 ; GFX9-NEXT:    v_writelane_b32 v40, s8, 4
-; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 5
 ; GFX9-NEXT:    s_mov_b32 s4, 1.0
 ; GFX9-NEXT:    s_mov_b32 s5, 2.0
@@ -9805,9 +10064,10 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 1.0
-; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v5f32_inreg@rel32@lo+4
@@ -9853,9 +10113,10 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
+; GFX11-NEXT:    ; implicit-def: $vgpr40
+; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_mov_b32 s4, 1.0
-; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v5f32_inreg@rel32@lo+4
@@ -9901,9 +10162,10 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 1.0
-; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v5f32_inreg@rel32@lo+4
@@ -9951,9 +10213,10 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX9-NEXT:    s_mov_b32 s4, 0
 ; GFX9-NEXT:    s_mov_b32 s5, 0x40100000
@@ -9988,9 +10251,10 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 0
-; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_f64_inreg@rel32@lo+4
@@ -10027,9 +10291,10 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
+; GFX11-NEXT:    ; implicit-def: $vgpr40
+; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_mov_b32 s4, 0
-; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_f64_inreg@rel32@lo+4
@@ -10066,9 +10331,10 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 0
-; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_f64_inreg@rel32@lo+4
@@ -10107,11 +10373,12 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX9-NEXT:    v_writelane_b32 v40, s7, 3
-; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 4
 ; GFX9-NEXT:    s_mov_b32 s4, 0
 ; GFX9-NEXT:    s_mov_b32 s5, 2.0
@@ -10150,9 +10417,10 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 0
-; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2f64_inreg@rel32@lo+4
@@ -10195,9 +10463,10 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
+; GFX11-NEXT:    ; implicit-def: $vgpr40
+; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_mov_b32 s4, 0
-; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v2f64_inreg@rel32@lo+4
@@ -10240,9 +10509,10 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 0
-; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v2f64_inreg@rel32@lo+4
@@ -10287,13 +10557,14 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX9-NEXT:    v_writelane_b32 v40, s7, 3
 ; GFX9-NEXT:    v_writelane_b32 v40, s8, 4
 ; GFX9-NEXT:    v_writelane_b32 v40, s9, 5
-; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 6
 ; GFX9-NEXT:    s_mov_b32 s4, 0
 ; GFX9-NEXT:    s_mov_b32 s5, 2.0
@@ -10336,9 +10607,10 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 0
-; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3f64_inreg@rel32@lo+4
@@ -10387,9 +10659,10 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
+; GFX11-NEXT:    ; implicit-def: $vgpr40
+; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_mov_b32 s4, 0
-; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v3f64_inreg@rel32@lo+4
@@ -10438,9 +10711,10 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 0
-; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v3f64_inreg@rel32@lo+4
@@ -10491,9 +10765,10 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    s_load_dword s4, s[34:35], 0x0
-; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 2
@@ -10525,9 +10800,10 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_load_dword s4, s[34:35], 0x0
-; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2i16_inreg@rel32@lo+4
@@ -10561,9 +10837,10 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
+; GFX11-NEXT:    ; implicit-def: $vgpr40
+; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_load_b32 s4, s[0:1], 0x0
-; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v2i16_inreg@rel32@lo+4
@@ -10597,9 +10874,10 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_load_dword s4, s[0:1], 0x0
-; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v2i16_inreg@rel32@lo+4
@@ -10636,10 +10914,11 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    s_load_dwordx2 s[4:5], s[34:35], 0x0
-; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 3
@@ -10672,8 +10951,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    s_load_dwordx2 s[4:5], s[34:35], 0x0
@@ -10710,8 +10990,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX11-NEXT:    s_load_b64 s[4:5], s[0:1], 0x0
@@ -10748,8 +11029,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-SCRATCH-NEXT:    s_load_dwordx2 s[4:5], s[0:1], 0x0
@@ -10789,10 +11071,11 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    s_load_dwordx2 s[4:5], s[34:35], 0x0
-; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 3
@@ -10825,8 +11108,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    s_load_dwordx2 s[4:5], s[34:35], 0x0
@@ -10863,8 +11147,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX11-NEXT:    s_load_b64 s[4:5], s[0:1], 0x0
@@ -10901,8 +11186,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-SCRATCH-NEXT:    s_load_dwordx2 s[4:5], s[0:1], 0x0
@@ -10942,9 +11228,10 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX9-NEXT:    s_mov_b32 s4, 0x20001
 ; GFX9-NEXT:    s_mov_b32 s5, 3
@@ -10979,9 +11266,10 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 0x20001
-; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3i16_inreg@rel32@lo+4
@@ -11018,9 +11306,10 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
+; GFX11-NEXT:    ; implicit-def: $vgpr40
+; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_mov_b32 s4, 0x20001
-; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v3i16_inreg@rel32@lo+4
@@ -11057,9 +11346,10 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 0x20001
-; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v3i16_inreg@rel32@lo+4
@@ -11098,9 +11388,10 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX9-NEXT:    s_mov_b32 s4, 0x40003c00
 ; GFX9-NEXT:    s_movk_i32 s5, 0x4400
@@ -11135,9 +11426,10 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 0x40003c00
-; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3f16_inreg@rel32@lo+4
@@ -11174,9 +11466,10 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
+; GFX11-NEXT:    ; implicit-def: $vgpr40
+; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_mov_b32 s4, 0x40003c00
-; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v3f16_inreg@rel32@lo+4
@@ -11213,9 +11506,10 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 0x40003c00
-; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v3f16_inreg@rel32@lo+4
@@ -11254,10 +11548,11 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    s_load_dwordx2 s[4:5], s[34:35], 0x0
-; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 3
@@ -11290,8 +11585,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    s_load_dwordx2 s[4:5], s[34:35], 0x0
@@ -11328,8 +11624,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX11-NEXT:    s_load_b64 s[4:5], s[0:1], 0x0
@@ -11366,8 +11663,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-SCRATCH-NEXT:    s_load_dwordx2 s[4:5], s[0:1], 0x0
@@ -11407,9 +11705,10 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX9-NEXT:    s_mov_b32 s4, 0x20001
 ; GFX9-NEXT:    s_mov_b32 s5, 0x40003
@@ -11444,9 +11743,10 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 0x20001
-; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v4i16_inreg@rel32@lo+4
@@ -11483,9 +11783,10 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
+; GFX11-NEXT:    ; implicit-def: $vgpr40
+; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_mov_b32 s4, 0x20001
-; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v4i16_inreg@rel32@lo+4
@@ -11522,9 +11823,10 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 0x20001
-; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v4i16_inreg@rel32@lo+4
@@ -11563,9 +11865,10 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    s_load_dword s4, s[34:35], 0x0
-; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 2
@@ -11597,9 +11900,10 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_load_dword s4, s[34:35], 0x0
-; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2f16_inreg@rel32@lo+4
@@ -11633,9 +11937,10 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
+; GFX11-NEXT:    ; implicit-def: $vgpr40
+; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_load_b32 s4, s[0:1], 0x0
-; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v2f16_inreg@rel32@lo+4
@@ -11669,9 +11974,10 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_load_dword s4, s[0:1], 0x0
-; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v2f16_inreg@rel32@lo+4
@@ -11708,10 +12014,11 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    s_load_dwordx2 s[4:5], s[34:35], 0x0
-; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 3
@@ -11744,8 +12051,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    s_load_dwordx2 s[4:5], s[34:35], 0x0
@@ -11782,8 +12090,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX11-NEXT:    s_load_b64 s[4:5], s[0:1], 0x0
@@ -11820,8 +12129,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-SCRATCH-NEXT:    s_load_dwordx2 s[4:5], s[0:1], 0x0
@@ -11861,9 +12171,10 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX9-NEXT:    s_mov_b32 s4, 1
 ; GFX9-NEXT:    s_mov_b32 s5, 2
@@ -11898,9 +12209,10 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 1
-; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2i32_inreg@rel32@lo+4
@@ -11937,9 +12249,10 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
+; GFX11-NEXT:    ; implicit-def: $vgpr40
+; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_mov_b32 s4, 1
-; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v2i32_inreg@rel32@lo+4
@@ -11976,9 +12289,10 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 1
-; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v2i32_inreg@rel32@lo+4
@@ -12017,10 +12331,11 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s6, 2
-; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 3
 ; GFX9-NEXT:    s_mov_b32 s4, 3
 ; GFX9-NEXT:    s_mov_b32 s5, 4
@@ -12057,9 +12372,10 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 3
-; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3i32_inreg@rel32@lo+4
@@ -12099,9 +12415,10 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
+; GFX11-NEXT:    ; implicit-def: $vgpr40
+; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_mov_b32 s4, 3
-; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v3i32_inreg@rel32@lo+4
@@ -12141,9 +12458,10 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 3
-; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v3i32_inreg@rel32@lo+4
@@ -12185,11 +12503,12 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX9-NEXT:    v_writelane_b32 v40, s7, 3
-; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 4
 ; GFX9-NEXT:    s_mov_b32 s4, 3
 ; GFX9-NEXT:    s_mov_b32 s5, 4
@@ -12228,9 +12547,10 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 3
-; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3i32_i32_inreg@rel32@lo+4
@@ -12273,9 +12593,10 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
+; GFX11-NEXT:    ; implicit-def: $vgpr40
+; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_mov_b32 s4, 3
-; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v3i32_i32_inreg@rel32@lo+4
@@ -12318,9 +12639,10 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 3
-; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v3i32_i32_inreg@rel32@lo+4
@@ -12365,12 +12687,13 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX9-NEXT:    v_writelane_b32 v40, s7, 3
 ; GFX9-NEXT:    s_load_dwordx4 s[4:7], s[34:35], 0x0
-; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 4
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 5
@@ -12405,8 +12728,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
@@ -12447,8 +12771,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX11-NEXT:    v_writelane_b32 v40, s6, 2
@@ -12489,8 +12814,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s6, 2
@@ -12534,11 +12860,12 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX9-NEXT:    v_writelane_b32 v40, s7, 3
-; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 4
 ; GFX9-NEXT:    s_mov_b32 s4, 1
 ; GFX9-NEXT:    s_mov_b32 s5, 2
@@ -12577,9 +12904,10 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 1
-; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v4i32_inreg@rel32@lo+4
@@ -12622,9 +12950,10 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
+; GFX11-NEXT:    ; implicit-def: $vgpr40
+; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_mov_b32 s4, 1
-; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v4i32_inreg@rel32@lo+4
@@ -12667,9 +12996,10 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 1
-; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v4i32_inreg@rel32@lo+4
@@ -12714,12 +13044,13 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX9-NEXT:    v_writelane_b32 v40, s7, 3
 ; GFX9-NEXT:    v_writelane_b32 v40, s8, 4
-; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 5
 ; GFX9-NEXT:    s_mov_b32 s4, 1
 ; GFX9-NEXT:    s_mov_b32 s5, 2
@@ -12760,9 +13091,10 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 1
-; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v5i32_inreg@rel32@lo+4
@@ -12808,9 +13140,10 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
+; GFX11-NEXT:    ; implicit-def: $vgpr40
+; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_mov_b32 s4, 1
-; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v5i32_inreg@rel32@lo+4
@@ -12856,9 +13189,10 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 1
-; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v5i32_inreg@rel32@lo+4
@@ -12906,9 +13240,10 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX9-NEXT:    v_writelane_b32 v40, s7, 3
@@ -12956,8 +13291,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
@@ -13008,8 +13344,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
+; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_load_b64 s[0:1], s[0:1], 0x0
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s5, 1
@@ -13060,8 +13397,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x0
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
@@ -13116,6 +13454,8 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s6, 2
@@ -13124,7 +13464,6 @@
 ; GFX9-NEXT:    v_writelane_b32 v40, s9, 5
 ; GFX9-NEXT:    v_writelane_b32 v40, s10, 6
 ; GFX9-NEXT:    v_writelane_b32 v40, s11, 7
-; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 8
 ; GFX9-NEXT:    s_mov_b32 s4, 1
 ; GFX9-NEXT:    s_mov_b32 s5, 2
@@ -13171,9 +13510,10 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 1
-; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v8i32_inreg@rel32@lo+4
@@ -13228,9 +13568,10 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
+; GFX11-NEXT:    ; implicit-def: $vgpr40
+; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_mov_b32 s4, 1
-; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_v8i32_inreg@rel32@lo+4
@@ -13285,9 +13626,10 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 1
-; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, external_void_func_v8i32_inreg@rel32@lo+4
@@ -13344,6 +13686,8 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s6, 2
@@ -13354,7 +13698,6 @@
 ; GFX9-NEXT:    v_writelane_b32 v40, s11, 7
 ; GFX9-NEXT:    v_writelane_b32 v40, s12, 8
 ; GFX9-NEXT:    v_writelane_b32 v40, s13, 9
-; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s14, 10
 ; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX9-NEXT:    v_writelane_b32 v40, s15, 11
@@ -13410,8 +13753,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
@@ -13478,8 +13822,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
+; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_load_b64 s[0:1], s[0:1], 0x0
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s5, 1
@@ -13546,8 +13891,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x0
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
@@ -13618,6 +13964,8 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s6, 2
@@ -13633,7 +13981,6 @@
 ; GFX9-NEXT:    v_writelane_b32 v40, s16, 12
 ; GFX9-NEXT:    v_writelane_b32 v40, s17, 13
 ; GFX9-NEXT:    v_writelane_b32 v40, s18, 14
-; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s19, 15
 ; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX9-NEXT:    v_writelane_b32 v40, s20, 16
@@ -13728,8 +14075,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
@@ -13841,8 +14189,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
+; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_load_b64 s[0:1], s[0:1], 0x0
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s5, 1
@@ -13948,8 +14297,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x0
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
@@ -14061,6 +14411,8 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s6, 2
@@ -14075,7 +14427,6 @@
 ; GFX9-NEXT:    v_writelane_b32 v40, s15, 11
 ; GFX9-NEXT:    v_writelane_b32 v40, s16, 12
 ; GFX9-NEXT:    v_writelane_b32 v40, s17, 13
-; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s18, 14
 ; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX9-NEXT:    v_writelane_b32 v40, s19, 15
@@ -14176,8 +14527,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
@@ -14294,8 +14646,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
+; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    s_load_b64 s[0:1], s[0:1], 0x0
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    v_writelane_b32 v40, s5, 1
@@ -14404,8 +14757,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x0
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
@@ -14525,6 +14879,7 @@
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
 ; GFX9-NEXT:    buffer_load_dword v32, off, s[0:3], s33
 ; GFX9-NEXT:    buffer_load_dword v33, off, s[0:3], s33 offset:4
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x800
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
@@ -14563,8 +14918,9 @@
 ; GFX10-NEXT:    s_clause 0x1
 ; GFX10-NEXT:    buffer_load_dword v32, off, s[0:3], s33
 ; GFX10-NEXT:    buffer_load_dword v33, off, s[0:3], s33 offset:4
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x400
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, stack_passed_f64_arg@rel32@lo+4
@@ -14601,8 +14957,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:12
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX11-NEXT:    scratch_load_b64 v[32:33], off, s33
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 32
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, stack_passed_f64_arg@rel32@lo+4
@@ -14636,8 +14993,9 @@
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX10-SCRATCH-NEXT:    scratch_load_dwordx2 v[32:33], off, s33
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 32
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, stack_passed_f64_arg@rel32@lo+4
@@ -14680,6 +15038,7 @@
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 13
 ; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:4
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 14
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:8
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 15
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
@@ -14750,11 +15109,12 @@
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 14
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_mov_b32_e32 v3, 15
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    buffer_store_dword v0, off, s[0:3], s32
 ; GFX10-NEXT:    buffer_store_dword v1, off, s[0:3], s32 offset:4
 ; GFX10-NEXT:    buffer_store_dword v2, off, s[0:3], s32 offset:8
 ; GFX10-NEXT:    buffer_store_dword v3, off, s[0:3], s32 offset:12
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 0
@@ -14821,8 +15181,9 @@
 ; GFX11-NEXT:    v_dual_mov_b32 v0, 12 :: v_dual_mov_b32 v1, 13
 ; GFX11-NEXT:    v_dual_mov_b32 v2, 14 :: v_dual_mov_b32 v3, 15
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    v_dual_mov_b32 v4, 1 :: v_dual_mov_b32 v5, 1
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    scratch_store_b128 off, v[0:3], s32
 ; GFX11-NEXT:    v_dual_mov_b32 v0, 0 :: v_dual_mov_b32 v1, 0
 ; GFX11-NEXT:    v_dual_mov_b32 v2, 0 :: v_dual_mov_b32 v3, 1
@@ -14875,8 +15236,9 @@
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v2, 14
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v3, 15
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v4, 1
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v5, 1
 ; GFX10-SCRATCH-NEXT:    scratch_store_dwordx4 off, v[0:3], s32
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 0
@@ -14969,6 +15331,7 @@
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 13
 ; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:20
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 14
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:24
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 15
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
@@ -15045,6 +15408,7 @@
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 11
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 12
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 13
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_mov_b32_e32 v4, 15
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:12
@@ -15120,10 +15484,11 @@
 ; GFX11-NEXT:    v_dual_mov_b32 v4, 8 :: v_dual_mov_b32 v5, 9
 ; GFX11-NEXT:    v_dual_mov_b32 v6, 10 :: v_dual_mov_b32 v7, 11
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_clause 0x1
 ; GFX11-NEXT:    scratch_store_b128 off, v[0:3], s32 offset:16
 ; GFX11-NEXT:    scratch_store_b128 off, v[4:7], s32
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_dual_mov_b32 v0, 0 :: v_dual_mov_b32 v1, 0
 ; GFX11-NEXT:    v_dual_mov_b32 v2, 0 :: v_dual_mov_b32 v3, 0
 ; GFX11-NEXT:    v_dual_mov_b32 v4, 0 :: v_dual_mov_b32 v5, 1
@@ -15180,9 +15545,10 @@
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v6, 10
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v7, 11
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    scratch_store_dwordx4 off, v[0:3], s32 offset:16
 ; GFX10-SCRATCH-NEXT:    scratch_store_dwordx4 off, v[4:7], s32
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v2, 0
@@ -15271,6 +15637,7 @@
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0x41500000
 ; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:20
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0x41600000
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:24
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0x41700000
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
@@ -15347,6 +15714,7 @@
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 0x41300000
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 0x41400000
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 0x41500000
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_mov_b32_e32 v4, 0x41700000
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:12
@@ -15426,11 +15794,12 @@
 ; GFX11-NEXT:    v_mov_b32_e32 v6, 0x41200000
 ; GFX11-NEXT:    v_mov_b32_e32 v7, 0x41300000
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_clause 0x1
 ; GFX11-NEXT:    scratch_store_b128 off, v[0:3], s32 offset:16
 ; GFX11-NEXT:    scratch_store_b128 off, v[4:7], s32
 ; GFX11-NEXT:    v_mov_b32_e32 v6, 1.0
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_dual_mov_b32 v0, 0 :: v_dual_mov_b32 v1, 0
 ; GFX11-NEXT:    v_dual_mov_b32 v2, 0 :: v_dual_mov_b32 v3, 0
 ; GFX11-NEXT:    v_dual_mov_b32 v4, 0 :: v_dual_mov_b32 v5, 1.0
@@ -15488,9 +15857,10 @@
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v6, 0x41200000
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v7, 0x41300000
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    scratch_store_dwordx4 off, v[0:3], s32 offset:16
 ; GFX10-SCRATCH-NEXT:    scratch_store_dwordx4 off, v[4:7], s32
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v0, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v2, 0
diff --git a/llvm/test/CodeGen/AMDGPU/gfx-callable-preserved-registers.ll b/llvm/test/CodeGen/AMDGPU/gfx-callable-preserved-registers.ll
--- a/llvm/test/CodeGen/AMDGPU/gfx-callable-preserved-registers.ll
+++ b/llvm/test/CodeGen/AMDGPU/gfx-callable-preserved-registers.ll
@@ -15,9 +15,10 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 3
@@ -53,8 +54,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    s_getpc_b64 s[4:5]
@@ -93,8 +95,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX11-NEXT:    s_getpc_b64 s[4:5]
@@ -134,6 +137,7 @@
 ; GFX9-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr0
 ; GFX9-NEXT:    v_writelane_b32 v0, s28, 0
 ; GFX9-NEXT:    v_writelane_b32 v0, s29, 1
 ; GFX9-NEXT:    v_writelane_b32 v0, s30, 2
@@ -162,6 +166,7 @@
 ; GFX10-NEXT:    buffer_store_dword v0, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
+; GFX10-NEXT:    ; implicit-def: $vgpr0
 ; GFX10-NEXT:    v_writelane_b32 v0, s28, 0
 ; GFX10-NEXT:    v_writelane_b32 v0, s29, 1
 ; GFX10-NEXT:    v_writelane_b32 v0, s30, 2
@@ -188,9 +193,10 @@
 ; GFX11:       ; %bb.0:
 ; GFX11-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX11-NEXT:    s_waitcnt_vscnt null, 0x0
-; GFX11-NEXT:    s_xor_saveexec_b32 s0, -1
+; GFX11-NEXT:    s_xor_saveexec_b32 s1, -1
 ; GFX11-NEXT:    scratch_store_b32 off, v0, s32 ; 4-byte Folded Spill
-; GFX11-NEXT:    s_mov_b32 exec_lo, s0
+; GFX11-NEXT:    s_mov_b32 exec_lo, s1
+; GFX11-NEXT:    ; implicit-def: $vgpr0
 ; GFX11-NEXT:    v_writelane_b32 v0, s28, 0
 ; GFX11-NEXT:    v_writelane_b32 v0, s29, 1
 ; GFX11-NEXT:    v_writelane_b32 v0, s30, 2
@@ -206,9 +212,9 @@
 ; GFX11-NEXT:    v_readlane_b32 s30, v0, 2
 ; GFX11-NEXT:    v_readlane_b32 s29, v0, 1
 ; GFX11-NEXT:    v_readlane_b32 s28, v0, 0
-; GFX11-NEXT:    s_xor_saveexec_b32 s0, -1
+; GFX11-NEXT:    s_xor_saveexec_b32 s1, -1
 ; GFX11-NEXT:    scratch_load_b32 v0, off, s32 ; 4-byte Folded Reload
-; GFX11-NEXT:    s_mov_b32 exec_lo, s0
+; GFX11-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX11-NEXT:    s_waitcnt vmcnt(0)
 ; GFX11-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX11-NEXT:    s_setpc_b64 s[30:31]
@@ -232,8 +238,9 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
-; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
+; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 2
@@ -273,8 +280,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_void@rel32@lo+4
@@ -316,8 +324,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_void@rel32@lo+4
@@ -359,32 +368,33 @@
 ; GFX9-NEXT:    s_mov_b32 s34, s33
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_or_saveexec_b64 s[36:37], -1
-; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
+; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v42, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr41
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX9-NEXT:    v_writelane_b32 v41, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v42, s34, 0
-; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 ; 4-byte Folded Spill
-; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
+; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
+; GFX9-NEXT:    v_writelane_b32 v41, s31, 1
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; def v31
 ; GFX9-NEXT:    ;;#ASMEND
-; GFX9-NEXT:    v_mov_b32_e32 v41, v31
+; GFX9-NEXT:    v_mov_b32_e32 v40, v31
 ; GFX9-NEXT:    s_getpc_b64 s[34:35]
 ; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_void@rel32@lo+4
 ; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_void@rel32@hi+12
 ; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
-; GFX9-NEXT:    v_mov_b32_e32 v31, v41
+; GFX9-NEXT:    v_mov_b32_e32 v31, v40
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; use v31
 ; GFX9-NEXT:    ;;#ASMEND
-; GFX9-NEXT:    buffer_load_dword v41, off, s[0:3], s33 ; 4-byte Folded Reload
-; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
-; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
+; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s33 ; 4-byte Folded Reload
+; GFX9-NEXT:    v_readlane_b32 s31, v41, 1
+; GFX9-NEXT:    v_readlane_b32 s30, v41, 0
 ; GFX9-NEXT:    v_readlane_b32 s34, v42, 0
 ; GFX9-NEXT:    s_or_saveexec_b64 s[36:37], -1
-; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
+; GFX9-NEXT:    buffer_load_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
 ; GFX9-NEXT:    buffer_load_dword v42, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
@@ -399,34 +409,35 @@
 ; GFX10-NEXT:    s_mov_b32 s34, s33
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_or_saveexec_b32 s35, -1
-; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
+; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    buffer_store_dword v42, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr41
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v41, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v42, s34, 0
-; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 ; 4-byte Folded Spill
+; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; def v31
 ; GFX10-NEXT:    ;;#ASMEND
-; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    v_mov_b32_e32 v41, v31
+; GFX10-NEXT:    v_mov_b32_e32 v40, v31
+; GFX10-NEXT:    v_writelane_b32 v41, s31, 1
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_void@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_void@rel32@hi+12
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
-; GFX10-NEXT:    v_mov_b32_e32 v31, v41
+; GFX10-NEXT:    v_mov_b32_e32 v31, v40
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; use v31
 ; GFX10-NEXT:    ;;#ASMEND
-; GFX10-NEXT:    buffer_load_dword v41, off, s[0:3], s33 ; 4-byte Folded Reload
-; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
-; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
+; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s33 ; 4-byte Folded Reload
+; GFX10-NEXT:    v_readlane_b32 s31, v41, 1
+; GFX10-NEXT:    v_readlane_b32 s30, v41, 0
 ; GFX10-NEXT:    v_readlane_b32 s34, v42, 0
 ; GFX10-NEXT:    s_or_saveexec_b32 s35, -1
 ; GFX10-NEXT:    s_clause 0x1
-; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s33 offset:4
+; GFX10-NEXT:    buffer_load_dword v41, off, s[0:3], s33 offset:4
 ; GFX10-NEXT:    buffer_load_dword v42, off, s[0:3], s33 offset:8
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
@@ -443,34 +454,35 @@
 ; GFX11-NEXT:    s_mov_b32 s33, s32
 ; GFX11-NEXT:    s_or_saveexec_b32 s1, -1
 ; GFX11-NEXT:    s_clause 0x1
-; GFX11-NEXT:    scratch_store_b32 off, v40, s33 offset:4
+; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    scratch_store_b32 off, v42, s33 offset:8
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr41
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v41, s30, 0
 ; GFX11-NEXT:    v_writelane_b32 v42, s0, 0
-; GFX11-NEXT:    scratch_store_b32 off, v41, s33 ; 4-byte Folded Spill
+; GFX11-NEXT:    scratch_store_b32 off, v40, s33 ; 4-byte Folded Spill
 ; GFX11-NEXT:    ;;#ASMSTART
 ; GFX11-NEXT:    ; def v31
 ; GFX11-NEXT:    ;;#ASMEND
-; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX11-NEXT:    v_mov_b32_e32 v41, v31
+; GFX11-NEXT:    v_mov_b32_e32 v40, v31
+; GFX11-NEXT:    v_writelane_b32 v41, s31, 1
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_void@rel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, external_void_func_void@rel32@hi+12
 ; GFX11-NEXT:    s_delay_alu instid0(SALU_CYCLE_1)
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
-; GFX11-NEXT:    v_mov_b32_e32 v31, v41
+; GFX11-NEXT:    v_mov_b32_e32 v31, v40
 ; GFX11-NEXT:    ;;#ASMSTART
 ; GFX11-NEXT:    ; use v31
 ; GFX11-NEXT:    ;;#ASMEND
-; GFX11-NEXT:    scratch_load_b32 v41, off, s33 ; 4-byte Folded Reload
-; GFX11-NEXT:    v_readlane_b32 s31, v40, 1
-; GFX11-NEXT:    v_readlane_b32 s30, v40, 0
+; GFX11-NEXT:    scratch_load_b32 v40, off, s33 ; 4-byte Folded Reload
+; GFX11-NEXT:    v_readlane_b32 s31, v41, 1
+; GFX11-NEXT:    v_readlane_b32 s30, v41, 0
 ; GFX11-NEXT:    v_readlane_b32 s0, v42, 0
 ; GFX11-NEXT:    s_or_saveexec_b32 s1, -1
 ; GFX11-NEXT:    s_clause 0x1
-; GFX11-NEXT:    scratch_load_b32 v40, off, s33 offset:4
+; GFX11-NEXT:    scratch_load_b32 v41, off, s33 offset:4
 ; GFX11-NEXT:    scratch_load_b32 v42, off, s33 offset:8
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX11-NEXT:    s_add_i32 s32, s32, -16
@@ -494,8 +506,9 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
-; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
+; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 2
@@ -535,8 +548,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_void@rel32@lo+4
@@ -544,8 +558,8 @@
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; def s33
 ; GFX10-NEXT:    ;;#ASMEND
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-NEXT:    s_mov_b32 s4, s33
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 2
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    s_mov_b32 s33, s4
@@ -578,8 +592,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_void@rel32@lo+4
@@ -587,8 +602,8 @@
 ; GFX11-NEXT:    ;;#ASMSTART
 ; GFX11-NEXT:    ; def s33
 ; GFX11-NEXT:    ;;#ASMEND
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX11-NEXT:    s_mov_b32 s4, s33
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX11-NEXT:    v_writelane_b32 v40, s31, 2
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX11-NEXT:    s_mov_b32 s33, s4
@@ -625,8 +640,9 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
-; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
+; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX9-NEXT:    ;;#ASMSTART
@@ -666,8 +682,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[36:37]
 ; GFX10-NEXT:    s_add_u32 s36, s36, external_void_func_void@rel32@lo+4
@@ -675,8 +692,8 @@
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; def s34
 ; GFX10-NEXT:    ;;#ASMEND
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-NEXT:    s_mov_b32 s4, s34
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 2
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[36:37]
 ; GFX10-NEXT:    s_mov_b32 s34, s4
@@ -709,8 +726,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_void@rel32@lo+4
@@ -718,8 +736,8 @@
 ; GFX11-NEXT:    ;;#ASMSTART
 ; GFX11-NEXT:    ; def s34
 ; GFX11-NEXT:    ;;#ASMEND
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX11-NEXT:    s_mov_b32 s4, s34
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX11-NEXT:    v_writelane_b32 v40, s31, 2
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX11-NEXT:    s_mov_b32 s34, s4
@@ -756,6 +774,7 @@
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v42, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr41
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v41, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v42, s34, 0
@@ -795,17 +814,18 @@
 ; GFX10-NEXT:    buffer_store_dword v42, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v41, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr41
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v41, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v42, s34, 0
 ; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; def v40
 ; GFX10-NEXT:    ;;#ASMEND
-; GFX10-NEXT:    v_writelane_b32 v41, s31, 1
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_void@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_void@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v41, s31, 1
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; use v40
@@ -836,18 +856,18 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    scratch_store_b32 off, v42, s33 offset:8
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v41, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr41
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v41, s30, 0
 ; GFX11-NEXT:    v_writelane_b32 v42, s0, 0
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33 ; 4-byte Folded Spill
 ; GFX11-NEXT:    ;;#ASMSTART
 ; GFX11-NEXT:    ; def v40
 ; GFX11-NEXT:    ;;#ASMEND
-; GFX11-NEXT:    v_writelane_b32 v41, s31, 1
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_void@rel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, external_void_func_void@rel32@hi+12
-; GFX11-NEXT:    s_delay_alu instid0(SALU_CYCLE_1)
+; GFX11-NEXT:    v_writelane_b32 v41, s31, 1
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX11-NEXT:    ;;#ASMSTART
 ; GFX11-NEXT:    ; use v40
@@ -875,17 +895,18 @@
 ; GFX9-LABEL: void_func_void_clobber_s33:
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX9-NEXT:    s_xor_saveexec_b64 s[4:5], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[6:7], -1
 ; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], s32 ; 4-byte Folded Spill
-; GFX9-NEXT:    s_mov_b64 exec, s[4:5]
+; GFX9-NEXT:    s_mov_b64 exec, s[6:7]
+; GFX9-NEXT:    ; implicit-def: $vgpr0
 ; GFX9-NEXT:    v_writelane_b32 v0, s33, 0
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; clobber
 ; GFX9-NEXT:    ;;#ASMEND
 ; GFX9-NEXT:    v_readlane_b32 s33, v0, 0
-; GFX9-NEXT:    s_xor_saveexec_b64 s[4:5], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[6:7], -1
 ; GFX9-NEXT:    buffer_load_dword v0, off, s[0:3], s32 ; 4-byte Folded Reload
-; GFX9-NEXT:    s_mov_b64 exec, s[4:5]
+; GFX9-NEXT:    s_mov_b64 exec, s[6:7]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
 ;
@@ -893,19 +914,20 @@
 ; GFX10:       ; %bb.0:
 ; GFX10-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
-; GFX10-NEXT:    s_xor_saveexec_b32 s4, -1
+; GFX10-NEXT:    s_xor_saveexec_b32 s5, -1
 ; GFX10-NEXT:    buffer_store_dword v0, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
-; GFX10-NEXT:    s_mov_b32 exec_lo, s4
+; GFX10-NEXT:    s_mov_b32 exec_lo, s5
+; GFX10-NEXT:    ; implicit-def: $vgpr0
 ; GFX10-NEXT:    v_writelane_b32 v0, s33, 0
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; clobber
 ; GFX10-NEXT:    ;;#ASMEND
 ; GFX10-NEXT:    v_readlane_b32 s33, v0, 0
-; GFX10-NEXT:    s_xor_saveexec_b32 s4, -1
+; GFX10-NEXT:    s_xor_saveexec_b32 s5, -1
 ; GFX10-NEXT:    buffer_load_dword v0, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
-; GFX10-NEXT:    s_mov_b32 exec_lo, s4
+; GFX10-NEXT:    s_mov_b32 exec_lo, s5
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    s_setpc_b64 s[30:31]
@@ -914,18 +936,19 @@
 ; GFX11:       ; %bb.0:
 ; GFX11-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX11-NEXT:    s_waitcnt_vscnt null, 0x0
-; GFX11-NEXT:    s_xor_saveexec_b32 s0, -1
+; GFX11-NEXT:    s_xor_saveexec_b32 s1, -1
 ; GFX11-NEXT:    scratch_store_b32 off, v0, s32 ; 4-byte Folded Spill
-; GFX11-NEXT:    s_mov_b32 exec_lo, s0
+; GFX11-NEXT:    s_mov_b32 exec_lo, s1
+; GFX11-NEXT:    ; implicit-def: $vgpr0
 ; GFX11-NEXT:    v_writelane_b32 v0, s33, 0
 ; GFX11-NEXT:    ;;#ASMSTART
 ; GFX11-NEXT:    ; clobber
 ; GFX11-NEXT:    ;;#ASMEND
 ; GFX11-NEXT:    s_delay_alu instid0(VALU_DEP_1)
 ; GFX11-NEXT:    v_readlane_b32 s33, v0, 0
-; GFX11-NEXT:    s_xor_saveexec_b32 s0, -1
+; GFX11-NEXT:    s_xor_saveexec_b32 s1, -1
 ; GFX11-NEXT:    scratch_load_b32 v0, off, s32 ; 4-byte Folded Reload
-; GFX11-NEXT:    s_mov_b32 exec_lo, s0
+; GFX11-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX11-NEXT:    s_waitcnt vmcnt(0)
 ; GFX11-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX11-NEXT:    s_setpc_b64 s[30:31]
@@ -937,17 +960,18 @@
 ; GFX9-LABEL: void_func_void_clobber_s34:
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX9-NEXT:    s_xor_saveexec_b64 s[4:5], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[6:7], -1
 ; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], s32 ; 4-byte Folded Spill
-; GFX9-NEXT:    s_mov_b64 exec, s[4:5]
+; GFX9-NEXT:    s_mov_b64 exec, s[6:7]
+; GFX9-NEXT:    ; implicit-def: $vgpr0
 ; GFX9-NEXT:    v_writelane_b32 v0, s34, 0
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; clobber
 ; GFX9-NEXT:    ;;#ASMEND
 ; GFX9-NEXT:    v_readlane_b32 s34, v0, 0
-; GFX9-NEXT:    s_xor_saveexec_b64 s[4:5], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[6:7], -1
 ; GFX9-NEXT:    buffer_load_dword v0, off, s[0:3], s32 ; 4-byte Folded Reload
-; GFX9-NEXT:    s_mov_b64 exec, s[4:5]
+; GFX9-NEXT:    s_mov_b64 exec, s[6:7]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
 ;
@@ -955,19 +979,20 @@
 ; GFX10:       ; %bb.0:
 ; GFX10-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
-; GFX10-NEXT:    s_xor_saveexec_b32 s4, -1
+; GFX10-NEXT:    s_xor_saveexec_b32 s5, -1
 ; GFX10-NEXT:    buffer_store_dword v0, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
-; GFX10-NEXT:    s_mov_b32 exec_lo, s4
+; GFX10-NEXT:    s_mov_b32 exec_lo, s5
+; GFX10-NEXT:    ; implicit-def: $vgpr0
 ; GFX10-NEXT:    v_writelane_b32 v0, s34, 0
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; clobber
 ; GFX10-NEXT:    ;;#ASMEND
 ; GFX10-NEXT:    v_readlane_b32 s34, v0, 0
-; GFX10-NEXT:    s_xor_saveexec_b32 s4, -1
+; GFX10-NEXT:    s_xor_saveexec_b32 s5, -1
 ; GFX10-NEXT:    buffer_load_dword v0, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
-; GFX10-NEXT:    s_mov_b32 exec_lo, s4
+; GFX10-NEXT:    s_mov_b32 exec_lo, s5
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    s_setpc_b64 s[30:31]
@@ -976,18 +1001,19 @@
 ; GFX11:       ; %bb.0:
 ; GFX11-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX11-NEXT:    s_waitcnt_vscnt null, 0x0
-; GFX11-NEXT:    s_xor_saveexec_b32 s0, -1
+; GFX11-NEXT:    s_xor_saveexec_b32 s1, -1
 ; GFX11-NEXT:    scratch_store_b32 off, v0, s32 ; 4-byte Folded Spill
-; GFX11-NEXT:    s_mov_b32 exec_lo, s0
+; GFX11-NEXT:    s_mov_b32 exec_lo, s1
+; GFX11-NEXT:    ; implicit-def: $vgpr0
 ; GFX11-NEXT:    v_writelane_b32 v0, s34, 0
 ; GFX11-NEXT:    ;;#ASMSTART
 ; GFX11-NEXT:    ; clobber
 ; GFX11-NEXT:    ;;#ASMEND
 ; GFX11-NEXT:    s_delay_alu instid0(VALU_DEP_1)
 ; GFX11-NEXT:    v_readlane_b32 s34, v0, 0
-; GFX11-NEXT:    s_xor_saveexec_b32 s0, -1
+; GFX11-NEXT:    s_xor_saveexec_b32 s1, -1
 ; GFX11-NEXT:    scratch_load_b32 v0, off, s32 ; 4-byte Folded Reload
-; GFX11-NEXT:    s_mov_b32 exec_lo, s0
+; GFX11-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX11-NEXT:    s_waitcnt vmcnt(0)
 ; GFX11-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX11-NEXT:    s_setpc_b64 s[30:31]
@@ -1005,6 +1031,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
@@ -1036,8 +1063,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, void_func_void_clobber_s33@rel32@lo+4
@@ -1069,8 +1097,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, void_func_void_clobber_s33@rel32@lo+4
@@ -1104,6 +1133,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
@@ -1135,8 +1165,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, void_func_void_clobber_s34@rel32@lo+4
@@ -1168,8 +1199,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, void_func_void_clobber_s34@rel32@lo+4
@@ -1203,8 +1235,9 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
-; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
+; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX9-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 2
@@ -1243,8 +1276,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s34, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_void@rel32@lo+4
@@ -1252,8 +1286,8 @@
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; def s40
 ; GFX10-NEXT:    ;;#ASMEND
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-NEXT:    s_mov_b32 s4, s40
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 2
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    ;;#ASMSTART
@@ -1285,8 +1319,9 @@
 ; GFX11-NEXT:    scratch_store_b32 off, v40, s33
 ; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr40
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX11-NEXT:    v_writelane_b32 v41, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_void@rel32@lo+4
@@ -1294,8 +1329,8 @@
 ; GFX11-NEXT:    ;;#ASMSTART
 ; GFX11-NEXT:    ; def s40
 ; GFX11-NEXT:    ;;#ASMEND
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX11-NEXT:    s_mov_b32 s4, s40
+; GFX11-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX11-NEXT:    v_writelane_b32 v40, s31, 2
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX11-NEXT:    ;;#ASMSTART
@@ -1328,15 +1363,16 @@
 ; GFX9-NEXT:    s_mov_b32 s34, s33
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_or_saveexec_b64 s[36:37], -1
-; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
+; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v42, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
-; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX9-NEXT:    ; implicit-def: $vgpr41
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
+; GFX9-NEXT:    v_writelane_b32 v41, s4, 0
+; GFX9-NEXT:    v_writelane_b32 v41, s30, 1
 ; GFX9-NEXT:    v_writelane_b32 v42, s34, 0
-; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 ; 4-byte Folded Spill
-; GFX9-NEXT:    v_writelane_b32 v40, s31, 2
+; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
+; GFX9-NEXT:    v_writelane_b32 v41, s31, 2
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; def s40
 ; GFX9-NEXT:    ;;#ASMEND
@@ -1344,7 +1380,7 @@
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; def v32
 ; GFX9-NEXT:    ;;#ASMEND
-; GFX9-NEXT:    v_mov_b32_e32 v41, v32
+; GFX9-NEXT:    v_mov_b32_e32 v40, v32
 ; GFX9-NEXT:    s_getpc_b64 s[34:35]
 ; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_void@rel32@lo+4
 ; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_void@rel32@hi+12
@@ -1353,15 +1389,15 @@
 ; GFX9-NEXT:    ; use s4
 ; GFX9-NEXT:    ;;#ASMEND
 ; GFX9-NEXT:    ;;#ASMSTART
-; GFX9-NEXT:    ; use v41
+; GFX9-NEXT:    ; use v40
 ; GFX9-NEXT:    ;;#ASMEND
-; GFX9-NEXT:    buffer_load_dword v41, off, s[0:3], s33 ; 4-byte Folded Reload
-; GFX9-NEXT:    v_readlane_b32 s31, v40, 2
-; GFX9-NEXT:    v_readlane_b32 s30, v40, 1
-; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
+; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s33 ; 4-byte Folded Reload
+; GFX9-NEXT:    v_readlane_b32 s31, v41, 2
+; GFX9-NEXT:    v_readlane_b32 s30, v41, 1
+; GFX9-NEXT:    v_readlane_b32 s4, v41, 0
 ; GFX9-NEXT:    v_readlane_b32 s34, v42, 0
 ; GFX9-NEXT:    s_or_saveexec_b64 s[36:37], -1
-; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
+; GFX9-NEXT:    buffer_load_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
 ; GFX9-NEXT:    buffer_load_dword v42, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[36:37]
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
@@ -1376,42 +1412,43 @@
 ; GFX10-NEXT:    s_mov_b32 s34, s33
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_or_saveexec_b32 s35, -1
-; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
+; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    buffer_store_dword v42, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr41
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    v_writelane_b32 v41, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v42, s34, 0
-; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 ; 4-byte Folded Spill
+; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; def s40
 ; GFX10-NEXT:    ;;#ASMEND
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-NEXT:    s_mov_b32 s4, s40
+; GFX10-NEXT:    v_writelane_b32 v41, s30, 1
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; def v32
 ; GFX10-NEXT:    ;;#ASMEND
-; GFX10-NEXT:    v_mov_b32_e32 v41, v32
+; GFX10-NEXT:    v_mov_b32_e32 v40, v32
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_void@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_void@rel32@hi+12
-; GFX10-NEXT:    v_writelane_b32 v40, s31, 2
+; GFX10-NEXT:    v_writelane_b32 v41, s31, 2
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; use s4
 ; GFX10-NEXT:    ;;#ASMEND
 ; GFX10-NEXT:    ;;#ASMSTART
-; GFX10-NEXT:    ; use v41
+; GFX10-NEXT:    ; use v40
 ; GFX10-NEXT:    ;;#ASMEND
-; GFX10-NEXT:    buffer_load_dword v41, off, s[0:3], s33 ; 4-byte Folded Reload
-; GFX10-NEXT:    v_readlane_b32 s31, v40, 2
-; GFX10-NEXT:    v_readlane_b32 s30, v40, 1
-; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
+; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s33 ; 4-byte Folded Reload
+; GFX10-NEXT:    v_readlane_b32 s31, v41, 2
+; GFX10-NEXT:    v_readlane_b32 s30, v41, 1
+; GFX10-NEXT:    v_readlane_b32 s4, v41, 0
 ; GFX10-NEXT:    v_readlane_b32 s34, v42, 0
 ; GFX10-NEXT:    s_or_saveexec_b32 s35, -1
 ; GFX10-NEXT:    s_clause 0x1
-; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s33 offset:4
+; GFX10-NEXT:    buffer_load_dword v41, off, s[0:3], s33 offset:4
 ; GFX10-NEXT:    buffer_load_dword v42, off, s[0:3], s33 offset:8
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s35
@@ -1428,41 +1465,42 @@
 ; GFX11-NEXT:    s_mov_b32 s33, s32
 ; GFX11-NEXT:    s_or_saveexec_b32 s1, -1
 ; GFX11-NEXT:    s_clause 0x1
-; GFX11-NEXT:    scratch_store_b32 off, v40, s33 offset:4
+; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:4
 ; GFX11-NEXT:    scratch_store_b32 off, v42, s33 offset:8
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
-; GFX11-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr41
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
+; GFX11-NEXT:    v_writelane_b32 v41, s4, 0
 ; GFX11-NEXT:    v_writelane_b32 v42, s0, 0
-; GFX11-NEXT:    scratch_store_b32 off, v41, s33 ; 4-byte Folded Spill
+; GFX11-NEXT:    scratch_store_b32 off, v40, s33 ; 4-byte Folded Spill
 ; GFX11-NEXT:    ;;#ASMSTART
 ; GFX11-NEXT:    ; def s40
 ; GFX11-NEXT:    ;;#ASMEND
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX11-NEXT:    s_mov_b32 s4, s40
+; GFX11-NEXT:    v_writelane_b32 v41, s30, 1
 ; GFX11-NEXT:    ;;#ASMSTART
 ; GFX11-NEXT:    ; def v32
 ; GFX11-NEXT:    ;;#ASMEND
-; GFX11-NEXT:    v_mov_b32_e32 v41, v32
+; GFX11-NEXT:    v_mov_b32_e32 v40, v32
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, external_void_func_void@rel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, external_void_func_void@rel32@hi+12
-; GFX11-NEXT:    v_writelane_b32 v40, s31, 2
+; GFX11-NEXT:    v_writelane_b32 v41, s31, 2
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX11-NEXT:    ;;#ASMSTART
 ; GFX11-NEXT:    ; use s4
 ; GFX11-NEXT:    ;;#ASMEND
 ; GFX11-NEXT:    ;;#ASMSTART
-; GFX11-NEXT:    ; use v41
+; GFX11-NEXT:    ; use v40
 ; GFX11-NEXT:    ;;#ASMEND
-; GFX11-NEXT:    scratch_load_b32 v41, off, s33 ; 4-byte Folded Reload
-; GFX11-NEXT:    v_readlane_b32 s31, v40, 2
-; GFX11-NEXT:    v_readlane_b32 s30, v40, 1
-; GFX11-NEXT:    v_readlane_b32 s4, v40, 0
+; GFX11-NEXT:    scratch_load_b32 v40, off, s33 ; 4-byte Folded Reload
+; GFX11-NEXT:    v_readlane_b32 s31, v41, 2
+; GFX11-NEXT:    v_readlane_b32 s30, v41, 1
+; GFX11-NEXT:    v_readlane_b32 s4, v41, 0
 ; GFX11-NEXT:    v_readlane_b32 s0, v42, 0
 ; GFX11-NEXT:    s_or_saveexec_b32 s1, -1
 ; GFX11-NEXT:    s_clause 0x1
-; GFX11-NEXT:    scratch_load_b32 v40, off, s33 offset:4
+; GFX11-NEXT:    scratch_load_b32 v41, off, s33 offset:4
 ; GFX11-NEXT:    scratch_load_b32 v42, off, s33 offset:8
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX11-NEXT:    s_add_i32 s32, s32, -16
diff --git a/llvm/test/CodeGen/AMDGPU/gfx-callable-return-types.ll b/llvm/test/CodeGen/AMDGPU/gfx-callable-return-types.ll
--- a/llvm/test/CodeGen/AMDGPU/gfx-callable-return-types.ll
+++ b/llvm/test/CodeGen/AMDGPU/gfx-callable-return-types.ll
@@ -34,6 +34,7 @@
 ; GFX9-NEXT:    s_add_u32 s34, s34, return_i1@gotpcrel32@lo+4
 ; GFX9-NEXT:    s_addc_u32 s35, s35, return_i1@gotpcrel32@hi+12
 ; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
+; GFX9-NEXT:    ; implicit-def: $vgpr1
 ; GFX9-NEXT:    v_writelane_b32 v1, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v1, s31, 1
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
@@ -62,6 +63,7 @@
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, return_i1@gotpcrel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, return_i1@gotpcrel32@hi+12
+; GFX10-NEXT:    ; implicit-def: $vgpr1
 ; GFX10-NEXT:    v_writelane_b32 v1, s30, 0
 ; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX10-NEXT:    v_writelane_b32 v1, s31, 1
@@ -82,7 +84,7 @@
 ; GFX11:       ; %bb.0: ; %entry
 ; GFX11-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX11-NEXT:    s_waitcnt_vscnt null, 0x0
-; GFX11-NEXT:    s_mov_b32 s2, s33
+; GFX11-NEXT:    s_mov_b32 s3, s33
 ; GFX11-NEXT:    s_mov_b32 s33, s32
 ; GFX11-NEXT:    s_xor_saveexec_b32 s0, -1
 ; GFX11-NEXT:    scratch_store_b32 off, v1, s33 ; 4-byte Folded Spill
@@ -91,6 +93,7 @@
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, return_i1@gotpcrel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, return_i1@gotpcrel32@hi+12
+; GFX11-NEXT:    ; implicit-def: $vgpr1
 ; GFX11-NEXT:    v_writelane_b32 v1, s30, 0
 ; GFX11-NEXT:    s_load_b64 s[0:1], s[0:1], 0x0
 ; GFX11-NEXT:    v_writelane_b32 v1, s31, 1
@@ -103,7 +106,7 @@
 ; GFX11-NEXT:    scratch_load_b32 v1, off, s33 ; 4-byte Folded Reload
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s0
 ; GFX11-NEXT:    s_add_i32 s32, s32, -16
-; GFX11-NEXT:    s_mov_b32 s33, s2
+; GFX11-NEXT:    s_mov_b32 s33, s3
 ; GFX11-NEXT:    s_waitcnt vmcnt(0)
 ; GFX11-NEXT:    s_setpc_b64 s[30:31]
 entry:
@@ -142,6 +145,7 @@
 ; GFX9-NEXT:    s_add_u32 s34, s34, return_i16@gotpcrel32@lo+4
 ; GFX9-NEXT:    s_addc_u32 s35, s35, return_i16@gotpcrel32@hi+12
 ; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
+; GFX9-NEXT:    ; implicit-def: $vgpr1
 ; GFX9-NEXT:    v_writelane_b32 v1, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v1, s31, 1
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
@@ -170,6 +174,7 @@
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, return_i16@gotpcrel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, return_i16@gotpcrel32@hi+12
+; GFX10-NEXT:    ; implicit-def: $vgpr1
 ; GFX10-NEXT:    v_writelane_b32 v1, s30, 0
 ; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX10-NEXT:    v_writelane_b32 v1, s31, 1
@@ -190,7 +195,7 @@
 ; GFX11:       ; %bb.0: ; %entry
 ; GFX11-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX11-NEXT:    s_waitcnt_vscnt null, 0x0
-; GFX11-NEXT:    s_mov_b32 s2, s33
+; GFX11-NEXT:    s_mov_b32 s3, s33
 ; GFX11-NEXT:    s_mov_b32 s33, s32
 ; GFX11-NEXT:    s_xor_saveexec_b32 s0, -1
 ; GFX11-NEXT:    scratch_store_b32 off, v1, s33 ; 4-byte Folded Spill
@@ -199,6 +204,7 @@
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, return_i16@gotpcrel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, return_i16@gotpcrel32@hi+12
+; GFX11-NEXT:    ; implicit-def: $vgpr1
 ; GFX11-NEXT:    v_writelane_b32 v1, s30, 0
 ; GFX11-NEXT:    s_load_b64 s[0:1], s[0:1], 0x0
 ; GFX11-NEXT:    v_writelane_b32 v1, s31, 1
@@ -211,7 +217,7 @@
 ; GFX11-NEXT:    scratch_load_b32 v1, off, s33 ; 4-byte Folded Reload
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s0
 ; GFX11-NEXT:    s_add_i32 s32, s32, -16
-; GFX11-NEXT:    s_mov_b32 s33, s2
+; GFX11-NEXT:    s_mov_b32 s33, s3
 ; GFX11-NEXT:    s_waitcnt vmcnt(0)
 ; GFX11-NEXT:    s_setpc_b64 s[30:31]
 entry:
@@ -250,6 +256,7 @@
 ; GFX9-NEXT:    s_add_u32 s34, s34, return_2xi16@gotpcrel32@lo+4
 ; GFX9-NEXT:    s_addc_u32 s35, s35, return_2xi16@gotpcrel32@hi+12
 ; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
+; GFX9-NEXT:    ; implicit-def: $vgpr1
 ; GFX9-NEXT:    v_writelane_b32 v1, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v1, s31, 1
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
@@ -278,6 +285,7 @@
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, return_2xi16@gotpcrel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, return_2xi16@gotpcrel32@hi+12
+; GFX10-NEXT:    ; implicit-def: $vgpr1
 ; GFX10-NEXT:    v_writelane_b32 v1, s30, 0
 ; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX10-NEXT:    v_writelane_b32 v1, s31, 1
@@ -298,7 +306,7 @@
 ; GFX11:       ; %bb.0: ; %entry
 ; GFX11-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX11-NEXT:    s_waitcnt_vscnt null, 0x0
-; GFX11-NEXT:    s_mov_b32 s2, s33
+; GFX11-NEXT:    s_mov_b32 s3, s33
 ; GFX11-NEXT:    s_mov_b32 s33, s32
 ; GFX11-NEXT:    s_xor_saveexec_b32 s0, -1
 ; GFX11-NEXT:    scratch_store_b32 off, v1, s33 ; 4-byte Folded Spill
@@ -307,6 +315,7 @@
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, return_2xi16@gotpcrel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, return_2xi16@gotpcrel32@hi+12
+; GFX11-NEXT:    ; implicit-def: $vgpr1
 ; GFX11-NEXT:    v_writelane_b32 v1, s30, 0
 ; GFX11-NEXT:    s_load_b64 s[0:1], s[0:1], 0x0
 ; GFX11-NEXT:    v_writelane_b32 v1, s31, 1
@@ -319,7 +328,7 @@
 ; GFX11-NEXT:    scratch_load_b32 v1, off, s33 ; 4-byte Folded Reload
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s0
 ; GFX11-NEXT:    s_add_i32 s32, s32, -16
-; GFX11-NEXT:    s_mov_b32 s33, s2
+; GFX11-NEXT:    s_mov_b32 s33, s3
 ; GFX11-NEXT:    s_waitcnt vmcnt(0)
 ; GFX11-NEXT:    s_setpc_b64 s[30:31]
 entry:
@@ -367,6 +376,7 @@
 ; GFX9-NEXT:    s_add_u32 s34, s34, return_3xi16@gotpcrel32@lo+4
 ; GFX9-NEXT:    s_addc_u32 s35, s35, return_3xi16@gotpcrel32@hi+12
 ; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
+; GFX9-NEXT:    ; implicit-def: $vgpr2
 ; GFX9-NEXT:    v_writelane_b32 v2, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v2, s31, 1
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
@@ -395,6 +405,7 @@
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, return_3xi16@gotpcrel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, return_3xi16@gotpcrel32@hi+12
+; GFX10-NEXT:    ; implicit-def: $vgpr2
 ; GFX10-NEXT:    v_writelane_b32 v2, s30, 0
 ; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX10-NEXT:    v_writelane_b32 v2, s31, 1
@@ -415,7 +426,7 @@
 ; GFX11:       ; %bb.0: ; %entry
 ; GFX11-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX11-NEXT:    s_waitcnt_vscnt null, 0x0
-; GFX11-NEXT:    s_mov_b32 s2, s33
+; GFX11-NEXT:    s_mov_b32 s3, s33
 ; GFX11-NEXT:    s_mov_b32 s33, s32
 ; GFX11-NEXT:    s_xor_saveexec_b32 s0, -1
 ; GFX11-NEXT:    scratch_store_b32 off, v2, s33 ; 4-byte Folded Spill
@@ -424,6 +435,7 @@
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, return_3xi16@gotpcrel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, return_3xi16@gotpcrel32@hi+12
+; GFX11-NEXT:    ; implicit-def: $vgpr2
 ; GFX11-NEXT:    v_writelane_b32 v2, s30, 0
 ; GFX11-NEXT:    s_load_b64 s[0:1], s[0:1], 0x0
 ; GFX11-NEXT:    v_writelane_b32 v2, s31, 1
@@ -436,7 +448,7 @@
 ; GFX11-NEXT:    scratch_load_b32 v2, off, s33 ; 4-byte Folded Reload
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s0
 ; GFX11-NEXT:    s_add_i32 s32, s32, -16
-; GFX11-NEXT:    s_mov_b32 s33, s2
+; GFX11-NEXT:    s_mov_b32 s33, s3
 ; GFX11-NEXT:    s_waitcnt vmcnt(0)
 ; GFX11-NEXT:    s_setpc_b64 s[30:31]
 entry:
@@ -1649,8 +1661,9 @@
 ; GFX9-NEXT:    s_add_u32 s34, s34, return_512xi32@gotpcrel32@lo+4
 ; GFX9-NEXT:    s_addc_u32 s35, s35, return_512xi32@gotpcrel32@hi+12
 ; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
-; GFX9-NEXT:    v_writelane_b32 v2, s30, 0
+; GFX9-NEXT:    ; implicit-def: $vgpr2
 ; GFX9-NEXT:    v_lshrrev_b32_e64 v0, 6, s33
+; GFX9-NEXT:    v_writelane_b32 v2, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v2, s31, 1
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
@@ -1679,9 +1692,10 @@
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, return_512xi32@gotpcrel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, return_512xi32@gotpcrel32@hi+12
-; GFX10-NEXT:    v_writelane_b32 v2, s30, 0
-; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
+; GFX10-NEXT:    ; implicit-def: $vgpr2
 ; GFX10-NEXT:    v_lshrrev_b32_e64 v0, 5, s33
+; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
+; GFX10-NEXT:    v_writelane_b32 v2, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v2, s31, 1
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
@@ -1711,9 +1725,10 @@
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, return_512xi32@gotpcrel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, return_512xi32@gotpcrel32@hi+12
-; GFX11-NEXT:    v_writelane_b32 v5, s30, 0
-; GFX11-NEXT:    s_load_b64 s[0:1], s[0:1], 0x0
+; GFX11-NEXT:    ; implicit-def: $vgpr5
 ; GFX11-NEXT:    v_mov_b32_e32 v0, s33
+; GFX11-NEXT:    s_load_b64 s[0:1], s[0:1], 0x0
+; GFX11-NEXT:    v_writelane_b32 v5, s30, 0
 ; GFX11-NEXT:    v_writelane_b32 v5, s31, 1
 ; GFX11-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
diff --git a/llvm/test/CodeGen/AMDGPU/indirect-call.ll b/llvm/test/CodeGen/AMDGPU/indirect-call.ll
--- a/llvm/test/CodeGen/AMDGPU/indirect-call.ll
+++ b/llvm/test/CodeGen/AMDGPU/indirect-call.ll
@@ -401,6 +401,7 @@
 ; GCN-NEXT:    s_mov_b64 exec, s[18:19]
 ; GCN-NEXT:    v_writelane_b32 v41, s16, 0
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
+; GCN-NEXT:    ; implicit-def: $vgpr40
 ; GCN-NEXT:    v_writelane_b32 v40, s30, 0
 ; GCN-NEXT:    v_writelane_b32 v40, s31, 1
 ; GCN-NEXT:    v_writelane_b32 v40, s34, 2
@@ -487,6 +488,7 @@
 ; GISEL-NEXT:    s_mov_b64 exec, s[18:19]
 ; GISEL-NEXT:    v_writelane_b32 v41, s16, 0
 ; GISEL-NEXT:    s_addk_i32 s32, 0x400
+; GISEL-NEXT:    ; implicit-def: $vgpr40
 ; GISEL-NEXT:    v_writelane_b32 v40, s30, 0
 ; GISEL-NEXT:    v_writelane_b32 v40, s31, 1
 ; GISEL-NEXT:    v_writelane_b32 v40, s34, 2
@@ -577,6 +579,7 @@
 ; GCN-NEXT:    s_mov_b64 exec, s[18:19]
 ; GCN-NEXT:    v_writelane_b32 v41, s16, 0
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
+; GCN-NEXT:    ; implicit-def: $vgpr40
 ; GCN-NEXT:    v_writelane_b32 v40, s30, 0
 ; GCN-NEXT:    v_writelane_b32 v40, s31, 1
 ; GCN-NEXT:    v_writelane_b32 v40, s34, 2
@@ -666,6 +669,7 @@
 ; GISEL-NEXT:    s_mov_b64 exec, s[18:19]
 ; GISEL-NEXT:    v_writelane_b32 v41, s16, 0
 ; GISEL-NEXT:    s_addk_i32 s32, 0x400
+; GISEL-NEXT:    ; implicit-def: $vgpr40
 ; GISEL-NEXT:    v_writelane_b32 v40, s30, 0
 ; GISEL-NEXT:    v_writelane_b32 v40, s31, 1
 ; GISEL-NEXT:    v_writelane_b32 v40, s34, 2
@@ -757,6 +761,7 @@
 ; GCN-NEXT:    s_mov_b64 exec, s[18:19]
 ; GCN-NEXT:    v_writelane_b32 v41, s16, 0
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
+; GCN-NEXT:    ; implicit-def: $vgpr40
 ; GCN-NEXT:    v_writelane_b32 v40, s30, 0
 ; GCN-NEXT:    v_writelane_b32 v40, s31, 1
 ; GCN-NEXT:    v_writelane_b32 v40, s34, 2
@@ -845,6 +850,7 @@
 ; GISEL-NEXT:    s_mov_b64 exec, s[18:19]
 ; GISEL-NEXT:    v_writelane_b32 v41, s16, 0
 ; GISEL-NEXT:    s_addk_i32 s32, 0x400
+; GISEL-NEXT:    ; implicit-def: $vgpr40
 ; GISEL-NEXT:    v_writelane_b32 v40, s30, 0
 ; GISEL-NEXT:    v_writelane_b32 v40, s31, 1
 ; GISEL-NEXT:    v_writelane_b32 v40, s34, 2
@@ -938,6 +944,7 @@
 ; GCN-NEXT:    s_mov_b64 exec, s[18:19]
 ; GCN-NEXT:    v_writelane_b32 v41, s16, 0
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
+; GCN-NEXT:    ; implicit-def: $vgpr40
 ; GCN-NEXT:    v_writelane_b32 v40, s30, 0
 ; GCN-NEXT:    v_writelane_b32 v40, s31, 1
 ; GCN-NEXT:    v_writelane_b32 v40, s34, 2
@@ -1035,6 +1042,7 @@
 ; GISEL-NEXT:    s_mov_b64 exec, s[18:19]
 ; GISEL-NEXT:    v_writelane_b32 v41, s16, 0
 ; GISEL-NEXT:    s_addk_i32 s32, 0x400
+; GISEL-NEXT:    ; implicit-def: $vgpr40
 ; GISEL-NEXT:    v_writelane_b32 v40, s30, 0
 ; GISEL-NEXT:    v_writelane_b32 v40, s31, 1
 ; GISEL-NEXT:    v_writelane_b32 v40, s34, 2
@@ -1141,6 +1149,7 @@
 ; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[6:7]
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
+; GCN-NEXT:    ; implicit-def: $vgpr40
 ; GCN-NEXT:    v_writelane_b32 v40, s30, 0
 ; GCN-NEXT:    v_writelane_b32 v40, s31, 1
 ; GCN-NEXT:    v_writelane_b32 v40, s34, 2
@@ -1235,6 +1244,7 @@
 ; GISEL-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GISEL-NEXT:    s_mov_b64 exec, s[6:7]
 ; GISEL-NEXT:    s_addk_i32 s32, 0x400
+; GISEL-NEXT:    ; implicit-def: $vgpr40
 ; GISEL-NEXT:    v_writelane_b32 v40, s30, 0
 ; GISEL-NEXT:    v_writelane_b32 v40, s31, 1
 ; GISEL-NEXT:    v_writelane_b32 v40, s34, 2
@@ -1327,196 +1337,198 @@
 ; GCN-LABEL: test_indirect_call_vgpr_ptr_arg_and_reuse:
 ; GCN:       ; %bb.0:
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GCN-NEXT:    s_mov_b32 s10, s33
+; GCN-NEXT:    s_mov_b32 s12, s33
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
-; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
+; GCN-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
-; GCN-NEXT:    buffer_store_dword v41, off, s[0:3], s33 ; 4-byte Folded Spill
-; GCN-NEXT:    v_writelane_b32 v40, s30, 0
-; GCN-NEXT:    v_writelane_b32 v40, s31, 1
-; GCN-NEXT:    v_writelane_b32 v40, s34, 2
-; GCN-NEXT:    v_writelane_b32 v40, s35, 3
-; GCN-NEXT:    v_writelane_b32 v40, s36, 4
-; GCN-NEXT:    v_writelane_b32 v40, s37, 5
-; GCN-NEXT:    v_writelane_b32 v40, s38, 6
-; GCN-NEXT:    v_writelane_b32 v40, s39, 7
-; GCN-NEXT:    v_writelane_b32 v40, s40, 8
-; GCN-NEXT:    v_writelane_b32 v40, s41, 9
-; GCN-NEXT:    v_writelane_b32 v40, s42, 10
-; GCN-NEXT:    v_writelane_b32 v40, s43, 11
-; GCN-NEXT:    v_writelane_b32 v40, s44, 12
-; GCN-NEXT:    v_writelane_b32 v40, s45, 13
-; GCN-NEXT:    v_writelane_b32 v40, s46, 14
-; GCN-NEXT:    v_writelane_b32 v40, s47, 15
-; GCN-NEXT:    v_writelane_b32 v40, s48, 16
-; GCN-NEXT:    v_writelane_b32 v40, s49, 17
-; GCN-NEXT:    v_writelane_b32 v40, s50, 18
-; GCN-NEXT:    v_writelane_b32 v40, s51, 19
-; GCN-NEXT:    v_writelane_b32 v40, s52, 20
-; GCN-NEXT:    v_writelane_b32 v40, s53, 21
-; GCN-NEXT:    v_writelane_b32 v40, s54, 22
-; GCN-NEXT:    v_writelane_b32 v40, s55, 23
-; GCN-NEXT:    v_writelane_b32 v40, s56, 24
-; GCN-NEXT:    v_writelane_b32 v40, s57, 25
-; GCN-NEXT:    v_writelane_b32 v40, s58, 26
-; GCN-NEXT:    v_writelane_b32 v40, s59, 27
-; GCN-NEXT:    v_writelane_b32 v40, s60, 28
-; GCN-NEXT:    v_writelane_b32 v40, s61, 29
-; GCN-NEXT:    v_writelane_b32 v40, s62, 30
-; GCN-NEXT:    v_writelane_b32 v40, s63, 31
-; GCN-NEXT:    v_mov_b32_e32 v41, v0
+; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
+; GCN-NEXT:    ; implicit-def: $vgpr41
+; GCN-NEXT:    v_writelane_b32 v41, s30, 0
+; GCN-NEXT:    v_writelane_b32 v41, s31, 1
+; GCN-NEXT:    v_writelane_b32 v41, s34, 2
+; GCN-NEXT:    v_writelane_b32 v41, s35, 3
+; GCN-NEXT:    v_writelane_b32 v41, s36, 4
+; GCN-NEXT:    v_writelane_b32 v41, s37, 5
+; GCN-NEXT:    v_writelane_b32 v41, s38, 6
+; GCN-NEXT:    v_writelane_b32 v41, s39, 7
+; GCN-NEXT:    v_writelane_b32 v41, s40, 8
+; GCN-NEXT:    v_writelane_b32 v41, s41, 9
+; GCN-NEXT:    v_writelane_b32 v41, s42, 10
+; GCN-NEXT:    v_writelane_b32 v41, s43, 11
+; GCN-NEXT:    v_writelane_b32 v41, s44, 12
+; GCN-NEXT:    v_writelane_b32 v41, s45, 13
+; GCN-NEXT:    v_writelane_b32 v41, s46, 14
+; GCN-NEXT:    v_writelane_b32 v41, s47, 15
+; GCN-NEXT:    v_writelane_b32 v41, s48, 16
+; GCN-NEXT:    v_writelane_b32 v41, s49, 17
+; GCN-NEXT:    v_writelane_b32 v41, s50, 18
+; GCN-NEXT:    v_writelane_b32 v41, s51, 19
+; GCN-NEXT:    v_writelane_b32 v41, s52, 20
+; GCN-NEXT:    v_writelane_b32 v41, s53, 21
+; GCN-NEXT:    v_writelane_b32 v41, s54, 22
+; GCN-NEXT:    v_writelane_b32 v41, s55, 23
+; GCN-NEXT:    v_writelane_b32 v41, s56, 24
+; GCN-NEXT:    v_writelane_b32 v41, s57, 25
+; GCN-NEXT:    v_writelane_b32 v41, s58, 26
+; GCN-NEXT:    v_writelane_b32 v41, s59, 27
+; GCN-NEXT:    v_writelane_b32 v41, s60, 28
+; GCN-NEXT:    v_writelane_b32 v41, s61, 29
+; GCN-NEXT:    v_writelane_b32 v41, s62, 30
+; GCN-NEXT:    v_writelane_b32 v41, s63, 31
+; GCN-NEXT:    v_mov_b32_e32 v40, v0
 ; GCN-NEXT:    s_mov_b64 s[4:5], exec
 ; GCN-NEXT:  .LBB7_1: ; =>This Inner Loop Header: Depth=1
 ; GCN-NEXT:    v_readfirstlane_b32 s6, v1
 ; GCN-NEXT:    v_readfirstlane_b32 s7, v2
 ; GCN-NEXT:    v_cmp_eq_u64_e32 vcc, s[6:7], v[1:2]
 ; GCN-NEXT:    s_and_saveexec_b64 s[8:9], vcc
-; GCN-NEXT:    v_mov_b32_e32 v0, v41
+; GCN-NEXT:    v_mov_b32_e32 v0, v40
 ; GCN-NEXT:    s_swappc_b64 s[30:31], s[6:7]
 ; GCN-NEXT:    ; implicit-def: $vgpr1_vgpr2
 ; GCN-NEXT:    s_xor_b64 exec, exec, s[8:9]
 ; GCN-NEXT:    s_cbranch_execnz .LBB7_1
 ; GCN-NEXT:  ; %bb.2:
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
-; GCN-NEXT:    v_mov_b32_e32 v0, v41
-; GCN-NEXT:    v_readlane_b32 s63, v40, 31
-; GCN-NEXT:    v_readlane_b32 s62, v40, 30
-; GCN-NEXT:    v_readlane_b32 s61, v40, 29
-; GCN-NEXT:    v_readlane_b32 s60, v40, 28
-; GCN-NEXT:    v_readlane_b32 s59, v40, 27
-; GCN-NEXT:    v_readlane_b32 s58, v40, 26
-; GCN-NEXT:    v_readlane_b32 s57, v40, 25
-; GCN-NEXT:    v_readlane_b32 s56, v40, 24
-; GCN-NEXT:    v_readlane_b32 s55, v40, 23
-; GCN-NEXT:    v_readlane_b32 s54, v40, 22
-; GCN-NEXT:    v_readlane_b32 s53, v40, 21
-; GCN-NEXT:    v_readlane_b32 s52, v40, 20
-; GCN-NEXT:    v_readlane_b32 s51, v40, 19
-; GCN-NEXT:    v_readlane_b32 s50, v40, 18
-; GCN-NEXT:    v_readlane_b32 s49, v40, 17
-; GCN-NEXT:    v_readlane_b32 s48, v40, 16
-; GCN-NEXT:    v_readlane_b32 s47, v40, 15
-; GCN-NEXT:    v_readlane_b32 s46, v40, 14
-; GCN-NEXT:    v_readlane_b32 s45, v40, 13
-; GCN-NEXT:    v_readlane_b32 s44, v40, 12
-; GCN-NEXT:    v_readlane_b32 s43, v40, 11
-; GCN-NEXT:    v_readlane_b32 s42, v40, 10
-; GCN-NEXT:    v_readlane_b32 s41, v40, 9
-; GCN-NEXT:    v_readlane_b32 s40, v40, 8
-; GCN-NEXT:    v_readlane_b32 s39, v40, 7
-; GCN-NEXT:    v_readlane_b32 s38, v40, 6
-; GCN-NEXT:    v_readlane_b32 s37, v40, 5
-; GCN-NEXT:    v_readlane_b32 s36, v40, 4
-; GCN-NEXT:    v_readlane_b32 s35, v40, 3
-; GCN-NEXT:    v_readlane_b32 s34, v40, 2
-; GCN-NEXT:    v_readlane_b32 s31, v40, 1
-; GCN-NEXT:    v_readlane_b32 s30, v40, 0
-; GCN-NEXT:    buffer_load_dword v41, off, s[0:3], s33 ; 4-byte Folded Reload
+; GCN-NEXT:    v_mov_b32_e32 v0, v40
+; GCN-NEXT:    v_readlane_b32 s63, v41, 31
+; GCN-NEXT:    v_readlane_b32 s62, v41, 30
+; GCN-NEXT:    v_readlane_b32 s61, v41, 29
+; GCN-NEXT:    v_readlane_b32 s60, v41, 28
+; GCN-NEXT:    v_readlane_b32 s59, v41, 27
+; GCN-NEXT:    v_readlane_b32 s58, v41, 26
+; GCN-NEXT:    v_readlane_b32 s57, v41, 25
+; GCN-NEXT:    v_readlane_b32 s56, v41, 24
+; GCN-NEXT:    v_readlane_b32 s55, v41, 23
+; GCN-NEXT:    v_readlane_b32 s54, v41, 22
+; GCN-NEXT:    v_readlane_b32 s53, v41, 21
+; GCN-NEXT:    v_readlane_b32 s52, v41, 20
+; GCN-NEXT:    v_readlane_b32 s51, v41, 19
+; GCN-NEXT:    v_readlane_b32 s50, v41, 18
+; GCN-NEXT:    v_readlane_b32 s49, v41, 17
+; GCN-NEXT:    v_readlane_b32 s48, v41, 16
+; GCN-NEXT:    v_readlane_b32 s47, v41, 15
+; GCN-NEXT:    v_readlane_b32 s46, v41, 14
+; GCN-NEXT:    v_readlane_b32 s45, v41, 13
+; GCN-NEXT:    v_readlane_b32 s44, v41, 12
+; GCN-NEXT:    v_readlane_b32 s43, v41, 11
+; GCN-NEXT:    v_readlane_b32 s42, v41, 10
+; GCN-NEXT:    v_readlane_b32 s41, v41, 9
+; GCN-NEXT:    v_readlane_b32 s40, v41, 8
+; GCN-NEXT:    v_readlane_b32 s39, v41, 7
+; GCN-NEXT:    v_readlane_b32 s38, v41, 6
+; GCN-NEXT:    v_readlane_b32 s37, v41, 5
+; GCN-NEXT:    v_readlane_b32 s36, v41, 4
+; GCN-NEXT:    v_readlane_b32 s35, v41, 3
+; GCN-NEXT:    v_readlane_b32 s34, v41, 2
+; GCN-NEXT:    v_readlane_b32 s31, v41, 1
+; GCN-NEXT:    v_readlane_b32 s30, v41, 0
+; GCN-NEXT:    buffer_load_dword v40, off, s[0:3], s33 ; 4-byte Folded Reload
 ; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
-; GCN-NEXT:    buffer_load_dword v40, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
+; GCN-NEXT:    buffer_load_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_addk_i32 s32, 0xfc00
-; GCN-NEXT:    s_mov_b32 s33, s10
+; GCN-NEXT:    s_mov_b32 s33, s12
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
 ; GCN-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GISEL-LABEL: test_indirect_call_vgpr_ptr_arg_and_reuse:
 ; GISEL:       ; %bb.0:
 ; GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GISEL-NEXT:    s_mov_b32 s10, s33
+; GISEL-NEXT:    s_mov_b32 s12, s33
 ; GISEL-NEXT:    s_mov_b32 s33, s32
 ; GISEL-NEXT:    s_or_saveexec_b64 s[4:5], -1
-; GISEL-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
+; GISEL-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GISEL-NEXT:    s_mov_b64 exec, s[4:5]
 ; GISEL-NEXT:    s_addk_i32 s32, 0x400
-; GISEL-NEXT:    buffer_store_dword v41, off, s[0:3], s33 ; 4-byte Folded Spill
-; GISEL-NEXT:    v_writelane_b32 v40, s30, 0
-; GISEL-NEXT:    v_writelane_b32 v40, s31, 1
-; GISEL-NEXT:    v_writelane_b32 v40, s34, 2
-; GISEL-NEXT:    v_writelane_b32 v40, s35, 3
-; GISEL-NEXT:    v_writelane_b32 v40, s36, 4
-; GISEL-NEXT:    v_writelane_b32 v40, s37, 5
-; GISEL-NEXT:    v_writelane_b32 v40, s38, 6
-; GISEL-NEXT:    v_writelane_b32 v40, s39, 7
-; GISEL-NEXT:    v_writelane_b32 v40, s40, 8
-; GISEL-NEXT:    v_writelane_b32 v40, s41, 9
-; GISEL-NEXT:    v_writelane_b32 v40, s42, 10
-; GISEL-NEXT:    v_writelane_b32 v40, s43, 11
-; GISEL-NEXT:    v_writelane_b32 v40, s44, 12
-; GISEL-NEXT:    v_writelane_b32 v40, s45, 13
-; GISEL-NEXT:    v_writelane_b32 v40, s46, 14
-; GISEL-NEXT:    v_writelane_b32 v40, s47, 15
-; GISEL-NEXT:    v_writelane_b32 v40, s48, 16
-; GISEL-NEXT:    v_writelane_b32 v40, s49, 17
-; GISEL-NEXT:    v_writelane_b32 v40, s50, 18
-; GISEL-NEXT:    v_writelane_b32 v40, s51, 19
-; GISEL-NEXT:    v_writelane_b32 v40, s52, 20
-; GISEL-NEXT:    v_writelane_b32 v40, s53, 21
-; GISEL-NEXT:    v_writelane_b32 v40, s54, 22
-; GISEL-NEXT:    v_writelane_b32 v40, s55, 23
-; GISEL-NEXT:    v_writelane_b32 v40, s56, 24
-; GISEL-NEXT:    v_writelane_b32 v40, s57, 25
-; GISEL-NEXT:    v_writelane_b32 v40, s58, 26
-; GISEL-NEXT:    v_writelane_b32 v40, s59, 27
-; GISEL-NEXT:    v_writelane_b32 v40, s60, 28
-; GISEL-NEXT:    v_writelane_b32 v40, s61, 29
-; GISEL-NEXT:    v_writelane_b32 v40, s62, 30
-; GISEL-NEXT:    v_writelane_b32 v40, s63, 31
-; GISEL-NEXT:    v_mov_b32_e32 v41, v0
+; GISEL-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
+; GISEL-NEXT:    ; implicit-def: $vgpr41
+; GISEL-NEXT:    v_writelane_b32 v41, s30, 0
+; GISEL-NEXT:    v_writelane_b32 v41, s31, 1
+; GISEL-NEXT:    v_writelane_b32 v41, s34, 2
+; GISEL-NEXT:    v_writelane_b32 v41, s35, 3
+; GISEL-NEXT:    v_writelane_b32 v41, s36, 4
+; GISEL-NEXT:    v_writelane_b32 v41, s37, 5
+; GISEL-NEXT:    v_writelane_b32 v41, s38, 6
+; GISEL-NEXT:    v_writelane_b32 v41, s39, 7
+; GISEL-NEXT:    v_writelane_b32 v41, s40, 8
+; GISEL-NEXT:    v_writelane_b32 v41, s41, 9
+; GISEL-NEXT:    v_writelane_b32 v41, s42, 10
+; GISEL-NEXT:    v_writelane_b32 v41, s43, 11
+; GISEL-NEXT:    v_writelane_b32 v41, s44, 12
+; GISEL-NEXT:    v_writelane_b32 v41, s45, 13
+; GISEL-NEXT:    v_writelane_b32 v41, s46, 14
+; GISEL-NEXT:    v_writelane_b32 v41, s47, 15
+; GISEL-NEXT:    v_writelane_b32 v41, s48, 16
+; GISEL-NEXT:    v_writelane_b32 v41, s49, 17
+; GISEL-NEXT:    v_writelane_b32 v41, s50, 18
+; GISEL-NEXT:    v_writelane_b32 v41, s51, 19
+; GISEL-NEXT:    v_writelane_b32 v41, s52, 20
+; GISEL-NEXT:    v_writelane_b32 v41, s53, 21
+; GISEL-NEXT:    v_writelane_b32 v41, s54, 22
+; GISEL-NEXT:    v_writelane_b32 v41, s55, 23
+; GISEL-NEXT:    v_writelane_b32 v41, s56, 24
+; GISEL-NEXT:    v_writelane_b32 v41, s57, 25
+; GISEL-NEXT:    v_writelane_b32 v41, s58, 26
+; GISEL-NEXT:    v_writelane_b32 v41, s59, 27
+; GISEL-NEXT:    v_writelane_b32 v41, s60, 28
+; GISEL-NEXT:    v_writelane_b32 v41, s61, 29
+; GISEL-NEXT:    v_writelane_b32 v41, s62, 30
+; GISEL-NEXT:    v_writelane_b32 v41, s63, 31
+; GISEL-NEXT:    v_mov_b32_e32 v40, v0
 ; GISEL-NEXT:    s_mov_b64 s[4:5], exec
 ; GISEL-NEXT:  .LBB7_1: ; =>This Inner Loop Header: Depth=1
 ; GISEL-NEXT:    v_readfirstlane_b32 s6, v1
 ; GISEL-NEXT:    v_readfirstlane_b32 s7, v2
 ; GISEL-NEXT:    v_cmp_eq_u64_e32 vcc, s[6:7], v[1:2]
 ; GISEL-NEXT:    s_and_saveexec_b64 s[8:9], vcc
-; GISEL-NEXT:    v_mov_b32_e32 v0, v41
+; GISEL-NEXT:    v_mov_b32_e32 v0, v40
 ; GISEL-NEXT:    s_swappc_b64 s[30:31], s[6:7]
 ; GISEL-NEXT:    ; implicit-def: $vgpr1
 ; GISEL-NEXT:    s_xor_b64 exec, exec, s[8:9]
 ; GISEL-NEXT:    s_cbranch_execnz .LBB7_1
 ; GISEL-NEXT:  ; %bb.2:
 ; GISEL-NEXT:    s_mov_b64 exec, s[4:5]
-; GISEL-NEXT:    v_mov_b32_e32 v0, v41
-; GISEL-NEXT:    v_readlane_b32 s63, v40, 31
-; GISEL-NEXT:    v_readlane_b32 s62, v40, 30
-; GISEL-NEXT:    v_readlane_b32 s61, v40, 29
-; GISEL-NEXT:    v_readlane_b32 s60, v40, 28
-; GISEL-NEXT:    v_readlane_b32 s59, v40, 27
-; GISEL-NEXT:    v_readlane_b32 s58, v40, 26
-; GISEL-NEXT:    v_readlane_b32 s57, v40, 25
-; GISEL-NEXT:    v_readlane_b32 s56, v40, 24
-; GISEL-NEXT:    v_readlane_b32 s55, v40, 23
-; GISEL-NEXT:    v_readlane_b32 s54, v40, 22
-; GISEL-NEXT:    v_readlane_b32 s53, v40, 21
-; GISEL-NEXT:    v_readlane_b32 s52, v40, 20
-; GISEL-NEXT:    v_readlane_b32 s51, v40, 19
-; GISEL-NEXT:    v_readlane_b32 s50, v40, 18
-; GISEL-NEXT:    v_readlane_b32 s49, v40, 17
-; GISEL-NEXT:    v_readlane_b32 s48, v40, 16
-; GISEL-NEXT:    v_readlane_b32 s47, v40, 15
-; GISEL-NEXT:    v_readlane_b32 s46, v40, 14
-; GISEL-NEXT:    v_readlane_b32 s45, v40, 13
-; GISEL-NEXT:    v_readlane_b32 s44, v40, 12
-; GISEL-NEXT:    v_readlane_b32 s43, v40, 11
-; GISEL-NEXT:    v_readlane_b32 s42, v40, 10
-; GISEL-NEXT:    v_readlane_b32 s41, v40, 9
-; GISEL-NEXT:    v_readlane_b32 s40, v40, 8
-; GISEL-NEXT:    v_readlane_b32 s39, v40, 7
-; GISEL-NEXT:    v_readlane_b32 s38, v40, 6
-; GISEL-NEXT:    v_readlane_b32 s37, v40, 5
-; GISEL-NEXT:    v_readlane_b32 s36, v40, 4
-; GISEL-NEXT:    v_readlane_b32 s35, v40, 3
-; GISEL-NEXT:    v_readlane_b32 s34, v40, 2
-; GISEL-NEXT:    v_readlane_b32 s31, v40, 1
-; GISEL-NEXT:    v_readlane_b32 s30, v40, 0
-; GISEL-NEXT:    buffer_load_dword v41, off, s[0:3], s33 ; 4-byte Folded Reload
+; GISEL-NEXT:    v_mov_b32_e32 v0, v40
+; GISEL-NEXT:    v_readlane_b32 s63, v41, 31
+; GISEL-NEXT:    v_readlane_b32 s62, v41, 30
+; GISEL-NEXT:    v_readlane_b32 s61, v41, 29
+; GISEL-NEXT:    v_readlane_b32 s60, v41, 28
+; GISEL-NEXT:    v_readlane_b32 s59, v41, 27
+; GISEL-NEXT:    v_readlane_b32 s58, v41, 26
+; GISEL-NEXT:    v_readlane_b32 s57, v41, 25
+; GISEL-NEXT:    v_readlane_b32 s56, v41, 24
+; GISEL-NEXT:    v_readlane_b32 s55, v41, 23
+; GISEL-NEXT:    v_readlane_b32 s54, v41, 22
+; GISEL-NEXT:    v_readlane_b32 s53, v41, 21
+; GISEL-NEXT:    v_readlane_b32 s52, v41, 20
+; GISEL-NEXT:    v_readlane_b32 s51, v41, 19
+; GISEL-NEXT:    v_readlane_b32 s50, v41, 18
+; GISEL-NEXT:    v_readlane_b32 s49, v41, 17
+; GISEL-NEXT:    v_readlane_b32 s48, v41, 16
+; GISEL-NEXT:    v_readlane_b32 s47, v41, 15
+; GISEL-NEXT:    v_readlane_b32 s46, v41, 14
+; GISEL-NEXT:    v_readlane_b32 s45, v41, 13
+; GISEL-NEXT:    v_readlane_b32 s44, v41, 12
+; GISEL-NEXT:    v_readlane_b32 s43, v41, 11
+; GISEL-NEXT:    v_readlane_b32 s42, v41, 10
+; GISEL-NEXT:    v_readlane_b32 s41, v41, 9
+; GISEL-NEXT:    v_readlane_b32 s40, v41, 8
+; GISEL-NEXT:    v_readlane_b32 s39, v41, 7
+; GISEL-NEXT:    v_readlane_b32 s38, v41, 6
+; GISEL-NEXT:    v_readlane_b32 s37, v41, 5
+; GISEL-NEXT:    v_readlane_b32 s36, v41, 4
+; GISEL-NEXT:    v_readlane_b32 s35, v41, 3
+; GISEL-NEXT:    v_readlane_b32 s34, v41, 2
+; GISEL-NEXT:    v_readlane_b32 s31, v41, 1
+; GISEL-NEXT:    v_readlane_b32 s30, v41, 0
+; GISEL-NEXT:    buffer_load_dword v40, off, s[0:3], s33 ; 4-byte Folded Reload
 ; GISEL-NEXT:    s_or_saveexec_b64 s[4:5], -1
-; GISEL-NEXT:    buffer_load_dword v40, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
+; GISEL-NEXT:    buffer_load_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
 ; GISEL-NEXT:    s_mov_b64 exec, s[4:5]
 ; GISEL-NEXT:    s_addk_i32 s32, 0xfc00
-; GISEL-NEXT:    s_mov_b32 s33, s10
+; GISEL-NEXT:    s_mov_b32 s33, s12
 ; GISEL-NEXT:    s_waitcnt vmcnt(0)
 ; GISEL-NEXT:    s_setpc_b64 s[30:31]
   call amdgpu_gfx void %fptr(i32 %i)
@@ -1531,12 +1543,13 @@
 ; GCN-LABEL: test_indirect_call_vgpr_ptr_arg_and_return:
 ; GCN:       ; %bb.0:
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GCN-NEXT:    s_mov_b32 s10, s33
+; GCN-NEXT:    s_mov_b32 s12, s33
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
+; GCN-NEXT:    ; implicit-def: $vgpr40
 ; GCN-NEXT:    v_writelane_b32 v40, s30, 0
 ; GCN-NEXT:    v_writelane_b32 v40, s31, 1
 ; GCN-NEXT:    v_writelane_b32 v40, s34, 2
@@ -1620,19 +1633,20 @@
 ; GCN-NEXT:    buffer_load_dword v40, off, s[0:3], s33 ; 4-byte Folded Reload
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_addk_i32 s32, 0xfc00
-; GCN-NEXT:    s_mov_b32 s33, s10
+; GCN-NEXT:    s_mov_b32 s33, s12
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
 ; GCN-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GISEL-LABEL: test_indirect_call_vgpr_ptr_arg_and_return:
 ; GISEL:       ; %bb.0:
 ; GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GISEL-NEXT:    s_mov_b32 s10, s33
+; GISEL-NEXT:    s_mov_b32 s12, s33
 ; GISEL-NEXT:    s_mov_b32 s33, s32
 ; GISEL-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GISEL-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GISEL-NEXT:    s_mov_b64 exec, s[4:5]
 ; GISEL-NEXT:    s_addk_i32 s32, 0x400
+; GISEL-NEXT:    ; implicit-def: $vgpr40
 ; GISEL-NEXT:    v_writelane_b32 v40, s30, 0
 ; GISEL-NEXT:    v_writelane_b32 v40, s31, 1
 ; GISEL-NEXT:    v_writelane_b32 v40, s34, 2
@@ -1716,7 +1730,7 @@
 ; GISEL-NEXT:    buffer_load_dword v40, off, s[0:3], s33 ; 4-byte Folded Reload
 ; GISEL-NEXT:    s_mov_b64 exec, s[4:5]
 ; GISEL-NEXT:    s_addk_i32 s32, 0xfc00
-; GISEL-NEXT:    s_mov_b32 s33, s10
+; GISEL-NEXT:    s_mov_b32 s33, s12
 ; GISEL-NEXT:    s_waitcnt vmcnt(0)
 ; GISEL-NEXT:    s_setpc_b64 s[30:31]
   %ret = call amdgpu_gfx i32 %fptr(i32 %i)
@@ -1728,12 +1742,13 @@
 ; GCN-LABEL: test_indirect_tail_call_vgpr_ptr:
 ; GCN:       ; %bb.0:
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GCN-NEXT:    s_mov_b32 s10, s33
+; GCN-NEXT:    s_mov_b32 s12, s33
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
+; GCN-NEXT:    ; implicit-def: $vgpr40
 ; GCN-NEXT:    v_writelane_b32 v40, s30, 0
 ; GCN-NEXT:    v_writelane_b32 v40, s31, 1
 ; GCN-NEXT:    v_writelane_b32 v40, s34, 2
@@ -1814,19 +1829,20 @@
 ; GCN-NEXT:    buffer_load_dword v40, off, s[0:3], s33 ; 4-byte Folded Reload
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_addk_i32 s32, 0xfc00
-; GCN-NEXT:    s_mov_b32 s33, s10
+; GCN-NEXT:    s_mov_b32 s33, s12
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
 ; GCN-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GISEL-LABEL: test_indirect_tail_call_vgpr_ptr:
 ; GISEL:       ; %bb.0:
 ; GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GISEL-NEXT:    s_mov_b32 s10, s33
+; GISEL-NEXT:    s_mov_b32 s12, s33
 ; GISEL-NEXT:    s_mov_b32 s33, s32
 ; GISEL-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GISEL-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GISEL-NEXT:    s_mov_b64 exec, s[4:5]
 ; GISEL-NEXT:    s_addk_i32 s32, 0x400
+; GISEL-NEXT:    ; implicit-def: $vgpr40
 ; GISEL-NEXT:    v_writelane_b32 v40, s30, 0
 ; GISEL-NEXT:    v_writelane_b32 v40, s31, 1
 ; GISEL-NEXT:    v_writelane_b32 v40, s34, 2
@@ -1907,7 +1923,7 @@
 ; GISEL-NEXT:    buffer_load_dword v40, off, s[0:3], s33 ; 4-byte Folded Reload
 ; GISEL-NEXT:    s_mov_b64 exec, s[4:5]
 ; GISEL-NEXT:    s_addk_i32 s32, 0xfc00
-; GISEL-NEXT:    s_mov_b32 s33, s10
+; GISEL-NEXT:    s_mov_b32 s33, s12
 ; GISEL-NEXT:    s_waitcnt vmcnt(0)
 ; GISEL-NEXT:    s_setpc_b64 s[30:31]
   tail call amdgpu_gfx void %fptr()
diff --git a/llvm/test/CodeGen/AMDGPU/kernel-vgpr-spill-mubuf-with-voffset.ll b/llvm/test/CodeGen/AMDGPU/kernel-vgpr-spill-mubuf-with-voffset.ll
--- a/llvm/test/CodeGen/AMDGPU/kernel-vgpr-spill-mubuf-with-voffset.ll
+++ b/llvm/test/CodeGen/AMDGPU/kernel-vgpr-spill-mubuf-with-voffset.ll
@@ -13,17 +13,30 @@
 ; CHECK-NEXT:    s_addc_u32 flat_scratch_hi, s13, 0
 ; CHECK-NEXT:    s_add_u32 s0, s0, s17
 ; CHECK-NEXT:    s_addc_u32 s1, s1, 0
-; CHECK-NEXT:    v_writelane_b32 v40, s16, 0
+; CHECK-NEXT:    ; implicit-def: $vgpr3
+; CHECK-NEXT:    v_writelane_b32 v3, s16, 0
+; CHECK-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; CHECK-NEXT:    s_add_i32 s12, s33, 0x100200
+; CHECK-NEXT:    buffer_store_dword v3, off, s[0:3], s12 ; 4-byte Folded Spill
+; CHECK-NEXT:    s_mov_b64 exec, s[34:35]
 ; CHECK-NEXT:    s_mov_b32 s13, s15
 ; CHECK-NEXT:    s_mov_b32 s12, s14
-; CHECK-NEXT:    v_readlane_b32 s14, v40, 0
+; CHECK-NEXT:    v_readlane_b32 s14, v3, 0
 ; CHECK-NEXT:    s_mov_b64 s[16:17], s[8:9]
 ; CHECK-NEXT:    v_mov_b32_e32 v3, v2
 ; CHECK-NEXT:    v_mov_b32_e32 v2, v1
 ; CHECK-NEXT:    v_mov_b32_e32 v1, v0
+; CHECK-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; CHECK-NEXT:    s_add_i32 s8, s33, 0x100200
+; CHECK-NEXT:    buffer_load_dword v0, off, s[0:3], s8 ; 4-byte Folded Reload
+; CHECK-NEXT:    s_mov_b64 exec, s[34:35]
 ; CHECK-NEXT:    s_load_dword s8, s[16:17], 0x0
-; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
-; CHECK-NEXT:    v_writelane_b32 v40, s8, 1
+; CHECK-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
+; CHECK-NEXT:    v_writelane_b32 v0, s8, 1
+; CHECK-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; CHECK-NEXT:    s_add_i32 s8, s33, 0x100200
+; CHECK-NEXT:    buffer_store_dword v0, off, s[0:3], s8 ; 4-byte Folded Spill
+; CHECK-NEXT:    s_mov_b64 exec, s[34:35]
 ; CHECK-NEXT:    ;;#ASMSTART
 ; CHECK-NEXT:    ; def vgpr10
 ; CHECK-NEXT:    ;;#ASMEND
@@ -56,9 +69,14 @@
 ; CHECK-NEXT:    s_mov_b64 s[2:3], s[22:23]
 ; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
 ; CHECK-NEXT:    s_swappc_b64 s[30:31], s[16:17]
+; CHECK-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; CHECK-NEXT:    s_add_i32 s4, s33, 0x100200
+; CHECK-NEXT:    buffer_load_dword v0, off, s[0:3], s4 ; 4-byte Folded Reload
+; CHECK-NEXT:    s_mov_b64 exec, s[34:35]
 ; CHECK-NEXT:    s_add_i32 s4, s33, 0x100100
 ; CHECK-NEXT:    buffer_load_dword v10, off, s[0:3], s4 ; 4-byte Folded Reload
-; CHECK-NEXT:    v_readlane_b32 s4, v40, 1
+; CHECK-NEXT:    s_waitcnt vmcnt(1)
+; CHECK-NEXT:    v_readlane_b32 s4, v0, 1
 ; CHECK-NEXT:    s_mov_b32 s5, 0
 ; CHECK-NEXT:    s_cmp_eq_u32 s4, s5
 ; CHECK-NEXT:    v_mov_b32_e32 v0, 0x4000
diff --git a/llvm/test/CodeGen/AMDGPU/load-constant-i16.ll b/llvm/test/CodeGen/AMDGPU/load-constant-i16.ll
--- a/llvm/test/CodeGen/AMDGPU/load-constant-i16.ll
+++ b/llvm/test/CodeGen/AMDGPU/load-constant-i16.ll
@@ -3037,17 +3037,17 @@
 ; GCN-HSA-NEXT:    s_lshr_b32 s35, s8, 16
 ; GCN-HSA-NEXT:    s_lshr_b32 s37, s11, 16
 ; GCN-HSA-NEXT:    s_lshr_b32 s38, s10, 16
-; GCN-HSA-NEXT:    s_lshr_b32 s39, s13, 16
-; GCN-HSA-NEXT:    s_lshr_b32 s40, s12, 16
-; GCN-HSA-NEXT:    s_lshr_b32 s41, s15, 16
-; GCN-HSA-NEXT:    s_lshr_b32 s42, s14, 16
+; GCN-HSA-NEXT:    s_lshr_b32 s40, s13, 16
+; GCN-HSA-NEXT:    s_lshr_b32 s41, s12, 16
+; GCN-HSA-NEXT:    s_lshr_b32 s42, s15, 16
+; GCN-HSA-NEXT:    s_lshr_b32 s43, s14, 16
 ; GCN-HSA-NEXT:    s_and_b32 s25, s1, 0xffff
 ; GCN-HSA-NEXT:    s_and_b32 s27, s0, 0xffff
 ; GCN-HSA-NEXT:    s_and_b32 s29, s3, 0xffff
 ; GCN-HSA-NEXT:    s_and_b32 s31, s2, 0xffff
 ; GCN-HSA-NEXT:    s_and_b32 s34, s5, 0xffff
 ; GCN-HSA-NEXT:    s_and_b32 s36, s4, 0xffff
-; GCN-HSA-NEXT:    s_and_b32 s43, s7, 0xffff
+; GCN-HSA-NEXT:    s_and_b32 s39, s7, 0xffff
 ; GCN-HSA-NEXT:    s_and_b32 s44, s6, 0xffff
 ; GCN-HSA-NEXT:    s_and_b32 s45, s9, 0xffff
 ; GCN-HSA-NEXT:    s_and_b32 s46, s8, 0xffff
@@ -3172,13 +3172,13 @@
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v8, s52
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v7, s18
 ; GCN-HSA-NEXT:    flat_store_dwordx4 v[9:10], v[0:3]
-; GCN-HSA-NEXT:    v_mov_b32_e32 v9, s42
+; GCN-HSA-NEXT:    v_mov_b32_e32 v9, s43
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v0, s50
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v10, s51
-; GCN-HSA-NEXT:    v_mov_b32_e32 v11, s41
-; GCN-HSA-NEXT:    v_mov_b32_e32 v1, s40
+; GCN-HSA-NEXT:    v_mov_b32_e32 v11, s42
+; GCN-HSA-NEXT:    v_mov_b32_e32 v1, s41
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v2, s49
-; GCN-HSA-NEXT:    v_mov_b32_e32 v3, s39
+; GCN-HSA-NEXT:    v_mov_b32_e32 v3, s40
 ; GCN-HSA-NEXT:    flat_store_dwordx4 v[12:13], v[20:23]
 ; GCN-HSA-NEXT:    flat_store_dwordx4 v[14:15], v[4:7]
 ; GCN-HSA-NEXT:    flat_store_dwordx4 v[16:17], v[8:11]
@@ -3206,7 +3206,7 @@
 ; GCN-HSA-NEXT:    s_add_u32 s0, s16, 32
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v0, s44
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v1, s30
-; GCN-HSA-NEXT:    v_mov_b32_e32 v2, s43
+; GCN-HSA-NEXT:    v_mov_b32_e32 v2, s39
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v3, s28
 ; GCN-HSA-NEXT:    s_addc_u32 s1, s17, 0
 ; GCN-HSA-NEXT:    flat_store_dwordx4 v[4:5], v[0:3]
@@ -6181,129 +6181,129 @@
 ; GCN-HSA-NEXT:    s_lshr_b32 s33, s4, 16
 ; GCN-HSA-NEXT:    s_lshr_b32 s34, s2, 16
 ; GCN-HSA-NEXT:    s_lshr_b32 s18, s0, 16
-; GCN-HSA-NEXT:    s_and_b32 s35, s0, 0xffff
-; GCN-HSA-NEXT:    s_and_b32 s2, s2, 0xffff
+; GCN-HSA-NEXT:    s_and_b32 s0, s0, 0xffff
+; GCN-HSA-NEXT:    s_and_b32 s35, s2, 0xffff
 ; GCN-HSA-NEXT:    s_and_b32 s4, s4, 0xffff
 ; GCN-HSA-NEXT:    s_and_b32 s6, s6, 0xffff
 ; GCN-HSA-NEXT:    s_and_b32 s8, s8, 0xffff
 ; GCN-HSA-NEXT:    s_and_b32 s10, s10, 0xffff
 ; GCN-HSA-NEXT:    s_and_b32 s12, s12, 0xffff
 ; GCN-HSA-NEXT:    s_and_b32 s14, s14, 0xffff
-; GCN-HSA-NEXT:    s_and_b32 s36, s1, 0xffff
-; GCN-HSA-NEXT:    s_and_b32 s3, s3, 0xffff
+; GCN-HSA-NEXT:    s_and_b32 s1, s1, 0xffff
+; GCN-HSA-NEXT:    s_and_b32 s36, s3, 0xffff
 ; GCN-HSA-NEXT:    s_and_b32 s5, s5, 0xffff
 ; GCN-HSA-NEXT:    s_and_b32 s7, s7, 0xffff
 ; GCN-HSA-NEXT:    s_and_b32 s9, s9, 0xffff
 ; GCN-HSA-NEXT:    s_and_b32 s11, s11, 0xffff
 ; GCN-HSA-NEXT:    s_and_b32 s13, s13, 0xffff
 ; GCN-HSA-NEXT:    s_and_b32 s15, s15, 0xffff
-; GCN-HSA-NEXT:    s_add_u32 s0, s16, 0xf0
-; GCN-HSA-NEXT:    s_addc_u32 s1, s17, 0
-; GCN-HSA-NEXT:    v_mov_b32_e32 v5, s1
-; GCN-HSA-NEXT:    v_mov_b32_e32 v4, s0
-; GCN-HSA-NEXT:    s_add_u32 s0, s16, 0xd0
-; GCN-HSA-NEXT:    s_addc_u32 s1, s17, 0
-; GCN-HSA-NEXT:    v_mov_b32_e32 v7, s1
-; GCN-HSA-NEXT:    v_mov_b32_e32 v6, s0
-; GCN-HSA-NEXT:    s_add_u32 s0, s16, 0xb0
-; GCN-HSA-NEXT:    s_addc_u32 s1, s17, 0
-; GCN-HSA-NEXT:    v_mov_b32_e32 v9, s1
-; GCN-HSA-NEXT:    v_mov_b32_e32 v8, s0
-; GCN-HSA-NEXT:    s_add_u32 s0, s16, 0x90
-; GCN-HSA-NEXT:    s_addc_u32 s1, s17, 0
-; GCN-HSA-NEXT:    v_mov_b32_e32 v11, s1
-; GCN-HSA-NEXT:    v_mov_b32_e32 v10, s0
+; GCN-HSA-NEXT:    s_add_u32 s2, s16, 0xf0
+; GCN-HSA-NEXT:    s_addc_u32 s3, s17, 0
+; GCN-HSA-NEXT:    v_mov_b32_e32 v5, s3
+; GCN-HSA-NEXT:    v_mov_b32_e32 v4, s2
+; GCN-HSA-NEXT:    s_add_u32 s2, s16, 0xd0
+; GCN-HSA-NEXT:    s_addc_u32 s3, s17, 0
+; GCN-HSA-NEXT:    v_mov_b32_e32 v7, s3
+; GCN-HSA-NEXT:    v_mov_b32_e32 v6, s2
+; GCN-HSA-NEXT:    s_add_u32 s2, s16, 0xb0
+; GCN-HSA-NEXT:    s_addc_u32 s3, s17, 0
+; GCN-HSA-NEXT:    v_mov_b32_e32 v9, s3
+; GCN-HSA-NEXT:    v_mov_b32_e32 v8, s2
+; GCN-HSA-NEXT:    s_add_u32 s2, s16, 0x90
+; GCN-HSA-NEXT:    s_addc_u32 s3, s17, 0
+; GCN-HSA-NEXT:    v_mov_b32_e32 v11, s3
+; GCN-HSA-NEXT:    v_mov_b32_e32 v10, s2
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v0, s15
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v2, s26
-; GCN-HSA-NEXT:    s_add_u32 s0, s16, 0x70
+; GCN-HSA-NEXT:    s_add_u32 s2, s16, 0x70
 ; GCN-HSA-NEXT:    flat_store_dwordx4 v[4:5], v[0:3]
-; GCN-HSA-NEXT:    s_addc_u32 s1, s17, 0
+; GCN-HSA-NEXT:    s_addc_u32 s3, s17, 0
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v0, s13
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v2, s25
 ; GCN-HSA-NEXT:    flat_store_dwordx4 v[6:7], v[0:3]
-; GCN-HSA-NEXT:    v_mov_b32_e32 v5, s1
+; GCN-HSA-NEXT:    v_mov_b32_e32 v5, s3
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v0, s11
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v2, s24
 ; GCN-HSA-NEXT:    flat_store_dwordx4 v[8:9], v[0:3]
-; GCN-HSA-NEXT:    v_mov_b32_e32 v4, s0
+; GCN-HSA-NEXT:    v_mov_b32_e32 v4, s2
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v0, s9
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v2, s23
-; GCN-HSA-NEXT:    s_add_u32 s0, s16, 0x50
+; GCN-HSA-NEXT:    s_add_u32 s2, s16, 0x50
 ; GCN-HSA-NEXT:    flat_store_dwordx4 v[10:11], v[0:3]
-; GCN-HSA-NEXT:    s_addc_u32 s1, s17, 0
+; GCN-HSA-NEXT:    s_addc_u32 s3, s17, 0
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v0, s7
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v2, s22
 ; GCN-HSA-NEXT:    flat_store_dwordx4 v[4:5], v[0:3]
-; GCN-HSA-NEXT:    v_mov_b32_e32 v5, s1
-; GCN-HSA-NEXT:    v_mov_b32_e32 v4, s0
-; GCN-HSA-NEXT:    s_add_u32 s0, s16, 48
+; GCN-HSA-NEXT:    v_mov_b32_e32 v5, s3
+; GCN-HSA-NEXT:    v_mov_b32_e32 v4, s2
+; GCN-HSA-NEXT:    s_add_u32 s2, s16, 48
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v0, s5
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v2, s21
-; GCN-HSA-NEXT:    s_addc_u32 s1, s17, 0
+; GCN-HSA-NEXT:    s_addc_u32 s3, s17, 0
 ; GCN-HSA-NEXT:    flat_store_dwordx4 v[4:5], v[0:3]
-; GCN-HSA-NEXT:    v_mov_b32_e32 v5, s1
-; GCN-HSA-NEXT:    v_mov_b32_e32 v4, s0
-; GCN-HSA-NEXT:    s_add_u32 s0, s16, 16
-; GCN-HSA-NEXT:    v_mov_b32_e32 v0, s3
+; GCN-HSA-NEXT:    v_mov_b32_e32 v5, s3
+; GCN-HSA-NEXT:    v_mov_b32_e32 v4, s2
+; GCN-HSA-NEXT:    s_add_u32 s2, s16, 16
+; GCN-HSA-NEXT:    v_mov_b32_e32 v0, s36
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v2, s20
-; GCN-HSA-NEXT:    s_addc_u32 s1, s17, 0
+; GCN-HSA-NEXT:    s_addc_u32 s3, s17, 0
 ; GCN-HSA-NEXT:    flat_store_dwordx4 v[4:5], v[0:3]
-; GCN-HSA-NEXT:    v_mov_b32_e32 v5, s1
-; GCN-HSA-NEXT:    v_mov_b32_e32 v4, s0
-; GCN-HSA-NEXT:    s_add_u32 s0, s16, 0xe0
-; GCN-HSA-NEXT:    v_mov_b32_e32 v0, s36
+; GCN-HSA-NEXT:    v_mov_b32_e32 v5, s3
+; GCN-HSA-NEXT:    v_mov_b32_e32 v4, s2
+; GCN-HSA-NEXT:    s_add_u32 s2, s16, 0xe0
+; GCN-HSA-NEXT:    v_mov_b32_e32 v0, s1
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v2, s19
-; GCN-HSA-NEXT:    s_addc_u32 s1, s17, 0
+; GCN-HSA-NEXT:    s_addc_u32 s3, s17, 0
 ; GCN-HSA-NEXT:    flat_store_dwordx4 v[4:5], v[0:3]
-; GCN-HSA-NEXT:    v_mov_b32_e32 v5, s1
-; GCN-HSA-NEXT:    v_mov_b32_e32 v4, s0
-; GCN-HSA-NEXT:    s_add_u32 s0, s16, 0xc0
+; GCN-HSA-NEXT:    v_mov_b32_e32 v5, s3
+; GCN-HSA-NEXT:    v_mov_b32_e32 v4, s2
+; GCN-HSA-NEXT:    s_add_u32 s2, s16, 0xc0
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v0, s14
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v2, s27
-; GCN-HSA-NEXT:    s_addc_u32 s1, s17, 0
+; GCN-HSA-NEXT:    s_addc_u32 s3, s17, 0
 ; GCN-HSA-NEXT:    flat_store_dwordx4 v[4:5], v[0:3]
-; GCN-HSA-NEXT:    v_mov_b32_e32 v5, s1
-; GCN-HSA-NEXT:    v_mov_b32_e32 v4, s0
-; GCN-HSA-NEXT:    s_add_u32 s0, s16, 0xa0
+; GCN-HSA-NEXT:    v_mov_b32_e32 v5, s3
+; GCN-HSA-NEXT:    v_mov_b32_e32 v4, s2
+; GCN-HSA-NEXT:    s_add_u32 s2, s16, 0xa0
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v0, s12
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v2, s28
-; GCN-HSA-NEXT:    s_addc_u32 s1, s17, 0
+; GCN-HSA-NEXT:    s_addc_u32 s3, s17, 0
 ; GCN-HSA-NEXT:    flat_store_dwordx4 v[4:5], v[0:3]
-; GCN-HSA-NEXT:    v_mov_b32_e32 v5, s1
-; GCN-HSA-NEXT:    v_mov_b32_e32 v4, s0
-; GCN-HSA-NEXT:    s_add_u32 s0, s16, 0x80
+; GCN-HSA-NEXT:    v_mov_b32_e32 v5, s3
+; GCN-HSA-NEXT:    v_mov_b32_e32 v4, s2
+; GCN-HSA-NEXT:    s_add_u32 s2, s16, 0x80
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v0, s10
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v2, s29
-; GCN-HSA-NEXT:    s_addc_u32 s1, s17, 0
+; GCN-HSA-NEXT:    s_addc_u32 s3, s17, 0
 ; GCN-HSA-NEXT:    flat_store_dwordx4 v[4:5], v[0:3]
-; GCN-HSA-NEXT:    v_mov_b32_e32 v5, s1
-; GCN-HSA-NEXT:    v_mov_b32_e32 v4, s0
-; GCN-HSA-NEXT:    s_add_u32 s0, s16, 0x60
+; GCN-HSA-NEXT:    v_mov_b32_e32 v5, s3
+; GCN-HSA-NEXT:    v_mov_b32_e32 v4, s2
+; GCN-HSA-NEXT:    s_add_u32 s2, s16, 0x60
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v0, s8
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v2, s30
-; GCN-HSA-NEXT:    s_addc_u32 s1, s17, 0
+; GCN-HSA-NEXT:    s_addc_u32 s3, s17, 0
 ; GCN-HSA-NEXT:    flat_store_dwordx4 v[4:5], v[0:3]
-; GCN-HSA-NEXT:    v_mov_b32_e32 v5, s1
-; GCN-HSA-NEXT:    v_mov_b32_e32 v4, s0
-; GCN-HSA-NEXT:    s_add_u32 s0, s16, 64
+; GCN-HSA-NEXT:    v_mov_b32_e32 v5, s3
+; GCN-HSA-NEXT:    v_mov_b32_e32 v4, s2
+; GCN-HSA-NEXT:    s_add_u32 s2, s16, 64
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v0, s6
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v2, s31
-; GCN-HSA-NEXT:    s_addc_u32 s1, s17, 0
+; GCN-HSA-NEXT:    s_addc_u32 s3, s17, 0
 ; GCN-HSA-NEXT:    flat_store_dwordx4 v[4:5], v[0:3]
-; GCN-HSA-NEXT:    v_mov_b32_e32 v5, s1
-; GCN-HSA-NEXT:    v_mov_b32_e32 v4, s0
-; GCN-HSA-NEXT:    s_add_u32 s0, s16, 32
+; GCN-HSA-NEXT:    v_mov_b32_e32 v5, s3
+; GCN-HSA-NEXT:    v_mov_b32_e32 v4, s2
+; GCN-HSA-NEXT:    s_add_u32 s2, s16, 32
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v0, s4
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v2, s33
-; GCN-HSA-NEXT:    s_addc_u32 s1, s17, 0
+; GCN-HSA-NEXT:    s_addc_u32 s3, s17, 0
 ; GCN-HSA-NEXT:    flat_store_dwordx4 v[4:5], v[0:3]
-; GCN-HSA-NEXT:    v_mov_b32_e32 v5, s1
-; GCN-HSA-NEXT:    v_mov_b32_e32 v0, s2
+; GCN-HSA-NEXT:    v_mov_b32_e32 v5, s3
+; GCN-HSA-NEXT:    v_mov_b32_e32 v0, s35
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v2, s34
-; GCN-HSA-NEXT:    v_mov_b32_e32 v4, s0
+; GCN-HSA-NEXT:    v_mov_b32_e32 v4, s2
 ; GCN-HSA-NEXT:    flat_store_dwordx4 v[4:5], v[0:3]
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v4, s16
-; GCN-HSA-NEXT:    v_mov_b32_e32 v0, s35
+; GCN-HSA-NEXT:    v_mov_b32_e32 v0, s0
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v2, s18
 ; GCN-HSA-NEXT:    v_mov_b32_e32 v5, s17
 ; GCN-HSA-NEXT:    flat_store_dwordx4 v[4:5], v[0:3]
@@ -6589,17 +6589,17 @@
 ; GCN-NOHSA-SI-NEXT:    s_lshr_b32 s26, s12, 16
 ; GCN-NOHSA-SI-NEXT:    s_lshr_b32 s28, s10, 16
 ; GCN-NOHSA-SI-NEXT:    s_lshr_b32 s30, s8, 16
-; GCN-NOHSA-SI-NEXT:    s_bfe_i64 s[48:49], s[20:21], 0x100000
-; GCN-NOHSA-SI-NEXT:    s_bfe_i64 s[50:51], s[18:19], 0x100000
-; GCN-NOHSA-SI-NEXT:    s_lshr_b32 s52, s6, 16
-; GCN-NOHSA-SI-NEXT:    s_lshr_b32 s54, s4, 16
-; GCN-NOHSA-SI-NEXT:    s_lshr_b32 s56, s2, 16
-; GCN-NOHSA-SI-NEXT:    s_lshr_b32 s58, s0, 16
+; GCN-NOHSA-SI-NEXT:    s_bfe_i64 s[50:51], s[20:21], 0x100000
+; GCN-NOHSA-SI-NEXT:    s_bfe_i64 s[52:53], s[18:19], 0x100000
+; GCN-NOHSA-SI-NEXT:    s_lshr_b32 s54, s6, 16
+; GCN-NOHSA-SI-NEXT:    s_lshr_b32 s56, s4, 16
+; GCN-NOHSA-SI-NEXT:    s_lshr_b32 s58, s2, 16
+; GCN-NOHSA-SI-NEXT:    s_lshr_b32 s60, s0, 16
 ; GCN-NOHSA-SI-NEXT:    s_bfe_i64 s[18:19], s[0:1], 0x100000
 ; GCN-NOHSA-SI-NEXT:    s_bfe_i64 s[20:21], s[2:3], 0x100000
 ; GCN-NOHSA-SI-NEXT:    s_bfe_i64 s[24:25], s[4:5], 0x100000
 ; GCN-NOHSA-SI-NEXT:    s_bfe_i64 s[34:35], s[6:7], 0x100000
-; GCN-NOHSA-SI-NEXT:    s_bfe_i64 s[60:61], s[8:9], 0x100000
+; GCN-NOHSA-SI-NEXT:    s_bfe_i64 s[48:49], s[8:9], 0x100000
 ; GCN-NOHSA-SI-NEXT:    s_bfe_i64 s[62:63], s[10:11], 0x100000
 ; GCN-NOHSA-SI-NEXT:    s_bfe_i64 s[64:65], s[12:13], 0x100000
 ; GCN-NOHSA-SI-NEXT:    s_bfe_i64 s[66:67], s[14:15], 0x100000
@@ -6613,12 +6613,12 @@
 ; GCN-NOHSA-SI-NEXT:    s_ashr_i64 s[4:5], s[4:5], 48
 ; GCN-NOHSA-SI-NEXT:    s_mov_b32 s0, s16
 ; GCN-NOHSA-SI-NEXT:    s_mov_b32 s1, s17
-; GCN-NOHSA-SI-NEXT:    v_mov_b32_e32 v0, s50
-; GCN-NOHSA-SI-NEXT:    v_mov_b32_e32 v1, s51
+; GCN-NOHSA-SI-NEXT:    v_mov_b32_e32 v0, s52
+; GCN-NOHSA-SI-NEXT:    v_mov_b32_e32 v1, s53
 ; GCN-NOHSA-SI-NEXT:    v_mov_b32_e32 v2, s12
 ; GCN-NOHSA-SI-NEXT:    v_mov_b32_e32 v3, s13
-; GCN-NOHSA-SI-NEXT:    v_mov_b32_e32 v4, s48
-; GCN-NOHSA-SI-NEXT:    v_mov_b32_e32 v5, s49
+; GCN-NOHSA-SI-NEXT:    v_mov_b32_e32 v4, s50
+; GCN-NOHSA-SI-NEXT:    v_mov_b32_e32 v5, s51
 ; GCN-NOHSA-SI-NEXT:    v_mov_b32_e32 v6, s2
 ; GCN-NOHSA-SI-NEXT:    v_mov_b32_e32 v7, s3
 ; GCN-NOHSA-SI-NEXT:    s_mov_b32 s3, 0xf000
@@ -6647,10 +6647,10 @@
 ; GCN-NOHSA-SI-NEXT:    v_mov_b32_e32 v22, s4
 ; GCN-NOHSA-SI-NEXT:    v_mov_b32_e32 v23, s5
 ; GCN-NOHSA-SI-NEXT:    buffer_store_dwordx4 v[4:7], off, s[0:3], 0 offset:208
-; GCN-NOHSA-SI-NEXT:    s_bfe_i64 s[4:5], s[58:59], 0x100000
-; GCN-NOHSA-SI-NEXT:    s_bfe_i64 s[6:7], s[56:57], 0x100000
-; GCN-NOHSA-SI-NEXT:    s_bfe_i64 s[8:9], s[54:55], 0x100000
-; GCN-NOHSA-SI-NEXT:    s_bfe_i64 s[10:11], s[52:53], 0x100000
+; GCN-NOHSA-SI-NEXT:    s_bfe_i64 s[4:5], s[60:61], 0x100000
+; GCN-NOHSA-SI-NEXT:    s_bfe_i64 s[6:7], s[58:59], 0x100000
+; GCN-NOHSA-SI-NEXT:    s_bfe_i64 s[8:9], s[56:57], 0x100000
+; GCN-NOHSA-SI-NEXT:    s_bfe_i64 s[10:11], s[54:55], 0x100000
 ; GCN-NOHSA-SI-NEXT:    s_bfe_i64 s[12:13], s[30:31], 0x100000
 ; GCN-NOHSA-SI-NEXT:    s_bfe_i64 s[14:15], s[28:29], 0x100000
 ; GCN-NOHSA-SI-NEXT:    s_bfe_i64 s[16:17], s[26:27], 0x100000
@@ -6678,8 +6678,8 @@
 ; GCN-NOHSA-SI-NEXT:    v_mov_b32_e32 v5, s65
 ; GCN-NOHSA-SI-NEXT:    v_mov_b32_e32 v8, s62
 ; GCN-NOHSA-SI-NEXT:    v_mov_b32_e32 v9, s63
-; GCN-NOHSA-SI-NEXT:    v_mov_b32_e32 v12, s60
-; GCN-NOHSA-SI-NEXT:    v_mov_b32_e32 v13, s61
+; GCN-NOHSA-SI-NEXT:    v_mov_b32_e32 v12, s48
+; GCN-NOHSA-SI-NEXT:    v_mov_b32_e32 v13, s49
 ; GCN-NOHSA-SI-NEXT:    v_mov_b32_e32 v16, s34
 ; GCN-NOHSA-SI-NEXT:    v_mov_b32_e32 v17, s35
 ; GCN-NOHSA-SI-NEXT:    v_mov_b32_e32 v20, s24
@@ -6923,123 +6923,123 @@
 ; GCN-NOHSA-VI-NEXT:    s_waitcnt lgkmcnt(0)
 ; GCN-NOHSA-VI-NEXT:    s_load_dwordx16 s[0:15], s[18:19], 0x0
 ; GCN-NOHSA-VI-NEXT:    s_waitcnt lgkmcnt(0)
-; GCN-NOHSA-VI-NEXT:    s_mov_b32 s36, s15
-; GCN-NOHSA-VI-NEXT:    s_mov_b32 s38, s13
+; GCN-NOHSA-VI-NEXT:    s_mov_b32 s38, s15
+; GCN-NOHSA-VI-NEXT:    s_mov_b32 s40, s13
 ; GCN-NOHSA-VI-NEXT:    s_ashr_i64 s[82:83], s[14:15], 48
-; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[36:37], s[36:37], 0x100000
-; GCN-NOHSA-VI-NEXT:    s_mov_b32 s40, s11
-; GCN-NOHSA-VI-NEXT:    s_mov_b32 s48, s3
-; GCN-NOHSA-VI-NEXT:    s_mov_b32 s50, s1
-; GCN-NOHSA-VI-NEXT:    s_lshr_b32 s64, s2, 16
-; GCN-NOHSA-VI-NEXT:    s_lshr_b32 s66, s0, 16
+; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[38:39], s[38:39], 0x100000
+; GCN-NOHSA-VI-NEXT:    s_mov_b32 s42, s11
+; GCN-NOHSA-VI-NEXT:    s_mov_b32 s50, s3
+; GCN-NOHSA-VI-NEXT:    s_mov_b32 s52, s1
+; GCN-NOHSA-VI-NEXT:    s_lshr_b32 s66, s2, 16
+; GCN-NOHSA-VI-NEXT:    s_lshr_b32 s68, s0, 16
 ; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[18:19], s[0:1], 0x100000
 ; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[20:21], s[2:3], 0x100000
-; GCN-NOHSA-VI-NEXT:    s_ashr_i64 s[68:69], s[0:1], 48
+; GCN-NOHSA-VI-NEXT:    s_ashr_i64 s[36:37], s[0:1], 48
 ; GCN-NOHSA-VI-NEXT:    s_ashr_i64 s[70:71], s[2:3], 48
 ; GCN-NOHSA-VI-NEXT:    s_ashr_i64 s[80:81], s[12:13], 48
 ; GCN-NOHSA-VI-NEXT:    s_mov_b32 s3, 0xf000
 ; GCN-NOHSA-VI-NEXT:    s_mov_b32 s2, -1
 ; GCN-NOHSA-VI-NEXT:    s_mov_b32 s0, s16
 ; GCN-NOHSA-VI-NEXT:    s_mov_b32 s1, s17
-; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[38:39], s[38:39], 0x100000
-; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v0, s36
-; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v1, s37
+; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[40:41], s[40:41], 0x100000
+; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v0, s38
+; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v1, s39
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v2, s82
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v3, s83
-; GCN-NOHSA-VI-NEXT:    s_mov_b32 s42, s9
+; GCN-NOHSA-VI-NEXT:    s_mov_b32 s44, s9
 ; GCN-NOHSA-VI-NEXT:    s_ashr_i64 s[78:79], s[10:11], 48
-; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[40:41], s[40:41], 0x100000
+; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[42:43], s[42:43], 0x100000
 ; GCN-NOHSA-VI-NEXT:    buffer_store_dwordx4 v[0:3], off, s[0:3], 0 offset:240
-; GCN-NOHSA-VI-NEXT:    s_mov_b32 s44, s7
-; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v0, s38
-; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v1, s39
+; GCN-NOHSA-VI-NEXT:    s_mov_b32 s46, s7
+; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v0, s40
+; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v1, s41
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v2, s80
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v3, s81
 ; GCN-NOHSA-VI-NEXT:    s_ashr_i64 s[76:77], s[8:9], 48
-; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[42:43], s[42:43], 0x100000
+; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[44:45], s[44:45], 0x100000
 ; GCN-NOHSA-VI-NEXT:    buffer_store_dwordx4 v[0:3], off, s[0:3], 0 offset:208
-; GCN-NOHSA-VI-NEXT:    s_mov_b32 s46, s5
-; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v0, s40
-; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v1, s41
+; GCN-NOHSA-VI-NEXT:    s_mov_b32 s48, s5
+; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v0, s42
+; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v1, s43
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v2, s78
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v3, s79
 ; GCN-NOHSA-VI-NEXT:    s_ashr_i64 s[74:75], s[6:7], 48
-; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[44:45], s[44:45], 0x100000
+; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[46:47], s[46:47], 0x100000
 ; GCN-NOHSA-VI-NEXT:    buffer_store_dwordx4 v[0:3], off, s[0:3], 0 offset:176
 ; GCN-NOHSA-VI-NEXT:    s_ashr_i64 s[72:73], s[4:5], 48
-; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v0, s42
-; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v1, s43
+; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v0, s44
+; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v1, s45
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v2, s76
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v3, s77
-; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[46:47], s[46:47], 0x100000
-; GCN-NOHSA-VI-NEXT:    buffer_store_dwordx4 v[0:3], off, s[0:3], 0 offset:144
 ; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[48:49], s[48:49], 0x100000
-; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v0, s44
-; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v1, s45
+; GCN-NOHSA-VI-NEXT:    buffer_store_dwordx4 v[0:3], off, s[0:3], 0 offset:144
+; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[50:51], s[50:51], 0x100000
+; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v0, s46
+; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v1, s47
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v2, s74
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v3, s75
 ; GCN-NOHSA-VI-NEXT:    buffer_store_dwordx4 v[0:3], off, s[0:3], 0 offset:112
-; GCN-NOHSA-VI-NEXT:    s_lshr_b32 s52, s14, 16
-; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v0, s46
-; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v1, s47
+; GCN-NOHSA-VI-NEXT:    s_lshr_b32 s54, s14, 16
+; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v0, s48
+; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v1, s49
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v2, s72
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v3, s73
-; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[50:51], s[50:51], 0x100000
+; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[52:53], s[52:53], 0x100000
 ; GCN-NOHSA-VI-NEXT:    buffer_store_dwordx4 v[0:3], off, s[0:3], 0 offset:80
-; GCN-NOHSA-VI-NEXT:    s_lshr_b32 s54, s12, 16
-; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v0, s48
-; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v1, s49
+; GCN-NOHSA-VI-NEXT:    s_lshr_b32 s56, s12, 16
+; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v0, s50
+; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v1, s51
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v2, s70
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v3, s71
 ; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[34:35], s[14:15], 0x100000
-; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[52:53], s[52:53], 0x100000
+; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[54:55], s[54:55], 0x100000
 ; GCN-NOHSA-VI-NEXT:    buffer_store_dwordx4 v[0:3], off, s[0:3], 0 offset:48
-; GCN-NOHSA-VI-NEXT:    s_lshr_b32 s56, s10, 16
-; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v0, s50
-; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v1, s51
-; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v2, s68
-; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v3, s69
+; GCN-NOHSA-VI-NEXT:    s_lshr_b32 s58, s10, 16
+; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v0, s52
+; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v1, s53
+; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v2, s36
+; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v3, s37
 ; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[30:31], s[12:13], 0x100000
-; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[16:17], s[54:55], 0x100000
+; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[16:17], s[56:57], 0x100000
 ; GCN-NOHSA-VI-NEXT:    buffer_store_dwordx4 v[0:3], off, s[0:3], 0 offset:16
-; GCN-NOHSA-VI-NEXT:    s_lshr_b32 s58, s8, 16
+; GCN-NOHSA-VI-NEXT:    s_lshr_b32 s60, s8, 16
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v0, s34
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v1, s35
-; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v2, s52
-; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v3, s53
+; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v2, s54
+; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v3, s55
 ; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[28:29], s[10:11], 0x100000
-; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[14:15], s[56:57], 0x100000
+; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[14:15], s[58:59], 0x100000
 ; GCN-NOHSA-VI-NEXT:    buffer_store_dwordx4 v[0:3], off, s[0:3], 0 offset:224
-; GCN-NOHSA-VI-NEXT:    s_lshr_b32 s60, s6, 16
+; GCN-NOHSA-VI-NEXT:    s_lshr_b32 s62, s6, 16
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v0, s30
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v1, s31
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v2, s16
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v3, s17
 ; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[26:27], s[8:9], 0x100000
-; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[12:13], s[58:59], 0x100000
+; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[12:13], s[60:61], 0x100000
 ; GCN-NOHSA-VI-NEXT:    buffer_store_dwordx4 v[0:3], off, s[0:3], 0 offset:192
-; GCN-NOHSA-VI-NEXT:    s_lshr_b32 s62, s4, 16
+; GCN-NOHSA-VI-NEXT:    s_lshr_b32 s64, s4, 16
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v0, s28
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v1, s29
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v2, s14
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v3, s15
 ; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[24:25], s[6:7], 0x100000
-; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[10:11], s[60:61], 0x100000
+; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[10:11], s[62:63], 0x100000
 ; GCN-NOHSA-VI-NEXT:    buffer_store_dwordx4 v[0:3], off, s[0:3], 0 offset:160
 ; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[22:23], s[4:5], 0x100000
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v0, s26
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v1, s27
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v2, s12
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v3, s13
-; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[8:9], s[62:63], 0x100000
+; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[8:9], s[64:65], 0x100000
 ; GCN-NOHSA-VI-NEXT:    buffer_store_dwordx4 v[0:3], off, s[0:3], 0 offset:128
-; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[6:7], s[64:65], 0x100000
+; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[6:7], s[66:67], 0x100000
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v0, s24
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v1, s25
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v2, s10
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v3, s11
 ; GCN-NOHSA-VI-NEXT:    buffer_store_dwordx4 v[0:3], off, s[0:3], 0 offset:96
-; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[4:5], s[66:67], 0x100000
+; GCN-NOHSA-VI-NEXT:    s_bfe_i64 s[4:5], s[68:69], 0x100000
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v0, s22
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v1, s23
 ; GCN-NOHSA-VI-NEXT:    v_mov_b32_e32 v2, s8
diff --git a/llvm/test/CodeGen/AMDGPU/mubuf-legalize-operands.ll b/llvm/test/CodeGen/AMDGPU/mubuf-legalize-operands.ll
--- a/llvm/test/CodeGen/AMDGPU/mubuf-legalize-operands.ll
+++ b/llvm/test/CodeGen/AMDGPU/mubuf-legalize-operands.ll
@@ -231,10 +231,10 @@
 ; W64-O0-DAG: s_mov_b64 [[SAVEEXEC:s\[[0-9]+:[0-9]+\]]], exec
 
 ; W64-O0: [[LOOPBB0:.LBB[0-9]+_[0-9]+]]: ; =>This Inner Loop Header: Depth=1
-; W64-O0: buffer_load_dword v[[VRSRC0:[0-9]+]], off, s[0:3], s32 offset:28 ; 4-byte Folded Reload
-; W64-O0: buffer_load_dword v[[VRSRC1:[0-9]+]], off, s[0:3], s32 offset:32 ; 4-byte Folded Reload
-; W64-O0: buffer_load_dword v[[VRSRC2:[0-9]+]], off, s[0:3], s32 offset:36 ; 4-byte Folded Reload
-; W64-O0: buffer_load_dword v[[VRSRC3:[0-9]+]], off, s[0:3], s32 offset:40 ; 4-byte Folded Reload
+; W64-O0: buffer_load_dword v[[VRSRC0:[0-9]+]], off, s[0:3], s32 offset:32 ; 4-byte Folded Reload
+; W64-O0: buffer_load_dword v[[VRSRC1:[0-9]+]], off, s[0:3], s32 offset:36 ; 4-byte Folded Reload
+; W64-O0: buffer_load_dword v[[VRSRC2:[0-9]+]], off, s[0:3], s32 offset:40 ; 4-byte Folded Reload
+; W64-O0: buffer_load_dword v[[VRSRC3:[0-9]+]], off, s[0:3], s32 offset:44 ; 4-byte Folded Reload
 ; W64-O0: s_waitcnt vmcnt(0)
 ; W64-O0-DAG: v_readfirstlane_b32 s[[S0:[0-9]+]], v[[VRSRC0]]
 ; W64-O0-DAG: v_readfirstlane_b32 s[[SRSRCTMP1:[0-9]+]], v[[VRSRC1]]
@@ -251,7 +251,7 @@
 ; W64-O0-DAG: s_mov_b32 s[[S2:[0-9]+]], s[[SRSRCTMP2]]
 ; W64-O0-DAG: s_mov_b32 s[[S3:[0-9]+]], s[[SRSRCTMP3]]
 ; W64-O0: s_and_saveexec_b64 [[SAVE:s\[[0-9]+:[0-9]+\]]], [[AND]]
-; W64-O0: buffer_load_dword [[IDX:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, s32 ; 4-byte Folded Reload
+; W64-O0: buffer_load_dword [[IDX:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, s32 offset:4 ; 4-byte Folded Reload
 ; W64-O0: buffer_load_format_x [[RES:v[0-9]+]], [[IDX]], s[[[S0]]:[[S3]]], {{.*}} idxen
 ; W64-O0: s_waitcnt vmcnt(0)
 ; W64-O0: buffer_store_dword [[RES]], off, s{{\[[0-9]+:[0-9]+\]}}, s{{[0-9]+}} offset:[[RES_OFF_TMP:[0-9]+]] ; 4-byte Folded Spill
@@ -264,16 +264,18 @@
 ; W64-O0: s_cbranch_execz [[TERMBB:.LBB[0-9]+_[0-9]+]]
 
 ; W64-O0: ; %bb.{{[0-9]+}}: ; %bb1
-; W64-O0-DAG: buffer_store_dword {{v[0-9]+}}, off, s{{\[[0-9]+:[0-9]+\]}}, s32 offset:[[IDX_OFF:[0-9]+]] ; 4-byte Folded Spill
+; W64-O0: buffer_load_dword
+; W64-O0: buffer_store_dword {{v[0-9]+}}, off, s{{\[[0-9]+:[0-9]+\]}}, s32 offset:[[IDX_OFF:[0-9]+]] ; 4-byte Folded Spill
 ; W64-O0-DAG: s_mov_b64 s[[[SAVEEXEC0:[0-9]+]]:[[SAVEEXEC1:[0-9]+]]], exec
 ; W64-O0: v_writelane_b32 [[VSAVEEXEC:v[0-9]+]], s[[SAVEEXEC0]], [[SAVEEXEC_IDX0:[0-9]+]]
 ; W64-O0: v_writelane_b32 [[VSAVEEXEC]], s[[SAVEEXEC1]], [[SAVEEXEC_IDX1:[0-9]+]]
+; W64-O0: buffer_store_dword [[VSAVEEXEC]], off, s{{\[[0-9]+:[0-9]+\]}}, s{{[0-9]+}} ; 4-byte Folded Spill
 
 ; W64-O0: [[LOOPBB1:.LBB[0-9]+_[0-9]+]]: ; =>This Inner Loop Header: Depth=1
-; W64-O0: buffer_load_dword v[[VRSRC0:[0-9]+]], off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
-; W64-O0: buffer_load_dword v[[VRSRC1:[0-9]+]], off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
-; W64-O0: buffer_load_dword v[[VRSRC2:[0-9]+]], off, s[0:3], s32 offset:12 ; 4-byte Folded Reload
-; W64-O0: buffer_load_dword v[[VRSRC3:[0-9]+]], off, s[0:3], s32 offset:16 ; 4-byte Folded Reload
+; W64-O0: buffer_load_dword v[[VRSRC0:[0-9]+]], off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
+; W64-O0: buffer_load_dword v[[VRSRC1:[0-9]+]], off, s[0:3], s32 offset:12 ; 4-byte Folded Reload
+; W64-O0: buffer_load_dword v[[VRSRC2:[0-9]+]], off, s[0:3], s32 offset:16 ; 4-byte Folded Reload
+; W64-O0: buffer_load_dword v[[VRSRC3:[0-9]+]], off, s[0:3], s32 offset:20 ; 4-byte Folded Reload
 ; W64-O0: s_waitcnt vmcnt(0)
 ; W64-O0-DAG: v_readfirstlane_b32 s[[S0:[0-9]+]], v[[VRSRC0]]
 ; W64-O0-DAG: v_readfirstlane_b32 s[[SRSRCTMP1:[0-9]+]], v[[VRSRC1]]
@@ -290,7 +292,9 @@
 ; W64-O0-DAG: s_mov_b32 s[[S2:[0-9]+]], s[[SRSRCTMP2]]
 ; W64-O0-DAG: s_mov_b32 s[[S3:[0-9]+]], s[[SRSRCTMP3]]
 ; W64-O0: s_and_saveexec_b64 [[SAVE:s\[[0-9]+:[0-9]+\]]], [[AND]]
+; W64-O0: buffer_store_dword
 ; W64-O0: buffer_load_dword [[IDX:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, s32  offset:[[IDX_OFF]] ; 4-byte Folded Reload
+; W64-O0: buffer_load_dword
 ; W64-O0: buffer_load_format_x [[RES:v[0-9]+]], [[IDX]], s[[[S0]]:[[S3]]], {{.*}} idxen
 ; W64-O0: s_waitcnt vmcnt(0)
 ; W64-O0: buffer_store_dword [[RES]], off, s{{\[[0-9]+:[0-9]+\]}}, s{{[0-9]+}} offset:[[RES_OFF_TMP:[0-9]+]] ; 4-byte Folded Spill
@@ -298,8 +302,9 @@
 ; W64-O0-NEXT: s_cbranch_execnz [[LOOPBB1]]
 
 ; W64-O0: buffer_load_dword [[RES:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, s{{[0-9]+}} offset:[[RES_OFF_TMP]] ; 4-byte Folded Reload
-; W64-O0: v_readlane_b32 s[[SAVEEXEC0:[0-9]+]], [[VSAVEEXEC]], [[SAVEEXEC_IDX0]]
-; W64-O0: v_readlane_b32 s[[SAVEEXEC1:[0-9]+]], [[VSAVEEXEC]], [[SAVEEXEC_IDX1]]
+; W64-O0: buffer_load_dword [[VSAVEEXEC1:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, s{{[0-9]+}} ; 4-byte Folded Reload
+; W64-O0: v_readlane_b32 s[[SAVEEXEC0:[0-9]+]], [[VSAVEEXEC1]], [[SAVEEXEC_IDX0]]
+; W64-O0: v_readlane_b32 s[[SAVEEXEC1:[0-9]+]], [[VSAVEEXEC1]], [[SAVEEXEC_IDX1]]
 ; W64-O0: s_mov_b64 exec, s[[[SAVEEXEC0]]:[[SAVEEXEC1]]]
 ; W64-O0: buffer_store_dword [[RES]], off, s{{\[[0-9]+:[0-9]+\]}}, s{{[0-9]+}} offset:[[RES_OFF]] ; 4-byte Folded Spill
 
diff --git a/llvm/test/CodeGen/AMDGPU/mul24-pass-ordering.ll b/llvm/test/CodeGen/AMDGPU/mul24-pass-ordering.ll
--- a/llvm/test/CodeGen/AMDGPU/mul24-pass-ordering.ll
+++ b/llvm/test/CodeGen/AMDGPU/mul24-pass-ordering.ll
@@ -191,48 +191,49 @@
 ; GFX9-NEXT:    s_mov_b32 s4, s33
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_or_saveexec_b64 s[6:7], -1
-; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:12 ; 4-byte Folded Spill
+; GFX9-NEXT:    buffer_store_dword v42, off, s[0:3], s33 offset:12 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v44, off, s[0:3], s33 offset:16 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[6:7]
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
-; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
+; GFX9-NEXT:    ; implicit-def: $vgpr42
 ; GFX9-NEXT:    s_addk_i32 s32, 0x800
-; GFX9-NEXT:    v_writelane_b32 v40, s34, 2
+; GFX9-NEXT:    v_writelane_b32 v42, s30, 0
+; GFX9-NEXT:    v_writelane_b32 v42, s31, 1
+; GFX9-NEXT:    v_writelane_b32 v42, s34, 2
 ; GFX9-NEXT:    v_writelane_b32 v44, s4, 0
-; GFX9-NEXT:    v_writelane_b32 v40, s36, 3
+; GFX9-NEXT:    v_writelane_b32 v42, s36, 3
 ; GFX9-NEXT:    s_getpc_b64 s[4:5]
 ; GFX9-NEXT:    s_add_u32 s4, s4, foo@gotpcrel32@lo+4
 ; GFX9-NEXT:    s_addc_u32 s5, s5, foo@gotpcrel32@hi+12
-; GFX9-NEXT:    v_writelane_b32 v40, s37, 4
+; GFX9-NEXT:    v_writelane_b32 v42, s37, 4
 ; GFX9-NEXT:    s_load_dwordx2 s[36:37], s[4:5], 0x0
-; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
-; GFX9-NEXT:    buffer_store_dword v42, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
+; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
+; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v43, off, s[0:3], s33 ; 4-byte Folded Spill
-; GFX9-NEXT:    v_mov_b32_e32 v41, v1
-; GFX9-NEXT:    v_mov_b32_e32 v42, v0
-; GFX9-NEXT:    v_mul_u32_u24_e32 v0, v42, v41
+; GFX9-NEXT:    v_mov_b32_e32 v40, v1
+; GFX9-NEXT:    v_mov_b32_e32 v41, v0
+; GFX9-NEXT:    v_mul_u32_u24_e32 v0, v41, v40
 ; GFX9-NEXT:    s_mov_b32 s34, s15
-; GFX9-NEXT:    v_and_b32_e32 v43, 0xffffff, v41
+; GFX9-NEXT:    v_and_b32_e32 v43, 0xffffff, v40
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX9-NEXT:    s_swappc_b64 s[30:31], s[36:37]
-; GFX9-NEXT:    v_mad_u32_u24 v41, v42, v41, v43
+; GFX9-NEXT:    v_mad_u32_u24 v40, v41, v40, v43
 ; GFX9-NEXT:    s_mov_b32 s15, s34
-; GFX9-NEXT:    v_mov_b32_e32 v0, v41
+; GFX9-NEXT:    v_mov_b32_e32 v0, v40
 ; GFX9-NEXT:    s_swappc_b64 s[30:31], s[36:37]
-; GFX9-NEXT:    v_add_u32_e32 v0, v41, v43
+; GFX9-NEXT:    v_add_u32_e32 v0, v40, v43
 ; GFX9-NEXT:    s_mov_b32 s15, s34
 ; GFX9-NEXT:    s_swappc_b64 s[30:31], s[36:37]
 ; GFX9-NEXT:    buffer_load_dword v43, off, s[0:3], s33 ; 4-byte Folded Reload
-; GFX9-NEXT:    buffer_load_dword v42, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
-; GFX9-NEXT:    buffer_load_dword v41, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload
-; GFX9-NEXT:    v_readlane_b32 s37, v40, 4
-; GFX9-NEXT:    v_readlane_b32 s36, v40, 3
-; GFX9-NEXT:    v_readlane_b32 s34, v40, 2
-; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
-; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
+; GFX9-NEXT:    buffer_load_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
+; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload
+; GFX9-NEXT:    v_readlane_b32 s37, v42, 4
+; GFX9-NEXT:    v_readlane_b32 s36, v42, 3
+; GFX9-NEXT:    v_readlane_b32 s34, v42, 2
+; GFX9-NEXT:    v_readlane_b32 s31, v42, 1
+; GFX9-NEXT:    v_readlane_b32 s30, v42, 0
 ; GFX9-NEXT:    v_readlane_b32 s4, v44, 0
 ; GFX9-NEXT:    s_or_saveexec_b64 s[6:7], -1
-; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s33 offset:12 ; 4-byte Folded Reload
+; GFX9-NEXT:    buffer_load_dword v42, off, s[0:3], s33 offset:12 ; 4-byte Folded Reload
 ; GFX9-NEXT:    buffer_load_dword v44, off, s[0:3], s33 offset:16 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[6:7]
 ; GFX9-NEXT:    s_addk_i32 s32, 0xf800
diff --git a/llvm/test/CodeGen/AMDGPU/need-fp-from-vgpr-spills.ll b/llvm/test/CodeGen/AMDGPU/need-fp-from-vgpr-spills.ll
--- a/llvm/test/CodeGen/AMDGPU/need-fp-from-vgpr-spills.ll
+++ b/llvm/test/CodeGen/AMDGPU/need-fp-from-vgpr-spills.ll
@@ -27,15 +27,19 @@
 ; CHECK-LABEL: csr_vgpr_spill_fp_callee:
 ; CHECK:       ; %bb.0: ; %bb
 ; CHECK-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; CHECK-NEXT:    s_mov_b32 s6, s33
+; CHECK-NEXT:    s_mov_b32 s14, s33
 ; CHECK-NEXT:    s_mov_b32 s33, s32
 ; CHECK-NEXT:    s_xor_saveexec_b64 s[4:5], -1
-; CHECK-NEXT:    buffer_store_dword v1, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
+; CHECK-NEXT:    buffer_store_dword v0, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
 ; CHECK-NEXT:    s_mov_b64 exec, s[4:5]
 ; CHECK-NEXT:    s_add_i32 s32, s32, 0x400
 ; CHECK-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
-; CHECK-NEXT:    v_writelane_b32 v1, s30, 0
-; CHECK-NEXT:    v_writelane_b32 v1, s31, 1
+; CHECK-NEXT:    ; implicit-def: $vgpr0
+; CHECK-NEXT:    v_writelane_b32 v0, s30, 0
+; CHECK-NEXT:    v_writelane_b32 v0, s31, 1
+; CHECK-NEXT:    s_or_saveexec_b64 s[12:13], -1
+; CHECK-NEXT:    buffer_store_dword v0, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
+; CHECK-NEXT:    s_mov_b64 exec, s[12:13]
 ; CHECK-NEXT:    s_getpc_b64 s[4:5]
 ; CHECK-NEXT:    s_add_u32 s4, s4, callee_has_fp@rel32@lo+4
 ; CHECK-NEXT:    s_addc_u32 s5, s5, callee_has_fp@rel32@hi+12
@@ -44,17 +48,21 @@
 ; CHECK-NEXT:    s_mov_b64 s[0:1], s[8:9]
 ; CHECK-NEXT:    s_mov_b64 s[2:3], s[10:11]
 ; CHECK-NEXT:    s_swappc_b64 s[30:31], s[4:5]
+; CHECK-NEXT:    s_or_saveexec_b64 s[12:13], -1
+; CHECK-NEXT:    buffer_load_dword v0, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
+; CHECK-NEXT:    s_mov_b64 exec, s[12:13]
 ; CHECK-NEXT:    ;;#ASMSTART
 ; CHECK-NEXT:    ; clobber csr v40
 ; CHECK-NEXT:    ;;#ASMEND
-; CHECK-NEXT:    v_readlane_b32 s31, v1, 1
-; CHECK-NEXT:    v_readlane_b32 s30, v1, 0
+; CHECK-NEXT:    s_waitcnt vmcnt(0)
+; CHECK-NEXT:    v_readlane_b32 s31, v0, 1
+; CHECK-NEXT:    v_readlane_b32 s30, v0, 0
 ; CHECK-NEXT:    buffer_load_dword v40, off, s[0:3], s33 ; 4-byte Folded Reload
 ; CHECK-NEXT:    s_xor_saveexec_b64 s[4:5], -1
-; CHECK-NEXT:    buffer_load_dword v1, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
+; CHECK-NEXT:    buffer_load_dword v0, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload
 ; CHECK-NEXT:    s_mov_b64 exec, s[4:5]
 ; CHECK-NEXT:    s_add_i32 s32, s32, 0xfffffc00
-; CHECK-NEXT:    s_mov_b32 s33, s6
+; CHECK-NEXT:    s_mov_b32 s33, s14
 ; CHECK-NEXT:    s_waitcnt vmcnt(0)
 ; CHECK-NEXT:    s_setpc_b64 s[30:31]
 bb:
@@ -91,21 +99,22 @@
 ; CHECK:       ; %bb.0: ; %bb
 ; CHECK-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; CHECK-NEXT:    s_xor_saveexec_b64 s[4:5], -1
-; CHECK-NEXT:    buffer_store_dword v1, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
+; CHECK-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; CHECK-NEXT:    s_mov_b64 exec, s[4:5]
 ; CHECK-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
-; CHECK-NEXT:    v_writelane_b32 v1, s33, 0
+; CHECK-NEXT:    ; implicit-def: $vgpr0
+; CHECK-NEXT:    v_writelane_b32 v0, s33, 0
 ; CHECK-NEXT:    ;;#ASMSTART
 ; CHECK-NEXT:    ; clobber csr v40
 ; CHECK-NEXT:    ;;#ASMEND
 ; CHECK-NEXT:    s_getpc_b64 s[4:5]
 ; CHECK-NEXT:    s_add_u32 s4, s4, callee_has_fp@rel32@lo+4
 ; CHECK-NEXT:    s_addc_u32 s5, s5, callee_has_fp@rel32@hi+12
-; CHECK-NEXT:    v_readlane_b32 s33, v1, 0
+; CHECK-NEXT:    v_readlane_b32 s33, v0, 0
 ; CHECK-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
-; CHECK-NEXT:    s_xor_saveexec_b64 s[6:7], -1
-; CHECK-NEXT:    buffer_load_dword v1, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
-; CHECK-NEXT:    s_mov_b64 exec, s[6:7]
+; CHECK-NEXT:    s_xor_saveexec_b64 s[8:9], -1
+; CHECK-NEXT:    buffer_load_dword v0, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
+; CHECK-NEXT:    s_mov_b64 exec, s[8:9]
 ; CHECK-NEXT:    s_setpc_b64 s[4:5]
 bb:
   call void asm sideeffect "; clobber csr v40", "~{v40}"()
@@ -152,12 +161,13 @@
 ; CHECK-LABEL: caller_save_vgpr_spill_fp_tail_call:
 ; CHECK:       ; %bb.0: ; %entry
 ; CHECK-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; CHECK-NEXT:    s_mov_b32 s6, s33
+; CHECK-NEXT:    s_mov_b32 s12, s33
 ; CHECK-NEXT:    s_mov_b32 s33, s32
 ; CHECK-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; CHECK-NEXT:    buffer_store_dword v1, off, s[0:3], s33 ; 4-byte Folded Spill
 ; CHECK-NEXT:    s_mov_b64 exec, s[4:5]
 ; CHECK-NEXT:    s_add_i32 s32, s32, 0x400
+; CHECK-NEXT:    ; implicit-def: $vgpr1
 ; CHECK-NEXT:    v_writelane_b32 v1, s30, 0
 ; CHECK-NEXT:    v_writelane_b32 v1, s31, 1
 ; CHECK-NEXT:    s_getpc_b64 s[4:5]
@@ -174,7 +184,7 @@
 ; CHECK-NEXT:    buffer_load_dword v1, off, s[0:3], s33 ; 4-byte Folded Reload
 ; CHECK-NEXT:    s_mov_b64 exec, s[4:5]
 ; CHECK-NEXT:    s_add_i32 s32, s32, 0xfffffc00
-; CHECK-NEXT:    s_mov_b32 s33, s6
+; CHECK-NEXT:    s_mov_b32 s33, s12
 ; CHECK-NEXT:    s_waitcnt vmcnt(0)
 ; CHECK-NEXT:    s_setpc_b64 s[30:31]
 entry:
@@ -186,14 +196,19 @@
 ; CHECK-LABEL: caller_save_vgpr_spill_fp:
 ; CHECK:       ; %bb.0: ; %entry
 ; CHECK-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; CHECK-NEXT:    s_mov_b32 s7, s33
+; CHECK-NEXT:    s_mov_b32 s13, s33
 ; CHECK-NEXT:    s_mov_b32 s33, s32
 ; CHECK-NEXT:    s_xor_saveexec_b64 s[4:5], -1
-; CHECK-NEXT:    buffer_store_dword v2, off, s[0:3], s33 ; 4-byte Folded Spill
+; CHECK-NEXT:    buffer_store_dword v0, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
+; CHECK-NEXT:    buffer_store_dword v1, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
 ; CHECK-NEXT:    s_mov_b64 exec, s[4:5]
 ; CHECK-NEXT:    s_add_i32 s32, s32, 0x400
-; CHECK-NEXT:    v_writelane_b32 v2, s30, 0
-; CHECK-NEXT:    v_writelane_b32 v2, s31, 1
+; CHECK-NEXT:    ; implicit-def: $vgpr0
+; CHECK-NEXT:    v_writelane_b32 v0, s30, 0
+; CHECK-NEXT:    v_writelane_b32 v0, s31, 1
+; CHECK-NEXT:    s_or_saveexec_b64 s[14:15], -1
+; CHECK-NEXT:    buffer_store_dword v0, off, s[0:3], s33 ; 4-byte Folded Spill
+; CHECK-NEXT:    s_mov_b64 exec, s[14:15]
 ; CHECK-NEXT:    s_getpc_b64 s[4:5]
 ; CHECK-NEXT:    s_add_u32 s4, s4, caller_save_vgpr_spill_fp_tail_call@rel32@lo+4
 ; CHECK-NEXT:    s_addc_u32 s5, s5, caller_save_vgpr_spill_fp_tail_call@rel32@hi+12
@@ -202,13 +217,18 @@
 ; CHECK-NEXT:    s_mov_b64 s[0:1], s[8:9]
 ; CHECK-NEXT:    s_mov_b64 s[2:3], s[10:11]
 ; CHECK-NEXT:    s_swappc_b64 s[30:31], s[4:5]
-; CHECK-NEXT:    v_readlane_b32 s31, v2, 1
-; CHECK-NEXT:    v_readlane_b32 s30, v2, 0
+; CHECK-NEXT:    s_or_saveexec_b64 s[14:15], -1
+; CHECK-NEXT:    buffer_load_dword v1, off, s[0:3], s33 ; 4-byte Folded Reload
+; CHECK-NEXT:    s_mov_b64 exec, s[14:15]
+; CHECK-NEXT:    s_waitcnt vmcnt(0)
+; CHECK-NEXT:    v_readlane_b32 s31, v1, 1
+; CHECK-NEXT:    v_readlane_b32 s30, v1, 0
 ; CHECK-NEXT:    s_xor_saveexec_b64 s[4:5], -1
-; CHECK-NEXT:    buffer_load_dword v2, off, s[0:3], s33 ; 4-byte Folded Reload
+; CHECK-NEXT:    buffer_load_dword v0, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
+; CHECK-NEXT:    buffer_load_dword v1, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload
 ; CHECK-NEXT:    s_mov_b64 exec, s[4:5]
 ; CHECK-NEXT:    s_add_i32 s32, s32, 0xfffffc00
-; CHECK-NEXT:    s_mov_b32 s33, s7
+; CHECK-NEXT:    s_mov_b32 s33, s13
 ; CHECK-NEXT:    s_waitcnt vmcnt(0)
 ; CHECK-NEXT:    s_setpc_b64 s[30:31]
 entry:
diff --git a/llvm/test/CodeGen/AMDGPU/no-source-locations-in-prologue.ll b/llvm/test/CodeGen/AMDGPU/no-source-locations-in-prologue.ll
--- a/llvm/test/CodeGen/AMDGPU/no-source-locations-in-prologue.ll
+++ b/llvm/test/CodeGen/AMDGPU/no-source-locations-in-prologue.ll
@@ -15,14 +15,21 @@
 ; CHECK-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; CHECK-NEXT:    s_mov_b32 s16, s33
 ; CHECK-NEXT:    s_mov_b32 s33, s32
-; CHECK-NEXT:    s_or_saveexec_b64 s[18:19], -1
-; CHECK-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
-; CHECK-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
+; CHECK-NEXT:    s_xor_saveexec_b64 s[18:19], -1
+; CHECK-NEXT:    buffer_store_dword v0, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
+; CHECK-NEXT:    s_mov_b64 exec, -1
+; CHECK-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
 ; CHECK-NEXT:    s_mov_b64 exec, s[18:19]
-; CHECK-NEXT:    v_writelane_b32 v41, s16, 0
+; CHECK-NEXT:    v_writelane_b32 v40, s34, 0
+; CHECK-NEXT:    v_writelane_b32 v40, s35, 1
+; CHECK-NEXT:    v_writelane_b32 v40, s16, 2
 ; CHECK-NEXT:    s_add_i32 s32, s32, 0x400
-; CHECK-NEXT:    v_writelane_b32 v40, s30, 0
-; CHECK-NEXT:    v_writelane_b32 v40, s31, 1
+; CHECK-NEXT:    ; implicit-def: $vgpr0
+; CHECK-NEXT:    v_writelane_b32 v0, s30, 0
+; CHECK-NEXT:    v_writelane_b32 v0, s31, 1
+; CHECK-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; CHECK-NEXT:    buffer_store_dword v0, off, s[0:3], s33 ; 4-byte Folded Spill
+; CHECK-NEXT:    s_mov_b64 exec, s[34:35]
 ; CHECK-NEXT:  .Ltmp0:
 ; CHECK-NEXT:    .loc 0 31 3 prologue_end ; lane-info.cpp:31:3
 ; CHECK-NEXT:    s_getpc_b64 s[16:17]
@@ -36,13 +43,20 @@
 ; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
 ; CHECK-NEXT:    s_swappc_b64 s[30:31], s[16:17]
 ; CHECK-NEXT:  .Ltmp1:
+; CHECK-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; CHECK-NEXT:    buffer_load_dword v0, off, s[0:3], s33 ; 4-byte Folded Reload
+; CHECK-NEXT:    s_mov_b64 exec, s[34:35]
 ; CHECK-NEXT:    .loc 0 32 1 ; lane-info.cpp:32:1
-; CHECK-NEXT:    v_readlane_b32 s31, v40, 1
-; CHECK-NEXT:    v_readlane_b32 s30, v40, 0
-; CHECK-NEXT:    v_readlane_b32 s4, v41, 0
-; CHECK-NEXT:    s_or_saveexec_b64 s[6:7], -1
-; CHECK-NEXT:    buffer_load_dword v40, off, s[0:3], s33 ; 4-byte Folded Reload
-; CHECK-NEXT:    buffer_load_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
+; CHECK-NEXT:    s_waitcnt vmcnt(0)
+; CHECK-NEXT:    v_readlane_b32 s31, v0, 1
+; CHECK-NEXT:    v_readlane_b32 s30, v0, 0
+; CHECK-NEXT:    v_readlane_b32 s34, v40, 0
+; CHECK-NEXT:    v_readlane_b32 s35, v40, 1
+; CHECK-NEXT:    v_readlane_b32 s4, v40, 2
+; CHECK-NEXT:    s_xor_saveexec_b64 s[6:7], -1
+; CHECK-NEXT:    buffer_load_dword v0, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
+; CHECK-NEXT:    s_mov_b64 exec, -1
+; CHECK-NEXT:    buffer_load_dword v40, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload
 ; CHECK-NEXT:    s_mov_b64 exec, s[6:7]
 ; CHECK-NEXT:    s_add_i32 s32, s32, 0xfffffc00
 ; CHECK-NEXT:    s_mov_b32 s33, s4
diff --git a/llvm/test/CodeGen/AMDGPU/partial-sgpr-to-vgpr-spills.ll b/llvm/test/CodeGen/AMDGPU/partial-sgpr-to-vgpr-spills.ll
--- a/llvm/test/CodeGen/AMDGPU/partial-sgpr-to-vgpr-spills.ll
+++ b/llvm/test/CodeGen/AMDGPU/partial-sgpr-to-vgpr-spills.ll
@@ -11,10 +11,17 @@
 define amdgpu_kernel void @spill_sgprs_to_multiple_vgprs(i32 addrspace(1)* %out, i32 %in) #0 {
 ; GCN-LABEL: spill_sgprs_to_multiple_vgprs:
 ; GCN:       ; %bb.0:
+; GCN-NEXT:    s_mov_b32 s92, SCRATCH_RSRC_DWORD0
+; GCN-NEXT:    s_mov_b32 s93, SCRATCH_RSRC_DWORD1
+; GCN-NEXT:    s_mov_b32 s94, -1
+; GCN-NEXT:    s_mov_b32 s95, 0xe8f000
+; GCN-NEXT:    s_add_u32 s92, s92, s3
+; GCN-NEXT:    s_addc_u32 s93, s93, 0
 ; GCN-NEXT:    s_load_dword s0, s[0:1], 0xb
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s[4:11]
 ; GCN-NEXT:    ;;#ASMEND
+; GCN-NEXT:    ; implicit-def: $vgpr0
 ; GCN-NEXT:    v_writelane_b32 v0, s4, 0
 ; GCN-NEXT:    v_writelane_b32 v0, s5, 1
 ; GCN-NEXT:    v_writelane_b32 v0, s6, 2
@@ -100,264 +107,285 @@
 ; GCN-NEXT:    v_writelane_b32 v0, s9, 61
 ; GCN-NEXT:    v_writelane_b32 v0, s10, 62
 ; GCN-NEXT:    v_writelane_b32 v0, s11, 63
+; GCN-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GCN-NEXT:    buffer_store_dword v0, off, s[92:95], 0 offset:12 ; 4-byte Folded Spill
+; GCN-NEXT:    s_mov_b64 exec, s[34:35]
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s[4:11]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_writelane_b32 v1, s4, 0
-; GCN-NEXT:    v_writelane_b32 v1, s5, 1
-; GCN-NEXT:    v_writelane_b32 v1, s6, 2
-; GCN-NEXT:    v_writelane_b32 v1, s7, 3
-; GCN-NEXT:    v_writelane_b32 v1, s8, 4
-; GCN-NEXT:    v_writelane_b32 v1, s9, 5
-; GCN-NEXT:    v_writelane_b32 v1, s10, 6
-; GCN-NEXT:    v_writelane_b32 v1, s11, 7
+; GCN-NEXT:    ; implicit-def: $vgpr0
+; GCN-NEXT:    v_writelane_b32 v0, s4, 0
+; GCN-NEXT:    v_writelane_b32 v0, s5, 1
+; GCN-NEXT:    v_writelane_b32 v0, s6, 2
+; GCN-NEXT:    v_writelane_b32 v0, s7, 3
+; GCN-NEXT:    v_writelane_b32 v0, s8, 4
+; GCN-NEXT:    v_writelane_b32 v0, s9, 5
+; GCN-NEXT:    v_writelane_b32 v0, s10, 6
+; GCN-NEXT:    v_writelane_b32 v0, s11, 7
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s[4:11]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_writelane_b32 v1, s4, 8
-; GCN-NEXT:    v_writelane_b32 v1, s5, 9
-; GCN-NEXT:    v_writelane_b32 v1, s6, 10
-; GCN-NEXT:    v_writelane_b32 v1, s7, 11
-; GCN-NEXT:    v_writelane_b32 v1, s8, 12
-; GCN-NEXT:    v_writelane_b32 v1, s9, 13
-; GCN-NEXT:    v_writelane_b32 v1, s10, 14
-; GCN-NEXT:    v_writelane_b32 v1, s11, 15
+; GCN-NEXT:    v_writelane_b32 v0, s4, 8
+; GCN-NEXT:    v_writelane_b32 v0, s5, 9
+; GCN-NEXT:    v_writelane_b32 v0, s6, 10
+; GCN-NEXT:    v_writelane_b32 v0, s7, 11
+; GCN-NEXT:    v_writelane_b32 v0, s8, 12
+; GCN-NEXT:    v_writelane_b32 v0, s9, 13
+; GCN-NEXT:    v_writelane_b32 v0, s10, 14
+; GCN-NEXT:    v_writelane_b32 v0, s11, 15
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s[4:11]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_writelane_b32 v1, s4, 16
-; GCN-NEXT:    v_writelane_b32 v1, s5, 17
-; GCN-NEXT:    v_writelane_b32 v1, s6, 18
-; GCN-NEXT:    v_writelane_b32 v1, s7, 19
-; GCN-NEXT:    v_writelane_b32 v1, s8, 20
-; GCN-NEXT:    v_writelane_b32 v1, s9, 21
-; GCN-NEXT:    v_writelane_b32 v1, s10, 22
-; GCN-NEXT:    v_writelane_b32 v1, s11, 23
+; GCN-NEXT:    v_writelane_b32 v0, s4, 16
+; GCN-NEXT:    v_writelane_b32 v0, s5, 17
+; GCN-NEXT:    v_writelane_b32 v0, s6, 18
+; GCN-NEXT:    v_writelane_b32 v0, s7, 19
+; GCN-NEXT:    v_writelane_b32 v0, s8, 20
+; GCN-NEXT:    v_writelane_b32 v0, s9, 21
+; GCN-NEXT:    v_writelane_b32 v0, s10, 22
+; GCN-NEXT:    v_writelane_b32 v0, s11, 23
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s[4:11]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_writelane_b32 v1, s4, 24
-; GCN-NEXT:    v_writelane_b32 v1, s5, 25
-; GCN-NEXT:    v_writelane_b32 v1, s6, 26
-; GCN-NEXT:    v_writelane_b32 v1, s7, 27
-; GCN-NEXT:    v_writelane_b32 v1, s8, 28
-; GCN-NEXT:    v_writelane_b32 v1, s9, 29
-; GCN-NEXT:    v_writelane_b32 v1, s10, 30
-; GCN-NEXT:    v_writelane_b32 v1, s11, 31
+; GCN-NEXT:    v_writelane_b32 v0, s4, 24
+; GCN-NEXT:    v_writelane_b32 v0, s5, 25
+; GCN-NEXT:    v_writelane_b32 v0, s6, 26
+; GCN-NEXT:    v_writelane_b32 v0, s7, 27
+; GCN-NEXT:    v_writelane_b32 v0, s8, 28
+; GCN-NEXT:    v_writelane_b32 v0, s9, 29
+; GCN-NEXT:    v_writelane_b32 v0, s10, 30
+; GCN-NEXT:    v_writelane_b32 v0, s11, 31
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s[4:11]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_writelane_b32 v1, s4, 32
-; GCN-NEXT:    v_writelane_b32 v1, s5, 33
-; GCN-NEXT:    v_writelane_b32 v1, s6, 34
-; GCN-NEXT:    v_writelane_b32 v1, s7, 35
-; GCN-NEXT:    v_writelane_b32 v1, s8, 36
-; GCN-NEXT:    v_writelane_b32 v1, s9, 37
-; GCN-NEXT:    v_writelane_b32 v1, s10, 38
-; GCN-NEXT:    v_writelane_b32 v1, s11, 39
+; GCN-NEXT:    v_writelane_b32 v0, s4, 32
+; GCN-NEXT:    v_writelane_b32 v0, s5, 33
+; GCN-NEXT:    v_writelane_b32 v0, s6, 34
+; GCN-NEXT:    v_writelane_b32 v0, s7, 35
+; GCN-NEXT:    v_writelane_b32 v0, s8, 36
+; GCN-NEXT:    v_writelane_b32 v0, s9, 37
+; GCN-NEXT:    v_writelane_b32 v0, s10, 38
+; GCN-NEXT:    v_writelane_b32 v0, s11, 39
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s[4:11]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_writelane_b32 v1, s4, 40
-; GCN-NEXT:    v_writelane_b32 v1, s5, 41
-; GCN-NEXT:    v_writelane_b32 v1, s6, 42
-; GCN-NEXT:    v_writelane_b32 v1, s7, 43
-; GCN-NEXT:    v_writelane_b32 v1, s8, 44
-; GCN-NEXT:    v_writelane_b32 v1, s9, 45
-; GCN-NEXT:    v_writelane_b32 v1, s10, 46
-; GCN-NEXT:    v_writelane_b32 v1, s11, 47
+; GCN-NEXT:    v_writelane_b32 v0, s4, 40
+; GCN-NEXT:    v_writelane_b32 v0, s5, 41
+; GCN-NEXT:    v_writelane_b32 v0, s6, 42
+; GCN-NEXT:    v_writelane_b32 v0, s7, 43
+; GCN-NEXT:    v_writelane_b32 v0, s8, 44
+; GCN-NEXT:    v_writelane_b32 v0, s9, 45
+; GCN-NEXT:    v_writelane_b32 v0, s10, 46
+; GCN-NEXT:    v_writelane_b32 v0, s11, 47
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s[4:11]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_writelane_b32 v1, s4, 48
-; GCN-NEXT:    v_writelane_b32 v1, s5, 49
-; GCN-NEXT:    v_writelane_b32 v1, s6, 50
-; GCN-NEXT:    v_writelane_b32 v1, s7, 51
-; GCN-NEXT:    v_writelane_b32 v1, s8, 52
-; GCN-NEXT:    v_writelane_b32 v1, s9, 53
-; GCN-NEXT:    v_writelane_b32 v1, s10, 54
-; GCN-NEXT:    v_writelane_b32 v1, s11, 55
+; GCN-NEXT:    v_writelane_b32 v0, s4, 48
+; GCN-NEXT:    v_writelane_b32 v0, s5, 49
+; GCN-NEXT:    v_writelane_b32 v0, s6, 50
+; GCN-NEXT:    v_writelane_b32 v0, s7, 51
+; GCN-NEXT:    v_writelane_b32 v0, s8, 52
+; GCN-NEXT:    v_writelane_b32 v0, s9, 53
+; GCN-NEXT:    v_writelane_b32 v0, s10, 54
+; GCN-NEXT:    v_writelane_b32 v0, s11, 55
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s[4:11]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_writelane_b32 v1, s4, 56
-; GCN-NEXT:    v_writelane_b32 v1, s5, 57
-; GCN-NEXT:    v_writelane_b32 v1, s6, 58
-; GCN-NEXT:    v_writelane_b32 v1, s7, 59
-; GCN-NEXT:    v_writelane_b32 v1, s8, 60
-; GCN-NEXT:    v_writelane_b32 v1, s9, 61
-; GCN-NEXT:    v_writelane_b32 v1, s10, 62
-; GCN-NEXT:    v_writelane_b32 v1, s11, 63
+; GCN-NEXT:    v_writelane_b32 v0, s4, 56
+; GCN-NEXT:    v_writelane_b32 v0, s5, 57
+; GCN-NEXT:    v_writelane_b32 v0, s6, 58
+; GCN-NEXT:    v_writelane_b32 v0, s7, 59
+; GCN-NEXT:    v_writelane_b32 v0, s8, 60
+; GCN-NEXT:    v_writelane_b32 v0, s9, 61
+; GCN-NEXT:    v_writelane_b32 v0, s10, 62
+; GCN-NEXT:    v_writelane_b32 v0, s11, 63
+; GCN-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GCN-NEXT:    buffer_store_dword v0, off, s[92:95], 0 offset:8 ; 4-byte Folded Spill
+; GCN-NEXT:    s_mov_b64 exec, s[34:35]
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s[4:11]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_writelane_b32 v2, s4, 0
-; GCN-NEXT:    v_writelane_b32 v2, s5, 1
-; GCN-NEXT:    v_writelane_b32 v2, s6, 2
-; GCN-NEXT:    v_writelane_b32 v2, s7, 3
-; GCN-NEXT:    v_writelane_b32 v2, s8, 4
-; GCN-NEXT:    v_writelane_b32 v2, s9, 5
-; GCN-NEXT:    v_writelane_b32 v2, s10, 6
-; GCN-NEXT:    v_writelane_b32 v2, s11, 7
+; GCN-NEXT:    ; implicit-def: $vgpr0
+; GCN-NEXT:    v_writelane_b32 v0, s4, 0
+; GCN-NEXT:    v_writelane_b32 v0, s5, 1
+; GCN-NEXT:    v_writelane_b32 v0, s6, 2
+; GCN-NEXT:    v_writelane_b32 v0, s7, 3
+; GCN-NEXT:    v_writelane_b32 v0, s8, 4
+; GCN-NEXT:    v_writelane_b32 v0, s9, 5
+; GCN-NEXT:    v_writelane_b32 v0, s10, 6
+; GCN-NEXT:    v_writelane_b32 v0, s11, 7
+; GCN-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GCN-NEXT:    buffer_store_dword v0, off, s[92:95], 0 offset:4 ; 4-byte Folded Spill
+; GCN-NEXT:    s_mov_b64 exec, s[34:35]
 ; GCN-NEXT:    s_mov_b32 s1, 0
 ; GCN-NEXT:    s_waitcnt lgkmcnt(0)
 ; GCN-NEXT:    s_cmp_lg_u32 s0, s1
 ; GCN-NEXT:    s_cbranch_scc1 .LBB0_2
 ; GCN-NEXT:  ; %bb.1: ; %bb0
-; GCN-NEXT:    v_readlane_b32 s8, v1, 56
-; GCN-NEXT:    v_readlane_b32 s9, v1, 57
-; GCN-NEXT:    v_readlane_b32 s10, v1, 58
-; GCN-NEXT:    v_readlane_b32 s11, v1, 59
-; GCN-NEXT:    v_readlane_b32 s12, v1, 60
-; GCN-NEXT:    v_readlane_b32 s13, v1, 61
-; GCN-NEXT:    v_readlane_b32 s14, v1, 62
-; GCN-NEXT:    v_readlane_b32 s15, v1, 63
-; GCN-NEXT:    v_readlane_b32 s16, v1, 48
-; GCN-NEXT:    v_readlane_b32 s17, v1, 49
-; GCN-NEXT:    v_readlane_b32 s18, v1, 50
-; GCN-NEXT:    v_readlane_b32 s19, v1, 51
-; GCN-NEXT:    v_readlane_b32 s20, v1, 52
-; GCN-NEXT:    v_readlane_b32 s21, v1, 53
-; GCN-NEXT:    v_readlane_b32 s22, v1, 54
-; GCN-NEXT:    v_readlane_b32 s23, v1, 55
-; GCN-NEXT:    v_readlane_b32 s24, v1, 40
-; GCN-NEXT:    v_readlane_b32 s25, v1, 41
-; GCN-NEXT:    v_readlane_b32 s26, v1, 42
-; GCN-NEXT:    v_readlane_b32 s27, v1, 43
-; GCN-NEXT:    v_readlane_b32 s28, v1, 44
-; GCN-NEXT:    v_readlane_b32 s29, v1, 45
-; GCN-NEXT:    v_readlane_b32 s30, v1, 46
-; GCN-NEXT:    v_readlane_b32 s31, v1, 47
-; GCN-NEXT:    v_readlane_b32 s36, v1, 32
-; GCN-NEXT:    v_readlane_b32 s37, v1, 33
-; GCN-NEXT:    v_readlane_b32 s38, v1, 34
-; GCN-NEXT:    v_readlane_b32 s39, v1, 35
-; GCN-NEXT:    v_readlane_b32 s40, v1, 36
-; GCN-NEXT:    v_readlane_b32 s41, v1, 37
-; GCN-NEXT:    v_readlane_b32 s42, v1, 38
-; GCN-NEXT:    v_readlane_b32 s43, v1, 39
-; GCN-NEXT:    v_readlane_b32 s44, v1, 24
-; GCN-NEXT:    v_readlane_b32 s45, v1, 25
-; GCN-NEXT:    v_readlane_b32 s46, v1, 26
-; GCN-NEXT:    v_readlane_b32 s47, v1, 27
-; GCN-NEXT:    v_readlane_b32 s48, v1, 28
-; GCN-NEXT:    v_readlane_b32 s49, v1, 29
-; GCN-NEXT:    v_readlane_b32 s50, v1, 30
-; GCN-NEXT:    v_readlane_b32 s51, v1, 31
-; GCN-NEXT:    v_readlane_b32 s52, v1, 16
-; GCN-NEXT:    v_readlane_b32 s53, v1, 17
-; GCN-NEXT:    v_readlane_b32 s54, v1, 18
-; GCN-NEXT:    v_readlane_b32 s55, v1, 19
-; GCN-NEXT:    v_readlane_b32 s56, v1, 20
-; GCN-NEXT:    v_readlane_b32 s57, v1, 21
-; GCN-NEXT:    v_readlane_b32 s58, v1, 22
-; GCN-NEXT:    v_readlane_b32 s59, v1, 23
-; GCN-NEXT:    v_readlane_b32 s60, v1, 8
-; GCN-NEXT:    v_readlane_b32 s61, v1, 9
-; GCN-NEXT:    v_readlane_b32 s62, v1, 10
-; GCN-NEXT:    v_readlane_b32 s63, v1, 11
-; GCN-NEXT:    v_readlane_b32 s64, v1, 12
-; GCN-NEXT:    v_readlane_b32 s65, v1, 13
-; GCN-NEXT:    v_readlane_b32 s66, v1, 14
-; GCN-NEXT:    v_readlane_b32 s67, v1, 15
-; GCN-NEXT:    v_readlane_b32 s68, v1, 0
-; GCN-NEXT:    v_readlane_b32 s69, v1, 1
-; GCN-NEXT:    v_readlane_b32 s70, v1, 2
-; GCN-NEXT:    v_readlane_b32 s71, v1, 3
-; GCN-NEXT:    v_readlane_b32 s72, v1, 4
-; GCN-NEXT:    v_readlane_b32 s73, v1, 5
-; GCN-NEXT:    v_readlane_b32 s74, v1, 6
-; GCN-NEXT:    v_readlane_b32 s75, v1, 7
-; GCN-NEXT:    v_readlane_b32 s76, v0, 56
-; GCN-NEXT:    v_readlane_b32 s77, v0, 57
-; GCN-NEXT:    v_readlane_b32 s78, v0, 58
-; GCN-NEXT:    v_readlane_b32 s79, v0, 59
-; GCN-NEXT:    v_readlane_b32 s80, v0, 60
-; GCN-NEXT:    v_readlane_b32 s81, v0, 61
-; GCN-NEXT:    v_readlane_b32 s82, v0, 62
-; GCN-NEXT:    v_readlane_b32 s83, v0, 63
-; GCN-NEXT:    v_readlane_b32 s84, v0, 48
-; GCN-NEXT:    v_readlane_b32 s85, v0, 49
-; GCN-NEXT:    v_readlane_b32 s86, v0, 50
-; GCN-NEXT:    v_readlane_b32 s87, v0, 51
-; GCN-NEXT:    v_readlane_b32 s88, v0, 52
-; GCN-NEXT:    v_readlane_b32 s89, v0, 53
-; GCN-NEXT:    v_readlane_b32 s90, v0, 54
-; GCN-NEXT:    v_readlane_b32 s91, v0, 55
-; GCN-NEXT:    v_readlane_b32 s0, v0, 0
-; GCN-NEXT:    v_readlane_b32 s1, v0, 1
-; GCN-NEXT:    v_readlane_b32 s2, v0, 2
-; GCN-NEXT:    v_readlane_b32 s3, v0, 3
-; GCN-NEXT:    v_readlane_b32 s4, v0, 4
-; GCN-NEXT:    v_readlane_b32 s5, v0, 5
-; GCN-NEXT:    v_readlane_b32 s6, v0, 6
-; GCN-NEXT:    v_readlane_b32 s7, v0, 7
+; GCN-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GCN-NEXT:    buffer_load_dword v0, off, s[92:95], 0 offset:4 ; 4-byte Folded Reload
+; GCN-NEXT:    s_mov_b64 exec, s[34:35]
+; GCN-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GCN-NEXT:    buffer_load_dword v1, off, s[92:95], 0 offset:12 ; 4-byte Folded Reload
+; GCN-NEXT:    s_mov_b64 exec, s[34:35]
+; GCN-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GCN-NEXT:    buffer_load_dword v2, off, s[92:95], 0 offset:8 ; 4-byte Folded Reload
+; GCN-NEXT:    s_mov_b64 exec, s[34:35]
+; GCN-NEXT:    s_waitcnt vmcnt(0)
+; GCN-NEXT:    v_readlane_b32 s8, v2, 56
+; GCN-NEXT:    v_readlane_b32 s9, v2, 57
+; GCN-NEXT:    v_readlane_b32 s10, v2, 58
+; GCN-NEXT:    v_readlane_b32 s11, v2, 59
+; GCN-NEXT:    v_readlane_b32 s12, v2, 60
+; GCN-NEXT:    v_readlane_b32 s13, v2, 61
+; GCN-NEXT:    v_readlane_b32 s14, v2, 62
+; GCN-NEXT:    v_readlane_b32 s15, v2, 63
+; GCN-NEXT:    v_readlane_b32 s16, v2, 48
+; GCN-NEXT:    v_readlane_b32 s17, v2, 49
+; GCN-NEXT:    v_readlane_b32 s18, v2, 50
+; GCN-NEXT:    v_readlane_b32 s19, v2, 51
+; GCN-NEXT:    v_readlane_b32 s20, v2, 52
+; GCN-NEXT:    v_readlane_b32 s21, v2, 53
+; GCN-NEXT:    v_readlane_b32 s22, v2, 54
+; GCN-NEXT:    v_readlane_b32 s23, v2, 55
+; GCN-NEXT:    v_readlane_b32 s24, v2, 40
+; GCN-NEXT:    v_readlane_b32 s25, v2, 41
+; GCN-NEXT:    v_readlane_b32 s26, v2, 42
+; GCN-NEXT:    v_readlane_b32 s27, v2, 43
+; GCN-NEXT:    v_readlane_b32 s28, v2, 44
+; GCN-NEXT:    v_readlane_b32 s29, v2, 45
+; GCN-NEXT:    v_readlane_b32 s30, v2, 46
+; GCN-NEXT:    v_readlane_b32 s31, v2, 47
+; GCN-NEXT:    v_readlane_b32 s36, v2, 32
+; GCN-NEXT:    v_readlane_b32 s37, v2, 33
+; GCN-NEXT:    v_readlane_b32 s38, v2, 34
+; GCN-NEXT:    v_readlane_b32 s39, v2, 35
+; GCN-NEXT:    v_readlane_b32 s40, v2, 36
+; GCN-NEXT:    v_readlane_b32 s41, v2, 37
+; GCN-NEXT:    v_readlane_b32 s42, v2, 38
+; GCN-NEXT:    v_readlane_b32 s43, v2, 39
+; GCN-NEXT:    v_readlane_b32 s44, v2, 24
+; GCN-NEXT:    v_readlane_b32 s45, v2, 25
+; GCN-NEXT:    v_readlane_b32 s46, v2, 26
+; GCN-NEXT:    v_readlane_b32 s47, v2, 27
+; GCN-NEXT:    v_readlane_b32 s48, v2, 28
+; GCN-NEXT:    v_readlane_b32 s49, v2, 29
+; GCN-NEXT:    v_readlane_b32 s50, v2, 30
+; GCN-NEXT:    v_readlane_b32 s51, v2, 31
+; GCN-NEXT:    v_readlane_b32 s52, v2, 16
+; GCN-NEXT:    v_readlane_b32 s53, v2, 17
+; GCN-NEXT:    v_readlane_b32 s54, v2, 18
+; GCN-NEXT:    v_readlane_b32 s55, v2, 19
+; GCN-NEXT:    v_readlane_b32 s56, v2, 20
+; GCN-NEXT:    v_readlane_b32 s57, v2, 21
+; GCN-NEXT:    v_readlane_b32 s58, v2, 22
+; GCN-NEXT:    v_readlane_b32 s59, v2, 23
+; GCN-NEXT:    v_readlane_b32 s60, v2, 8
+; GCN-NEXT:    v_readlane_b32 s61, v2, 9
+; GCN-NEXT:    v_readlane_b32 s62, v2, 10
+; GCN-NEXT:    v_readlane_b32 s63, v2, 11
+; GCN-NEXT:    v_readlane_b32 s64, v2, 12
+; GCN-NEXT:    v_readlane_b32 s65, v2, 13
+; GCN-NEXT:    v_readlane_b32 s66, v2, 14
+; GCN-NEXT:    v_readlane_b32 s67, v2, 15
+; GCN-NEXT:    v_readlane_b32 s68, v2, 0
+; GCN-NEXT:    v_readlane_b32 s69, v2, 1
+; GCN-NEXT:    v_readlane_b32 s70, v2, 2
+; GCN-NEXT:    v_readlane_b32 s71, v2, 3
+; GCN-NEXT:    v_readlane_b32 s72, v2, 4
+; GCN-NEXT:    v_readlane_b32 s73, v2, 5
+; GCN-NEXT:    v_readlane_b32 s74, v2, 6
+; GCN-NEXT:    v_readlane_b32 s75, v2, 7
+; GCN-NEXT:    v_readlane_b32 s76, v1, 56
+; GCN-NEXT:    v_readlane_b32 s77, v1, 57
+; GCN-NEXT:    v_readlane_b32 s78, v1, 58
+; GCN-NEXT:    v_readlane_b32 s79, v1, 59
+; GCN-NEXT:    v_readlane_b32 s80, v1, 60
+; GCN-NEXT:    v_readlane_b32 s81, v1, 61
+; GCN-NEXT:    v_readlane_b32 s82, v1, 62
+; GCN-NEXT:    v_readlane_b32 s83, v1, 63
+; GCN-NEXT:    v_readlane_b32 s84, v1, 48
+; GCN-NEXT:    v_readlane_b32 s85, v1, 49
+; GCN-NEXT:    v_readlane_b32 s86, v1, 50
+; GCN-NEXT:    v_readlane_b32 s87, v1, 51
+; GCN-NEXT:    v_readlane_b32 s88, v1, 52
+; GCN-NEXT:    v_readlane_b32 s89, v1, 53
+; GCN-NEXT:    v_readlane_b32 s90, v1, 54
+; GCN-NEXT:    v_readlane_b32 s91, v1, 55
+; GCN-NEXT:    v_readlane_b32 s0, v1, 0
+; GCN-NEXT:    v_readlane_b32 s1, v1, 1
+; GCN-NEXT:    v_readlane_b32 s2, v1, 2
+; GCN-NEXT:    v_readlane_b32 s3, v1, 3
+; GCN-NEXT:    v_readlane_b32 s4, v1, 4
+; GCN-NEXT:    v_readlane_b32 s5, v1, 5
+; GCN-NEXT:    v_readlane_b32 s6, v1, 6
+; GCN-NEXT:    v_readlane_b32 s7, v1, 7
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; use s[0:7]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_readlane_b32 s0, v0, 8
-; GCN-NEXT:    v_readlane_b32 s1, v0, 9
-; GCN-NEXT:    v_readlane_b32 s2, v0, 10
-; GCN-NEXT:    v_readlane_b32 s3, v0, 11
-; GCN-NEXT:    v_readlane_b32 s4, v0, 12
-; GCN-NEXT:    v_readlane_b32 s5, v0, 13
-; GCN-NEXT:    v_readlane_b32 s6, v0, 14
-; GCN-NEXT:    v_readlane_b32 s7, v0, 15
+; GCN-NEXT:    v_readlane_b32 s0, v1, 8
+; GCN-NEXT:    v_readlane_b32 s1, v1, 9
+; GCN-NEXT:    v_readlane_b32 s2, v1, 10
+; GCN-NEXT:    v_readlane_b32 s3, v1, 11
+; GCN-NEXT:    v_readlane_b32 s4, v1, 12
+; GCN-NEXT:    v_readlane_b32 s5, v1, 13
+; GCN-NEXT:    v_readlane_b32 s6, v1, 14
+; GCN-NEXT:    v_readlane_b32 s7, v1, 15
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; use s[0:7]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_readlane_b32 s0, v0, 16
-; GCN-NEXT:    v_readlane_b32 s1, v0, 17
-; GCN-NEXT:    v_readlane_b32 s2, v0, 18
-; GCN-NEXT:    v_readlane_b32 s3, v0, 19
-; GCN-NEXT:    v_readlane_b32 s4, v0, 20
-; GCN-NEXT:    v_readlane_b32 s5, v0, 21
-; GCN-NEXT:    v_readlane_b32 s6, v0, 22
-; GCN-NEXT:    v_readlane_b32 s7, v0, 23
+; GCN-NEXT:    v_readlane_b32 s0, v1, 16
+; GCN-NEXT:    v_readlane_b32 s1, v1, 17
+; GCN-NEXT:    v_readlane_b32 s2, v1, 18
+; GCN-NEXT:    v_readlane_b32 s3, v1, 19
+; GCN-NEXT:    v_readlane_b32 s4, v1, 20
+; GCN-NEXT:    v_readlane_b32 s5, v1, 21
+; GCN-NEXT:    v_readlane_b32 s6, v1, 22
+; GCN-NEXT:    v_readlane_b32 s7, v1, 23
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; use s[0:7]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_readlane_b32 s0, v0, 24
-; GCN-NEXT:    v_readlane_b32 s1, v0, 25
-; GCN-NEXT:    v_readlane_b32 s2, v0, 26
-; GCN-NEXT:    v_readlane_b32 s3, v0, 27
-; GCN-NEXT:    v_readlane_b32 s4, v0, 28
-; GCN-NEXT:    v_readlane_b32 s5, v0, 29
-; GCN-NEXT:    v_readlane_b32 s6, v0, 30
-; GCN-NEXT:    v_readlane_b32 s7, v0, 31
+; GCN-NEXT:    v_readlane_b32 s0, v1, 24
+; GCN-NEXT:    v_readlane_b32 s1, v1, 25
+; GCN-NEXT:    v_readlane_b32 s2, v1, 26
+; GCN-NEXT:    v_readlane_b32 s3, v1, 27
+; GCN-NEXT:    v_readlane_b32 s4, v1, 28
+; GCN-NEXT:    v_readlane_b32 s5, v1, 29
+; GCN-NEXT:    v_readlane_b32 s6, v1, 30
+; GCN-NEXT:    v_readlane_b32 s7, v1, 31
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; use s[0:7]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_readlane_b32 s0, v0, 32
-; GCN-NEXT:    v_readlane_b32 s1, v0, 33
-; GCN-NEXT:    v_readlane_b32 s2, v0, 34
-; GCN-NEXT:    v_readlane_b32 s3, v0, 35
-; GCN-NEXT:    v_readlane_b32 s4, v0, 36
-; GCN-NEXT:    v_readlane_b32 s5, v0, 37
-; GCN-NEXT:    v_readlane_b32 s6, v0, 38
-; GCN-NEXT:    v_readlane_b32 s7, v0, 39
+; GCN-NEXT:    v_readlane_b32 s0, v1, 32
+; GCN-NEXT:    v_readlane_b32 s1, v1, 33
+; GCN-NEXT:    v_readlane_b32 s2, v1, 34
+; GCN-NEXT:    v_readlane_b32 s3, v1, 35
+; GCN-NEXT:    v_readlane_b32 s4, v1, 36
+; GCN-NEXT:    v_readlane_b32 s5, v1, 37
+; GCN-NEXT:    v_readlane_b32 s6, v1, 38
+; GCN-NEXT:    v_readlane_b32 s7, v1, 39
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; use s[0:7]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_readlane_b32 s0, v0, 40
-; GCN-NEXT:    v_readlane_b32 s1, v0, 41
-; GCN-NEXT:    v_readlane_b32 s2, v0, 42
-; GCN-NEXT:    v_readlane_b32 s3, v0, 43
-; GCN-NEXT:    v_readlane_b32 s4, v0, 44
-; GCN-NEXT:    v_readlane_b32 s5, v0, 45
-; GCN-NEXT:    v_readlane_b32 s6, v0, 46
-; GCN-NEXT:    v_readlane_b32 s7, v0, 47
+; GCN-NEXT:    v_readlane_b32 s0, v1, 40
+; GCN-NEXT:    v_readlane_b32 s1, v1, 41
+; GCN-NEXT:    v_readlane_b32 s2, v1, 42
+; GCN-NEXT:    v_readlane_b32 s3, v1, 43
+; GCN-NEXT:    v_readlane_b32 s4, v1, 44
+; GCN-NEXT:    v_readlane_b32 s5, v1, 45
+; GCN-NEXT:    v_readlane_b32 s6, v1, 46
+; GCN-NEXT:    v_readlane_b32 s7, v1, 47
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; use s[0:7]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_readlane_b32 s0, v2, 0
-; GCN-NEXT:    v_readlane_b32 s1, v2, 1
-; GCN-NEXT:    v_readlane_b32 s2, v2, 2
-; GCN-NEXT:    v_readlane_b32 s3, v2, 3
-; GCN-NEXT:    v_readlane_b32 s4, v2, 4
-; GCN-NEXT:    v_readlane_b32 s5, v2, 5
-; GCN-NEXT:    v_readlane_b32 s6, v2, 6
-; GCN-NEXT:    v_readlane_b32 s7, v2, 7
+; GCN-NEXT:    v_readlane_b32 s0, v0, 0
+; GCN-NEXT:    v_readlane_b32 s1, v0, 1
+; GCN-NEXT:    v_readlane_b32 s2, v0, 2
+; GCN-NEXT:    v_readlane_b32 s3, v0, 3
+; GCN-NEXT:    v_readlane_b32 s4, v0, 4
+; GCN-NEXT:    v_readlane_b32 s5, v0, 5
+; GCN-NEXT:    v_readlane_b32 s6, v0, 6
+; GCN-NEXT:    v_readlane_b32 s7, v0, 7
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; use s[84:91]
 ; GCN-NEXT:    ;;#ASMEND
@@ -442,10 +470,17 @@
 define amdgpu_kernel void @split_sgpr_spill_2_vgprs(i32 addrspace(1)* %out, i32 %in) #1 {
 ; GCN-LABEL: split_sgpr_spill_2_vgprs:
 ; GCN:       ; %bb.0:
+; GCN-NEXT:    s_mov_b32 s52, SCRATCH_RSRC_DWORD0
+; GCN-NEXT:    s_mov_b32 s53, SCRATCH_RSRC_DWORD1
+; GCN-NEXT:    s_mov_b32 s54, -1
+; GCN-NEXT:    s_mov_b32 s55, 0xe8f000
+; GCN-NEXT:    s_add_u32 s52, s52, s3
+; GCN-NEXT:    s_addc_u32 s53, s53, 0
 ; GCN-NEXT:    s_load_dword s0, s[0:1], 0xb
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s[4:19]
 ; GCN-NEXT:    ;;#ASMEND
+; GCN-NEXT:    ; implicit-def: $vgpr0
 ; GCN-NEXT:    v_writelane_b32 v0, s4, 0
 ; GCN-NEXT:    v_writelane_b32 v0, s5, 1
 ; GCN-NEXT:    v_writelane_b32 v0, s6, 2
@@ -519,27 +554,41 @@
 ; GCN-NEXT:    v_writelane_b32 v0, s17, 61
 ; GCN-NEXT:    v_writelane_b32 v0, s18, 62
 ; GCN-NEXT:    v_writelane_b32 v0, s19, 63
+; GCN-NEXT:    s_or_saveexec_b64 s[28:29], -1
+; GCN-NEXT:    buffer_store_dword v0, off, s[52:55], 0 offset:8 ; 4-byte Folded Spill
+; GCN-NEXT:    s_mov_b64 exec, s[28:29]
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s[4:11]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_writelane_b32 v1, s4, 0
-; GCN-NEXT:    v_writelane_b32 v1, s5, 1
-; GCN-NEXT:    v_writelane_b32 v1, s6, 2
-; GCN-NEXT:    v_writelane_b32 v1, s7, 3
-; GCN-NEXT:    v_writelane_b32 v1, s8, 4
-; GCN-NEXT:    v_writelane_b32 v1, s9, 5
-; GCN-NEXT:    v_writelane_b32 v1, s10, 6
-; GCN-NEXT:    v_writelane_b32 v1, s11, 7
+; GCN-NEXT:    ; implicit-def: $vgpr0
+; GCN-NEXT:    v_writelane_b32 v0, s4, 0
+; GCN-NEXT:    v_writelane_b32 v0, s5, 1
+; GCN-NEXT:    v_writelane_b32 v0, s6, 2
+; GCN-NEXT:    v_writelane_b32 v0, s7, 3
+; GCN-NEXT:    v_writelane_b32 v0, s8, 4
+; GCN-NEXT:    v_writelane_b32 v0, s9, 5
+; GCN-NEXT:    v_writelane_b32 v0, s10, 6
+; GCN-NEXT:    v_writelane_b32 v0, s11, 7
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s[2:3]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_writelane_b32 v1, s2, 8
-; GCN-NEXT:    v_writelane_b32 v1, s3, 9
+; GCN-NEXT:    v_writelane_b32 v0, s2, 8
+; GCN-NEXT:    v_writelane_b32 v0, s3, 9
+; GCN-NEXT:    s_or_saveexec_b64 s[28:29], -1
+; GCN-NEXT:    buffer_store_dword v0, off, s[52:55], 0 offset:4 ; 4-byte Folded Spill
+; GCN-NEXT:    s_mov_b64 exec, s[28:29]
 ; GCN-NEXT:    s_mov_b32 s1, 0
 ; GCN-NEXT:    s_waitcnt lgkmcnt(0)
 ; GCN-NEXT:    s_cmp_lg_u32 s0, s1
 ; GCN-NEXT:    s_cbranch_scc1 .LBB1_2
 ; GCN-NEXT:  ; %bb.1: ; %bb0
+; GCN-NEXT:    s_or_saveexec_b64 s[28:29], -1
+; GCN-NEXT:    buffer_load_dword v0, off, s[52:55], 0 offset:8 ; 4-byte Folded Reload
+; GCN-NEXT:    s_mov_b64 exec, s[28:29]
+; GCN-NEXT:    s_or_saveexec_b64 s[28:29], -1
+; GCN-NEXT:    buffer_load_dword v1, off, s[52:55], 0 offset:4 ; 4-byte Folded Reload
+; GCN-NEXT:    s_mov_b64 exec, s[28:29]
+; GCN-NEXT:    s_waitcnt vmcnt(0)
 ; GCN-NEXT:    v_readlane_b32 s16, v1, 8
 ; GCN-NEXT:    v_readlane_b32 s17, v1, 9
 ; GCN-NEXT:    v_readlane_b32 s20, v1, 0
@@ -685,176 +734,176 @@
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s[4:19]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_writelane_b32 v31, s4, 0
-; GCN-NEXT:    v_writelane_b32 v31, s5, 1
-; GCN-NEXT:    v_writelane_b32 v31, s6, 2
-; GCN-NEXT:    v_writelane_b32 v31, s7, 3
-; GCN-NEXT:    v_writelane_b32 v31, s8, 4
-; GCN-NEXT:    v_writelane_b32 v31, s9, 5
-; GCN-NEXT:    v_writelane_b32 v31, s10, 6
-; GCN-NEXT:    v_writelane_b32 v31, s11, 7
-; GCN-NEXT:    v_writelane_b32 v31, s12, 8
-; GCN-NEXT:    v_writelane_b32 v31, s13, 9
-; GCN-NEXT:    v_writelane_b32 v31, s14, 10
-; GCN-NEXT:    v_writelane_b32 v31, s15, 11
-; GCN-NEXT:    v_writelane_b32 v31, s16, 12
-; GCN-NEXT:    v_writelane_b32 v31, s17, 13
-; GCN-NEXT:    v_writelane_b32 v31, s18, 14
-; GCN-NEXT:    v_writelane_b32 v31, s19, 15
+; GCN-NEXT:    ; implicit-def: $vgpr0
+; GCN-NEXT:    v_writelane_b32 v0, s4, 0
+; GCN-NEXT:    v_writelane_b32 v0, s5, 1
+; GCN-NEXT:    v_writelane_b32 v0, s6, 2
+; GCN-NEXT:    v_writelane_b32 v0, s7, 3
+; GCN-NEXT:    v_writelane_b32 v0, s8, 4
+; GCN-NEXT:    v_writelane_b32 v0, s9, 5
+; GCN-NEXT:    v_writelane_b32 v0, s10, 6
+; GCN-NEXT:    v_writelane_b32 v0, s11, 7
+; GCN-NEXT:    v_writelane_b32 v0, s12, 8
+; GCN-NEXT:    v_writelane_b32 v0, s13, 9
+; GCN-NEXT:    v_writelane_b32 v0, s14, 10
+; GCN-NEXT:    v_writelane_b32 v0, s15, 11
+; GCN-NEXT:    v_writelane_b32 v0, s16, 12
+; GCN-NEXT:    v_writelane_b32 v0, s17, 13
+; GCN-NEXT:    v_writelane_b32 v0, s18, 14
+; GCN-NEXT:    v_writelane_b32 v0, s19, 15
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s[4:19]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_writelane_b32 v31, s4, 16
-; GCN-NEXT:    v_writelane_b32 v31, s5, 17
-; GCN-NEXT:    v_writelane_b32 v31, s6, 18
-; GCN-NEXT:    v_writelane_b32 v31, s7, 19
-; GCN-NEXT:    v_writelane_b32 v31, s8, 20
-; GCN-NEXT:    v_writelane_b32 v31, s9, 21
-; GCN-NEXT:    v_writelane_b32 v31, s10, 22
-; GCN-NEXT:    v_writelane_b32 v31, s11, 23
-; GCN-NEXT:    v_writelane_b32 v31, s12, 24
-; GCN-NEXT:    v_writelane_b32 v31, s13, 25
-; GCN-NEXT:    v_writelane_b32 v31, s14, 26
-; GCN-NEXT:    v_writelane_b32 v31, s15, 27
-; GCN-NEXT:    v_writelane_b32 v31, s16, 28
-; GCN-NEXT:    v_writelane_b32 v31, s17, 29
-; GCN-NEXT:    v_writelane_b32 v31, s18, 30
-; GCN-NEXT:    v_writelane_b32 v31, s19, 31
+; GCN-NEXT:    v_writelane_b32 v0, s4, 16
+; GCN-NEXT:    v_writelane_b32 v0, s5, 17
+; GCN-NEXT:    v_writelane_b32 v0, s6, 18
+; GCN-NEXT:    v_writelane_b32 v0, s7, 19
+; GCN-NEXT:    v_writelane_b32 v0, s8, 20
+; GCN-NEXT:    v_writelane_b32 v0, s9, 21
+; GCN-NEXT:    v_writelane_b32 v0, s10, 22
+; GCN-NEXT:    v_writelane_b32 v0, s11, 23
+; GCN-NEXT:    v_writelane_b32 v0, s12, 24
+; GCN-NEXT:    v_writelane_b32 v0, s13, 25
+; GCN-NEXT:    v_writelane_b32 v0, s14, 26
+; GCN-NEXT:    v_writelane_b32 v0, s15, 27
+; GCN-NEXT:    v_writelane_b32 v0, s16, 28
+; GCN-NEXT:    v_writelane_b32 v0, s17, 29
+; GCN-NEXT:    v_writelane_b32 v0, s18, 30
+; GCN-NEXT:    v_writelane_b32 v0, s19, 31
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s[4:19]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_writelane_b32 v31, s4, 32
-; GCN-NEXT:    v_writelane_b32 v31, s5, 33
-; GCN-NEXT:    v_writelane_b32 v31, s6, 34
-; GCN-NEXT:    v_writelane_b32 v31, s7, 35
-; GCN-NEXT:    v_writelane_b32 v31, s8, 36
-; GCN-NEXT:    v_writelane_b32 v31, s9, 37
-; GCN-NEXT:    v_writelane_b32 v31, s10, 38
-; GCN-NEXT:    v_writelane_b32 v31, s11, 39
-; GCN-NEXT:    v_writelane_b32 v31, s12, 40
-; GCN-NEXT:    v_writelane_b32 v31, s13, 41
-; GCN-NEXT:    v_writelane_b32 v31, s14, 42
-; GCN-NEXT:    v_writelane_b32 v31, s15, 43
-; GCN-NEXT:    v_writelane_b32 v31, s16, 44
-; GCN-NEXT:    v_writelane_b32 v31, s17, 45
-; GCN-NEXT:    v_writelane_b32 v31, s18, 46
-; GCN-NEXT:    v_writelane_b32 v31, s19, 47
+; GCN-NEXT:    v_writelane_b32 v0, s4, 32
+; GCN-NEXT:    v_writelane_b32 v0, s5, 33
+; GCN-NEXT:    v_writelane_b32 v0, s6, 34
+; GCN-NEXT:    v_writelane_b32 v0, s7, 35
+; GCN-NEXT:    v_writelane_b32 v0, s8, 36
+; GCN-NEXT:    v_writelane_b32 v0, s9, 37
+; GCN-NEXT:    v_writelane_b32 v0, s10, 38
+; GCN-NEXT:    v_writelane_b32 v0, s11, 39
+; GCN-NEXT:    v_writelane_b32 v0, s12, 40
+; GCN-NEXT:    v_writelane_b32 v0, s13, 41
+; GCN-NEXT:    v_writelane_b32 v0, s14, 42
+; GCN-NEXT:    v_writelane_b32 v0, s15, 43
+; GCN-NEXT:    v_writelane_b32 v0, s16, 44
+; GCN-NEXT:    v_writelane_b32 v0, s17, 45
+; GCN-NEXT:    v_writelane_b32 v0, s18, 46
+; GCN-NEXT:    v_writelane_b32 v0, s19, 47
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s[4:19]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_writelane_b32 v31, s4, 48
-; GCN-NEXT:    v_writelane_b32 v31, s5, 49
-; GCN-NEXT:    v_writelane_b32 v31, s6, 50
-; GCN-NEXT:    v_writelane_b32 v31, s7, 51
-; GCN-NEXT:    v_writelane_b32 v31, s8, 52
-; GCN-NEXT:    v_writelane_b32 v31, s9, 53
-; GCN-NEXT:    v_writelane_b32 v31, s10, 54
-; GCN-NEXT:    v_writelane_b32 v31, s11, 55
-; GCN-NEXT:    v_writelane_b32 v31, s12, 56
-; GCN-NEXT:    v_writelane_b32 v31, s13, 57
-; GCN-NEXT:    v_writelane_b32 v31, s14, 58
-; GCN-NEXT:    v_writelane_b32 v31, s15, 59
-; GCN-NEXT:    v_writelane_b32 v31, s16, 60
-; GCN-NEXT:    v_writelane_b32 v31, s17, 61
-; GCN-NEXT:    v_writelane_b32 v31, s18, 62
-; GCN-NEXT:    v_writelane_b32 v31, s19, 63
+; GCN-NEXT:    v_writelane_b32 v0, s4, 48
+; GCN-NEXT:    v_writelane_b32 v0, s5, 49
+; GCN-NEXT:    v_writelane_b32 v0, s6, 50
+; GCN-NEXT:    v_writelane_b32 v0, s7, 51
+; GCN-NEXT:    v_writelane_b32 v0, s8, 52
+; GCN-NEXT:    v_writelane_b32 v0, s9, 53
+; GCN-NEXT:    v_writelane_b32 v0, s10, 54
+; GCN-NEXT:    v_writelane_b32 v0, s11, 55
+; GCN-NEXT:    v_writelane_b32 v0, s12, 56
+; GCN-NEXT:    v_writelane_b32 v0, s13, 57
+; GCN-NEXT:    v_writelane_b32 v0, s14, 58
+; GCN-NEXT:    v_writelane_b32 v0, s15, 59
+; GCN-NEXT:    v_writelane_b32 v0, s16, 60
+; GCN-NEXT:    v_writelane_b32 v0, s17, 61
+; GCN-NEXT:    v_writelane_b32 v0, s18, 62
+; GCN-NEXT:    v_writelane_b32 v0, s19, 63
+; GCN-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GCN-NEXT:    buffer_store_dword v0, off, s[52:55], 0 offset:8 ; 4-byte Folded Spill
+; GCN-NEXT:    s_mov_b64 exec, s[34:35]
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s[2:3]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    s_mov_b64 s[4:5], exec
-; GCN-NEXT:    s_mov_b64 exec, 3
-; GCN-NEXT:    buffer_store_dword v0, off, s[52:55], 0
+; GCN-NEXT:    ; implicit-def: $vgpr0
 ; GCN-NEXT:    v_writelane_b32 v0, s2, 0
 ; GCN-NEXT:    v_writelane_b32 v0, s3, 1
+; GCN-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GCN-NEXT:    buffer_store_dword v0, off, s[52:55], 0 offset:4 ; 4-byte Folded Spill
-; GCN-NEXT:    buffer_load_dword v0, off, s[52:55], 0
-; GCN-NEXT:    s_waitcnt vmcnt(0)
-; GCN-NEXT:    s_mov_b64 exec, s[4:5]
+; GCN-NEXT:    s_mov_b64 exec, s[34:35]
 ; GCN-NEXT:    s_mov_b32 s1, 0
 ; GCN-NEXT:    s_waitcnt lgkmcnt(0)
 ; GCN-NEXT:    s_cmp_lg_u32 s0, s1
 ; GCN-NEXT:    s_cbranch_scc1 .LBB2_2
 ; GCN-NEXT:  ; %bb.1: ; %bb0
-; GCN-NEXT:    v_readlane_b32 s36, v31, 32
-; GCN-NEXT:    v_readlane_b32 s37, v31, 33
-; GCN-NEXT:    v_readlane_b32 s38, v31, 34
-; GCN-NEXT:    v_readlane_b32 s39, v31, 35
-; GCN-NEXT:    v_readlane_b32 s40, v31, 36
-; GCN-NEXT:    v_readlane_b32 s41, v31, 37
-; GCN-NEXT:    v_readlane_b32 s42, v31, 38
-; GCN-NEXT:    v_readlane_b32 s43, v31, 39
-; GCN-NEXT:    v_readlane_b32 s44, v31, 40
-; GCN-NEXT:    v_readlane_b32 s45, v31, 41
-; GCN-NEXT:    v_readlane_b32 s46, v31, 42
-; GCN-NEXT:    v_readlane_b32 s47, v31, 43
-; GCN-NEXT:    v_readlane_b32 s48, v31, 44
-; GCN-NEXT:    v_readlane_b32 s49, v31, 45
-; GCN-NEXT:    v_readlane_b32 s50, v31, 46
-; GCN-NEXT:    v_readlane_b32 s51, v31, 47
-; GCN-NEXT:    v_readlane_b32 s0, v31, 16
-; GCN-NEXT:    v_readlane_b32 s1, v31, 17
-; GCN-NEXT:    v_readlane_b32 s2, v31, 18
-; GCN-NEXT:    v_readlane_b32 s3, v31, 19
-; GCN-NEXT:    v_readlane_b32 s4, v31, 20
-; GCN-NEXT:    v_readlane_b32 s5, v31, 21
-; GCN-NEXT:    v_readlane_b32 s6, v31, 22
-; GCN-NEXT:    v_readlane_b32 s7, v31, 23
-; GCN-NEXT:    v_readlane_b32 s8, v31, 24
-; GCN-NEXT:    v_readlane_b32 s9, v31, 25
-; GCN-NEXT:    v_readlane_b32 s10, v31, 26
-; GCN-NEXT:    v_readlane_b32 s11, v31, 27
-; GCN-NEXT:    v_readlane_b32 s12, v31, 28
-; GCN-NEXT:    v_readlane_b32 s13, v31, 29
-; GCN-NEXT:    v_readlane_b32 s14, v31, 30
-; GCN-NEXT:    v_readlane_b32 s15, v31, 31
-; GCN-NEXT:    v_readlane_b32 s16, v31, 0
-; GCN-NEXT:    v_readlane_b32 s17, v31, 1
-; GCN-NEXT:    v_readlane_b32 s18, v31, 2
-; GCN-NEXT:    v_readlane_b32 s19, v31, 3
-; GCN-NEXT:    v_readlane_b32 s20, v31, 4
-; GCN-NEXT:    v_readlane_b32 s21, v31, 5
-; GCN-NEXT:    v_readlane_b32 s22, v31, 6
-; GCN-NEXT:    v_readlane_b32 s23, v31, 7
-; GCN-NEXT:    v_readlane_b32 s24, v31, 8
-; GCN-NEXT:    v_readlane_b32 s25, v31, 9
-; GCN-NEXT:    v_readlane_b32 s26, v31, 10
-; GCN-NEXT:    v_readlane_b32 s27, v31, 11
-; GCN-NEXT:    v_readlane_b32 s28, v31, 12
-; GCN-NEXT:    v_readlane_b32 s29, v31, 13
-; GCN-NEXT:    v_readlane_b32 s30, v31, 14
-; GCN-NEXT:    v_readlane_b32 s31, v31, 15
+; GCN-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GCN-NEXT:    buffer_load_dword v0, off, s[52:55], 0 offset:4 ; 4-byte Folded Reload
+; GCN-NEXT:    s_mov_b64 exec, s[34:35]
+; GCN-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GCN-NEXT:    buffer_load_dword v1, off, s[52:55], 0 offset:8 ; 4-byte Folded Reload
+; GCN-NEXT:    s_mov_b64 exec, s[34:35]
+; GCN-NEXT:    s_waitcnt vmcnt(0)
+; GCN-NEXT:    v_readlane_b32 s36, v1, 32
+; GCN-NEXT:    v_readlane_b32 s37, v1, 33
+; GCN-NEXT:    v_readlane_b32 s38, v1, 34
+; GCN-NEXT:    v_readlane_b32 s39, v1, 35
+; GCN-NEXT:    v_readlane_b32 s40, v1, 36
+; GCN-NEXT:    v_readlane_b32 s41, v1, 37
+; GCN-NEXT:    v_readlane_b32 s42, v1, 38
+; GCN-NEXT:    v_readlane_b32 s43, v1, 39
+; GCN-NEXT:    v_readlane_b32 s44, v1, 40
+; GCN-NEXT:    v_readlane_b32 s45, v1, 41
+; GCN-NEXT:    v_readlane_b32 s46, v1, 42
+; GCN-NEXT:    v_readlane_b32 s47, v1, 43
+; GCN-NEXT:    v_readlane_b32 s48, v1, 44
+; GCN-NEXT:    v_readlane_b32 s49, v1, 45
+; GCN-NEXT:    v_readlane_b32 s50, v1, 46
+; GCN-NEXT:    v_readlane_b32 s51, v1, 47
+; GCN-NEXT:    v_readlane_b32 s0, v1, 16
+; GCN-NEXT:    v_readlane_b32 s1, v1, 17
+; GCN-NEXT:    v_readlane_b32 s2, v1, 18
+; GCN-NEXT:    v_readlane_b32 s3, v1, 19
+; GCN-NEXT:    v_readlane_b32 s4, v1, 20
+; GCN-NEXT:    v_readlane_b32 s5, v1, 21
+; GCN-NEXT:    v_readlane_b32 s6, v1, 22
+; GCN-NEXT:    v_readlane_b32 s7, v1, 23
+; GCN-NEXT:    v_readlane_b32 s8, v1, 24
+; GCN-NEXT:    v_readlane_b32 s9, v1, 25
+; GCN-NEXT:    v_readlane_b32 s10, v1, 26
+; GCN-NEXT:    v_readlane_b32 s11, v1, 27
+; GCN-NEXT:    v_readlane_b32 s12, v1, 28
+; GCN-NEXT:    v_readlane_b32 s13, v1, 29
+; GCN-NEXT:    v_readlane_b32 s14, v1, 30
+; GCN-NEXT:    v_readlane_b32 s15, v1, 31
+; GCN-NEXT:    v_readlane_b32 s16, v1, 0
+; GCN-NEXT:    v_readlane_b32 s17, v1, 1
+; GCN-NEXT:    v_readlane_b32 s18, v1, 2
+; GCN-NEXT:    v_readlane_b32 s19, v1, 3
+; GCN-NEXT:    v_readlane_b32 s20, v1, 4
+; GCN-NEXT:    v_readlane_b32 s21, v1, 5
+; GCN-NEXT:    v_readlane_b32 s22, v1, 6
+; GCN-NEXT:    v_readlane_b32 s23, v1, 7
+; GCN-NEXT:    v_readlane_b32 s24, v1, 8
+; GCN-NEXT:    v_readlane_b32 s25, v1, 9
+; GCN-NEXT:    v_readlane_b32 s26, v1, 10
+; GCN-NEXT:    v_readlane_b32 s27, v1, 11
+; GCN-NEXT:    v_readlane_b32 s28, v1, 12
+; GCN-NEXT:    v_readlane_b32 s29, v1, 13
+; GCN-NEXT:    v_readlane_b32 s30, v1, 14
+; GCN-NEXT:    v_readlane_b32 s31, v1, 15
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; use s[16:31]
 ; GCN-NEXT:    ;;#ASMEND
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; use s[0:15]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_readlane_b32 s4, v31, 48
-; GCN-NEXT:    v_readlane_b32 s5, v31, 49
-; GCN-NEXT:    v_readlane_b32 s6, v31, 50
-; GCN-NEXT:    v_readlane_b32 s7, v31, 51
-; GCN-NEXT:    v_readlane_b32 s8, v31, 52
-; GCN-NEXT:    v_readlane_b32 s9, v31, 53
-; GCN-NEXT:    v_readlane_b32 s10, v31, 54
-; GCN-NEXT:    v_readlane_b32 s11, v31, 55
-; GCN-NEXT:    v_readlane_b32 s12, v31, 56
-; GCN-NEXT:    v_readlane_b32 s13, v31, 57
-; GCN-NEXT:    v_readlane_b32 s14, v31, 58
-; GCN-NEXT:    v_readlane_b32 s15, v31, 59
-; GCN-NEXT:    v_readlane_b32 s16, v31, 60
-; GCN-NEXT:    v_readlane_b32 s17, v31, 61
-; GCN-NEXT:    v_readlane_b32 s18, v31, 62
-; GCN-NEXT:    v_readlane_b32 s19, v31, 63
-; GCN-NEXT:    s_mov_b64 s[2:3], exec
-; GCN-NEXT:    s_mov_b64 exec, 3
-; GCN-NEXT:    buffer_store_dword v0, off, s[52:55], 0
-; GCN-NEXT:    buffer_load_dword v0, off, s[52:55], 0 offset:4 ; 4-byte Folded Reload
-; GCN-NEXT:    s_waitcnt vmcnt(0)
+; GCN-NEXT:    v_readlane_b32 s4, v1, 48
+; GCN-NEXT:    v_readlane_b32 s5, v1, 49
+; GCN-NEXT:    v_readlane_b32 s6, v1, 50
+; GCN-NEXT:    v_readlane_b32 s7, v1, 51
+; GCN-NEXT:    v_readlane_b32 s8, v1, 52
+; GCN-NEXT:    v_readlane_b32 s9, v1, 53
+; GCN-NEXT:    v_readlane_b32 s10, v1, 54
+; GCN-NEXT:    v_readlane_b32 s11, v1, 55
+; GCN-NEXT:    v_readlane_b32 s12, v1, 56
+; GCN-NEXT:    v_readlane_b32 s13, v1, 57
+; GCN-NEXT:    v_readlane_b32 s14, v1, 58
+; GCN-NEXT:    v_readlane_b32 s15, v1, 59
+; GCN-NEXT:    v_readlane_b32 s16, v1, 60
+; GCN-NEXT:    v_readlane_b32 s17, v1, 61
+; GCN-NEXT:    v_readlane_b32 s18, v1, 62
+; GCN-NEXT:    v_readlane_b32 s19, v1, 63
 ; GCN-NEXT:    v_readlane_b32 s0, v0, 0
 ; GCN-NEXT:    v_readlane_b32 s1, v0, 1
-; GCN-NEXT:    buffer_load_dword v0, off, s[52:55], 0
-; GCN-NEXT:    s_waitcnt vmcnt(0)
-; GCN-NEXT:    s_mov_b64 exec, s[2:3]
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; use s[36:51]
 ; GCN-NEXT:    ;;#ASMEND
@@ -920,144 +969,152 @@
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s[4:19]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_writelane_b32 v31, s4, 0
-; GCN-NEXT:    v_writelane_b32 v31, s5, 1
-; GCN-NEXT:    v_writelane_b32 v31, s6, 2
-; GCN-NEXT:    v_writelane_b32 v31, s7, 3
-; GCN-NEXT:    v_writelane_b32 v31, s8, 4
-; GCN-NEXT:    v_writelane_b32 v31, s9, 5
-; GCN-NEXT:    v_writelane_b32 v31, s10, 6
-; GCN-NEXT:    v_writelane_b32 v31, s11, 7
-; GCN-NEXT:    v_writelane_b32 v31, s12, 8
-; GCN-NEXT:    v_writelane_b32 v31, s13, 9
-; GCN-NEXT:    v_writelane_b32 v31, s14, 10
-; GCN-NEXT:    v_writelane_b32 v31, s15, 11
-; GCN-NEXT:    v_writelane_b32 v31, s16, 12
-; GCN-NEXT:    v_writelane_b32 v31, s17, 13
-; GCN-NEXT:    v_writelane_b32 v31, s18, 14
-; GCN-NEXT:    v_writelane_b32 v31, s19, 15
+; GCN-NEXT:    ; implicit-def: $vgpr0
+; GCN-NEXT:    v_writelane_b32 v0, s4, 0
+; GCN-NEXT:    v_writelane_b32 v0, s5, 1
+; GCN-NEXT:    v_writelane_b32 v0, s6, 2
+; GCN-NEXT:    v_writelane_b32 v0, s7, 3
+; GCN-NEXT:    v_writelane_b32 v0, s8, 4
+; GCN-NEXT:    v_writelane_b32 v0, s9, 5
+; GCN-NEXT:    v_writelane_b32 v0, s10, 6
+; GCN-NEXT:    v_writelane_b32 v0, s11, 7
+; GCN-NEXT:    v_writelane_b32 v0, s12, 8
+; GCN-NEXT:    v_writelane_b32 v0, s13, 9
+; GCN-NEXT:    v_writelane_b32 v0, s14, 10
+; GCN-NEXT:    v_writelane_b32 v0, s15, 11
+; GCN-NEXT:    v_writelane_b32 v0, s16, 12
+; GCN-NEXT:    v_writelane_b32 v0, s17, 13
+; GCN-NEXT:    v_writelane_b32 v0, s18, 14
+; GCN-NEXT:    v_writelane_b32 v0, s19, 15
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s[4:19]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_writelane_b32 v31, s4, 16
-; GCN-NEXT:    v_writelane_b32 v31, s5, 17
-; GCN-NEXT:    v_writelane_b32 v31, s6, 18
-; GCN-NEXT:    v_writelane_b32 v31, s7, 19
-; GCN-NEXT:    v_writelane_b32 v31, s8, 20
-; GCN-NEXT:    v_writelane_b32 v31, s9, 21
-; GCN-NEXT:    v_writelane_b32 v31, s10, 22
-; GCN-NEXT:    v_writelane_b32 v31, s11, 23
-; GCN-NEXT:    v_writelane_b32 v31, s12, 24
-; GCN-NEXT:    v_writelane_b32 v31, s13, 25
-; GCN-NEXT:    v_writelane_b32 v31, s14, 26
-; GCN-NEXT:    v_writelane_b32 v31, s15, 27
-; GCN-NEXT:    v_writelane_b32 v31, s16, 28
-; GCN-NEXT:    v_writelane_b32 v31, s17, 29
-; GCN-NEXT:    v_writelane_b32 v31, s18, 30
-; GCN-NEXT:    v_writelane_b32 v31, s19, 31
+; GCN-NEXT:    v_writelane_b32 v0, s4, 16
+; GCN-NEXT:    v_writelane_b32 v0, s5, 17
+; GCN-NEXT:    v_writelane_b32 v0, s6, 18
+; GCN-NEXT:    v_writelane_b32 v0, s7, 19
+; GCN-NEXT:    v_writelane_b32 v0, s8, 20
+; GCN-NEXT:    v_writelane_b32 v0, s9, 21
+; GCN-NEXT:    v_writelane_b32 v0, s10, 22
+; GCN-NEXT:    v_writelane_b32 v0, s11, 23
+; GCN-NEXT:    v_writelane_b32 v0, s12, 24
+; GCN-NEXT:    v_writelane_b32 v0, s13, 25
+; GCN-NEXT:    v_writelane_b32 v0, s14, 26
+; GCN-NEXT:    v_writelane_b32 v0, s15, 27
+; GCN-NEXT:    v_writelane_b32 v0, s16, 28
+; GCN-NEXT:    v_writelane_b32 v0, s17, 29
+; GCN-NEXT:    v_writelane_b32 v0, s18, 30
+; GCN-NEXT:    v_writelane_b32 v0, s19, 31
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s[4:19]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_writelane_b32 v31, s4, 32
-; GCN-NEXT:    v_writelane_b32 v31, s5, 33
-; GCN-NEXT:    v_writelane_b32 v31, s6, 34
-; GCN-NEXT:    v_writelane_b32 v31, s7, 35
-; GCN-NEXT:    v_writelane_b32 v31, s8, 36
-; GCN-NEXT:    v_writelane_b32 v31, s9, 37
-; GCN-NEXT:    v_writelane_b32 v31, s10, 38
-; GCN-NEXT:    v_writelane_b32 v31, s11, 39
-; GCN-NEXT:    v_writelane_b32 v31, s12, 40
-; GCN-NEXT:    v_writelane_b32 v31, s13, 41
-; GCN-NEXT:    v_writelane_b32 v31, s14, 42
-; GCN-NEXT:    v_writelane_b32 v31, s15, 43
-; GCN-NEXT:    v_writelane_b32 v31, s16, 44
-; GCN-NEXT:    v_writelane_b32 v31, s17, 45
-; GCN-NEXT:    v_writelane_b32 v31, s18, 46
-; GCN-NEXT:    v_writelane_b32 v31, s19, 47
+; GCN-NEXT:    v_writelane_b32 v0, s4, 32
+; GCN-NEXT:    v_writelane_b32 v0, s5, 33
+; GCN-NEXT:    v_writelane_b32 v0, s6, 34
+; GCN-NEXT:    v_writelane_b32 v0, s7, 35
+; GCN-NEXT:    v_writelane_b32 v0, s8, 36
+; GCN-NEXT:    v_writelane_b32 v0, s9, 37
+; GCN-NEXT:    v_writelane_b32 v0, s10, 38
+; GCN-NEXT:    v_writelane_b32 v0, s11, 39
+; GCN-NEXT:    v_writelane_b32 v0, s12, 40
+; GCN-NEXT:    v_writelane_b32 v0, s13, 41
+; GCN-NEXT:    v_writelane_b32 v0, s14, 42
+; GCN-NEXT:    v_writelane_b32 v0, s15, 43
+; GCN-NEXT:    v_writelane_b32 v0, s16, 44
+; GCN-NEXT:    v_writelane_b32 v0, s17, 45
+; GCN-NEXT:    v_writelane_b32 v0, s18, 46
+; GCN-NEXT:    v_writelane_b32 v0, s19, 47
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s[4:19]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_writelane_b32 v31, s4, 48
-; GCN-NEXT:    v_writelane_b32 v31, s5, 49
-; GCN-NEXT:    v_writelane_b32 v31, s6, 50
-; GCN-NEXT:    v_writelane_b32 v31, s7, 51
-; GCN-NEXT:    v_writelane_b32 v31, s8, 52
-; GCN-NEXT:    v_writelane_b32 v31, s9, 53
-; GCN-NEXT:    v_writelane_b32 v31, s10, 54
-; GCN-NEXT:    v_writelane_b32 v31, s11, 55
-; GCN-NEXT:    v_writelane_b32 v31, s12, 56
-; GCN-NEXT:    v_writelane_b32 v31, s13, 57
-; GCN-NEXT:    v_writelane_b32 v31, s14, 58
-; GCN-NEXT:    v_writelane_b32 v31, s15, 59
-; GCN-NEXT:    v_writelane_b32 v31, s16, 60
-; GCN-NEXT:    v_writelane_b32 v31, s17, 61
-; GCN-NEXT:    v_writelane_b32 v31, s18, 62
-; GCN-NEXT:    v_writelane_b32 v31, s19, 63
+; GCN-NEXT:    v_writelane_b32 v0, s4, 48
+; GCN-NEXT:    v_writelane_b32 v0, s5, 49
+; GCN-NEXT:    v_writelane_b32 v0, s6, 50
+; GCN-NEXT:    v_writelane_b32 v0, s7, 51
+; GCN-NEXT:    v_writelane_b32 v0, s8, 52
+; GCN-NEXT:    v_writelane_b32 v0, s9, 53
+; GCN-NEXT:    v_writelane_b32 v0, s10, 54
+; GCN-NEXT:    v_writelane_b32 v0, s11, 55
+; GCN-NEXT:    v_writelane_b32 v0, s12, 56
+; GCN-NEXT:    v_writelane_b32 v0, s13, 57
+; GCN-NEXT:    v_writelane_b32 v0, s14, 58
+; GCN-NEXT:    v_writelane_b32 v0, s15, 59
+; GCN-NEXT:    v_writelane_b32 v0, s16, 60
+; GCN-NEXT:    v_writelane_b32 v0, s17, 61
+; GCN-NEXT:    v_writelane_b32 v0, s18, 62
+; GCN-NEXT:    v_writelane_b32 v0, s19, 63
+; GCN-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GCN-NEXT:    buffer_store_dword v0, off, s[52:55], 0 offset:8 ; 4-byte Folded Spill
+; GCN-NEXT:    s_mov_b64 exec, s[34:35]
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s[2:3]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    s_mov_b64 s[4:5], exec
-; GCN-NEXT:    s_mov_b64 exec, 3
-; GCN-NEXT:    buffer_store_dword v0, off, s[52:55], 0
+; GCN-NEXT:    ; implicit-def: $vgpr0
 ; GCN-NEXT:    v_writelane_b32 v0, s2, 0
 ; GCN-NEXT:    v_writelane_b32 v0, s3, 1
+; GCN-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GCN-NEXT:    buffer_store_dword v0, off, s[52:55], 0 offset:4 ; 4-byte Folded Spill
-; GCN-NEXT:    buffer_load_dword v0, off, s[52:55], 0
-; GCN-NEXT:    s_waitcnt vmcnt(0)
-; GCN-NEXT:    s_mov_b64 exec, s[4:5]
+; GCN-NEXT:    s_mov_b64 exec, s[34:35]
 ; GCN-NEXT:    s_mov_b32 s1, 0
 ; GCN-NEXT:    s_waitcnt lgkmcnt(0)
 ; GCN-NEXT:    s_cmp_lg_u32 s0, s1
 ; GCN-NEXT:    s_cbranch_scc1 .LBB3_2
 ; GCN-NEXT:  ; %bb.1: ; %bb0
-; GCN-NEXT:    v_readlane_b32 s36, v31, 32
-; GCN-NEXT:    v_readlane_b32 s37, v31, 33
-; GCN-NEXT:    v_readlane_b32 s38, v31, 34
-; GCN-NEXT:    v_readlane_b32 s39, v31, 35
-; GCN-NEXT:    v_readlane_b32 s40, v31, 36
-; GCN-NEXT:    v_readlane_b32 s41, v31, 37
-; GCN-NEXT:    v_readlane_b32 s42, v31, 38
-; GCN-NEXT:    v_readlane_b32 s43, v31, 39
-; GCN-NEXT:    v_readlane_b32 s44, v31, 40
-; GCN-NEXT:    v_readlane_b32 s45, v31, 41
-; GCN-NEXT:    v_readlane_b32 s46, v31, 42
-; GCN-NEXT:    v_readlane_b32 s47, v31, 43
-; GCN-NEXT:    v_readlane_b32 s48, v31, 44
-; GCN-NEXT:    v_readlane_b32 s49, v31, 45
-; GCN-NEXT:    v_readlane_b32 s50, v31, 46
-; GCN-NEXT:    v_readlane_b32 s51, v31, 47
-; GCN-NEXT:    v_readlane_b32 s0, v31, 16
-; GCN-NEXT:    v_readlane_b32 s1, v31, 17
-; GCN-NEXT:    v_readlane_b32 s2, v31, 18
-; GCN-NEXT:    v_readlane_b32 s3, v31, 19
-; GCN-NEXT:    v_readlane_b32 s4, v31, 20
-; GCN-NEXT:    v_readlane_b32 s5, v31, 21
-; GCN-NEXT:    v_readlane_b32 s6, v31, 22
-; GCN-NEXT:    v_readlane_b32 s7, v31, 23
-; GCN-NEXT:    v_readlane_b32 s8, v31, 24
-; GCN-NEXT:    v_readlane_b32 s9, v31, 25
-; GCN-NEXT:    v_readlane_b32 s10, v31, 26
-; GCN-NEXT:    v_readlane_b32 s11, v31, 27
-; GCN-NEXT:    v_readlane_b32 s12, v31, 28
-; GCN-NEXT:    v_readlane_b32 s13, v31, 29
-; GCN-NEXT:    v_readlane_b32 s14, v31, 30
-; GCN-NEXT:    v_readlane_b32 s15, v31, 31
-; GCN-NEXT:    v_readlane_b32 s16, v31, 0
-; GCN-NEXT:    v_readlane_b32 s17, v31, 1
-; GCN-NEXT:    v_readlane_b32 s18, v31, 2
-; GCN-NEXT:    v_readlane_b32 s19, v31, 3
-; GCN-NEXT:    v_readlane_b32 s20, v31, 4
-; GCN-NEXT:    v_readlane_b32 s21, v31, 5
-; GCN-NEXT:    v_readlane_b32 s22, v31, 6
-; GCN-NEXT:    v_readlane_b32 s23, v31, 7
-; GCN-NEXT:    v_readlane_b32 s24, v31, 8
-; GCN-NEXT:    v_readlane_b32 s25, v31, 9
-; GCN-NEXT:    v_readlane_b32 s26, v31, 10
-; GCN-NEXT:    v_readlane_b32 s27, v31, 11
-; GCN-NEXT:    v_readlane_b32 s28, v31, 12
-; GCN-NEXT:    v_readlane_b32 s29, v31, 13
-; GCN-NEXT:    v_readlane_b32 s30, v31, 14
-; GCN-NEXT:    v_readlane_b32 s31, v31, 15
+; GCN-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GCN-NEXT:    buffer_load_dword v1, off, s[52:55], 0 offset:4 ; 4-byte Folded Reload
+; GCN-NEXT:    s_mov_b64 exec, s[34:35]
+; GCN-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GCN-NEXT:    buffer_load_dword v2, off, s[52:55], 0 offset:8 ; 4-byte Folded Reload
+; GCN-NEXT:    s_mov_b64 exec, s[34:35]
+; GCN-NEXT:    s_waitcnt vmcnt(0)
+; GCN-NEXT:    v_readlane_b32 s36, v2, 32
+; GCN-NEXT:    v_readlane_b32 s37, v2, 33
+; GCN-NEXT:    v_readlane_b32 s38, v2, 34
+; GCN-NEXT:    v_readlane_b32 s39, v2, 35
+; GCN-NEXT:    v_readlane_b32 s40, v2, 36
+; GCN-NEXT:    v_readlane_b32 s41, v2, 37
+; GCN-NEXT:    v_readlane_b32 s42, v2, 38
+; GCN-NEXT:    v_readlane_b32 s43, v2, 39
+; GCN-NEXT:    v_readlane_b32 s44, v2, 40
+; GCN-NEXT:    v_readlane_b32 s45, v2, 41
+; GCN-NEXT:    v_readlane_b32 s46, v2, 42
+; GCN-NEXT:    v_readlane_b32 s47, v2, 43
+; GCN-NEXT:    v_readlane_b32 s48, v2, 44
+; GCN-NEXT:    v_readlane_b32 s49, v2, 45
+; GCN-NEXT:    v_readlane_b32 s50, v2, 46
+; GCN-NEXT:    v_readlane_b32 s51, v2, 47
+; GCN-NEXT:    v_readlane_b32 s0, v2, 16
+; GCN-NEXT:    v_readlane_b32 s1, v2, 17
+; GCN-NEXT:    v_readlane_b32 s2, v2, 18
+; GCN-NEXT:    v_readlane_b32 s3, v2, 19
+; GCN-NEXT:    v_readlane_b32 s4, v2, 20
+; GCN-NEXT:    v_readlane_b32 s5, v2, 21
+; GCN-NEXT:    v_readlane_b32 s6, v2, 22
+; GCN-NEXT:    v_readlane_b32 s7, v2, 23
+; GCN-NEXT:    v_readlane_b32 s8, v2, 24
+; GCN-NEXT:    v_readlane_b32 s9, v2, 25
+; GCN-NEXT:    v_readlane_b32 s10, v2, 26
+; GCN-NEXT:    v_readlane_b32 s11, v2, 27
+; GCN-NEXT:    v_readlane_b32 s12, v2, 28
+; GCN-NEXT:    v_readlane_b32 s13, v2, 29
+; GCN-NEXT:    v_readlane_b32 s14, v2, 30
+; GCN-NEXT:    v_readlane_b32 s15, v2, 31
+; GCN-NEXT:    v_readlane_b32 s16, v2, 0
+; GCN-NEXT:    v_readlane_b32 s17, v2, 1
+; GCN-NEXT:    v_readlane_b32 s18, v2, 2
+; GCN-NEXT:    v_readlane_b32 s19, v2, 3
+; GCN-NEXT:    v_readlane_b32 s20, v2, 4
+; GCN-NEXT:    v_readlane_b32 s21, v2, 5
+; GCN-NEXT:    v_readlane_b32 s22, v2, 6
+; GCN-NEXT:    v_readlane_b32 s23, v2, 7
+; GCN-NEXT:    v_readlane_b32 s24, v2, 8
+; GCN-NEXT:    v_readlane_b32 s25, v2, 9
+; GCN-NEXT:    v_readlane_b32 s26, v2, 10
+; GCN-NEXT:    v_readlane_b32 s27, v2, 11
+; GCN-NEXT:    v_readlane_b32 s28, v2, 12
+; GCN-NEXT:    v_readlane_b32 s29, v2, 13
+; GCN-NEXT:    v_readlane_b32 s30, v2, 14
+; GCN-NEXT:    v_readlane_b32 s31, v2, 15
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def v0
 ; GCN-NEXT:    ;;#ASMEND
@@ -1067,32 +1124,24 @@
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; use s[0:15]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_readlane_b32 s4, v31, 48
-; GCN-NEXT:    v_readlane_b32 s5, v31, 49
-; GCN-NEXT:    v_readlane_b32 s6, v31, 50
-; GCN-NEXT:    v_readlane_b32 s7, v31, 51
-; GCN-NEXT:    v_readlane_b32 s8, v31, 52
-; GCN-NEXT:    v_readlane_b32 s9, v31, 53
-; GCN-NEXT:    v_readlane_b32 s10, v31, 54
-; GCN-NEXT:    v_readlane_b32 s11, v31, 55
-; GCN-NEXT:    v_readlane_b32 s12, v31, 56
-; GCN-NEXT:    v_readlane_b32 s13, v31, 57
-; GCN-NEXT:    v_readlane_b32 s14, v31, 58
-; GCN-NEXT:    v_readlane_b32 s15, v31, 59
-; GCN-NEXT:    v_readlane_b32 s16, v31, 60
-; GCN-NEXT:    v_readlane_b32 s17, v31, 61
-; GCN-NEXT:    v_readlane_b32 s18, v31, 62
-; GCN-NEXT:    v_readlane_b32 s19, v31, 63
-; GCN-NEXT:    s_mov_b64 s[2:3], exec
-; GCN-NEXT:    s_mov_b64 exec, 3
-; GCN-NEXT:    buffer_store_dword v1, off, s[52:55], 0
-; GCN-NEXT:    buffer_load_dword v1, off, s[52:55], 0 offset:4 ; 4-byte Folded Reload
-; GCN-NEXT:    s_waitcnt vmcnt(0)
+; GCN-NEXT:    v_readlane_b32 s4, v2, 48
+; GCN-NEXT:    v_readlane_b32 s5, v2, 49
+; GCN-NEXT:    v_readlane_b32 s6, v2, 50
+; GCN-NEXT:    v_readlane_b32 s7, v2, 51
+; GCN-NEXT:    v_readlane_b32 s8, v2, 52
+; GCN-NEXT:    v_readlane_b32 s9, v2, 53
+; GCN-NEXT:    v_readlane_b32 s10, v2, 54
+; GCN-NEXT:    v_readlane_b32 s11, v2, 55
+; GCN-NEXT:    v_readlane_b32 s12, v2, 56
+; GCN-NEXT:    v_readlane_b32 s13, v2, 57
+; GCN-NEXT:    v_readlane_b32 s14, v2, 58
+; GCN-NEXT:    v_readlane_b32 s15, v2, 59
+; GCN-NEXT:    v_readlane_b32 s16, v2, 60
+; GCN-NEXT:    v_readlane_b32 s17, v2, 61
+; GCN-NEXT:    v_readlane_b32 s18, v2, 62
+; GCN-NEXT:    v_readlane_b32 s19, v2, 63
 ; GCN-NEXT:    v_readlane_b32 s0, v1, 0
 ; GCN-NEXT:    v_readlane_b32 s1, v1, 1
-; GCN-NEXT:    buffer_load_dword v1, off, s[52:55], 0
-; GCN-NEXT:    s_waitcnt vmcnt(0)
-; GCN-NEXT:    s_mov_b64 exec, s[2:3]
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; use s[36:51]
 ; GCN-NEXT:    ;;#ASMEND
diff --git a/llvm/test/CodeGen/AMDGPU/scc-clobbered-sgpr-to-vmem-spill.ll b/llvm/test/CodeGen/AMDGPU/scc-clobbered-sgpr-to-vmem-spill.ll
--- a/llvm/test/CodeGen/AMDGPU/scc-clobbered-sgpr-to-vmem-spill.ll
+++ b/llvm/test/CodeGen/AMDGPU/scc-clobbered-sgpr-to-vmem-spill.ll
@@ -1,22 +1,377 @@
-; RUN: not --crash llc -mtriple=amdgcn--amdhsa -mcpu=gfx900 -verify-machineinstrs -o /dev/null %s 2>&1 | FileCheck %s
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=amdgcn--amdhsa -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck %s
 
-; This ends up needing to spill SGPRs to memory, and also does not
-; have any free SGPRs available to save the exec mask when doing so.
-; The register scavenger also needs to use the emergency stack slot,
-; which tries to place the scavenged register restore instruction as
-; far the block as possible, near the terminator. This places a
-; restore instruction between the condition and the conditional
-; branch, which gets expanded into a sequence involving s_not_b64 on
-; the exec mask, clobbering SCC value before the branch. We probably
-; have to stop relying on being able to flip and restore the exec
-; mask, and always require a free SGPR for saving exec.
+; This was a negative test to catch an extreme case when all options are exhausted
+; while trying to spill SGPRs to memory. After we enabled SGPR spills into virtual VGPRs
+; the edge case won't arise and the test would always compile.
 
-; CHECK: *** Bad machine code: Using an undefined physical register ***
-; CHECK-NEXT: - function:    kernel0
-; CHECK-NEXT: - basic block: %bb.0
-; CHECK-NEXT: - instruction: S_CBRANCH_SCC1 %bb.2, implicit killed $scc
-; CHECK-NEXT: - operand 1:   implicit killed $scc
 define amdgpu_kernel void @kernel0(i32 addrspace(1)* %out, i32 %in) #1 {
+; CHECK-LABEL: kernel0:
+; CHECK:       ; %bb.0:
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; def s[2:3]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    ; implicit-def: $vgpr0
+; CHECK-NEXT:    s_load_dword s0, s[4:5], 0x8
+; CHECK-NEXT:    v_writelane_b32 v0, s2, 0
+; CHECK-NEXT:    v_writelane_b32 v0, s3, 1
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; def s[4:7]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    v_writelane_b32 v0, s4, 2
+; CHECK-NEXT:    v_writelane_b32 v0, s5, 3
+; CHECK-NEXT:    v_writelane_b32 v0, s6, 4
+; CHECK-NEXT:    v_writelane_b32 v0, s7, 5
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; def s[4:11]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    v_writelane_b32 v0, s4, 6
+; CHECK-NEXT:    v_writelane_b32 v0, s5, 7
+; CHECK-NEXT:    v_writelane_b32 v0, s6, 8
+; CHECK-NEXT:    v_writelane_b32 v0, s7, 9
+; CHECK-NEXT:    v_writelane_b32 v0, s8, 10
+; CHECK-NEXT:    v_writelane_b32 v0, s9, 11
+; CHECK-NEXT:    v_writelane_b32 v0, s10, 12
+; CHECK-NEXT:    v_writelane_b32 v0, s11, 13
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; def s[4:19]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    v_writelane_b32 v0, s4, 14
+; CHECK-NEXT:    v_writelane_b32 v0, s5, 15
+; CHECK-NEXT:    v_writelane_b32 v0, s6, 16
+; CHECK-NEXT:    v_writelane_b32 v0, s7, 17
+; CHECK-NEXT:    v_writelane_b32 v0, s8, 18
+; CHECK-NEXT:    v_writelane_b32 v0, s9, 19
+; CHECK-NEXT:    v_writelane_b32 v0, s10, 20
+; CHECK-NEXT:    v_writelane_b32 v0, s11, 21
+; CHECK-NEXT:    v_writelane_b32 v0, s12, 22
+; CHECK-NEXT:    v_writelane_b32 v0, s13, 23
+; CHECK-NEXT:    v_writelane_b32 v0, s14, 24
+; CHECK-NEXT:    v_writelane_b32 v0, s15, 25
+; CHECK-NEXT:    v_writelane_b32 v0, s16, 26
+; CHECK-NEXT:    v_writelane_b32 v0, s17, 27
+; CHECK-NEXT:    v_writelane_b32 v0, s18, 28
+; CHECK-NEXT:    v_writelane_b32 v0, s19, 29
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; def s[2:3]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    v_writelane_b32 v0, s2, 30
+; CHECK-NEXT:    v_writelane_b32 v0, s3, 31
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; def s[4:7]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    v_writelane_b32 v0, s4, 32
+; CHECK-NEXT:    v_writelane_b32 v0, s5, 33
+; CHECK-NEXT:    v_writelane_b32 v0, s6, 34
+; CHECK-NEXT:    v_writelane_b32 v0, s7, 35
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; def s[4:11]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    v_writelane_b32 v0, s4, 36
+; CHECK-NEXT:    v_writelane_b32 v0, s5, 37
+; CHECK-NEXT:    v_writelane_b32 v0, s6, 38
+; CHECK-NEXT:    v_writelane_b32 v0, s7, 39
+; CHECK-NEXT:    v_writelane_b32 v0, s8, 40
+; CHECK-NEXT:    v_writelane_b32 v0, s9, 41
+; CHECK-NEXT:    v_writelane_b32 v0, s10, 42
+; CHECK-NEXT:    v_writelane_b32 v0, s11, 43
+; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
+; CHECK-NEXT:    s_cmp_lg_u32 s0, 0
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; def s[16:31]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; def s[52:53]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; def s[48:51]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; def s[36:43]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; def s[0:15]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    v_writelane_b32 v0, s0, 44
+; CHECK-NEXT:    v_writelane_b32 v0, s1, 45
+; CHECK-NEXT:    v_writelane_b32 v0, s2, 46
+; CHECK-NEXT:    v_writelane_b32 v0, s3, 47
+; CHECK-NEXT:    v_writelane_b32 v0, s4, 48
+; CHECK-NEXT:    v_writelane_b32 v0, s5, 49
+; CHECK-NEXT:    v_writelane_b32 v0, s6, 50
+; CHECK-NEXT:    v_writelane_b32 v0, s7, 51
+; CHECK-NEXT:    v_writelane_b32 v0, s8, 52
+; CHECK-NEXT:    v_writelane_b32 v0, s9, 53
+; CHECK-NEXT:    v_writelane_b32 v0, s10, 54
+; CHECK-NEXT:    v_writelane_b32 v0, s11, 55
+; CHECK-NEXT:    v_writelane_b32 v0, s12, 56
+; CHECK-NEXT:    v_writelane_b32 v0, s13, 57
+; CHECK-NEXT:    v_writelane_b32 v0, s14, 58
+; CHECK-NEXT:    v_writelane_b32 v0, s15, 59
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; def s[34:35]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; def s[44:47]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; def s[0:7]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    ; implicit-def: $vgpr1
+; CHECK-NEXT:    v_writelane_b32 v0, s0, 60
+; CHECK-NEXT:    v_writelane_b32 v1, s4, 0
+; CHECK-NEXT:    v_writelane_b32 v0, s1, 61
+; CHECK-NEXT:    v_writelane_b32 v1, s5, 1
+; CHECK-NEXT:    v_writelane_b32 v0, s2, 62
+; CHECK-NEXT:    v_writelane_b32 v1, s6, 2
+; CHECK-NEXT:    v_writelane_b32 v0, s3, 63
+; CHECK-NEXT:    v_writelane_b32 v1, s7, 3
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; def s[0:15]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    v_writelane_b32 v1, s0, 4
+; CHECK-NEXT:    v_writelane_b32 v1, s1, 5
+; CHECK-NEXT:    v_writelane_b32 v1, s2, 6
+; CHECK-NEXT:    v_writelane_b32 v1, s3, 7
+; CHECK-NEXT:    v_writelane_b32 v1, s4, 8
+; CHECK-NEXT:    v_writelane_b32 v1, s5, 9
+; CHECK-NEXT:    v_writelane_b32 v1, s6, 10
+; CHECK-NEXT:    v_writelane_b32 v1, s7, 11
+; CHECK-NEXT:    v_writelane_b32 v1, s8, 12
+; CHECK-NEXT:    v_writelane_b32 v1, s9, 13
+; CHECK-NEXT:    v_writelane_b32 v1, s10, 14
+; CHECK-NEXT:    v_writelane_b32 v1, s11, 15
+; CHECK-NEXT:    v_writelane_b32 v1, s12, 16
+; CHECK-NEXT:    v_writelane_b32 v1, s13, 17
+; CHECK-NEXT:    v_writelane_b32 v1, s14, 18
+; CHECK-NEXT:    v_writelane_b32 v1, s15, 19
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; def s[54:55]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; def s[0:3]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    v_writelane_b32 v1, s0, 20
+; CHECK-NEXT:    v_writelane_b32 v1, s1, 21
+; CHECK-NEXT:    v_writelane_b32 v1, s2, 22
+; CHECK-NEXT:    v_writelane_b32 v1, s3, 23
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; def s[0:7]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    v_writelane_b32 v1, s0, 24
+; CHECK-NEXT:    v_writelane_b32 v1, s1, 25
+; CHECK-NEXT:    v_writelane_b32 v1, s2, 26
+; CHECK-NEXT:    v_writelane_b32 v1, s3, 27
+; CHECK-NEXT:    v_writelane_b32 v1, s4, 28
+; CHECK-NEXT:    v_writelane_b32 v1, s5, 29
+; CHECK-NEXT:    v_writelane_b32 v1, s6, 30
+; CHECK-NEXT:    v_writelane_b32 v1, s7, 31
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; def s[0:15]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    v_writelane_b32 v1, s0, 32
+; CHECK-NEXT:    v_writelane_b32 v1, s1, 33
+; CHECK-NEXT:    v_writelane_b32 v1, s2, 34
+; CHECK-NEXT:    v_writelane_b32 v1, s3, 35
+; CHECK-NEXT:    v_writelane_b32 v1, s4, 36
+; CHECK-NEXT:    v_writelane_b32 v1, s5, 37
+; CHECK-NEXT:    v_writelane_b32 v1, s6, 38
+; CHECK-NEXT:    v_writelane_b32 v1, s7, 39
+; CHECK-NEXT:    v_writelane_b32 v1, s8, 40
+; CHECK-NEXT:    v_writelane_b32 v1, s9, 41
+; CHECK-NEXT:    v_writelane_b32 v1, s10, 42
+; CHECK-NEXT:    v_writelane_b32 v1, s11, 43
+; CHECK-NEXT:    v_writelane_b32 v1, s12, 44
+; CHECK-NEXT:    v_writelane_b32 v1, s13, 45
+; CHECK-NEXT:    v_writelane_b32 v1, s14, 46
+; CHECK-NEXT:    v_writelane_b32 v1, s15, 47
+; CHECK-NEXT:    s_cbranch_scc0 .LBB0_2
+; CHECK-NEXT:  ; %bb.1: ; %ret
+; CHECK-NEXT:    s_endpgm
+; CHECK-NEXT:  .LBB0_2: ; %bb0
+; CHECK-NEXT:    v_readlane_b32 s0, v0, 0
+; CHECK-NEXT:    v_readlane_b32 s1, v0, 1
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; use s[0:1]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    v_readlane_b32 s0, v0, 2
+; CHECK-NEXT:    v_readlane_b32 s1, v0, 3
+; CHECK-NEXT:    v_readlane_b32 s2, v0, 4
+; CHECK-NEXT:    v_readlane_b32 s3, v0, 5
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; use s[0:3]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    v_readlane_b32 s0, v0, 6
+; CHECK-NEXT:    v_readlane_b32 s1, v0, 7
+; CHECK-NEXT:    v_readlane_b32 s2, v0, 8
+; CHECK-NEXT:    v_readlane_b32 s3, v0, 9
+; CHECK-NEXT:    v_readlane_b32 s4, v0, 10
+; CHECK-NEXT:    v_readlane_b32 s5, v0, 11
+; CHECK-NEXT:    v_readlane_b32 s6, v0, 12
+; CHECK-NEXT:    v_readlane_b32 s7, v0, 13
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; use s[0:7]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    v_readlane_b32 s0, v0, 14
+; CHECK-NEXT:    v_readlane_b32 s1, v0, 15
+; CHECK-NEXT:    v_readlane_b32 s2, v0, 16
+; CHECK-NEXT:    v_readlane_b32 s3, v0, 17
+; CHECK-NEXT:    v_readlane_b32 s4, v0, 18
+; CHECK-NEXT:    v_readlane_b32 s5, v0, 19
+; CHECK-NEXT:    v_readlane_b32 s6, v0, 20
+; CHECK-NEXT:    v_readlane_b32 s7, v0, 21
+; CHECK-NEXT:    v_readlane_b32 s8, v0, 22
+; CHECK-NEXT:    v_readlane_b32 s9, v0, 23
+; CHECK-NEXT:    v_readlane_b32 s10, v0, 24
+; CHECK-NEXT:    v_readlane_b32 s11, v0, 25
+; CHECK-NEXT:    v_readlane_b32 s12, v0, 26
+; CHECK-NEXT:    v_readlane_b32 s13, v0, 27
+; CHECK-NEXT:    v_readlane_b32 s14, v0, 28
+; CHECK-NEXT:    v_readlane_b32 s15, v0, 29
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; use s[0:15]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    v_readlane_b32 s0, v0, 30
+; CHECK-NEXT:    v_readlane_b32 s1, v0, 31
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; use s[0:1]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    v_readlane_b32 s0, v0, 32
+; CHECK-NEXT:    v_readlane_b32 s1, v0, 33
+; CHECK-NEXT:    v_readlane_b32 s2, v0, 34
+; CHECK-NEXT:    v_readlane_b32 s3, v0, 35
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; use s[0:3]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    v_readlane_b32 s0, v0, 36
+; CHECK-NEXT:    v_readlane_b32 s1, v0, 37
+; CHECK-NEXT:    v_readlane_b32 s2, v0, 38
+; CHECK-NEXT:    v_readlane_b32 s3, v0, 39
+; CHECK-NEXT:    v_readlane_b32 s4, v0, 40
+; CHECK-NEXT:    v_readlane_b32 s5, v0, 41
+; CHECK-NEXT:    v_readlane_b32 s6, v0, 42
+; CHECK-NEXT:    v_readlane_b32 s7, v0, 43
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; use s[0:7]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    v_readlane_b32 s0, v0, 44
+; CHECK-NEXT:    v_readlane_b32 s1, v0, 45
+; CHECK-NEXT:    v_readlane_b32 s2, v0, 46
+; CHECK-NEXT:    v_readlane_b32 s3, v0, 47
+; CHECK-NEXT:    v_readlane_b32 s4, v0, 48
+; CHECK-NEXT:    v_readlane_b32 s5, v0, 49
+; CHECK-NEXT:    v_readlane_b32 s6, v0, 50
+; CHECK-NEXT:    v_readlane_b32 s7, v0, 51
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; use s[16:31]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; use s[52:53]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; use s[48:51]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; use s[36:43]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    v_readlane_b32 s8, v0, 52
+; CHECK-NEXT:    v_readlane_b32 s9, v0, 53
+; CHECK-NEXT:    v_readlane_b32 s10, v0, 54
+; CHECK-NEXT:    v_readlane_b32 s11, v0, 55
+; CHECK-NEXT:    v_readlane_b32 s12, v0, 56
+; CHECK-NEXT:    v_readlane_b32 s13, v0, 57
+; CHECK-NEXT:    v_readlane_b32 s14, v0, 58
+; CHECK-NEXT:    v_readlane_b32 s15, v0, 59
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; use s[0:15]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    v_readlane_b32 s0, v0, 60
+; CHECK-NEXT:    v_readlane_b32 s1, v0, 61
+; CHECK-NEXT:    v_readlane_b32 s2, v0, 62
+; CHECK-NEXT:    v_readlane_b32 s3, v0, 63
+; CHECK-NEXT:    v_readlane_b32 s4, v1, 0
+; CHECK-NEXT:    v_readlane_b32 s5, v1, 1
+; CHECK-NEXT:    v_readlane_b32 s6, v1, 2
+; CHECK-NEXT:    v_readlane_b32 s7, v1, 3
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; use s[34:35]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; use s[44:47]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; use s[0:7]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    v_readlane_b32 s0, v1, 4
+; CHECK-NEXT:    v_readlane_b32 s1, v1, 5
+; CHECK-NEXT:    v_readlane_b32 s2, v1, 6
+; CHECK-NEXT:    v_readlane_b32 s3, v1, 7
+; CHECK-NEXT:    v_readlane_b32 s4, v1, 8
+; CHECK-NEXT:    v_readlane_b32 s5, v1, 9
+; CHECK-NEXT:    v_readlane_b32 s6, v1, 10
+; CHECK-NEXT:    v_readlane_b32 s7, v1, 11
+; CHECK-NEXT:    v_readlane_b32 s8, v1, 12
+; CHECK-NEXT:    v_readlane_b32 s9, v1, 13
+; CHECK-NEXT:    v_readlane_b32 s10, v1, 14
+; CHECK-NEXT:    v_readlane_b32 s11, v1, 15
+; CHECK-NEXT:    v_readlane_b32 s12, v1, 16
+; CHECK-NEXT:    v_readlane_b32 s13, v1, 17
+; CHECK-NEXT:    v_readlane_b32 s14, v1, 18
+; CHECK-NEXT:    v_readlane_b32 s15, v1, 19
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; use s[0:15]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    v_readlane_b32 s0, v1, 20
+; CHECK-NEXT:    v_readlane_b32 s1, v1, 21
+; CHECK-NEXT:    v_readlane_b32 s2, v1, 22
+; CHECK-NEXT:    v_readlane_b32 s3, v1, 23
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; use s[54:55]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; use s[0:3]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    v_readlane_b32 s0, v1, 24
+; CHECK-NEXT:    v_readlane_b32 s1, v1, 25
+; CHECK-NEXT:    v_readlane_b32 s2, v1, 26
+; CHECK-NEXT:    v_readlane_b32 s3, v1, 27
+; CHECK-NEXT:    v_readlane_b32 s4, v1, 28
+; CHECK-NEXT:    v_readlane_b32 s5, v1, 29
+; CHECK-NEXT:    v_readlane_b32 s6, v1, 30
+; CHECK-NEXT:    v_readlane_b32 s7, v1, 31
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; use s[0:7]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    v_readlane_b32 s0, v1, 32
+; CHECK-NEXT:    v_readlane_b32 s1, v1, 33
+; CHECK-NEXT:    v_readlane_b32 s2, v1, 34
+; CHECK-NEXT:    v_readlane_b32 s3, v1, 35
+; CHECK-NEXT:    v_readlane_b32 s4, v1, 36
+; CHECK-NEXT:    v_readlane_b32 s5, v1, 37
+; CHECK-NEXT:    v_readlane_b32 s6, v1, 38
+; CHECK-NEXT:    v_readlane_b32 s7, v1, 39
+; CHECK-NEXT:    v_readlane_b32 s8, v1, 40
+; CHECK-NEXT:    v_readlane_b32 s9, v1, 41
+; CHECK-NEXT:    v_readlane_b32 s10, v1, 42
+; CHECK-NEXT:    v_readlane_b32 s11, v1, 43
+; CHECK-NEXT:    v_readlane_b32 s12, v1, 44
+; CHECK-NEXT:    v_readlane_b32 s13, v1, 45
+; CHECK-NEXT:    v_readlane_b32 s14, v1, 46
+; CHECK-NEXT:    v_readlane_b32 s15, v1, 47
+; CHECK-NEXT:    ;;#ASMSTART
+; CHECK-NEXT:    ; use s[0:15]
+; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    s_endpgm
   call void asm sideeffect "", "~{v[0:7]}" () #0
   call void asm sideeffect "", "~{v[8:15]}" () #0
   call void asm sideeffect "", "~{v[16:19]}"() #0
diff --git a/llvm/test/CodeGen/AMDGPU/sgpr-spill-dead-frame-in-dbg-value.mir b/llvm/test/CodeGen/AMDGPU/sgpr-spill-dead-frame-in-dbg-value.mir
--- a/llvm/test/CodeGen/AMDGPU/sgpr-spill-dead-frame-in-dbg-value.mir
+++ b/llvm/test/CodeGen/AMDGPU/sgpr-spill-dead-frame-in-dbg-value.mir
@@ -1,4 +1,6 @@
-# RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx908 -amdgpu-spill-sgpr-to-vgpr=true -verify-machineinstrs -run-pass=si-lower-sgpr-spills,prologepilog -o - %s | FileCheck %s
+# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
+# RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx908 -amdgpu-spill-sgpr-to-vgpr=true -verify-machineinstrs -run-pass=si-lower-sgpr-spills -o - %s | FileCheck -check-prefix=SGPR_SPILL %s
+# RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx908 -amdgpu-spill-sgpr-to-vgpr=true -verify-machineinstrs --start-before=si-lower-sgpr-spills --stop-after=prologepilog -o - %s | FileCheck -check-prefix=PEI %s
 
 # After handling the SGPR spill to VGPR in SILowerSGPRSpills pass, replace the dead frame index in the DBG_VALUE instruction with reg 0.
 # Otherwise, the test would crash during PEI while trying to replace the dead frame index.
@@ -39,13 +41,21 @@
     workGroupIDX:    { reg: '$sgpr8' }
     privateSegmentWaveByteOffset: { reg: '$sgpr9' }
 body:             |
-  ; CHECK-LABEL: name: test
-  ; CHECK: bb.0:
-  ; CHECK:   $vgpr0 = V_WRITELANE_B32 killed $sgpr10, 0, $vgpr0
-  ; CHECK:   DBG_VALUE $noreg, 0
-  ; CHECK: bb.1:
-  ; CHECK:   $sgpr10 = V_READLANE_B32 $vgpr0, 0
-  ; CHECK:   S_ENDPGM 0
+  ; SGPR_SPILL-LABEL: name: test
+  ; SGPR_SPILL: bb.0:
+  ; SGPR_SPILL:   [[VGPR:%[0-9]+]]:vgpr_32 = IMPLICIT_DEF
+  ; SGPR_SPILL:   [[VGPR]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr10, 0, [[VGPR]]
+  ; SGPR_SPILL:   DBG_VALUE $noreg, 0
+  ; SGPR_SPILL: bb.1:
+  ; SGPR_SPILL:   $sgpr10 = V_READLANE_B32 [[VGPR]], 0
+  ; SGPR_SPILL:   S_ENDPGM 0
+  ; PEI-LABEL: name: test
+  ; PEI: bb.0:
+  ; PEI:   renamable $[[VGPR:vgpr[0-9]+]] = IMPLICIT_DEF
+  ; PEI:   renamable $[[VGPR]] = V_WRITELANE_B32 killed $sgpr10, 0, killed $[[VGPR]]
+  ; PEI: bb.1:
+  ; PEI:   $sgpr10 = V_READLANE_B32 killed $[[VGPR]], 0
+  ; PEI:   S_ENDPGM 0
   bb.0:
     renamable $sgpr10 = IMPLICIT_DEF
     SI_SPILL_S32_SAVE killed $sgpr10, %stack.0, implicit $exec, implicit $sgpr96_sgpr97_sgpr98_sgpr99, implicit $sgpr32
diff --git a/llvm/test/CodeGen/AMDGPU/sgpr-spill-fi-skip-processing-stack-arg-dbg-value.mir b/llvm/test/CodeGen/AMDGPU/sgpr-spill-fi-skip-processing-stack-arg-dbg-value.mir
--- a/llvm/test/CodeGen/AMDGPU/sgpr-spill-fi-skip-processing-stack-arg-dbg-value.mir
+++ b/llvm/test/CodeGen/AMDGPU/sgpr-spill-fi-skip-processing-stack-arg-dbg-value.mir
@@ -1,4 +1,4 @@
-# RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx908 -amdgpu-spill-sgpr-to-vgpr=true -verify-machineinstrs -run-pass=si-lower-sgpr-spills,prologepilog -o - %s | FileCheck %s
+# RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx908 -amdgpu-spill-sgpr-to-vgpr=true -verify-machineinstrs -run-pass=si-lower-sgpr-spills -o - %s | FileCheck %s
 
 # After handling the SGPR spill to VGPR in SILowerSGPRSpills pass, we replace the dead frame index in the DBG_VALUE instruction with reg 0.
 # Skip looking for frame indices in the debug value instruction for incoming arguments passed via stack. The test would crash otherwise.
@@ -45,7 +45,7 @@
 body:             |
   ; CHECK-LABEL: name: test
   ; CHECK: bb.0:
-  ; CHECK:   DBG_VALUE $noreg, 0
+  ; CHECK:   DBG_VALUE
   bb.0:
     renamable $sgpr10 = IMPLICIT_DEF
     SI_SPILL_S32_SAVE killed $sgpr10, %stack.0, implicit $exec, implicit $sgpr96_sgpr97_sgpr98_sgpr99, implicit $sgpr32
diff --git a/llvm/test/CodeGen/AMDGPU/sgpr-spill-no-vgprs.ll b/llvm/test/CodeGen/AMDGPU/sgpr-spill-no-vgprs.ll
--- a/llvm/test/CodeGen/AMDGPU/sgpr-spill-no-vgprs.ll
+++ b/llvm/test/CodeGen/AMDGPU/sgpr-spill-no-vgprs.ll
@@ -1,8 +1,10 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc -O0 -mtriple=amdgcn-amd-amdhsa -mcpu=hawaii -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
 
-; The first 64 SGPR spills can go to a VGPR, but there isn't a second
-; so some spills must be to memory. The last 16 element spill runs out of lanes at the 15th element.
+; This test was originally written when SGPRs are spilled directly to physical VGPRs and
+; stressed a case when there wasn't enough VGPRs to accommodate all spills.
+; When we started spilling them into virtual VGPR lanes, we always succeed in doing so.
+; The regalloc pass later takes care of allocating VGPRs to these virtual registers.
 
 define amdgpu_kernel void @partial_no_vgprs_last_sgpr_spill(i32 addrspace(1)* %out, i32 %in) #1 {
 ; GCN-LABEL: partial_no_vgprs_last_sgpr_spill:
@@ -23,179 +25,179 @@
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s[8:23]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_writelane_b32 v23, s8, 0
-; GCN-NEXT:    v_writelane_b32 v23, s9, 1
-; GCN-NEXT:    v_writelane_b32 v23, s10, 2
-; GCN-NEXT:    v_writelane_b32 v23, s11, 3
-; GCN-NEXT:    v_writelane_b32 v23, s12, 4
-; GCN-NEXT:    v_writelane_b32 v23, s13, 5
-; GCN-NEXT:    v_writelane_b32 v23, s14, 6
-; GCN-NEXT:    v_writelane_b32 v23, s15, 7
-; GCN-NEXT:    v_writelane_b32 v23, s16, 8
-; GCN-NEXT:    v_writelane_b32 v23, s17, 9
-; GCN-NEXT:    v_writelane_b32 v23, s18, 10
-; GCN-NEXT:    v_writelane_b32 v23, s19, 11
-; GCN-NEXT:    v_writelane_b32 v23, s20, 12
-; GCN-NEXT:    v_writelane_b32 v23, s21, 13
-; GCN-NEXT:    v_writelane_b32 v23, s22, 14
-; GCN-NEXT:    v_writelane_b32 v23, s23, 15
+; GCN-NEXT:    ; implicit-def: $vgpr0
+; GCN-NEXT:    v_writelane_b32 v0, s8, 0
+; GCN-NEXT:    v_writelane_b32 v0, s9, 1
+; GCN-NEXT:    v_writelane_b32 v0, s10, 2
+; GCN-NEXT:    v_writelane_b32 v0, s11, 3
+; GCN-NEXT:    v_writelane_b32 v0, s12, 4
+; GCN-NEXT:    v_writelane_b32 v0, s13, 5
+; GCN-NEXT:    v_writelane_b32 v0, s14, 6
+; GCN-NEXT:    v_writelane_b32 v0, s15, 7
+; GCN-NEXT:    v_writelane_b32 v0, s16, 8
+; GCN-NEXT:    v_writelane_b32 v0, s17, 9
+; GCN-NEXT:    v_writelane_b32 v0, s18, 10
+; GCN-NEXT:    v_writelane_b32 v0, s19, 11
+; GCN-NEXT:    v_writelane_b32 v0, s20, 12
+; GCN-NEXT:    v_writelane_b32 v0, s21, 13
+; GCN-NEXT:    v_writelane_b32 v0, s22, 14
+; GCN-NEXT:    v_writelane_b32 v0, s23, 15
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s[8:23]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_writelane_b32 v23, s8, 16
-; GCN-NEXT:    v_writelane_b32 v23, s9, 17
-; GCN-NEXT:    v_writelane_b32 v23, s10, 18
-; GCN-NEXT:    v_writelane_b32 v23, s11, 19
-; GCN-NEXT:    v_writelane_b32 v23, s12, 20
-; GCN-NEXT:    v_writelane_b32 v23, s13, 21
-; GCN-NEXT:    v_writelane_b32 v23, s14, 22
-; GCN-NEXT:    v_writelane_b32 v23, s15, 23
-; GCN-NEXT:    v_writelane_b32 v23, s16, 24
-; GCN-NEXT:    v_writelane_b32 v23, s17, 25
-; GCN-NEXT:    v_writelane_b32 v23, s18, 26
-; GCN-NEXT:    v_writelane_b32 v23, s19, 27
-; GCN-NEXT:    v_writelane_b32 v23, s20, 28
-; GCN-NEXT:    v_writelane_b32 v23, s21, 29
-; GCN-NEXT:    v_writelane_b32 v23, s22, 30
-; GCN-NEXT:    v_writelane_b32 v23, s23, 31
+; GCN-NEXT:    v_writelane_b32 v0, s8, 16
+; GCN-NEXT:    v_writelane_b32 v0, s9, 17
+; GCN-NEXT:    v_writelane_b32 v0, s10, 18
+; GCN-NEXT:    v_writelane_b32 v0, s11, 19
+; GCN-NEXT:    v_writelane_b32 v0, s12, 20
+; GCN-NEXT:    v_writelane_b32 v0, s13, 21
+; GCN-NEXT:    v_writelane_b32 v0, s14, 22
+; GCN-NEXT:    v_writelane_b32 v0, s15, 23
+; GCN-NEXT:    v_writelane_b32 v0, s16, 24
+; GCN-NEXT:    v_writelane_b32 v0, s17, 25
+; GCN-NEXT:    v_writelane_b32 v0, s18, 26
+; GCN-NEXT:    v_writelane_b32 v0, s19, 27
+; GCN-NEXT:    v_writelane_b32 v0, s20, 28
+; GCN-NEXT:    v_writelane_b32 v0, s21, 29
+; GCN-NEXT:    v_writelane_b32 v0, s22, 30
+; GCN-NEXT:    v_writelane_b32 v0, s23, 31
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s[8:23]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_writelane_b32 v23, s8, 32
-; GCN-NEXT:    v_writelane_b32 v23, s9, 33
-; GCN-NEXT:    v_writelane_b32 v23, s10, 34
-; GCN-NEXT:    v_writelane_b32 v23, s11, 35
-; GCN-NEXT:    v_writelane_b32 v23, s12, 36
-; GCN-NEXT:    v_writelane_b32 v23, s13, 37
-; GCN-NEXT:    v_writelane_b32 v23, s14, 38
-; GCN-NEXT:    v_writelane_b32 v23, s15, 39
-; GCN-NEXT:    v_writelane_b32 v23, s16, 40
-; GCN-NEXT:    v_writelane_b32 v23, s17, 41
-; GCN-NEXT:    v_writelane_b32 v23, s18, 42
-; GCN-NEXT:    v_writelane_b32 v23, s19, 43
-; GCN-NEXT:    v_writelane_b32 v23, s20, 44
-; GCN-NEXT:    v_writelane_b32 v23, s21, 45
-; GCN-NEXT:    v_writelane_b32 v23, s22, 46
-; GCN-NEXT:    v_writelane_b32 v23, s23, 47
+; GCN-NEXT:    v_writelane_b32 v0, s8, 32
+; GCN-NEXT:    v_writelane_b32 v0, s9, 33
+; GCN-NEXT:    v_writelane_b32 v0, s10, 34
+; GCN-NEXT:    v_writelane_b32 v0, s11, 35
+; GCN-NEXT:    v_writelane_b32 v0, s12, 36
+; GCN-NEXT:    v_writelane_b32 v0, s13, 37
+; GCN-NEXT:    v_writelane_b32 v0, s14, 38
+; GCN-NEXT:    v_writelane_b32 v0, s15, 39
+; GCN-NEXT:    v_writelane_b32 v0, s16, 40
+; GCN-NEXT:    v_writelane_b32 v0, s17, 41
+; GCN-NEXT:    v_writelane_b32 v0, s18, 42
+; GCN-NEXT:    v_writelane_b32 v0, s19, 43
+; GCN-NEXT:    v_writelane_b32 v0, s20, 44
+; GCN-NEXT:    v_writelane_b32 v0, s21, 45
+; GCN-NEXT:    v_writelane_b32 v0, s22, 46
+; GCN-NEXT:    v_writelane_b32 v0, s23, 47
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s[8:23]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_writelane_b32 v23, s8, 48
-; GCN-NEXT:    v_writelane_b32 v23, s9, 49
-; GCN-NEXT:    v_writelane_b32 v23, s10, 50
-; GCN-NEXT:    v_writelane_b32 v23, s11, 51
-; GCN-NEXT:    v_writelane_b32 v23, s12, 52
-; GCN-NEXT:    v_writelane_b32 v23, s13, 53
-; GCN-NEXT:    v_writelane_b32 v23, s14, 54
-; GCN-NEXT:    v_writelane_b32 v23, s15, 55
-; GCN-NEXT:    v_writelane_b32 v23, s16, 56
-; GCN-NEXT:    v_writelane_b32 v23, s17, 57
-; GCN-NEXT:    v_writelane_b32 v23, s18, 58
-; GCN-NEXT:    v_writelane_b32 v23, s19, 59
-; GCN-NEXT:    v_writelane_b32 v23, s20, 60
-; GCN-NEXT:    v_writelane_b32 v23, s21, 61
-; GCN-NEXT:    v_writelane_b32 v23, s22, 62
-; GCN-NEXT:    v_writelane_b32 v23, s23, 63
+; GCN-NEXT:    v_writelane_b32 v0, s8, 48
+; GCN-NEXT:    v_writelane_b32 v0, s9, 49
+; GCN-NEXT:    v_writelane_b32 v0, s10, 50
+; GCN-NEXT:    v_writelane_b32 v0, s11, 51
+; GCN-NEXT:    v_writelane_b32 v0, s12, 52
+; GCN-NEXT:    v_writelane_b32 v0, s13, 53
+; GCN-NEXT:    v_writelane_b32 v0, s14, 54
+; GCN-NEXT:    v_writelane_b32 v0, s15, 55
+; GCN-NEXT:    v_writelane_b32 v0, s16, 56
+; GCN-NEXT:    v_writelane_b32 v0, s17, 57
+; GCN-NEXT:    v_writelane_b32 v0, s18, 58
+; GCN-NEXT:    v_writelane_b32 v0, s19, 59
+; GCN-NEXT:    v_writelane_b32 v0, s20, 60
+; GCN-NEXT:    v_writelane_b32 v0, s21, 61
+; GCN-NEXT:    v_writelane_b32 v0, s22, 62
+; GCN-NEXT:    v_writelane_b32 v0, s23, 63
+; GCN-NEXT:    s_or_saveexec_b64 s[24:25], -1
+; GCN-NEXT:    buffer_store_dword v0, off, s[0:3], 0 offset:8 ; 4-byte Folded Spill
+; GCN-NEXT:    s_mov_b64 exec, s[24:25]
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s[6:7]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    s_mov_b64 s[8:9], exec
-; GCN-NEXT:    s_mov_b64 exec, 3
-; GCN-NEXT:    buffer_store_dword v0, off, s[0:3], 0
+; GCN-NEXT:    ; implicit-def: $vgpr0
 ; GCN-NEXT:    v_writelane_b32 v0, s6, 0
 ; GCN-NEXT:    v_writelane_b32 v0, s7, 1
+; GCN-NEXT:    s_or_saveexec_b64 s[24:25], -1
 ; GCN-NEXT:    buffer_store_dword v0, off, s[0:3], 0 offset:4 ; 4-byte Folded Spill
-; GCN-NEXT:    buffer_load_dword v0, off, s[0:3], 0
-; GCN-NEXT:    s_waitcnt vmcnt(0)
-; GCN-NEXT:    s_mov_b64 exec, s[8:9]
+; GCN-NEXT:    s_mov_b64 exec, s[24:25]
 ; GCN-NEXT:    s_mov_b32 s5, 0
 ; GCN-NEXT:    s_waitcnt lgkmcnt(0)
 ; GCN-NEXT:    s_cmp_lg_u32 s4, s5
 ; GCN-NEXT:    s_cbranch_scc1 .LBB0_2
 ; GCN-NEXT:  ; %bb.1: ; %bb0
-; GCN-NEXT:    v_readlane_b32 s4, v23, 0
-; GCN-NEXT:    v_readlane_b32 s5, v23, 1
-; GCN-NEXT:    v_readlane_b32 s6, v23, 2
-; GCN-NEXT:    v_readlane_b32 s7, v23, 3
-; GCN-NEXT:    v_readlane_b32 s8, v23, 4
-; GCN-NEXT:    v_readlane_b32 s9, v23, 5
-; GCN-NEXT:    v_readlane_b32 s10, v23, 6
-; GCN-NEXT:    v_readlane_b32 s11, v23, 7
-; GCN-NEXT:    v_readlane_b32 s12, v23, 8
-; GCN-NEXT:    v_readlane_b32 s13, v23, 9
-; GCN-NEXT:    v_readlane_b32 s14, v23, 10
-; GCN-NEXT:    v_readlane_b32 s15, v23, 11
-; GCN-NEXT:    v_readlane_b32 s16, v23, 12
-; GCN-NEXT:    v_readlane_b32 s17, v23, 13
-; GCN-NEXT:    v_readlane_b32 s18, v23, 14
-; GCN-NEXT:    v_readlane_b32 s19, v23, 15
+; GCN-NEXT:    s_or_saveexec_b64 s[24:25], -1
+; GCN-NEXT:    buffer_load_dword v0, off, s[0:3], 0 offset:4 ; 4-byte Folded Reload
+; GCN-NEXT:    s_mov_b64 exec, s[24:25]
+; GCN-NEXT:    s_or_saveexec_b64 s[24:25], -1
+; GCN-NEXT:    buffer_load_dword v1, off, s[0:3], 0 offset:8 ; 4-byte Folded Reload
+; GCN-NEXT:    s_mov_b64 exec, s[24:25]
+; GCN-NEXT:    s_waitcnt vmcnt(0)
+; GCN-NEXT:    v_readlane_b32 s4, v1, 0
+; GCN-NEXT:    v_readlane_b32 s5, v1, 1
+; GCN-NEXT:    v_readlane_b32 s6, v1, 2
+; GCN-NEXT:    v_readlane_b32 s7, v1, 3
+; GCN-NEXT:    v_readlane_b32 s8, v1, 4
+; GCN-NEXT:    v_readlane_b32 s9, v1, 5
+; GCN-NEXT:    v_readlane_b32 s10, v1, 6
+; GCN-NEXT:    v_readlane_b32 s11, v1, 7
+; GCN-NEXT:    v_readlane_b32 s12, v1, 8
+; GCN-NEXT:    v_readlane_b32 s13, v1, 9
+; GCN-NEXT:    v_readlane_b32 s14, v1, 10
+; GCN-NEXT:    v_readlane_b32 s15, v1, 11
+; GCN-NEXT:    v_readlane_b32 s16, v1, 12
+; GCN-NEXT:    v_readlane_b32 s17, v1, 13
+; GCN-NEXT:    v_readlane_b32 s18, v1, 14
+; GCN-NEXT:    v_readlane_b32 s19, v1, 15
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; use s[4:19]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_readlane_b32 s4, v23, 16
-; GCN-NEXT:    v_readlane_b32 s5, v23, 17
-; GCN-NEXT:    v_readlane_b32 s6, v23, 18
-; GCN-NEXT:    v_readlane_b32 s7, v23, 19
-; GCN-NEXT:    v_readlane_b32 s8, v23, 20
-; GCN-NEXT:    v_readlane_b32 s9, v23, 21
-; GCN-NEXT:    v_readlane_b32 s10, v23, 22
-; GCN-NEXT:    v_readlane_b32 s11, v23, 23
-; GCN-NEXT:    v_readlane_b32 s12, v23, 24
-; GCN-NEXT:    v_readlane_b32 s13, v23, 25
-; GCN-NEXT:    v_readlane_b32 s14, v23, 26
-; GCN-NEXT:    v_readlane_b32 s15, v23, 27
-; GCN-NEXT:    v_readlane_b32 s16, v23, 28
-; GCN-NEXT:    v_readlane_b32 s17, v23, 29
-; GCN-NEXT:    v_readlane_b32 s18, v23, 30
-; GCN-NEXT:    v_readlane_b32 s19, v23, 31
+; GCN-NEXT:    v_readlane_b32 s4, v1, 16
+; GCN-NEXT:    v_readlane_b32 s5, v1, 17
+; GCN-NEXT:    v_readlane_b32 s6, v1, 18
+; GCN-NEXT:    v_readlane_b32 s7, v1, 19
+; GCN-NEXT:    v_readlane_b32 s8, v1, 20
+; GCN-NEXT:    v_readlane_b32 s9, v1, 21
+; GCN-NEXT:    v_readlane_b32 s10, v1, 22
+; GCN-NEXT:    v_readlane_b32 s11, v1, 23
+; GCN-NEXT:    v_readlane_b32 s12, v1, 24
+; GCN-NEXT:    v_readlane_b32 s13, v1, 25
+; GCN-NEXT:    v_readlane_b32 s14, v1, 26
+; GCN-NEXT:    v_readlane_b32 s15, v1, 27
+; GCN-NEXT:    v_readlane_b32 s16, v1, 28
+; GCN-NEXT:    v_readlane_b32 s17, v1, 29
+; GCN-NEXT:    v_readlane_b32 s18, v1, 30
+; GCN-NEXT:    v_readlane_b32 s19, v1, 31
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; use s[4:19]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_readlane_b32 s4, v23, 32
-; GCN-NEXT:    v_readlane_b32 s5, v23, 33
-; GCN-NEXT:    v_readlane_b32 s6, v23, 34
-; GCN-NEXT:    v_readlane_b32 s7, v23, 35
-; GCN-NEXT:    v_readlane_b32 s8, v23, 36
-; GCN-NEXT:    v_readlane_b32 s9, v23, 37
-; GCN-NEXT:    v_readlane_b32 s10, v23, 38
-; GCN-NEXT:    v_readlane_b32 s11, v23, 39
-; GCN-NEXT:    v_readlane_b32 s12, v23, 40
-; GCN-NEXT:    v_readlane_b32 s13, v23, 41
-; GCN-NEXT:    v_readlane_b32 s14, v23, 42
-; GCN-NEXT:    v_readlane_b32 s15, v23, 43
-; GCN-NEXT:    v_readlane_b32 s16, v23, 44
-; GCN-NEXT:    v_readlane_b32 s17, v23, 45
-; GCN-NEXT:    v_readlane_b32 s18, v23, 46
-; GCN-NEXT:    v_readlane_b32 s19, v23, 47
+; GCN-NEXT:    v_readlane_b32 s4, v1, 32
+; GCN-NEXT:    v_readlane_b32 s5, v1, 33
+; GCN-NEXT:    v_readlane_b32 s6, v1, 34
+; GCN-NEXT:    v_readlane_b32 s7, v1, 35
+; GCN-NEXT:    v_readlane_b32 s8, v1, 36
+; GCN-NEXT:    v_readlane_b32 s9, v1, 37
+; GCN-NEXT:    v_readlane_b32 s10, v1, 38
+; GCN-NEXT:    v_readlane_b32 s11, v1, 39
+; GCN-NEXT:    v_readlane_b32 s12, v1, 40
+; GCN-NEXT:    v_readlane_b32 s13, v1, 41
+; GCN-NEXT:    v_readlane_b32 s14, v1, 42
+; GCN-NEXT:    v_readlane_b32 s15, v1, 43
+; GCN-NEXT:    v_readlane_b32 s16, v1, 44
+; GCN-NEXT:    v_readlane_b32 s17, v1, 45
+; GCN-NEXT:    v_readlane_b32 s18, v1, 46
+; GCN-NEXT:    v_readlane_b32 s19, v1, 47
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; use s[4:19]
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_readlane_b32 s8, v23, 48
-; GCN-NEXT:    v_readlane_b32 s9, v23, 49
-; GCN-NEXT:    v_readlane_b32 s10, v23, 50
-; GCN-NEXT:    v_readlane_b32 s11, v23, 51
-; GCN-NEXT:    v_readlane_b32 s12, v23, 52
-; GCN-NEXT:    v_readlane_b32 s13, v23, 53
-; GCN-NEXT:    v_readlane_b32 s14, v23, 54
-; GCN-NEXT:    v_readlane_b32 s15, v23, 55
-; GCN-NEXT:    v_readlane_b32 s16, v23, 56
-; GCN-NEXT:    v_readlane_b32 s17, v23, 57
-; GCN-NEXT:    v_readlane_b32 s18, v23, 58
-; GCN-NEXT:    v_readlane_b32 s19, v23, 59
-; GCN-NEXT:    v_readlane_b32 s20, v23, 60
-; GCN-NEXT:    v_readlane_b32 s21, v23, 61
-; GCN-NEXT:    v_readlane_b32 s22, v23, 62
-; GCN-NEXT:    v_readlane_b32 s23, v23, 63
-; GCN-NEXT:    s_mov_b64 s[6:7], exec
-; GCN-NEXT:    s_mov_b64 exec, 3
-; GCN-NEXT:    buffer_store_dword v0, off, s[0:3], 0
-; GCN-NEXT:    buffer_load_dword v0, off, s[0:3], 0 offset:4 ; 4-byte Folded Reload
-; GCN-NEXT:    s_waitcnt vmcnt(0)
+; GCN-NEXT:    v_readlane_b32 s8, v1, 48
+; GCN-NEXT:    v_readlane_b32 s9, v1, 49
+; GCN-NEXT:    v_readlane_b32 s10, v1, 50
+; GCN-NEXT:    v_readlane_b32 s11, v1, 51
+; GCN-NEXT:    v_readlane_b32 s12, v1, 52
+; GCN-NEXT:    v_readlane_b32 s13, v1, 53
+; GCN-NEXT:    v_readlane_b32 s14, v1, 54
+; GCN-NEXT:    v_readlane_b32 s15, v1, 55
+; GCN-NEXT:    v_readlane_b32 s16, v1, 56
+; GCN-NEXT:    v_readlane_b32 s17, v1, 57
+; GCN-NEXT:    v_readlane_b32 s18, v1, 58
+; GCN-NEXT:    v_readlane_b32 s19, v1, 59
+; GCN-NEXT:    v_readlane_b32 s20, v1, 60
+; GCN-NEXT:    v_readlane_b32 s21, v1, 61
+; GCN-NEXT:    v_readlane_b32 s22, v1, 62
+; GCN-NEXT:    v_readlane_b32 s23, v1, 63
 ; GCN-NEXT:    v_readlane_b32 s4, v0, 0
 ; GCN-NEXT:    v_readlane_b32 s5, v0, 1
-; GCN-NEXT:    buffer_load_dword v0, off, s[0:3], 0
-; GCN-NEXT:    s_waitcnt vmcnt(0)
-; GCN-NEXT:    s_mov_b64 exec, s[6:7]
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; use s[8:23]
 ; GCN-NEXT:    ;;#ASMEND
diff --git a/llvm/test/CodeGen/AMDGPU/sgpr-spill-partially-undef.mir b/llvm/test/CodeGen/AMDGPU/sgpr-spill-partially-undef.mir
--- a/llvm/test/CodeGen/AMDGPU/sgpr-spill-partially-undef.mir
+++ b/llvm/test/CodeGen/AMDGPU/sgpr-spill-partially-undef.mir
@@ -20,10 +20,11 @@
     liveins: $sgpr4
 
     ; CHECK-LABEL: name: sgpr_spill_s64_undef_high32
-    ; CHECK: liveins: $sgpr4, $vgpr0
+    ; CHECK: liveins: $sgpr4
     ; CHECK-NEXT: {{  $}}
-    ; CHECK-NEXT: $vgpr0 = V_WRITELANE_B32 $sgpr4, 0, $vgpr0, implicit-def $sgpr4_sgpr5, implicit $sgpr4_sgpr5
-    ; CHECK-NEXT: $vgpr0 = V_WRITELANE_B32 $sgpr5, 1, $vgpr0, implicit $sgpr4_sgpr5
+    ; CHECK-NEXT: [[DEF:%[0-9]+]]:vgpr_32 = IMPLICIT_DEF
+    ; CHECK-NEXT: [[V_WRITELANE_B32_:%[0-9]+]]:vgpr_32 = V_WRITELANE_B32 $sgpr4, 0, [[V_WRITELANE_B32_]], implicit-def $sgpr4_sgpr5, implicit $sgpr4_sgpr5
+    ; CHECK-NEXT: [[V_WRITELANE_B32_1:%[0-9]+]]:vgpr_32 = V_WRITELANE_B32 $sgpr5, 1, [[V_WRITELANE_B32_1]], implicit $sgpr4_sgpr5
     SI_SPILL_S64_SAVE renamable $sgpr4_sgpr5, %stack.0, implicit $exec, implicit $sgpr96_sgpr97_sgpr98_sgpr99, implicit $sgpr32 :: (store (s64) into %stack.0, align 4, addrspace 5)
 
 ...
@@ -45,10 +46,11 @@
     liveins: $sgpr5
 
     ; CHECK-LABEL: name: sgpr_spill_s64_undef_low32
-    ; CHECK: liveins: $sgpr5, $vgpr0
+    ; CHECK: liveins: $sgpr5
     ; CHECK-NEXT: {{  $}}
-    ; CHECK-NEXT: $vgpr0 = V_WRITELANE_B32 $sgpr4, 0, $vgpr0, implicit-def $sgpr4_sgpr5, implicit $sgpr4_sgpr5
-    ; CHECK-NEXT: $vgpr0 = V_WRITELANE_B32 $sgpr5, 1, $vgpr0, implicit $sgpr4_sgpr5
+    ; CHECK-NEXT: [[DEF:%[0-9]+]]:vgpr_32 = IMPLICIT_DEF
+    ; CHECK-NEXT: [[V_WRITELANE_B32_:%[0-9]+]]:vgpr_32 = V_WRITELANE_B32 $sgpr4, 0, [[V_WRITELANE_B32_]], implicit-def $sgpr4_sgpr5, implicit $sgpr4_sgpr5
+    ; CHECK-NEXT: [[V_WRITELANE_B32_1:%[0-9]+]]:vgpr_32 = V_WRITELANE_B32 $sgpr5, 1, [[V_WRITELANE_B32_1]], implicit $sgpr4_sgpr5
     SI_SPILL_S64_SAVE renamable $sgpr4_sgpr5, %stack.0, implicit $exec, implicit $sgpr96_sgpr97_sgpr98_sgpr99, implicit $sgpr32 :: (store (s64) into %stack.0, align 4, addrspace 5)
 
 ...
diff --git a/llvm/test/CodeGen/AMDGPU/sgpr-spill-update-only-slot-indexes.ll b/llvm/test/CodeGen/AMDGPU/sgpr-spill-update-only-slot-indexes.ll
--- a/llvm/test/CodeGen/AMDGPU/sgpr-spill-update-only-slot-indexes.ll
+++ b/llvm/test/CodeGen/AMDGPU/sgpr-spill-update-only-slot-indexes.ll
@@ -12,16 +12,17 @@
 ; GCN-NEXT:    s_mov_b32 s36, SCRATCH_RSRC_DWORD0
 ; GCN-NEXT:    s_mov_b32 s37, SCRATCH_RSRC_DWORD1
 ; GCN-NEXT:    s_mov_b32 s38, -1
+; GCN-NEXT:    ; implicit-def: $vgpr3
 ; GCN-NEXT:    s_mov_b32 s39, 0xe00000
-; GCN-NEXT:    v_writelane_b32 v40, s4, 0
+; GCN-NEXT:    v_writelane_b32 v3, s4, 0
 ; GCN-NEXT:    s_add_u32 s36, s36, s11
-; GCN-NEXT:    v_writelane_b32 v40, s5, 1
+; GCN-NEXT:    v_writelane_b32 v3, s5, 1
 ; GCN-NEXT:    s_addc_u32 s37, s37, 0
 ; GCN-NEXT:    s_mov_b64 s[4:5], s[0:1]
-; GCN-NEXT:    v_readlane_b32 s0, v40, 0
+; GCN-NEXT:    v_readlane_b32 s0, v3, 0
 ; GCN-NEXT:    s_mov_b32 s13, s9
 ; GCN-NEXT:    s_mov_b32 s12, s8
-; GCN-NEXT:    v_readlane_b32 s1, v40, 1
+; GCN-NEXT:    v_readlane_b32 s1, v3, 1
 ; GCN-NEXT:    s_add_u32 s8, s0, 36
 ; GCN-NEXT:    s_addc_u32 s9, s1, 0
 ; GCN-NEXT:    s_getpc_b64 s[0:1]
diff --git a/llvm/test/CodeGen/AMDGPU/sgpr-spill-vmem-large-frame.mir b/llvm/test/CodeGen/AMDGPU/sgpr-spill-vmem-large-frame.mir
--- a/llvm/test/CodeGen/AMDGPU/sgpr-spill-vmem-large-frame.mir
+++ b/llvm/test/CodeGen/AMDGPU/sgpr-spill-vmem-large-frame.mir
@@ -1,5 +1,5 @@
 # NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
-# RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx908 -amdgpu-spill-sgpr-to-vgpr=false -verify-machineinstrs -run-pass=si-lower-sgpr-spills,prologepilog -o - %s | FileCheck %s
+# RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx908 -amdgpu-spill-sgpr-to-vgpr=false -verify-machineinstrs -start-before=si-lower-sgpr-spills -stop-after=prologepilog -o - %s | FileCheck %s
 
 # Check that we allocate 2 emergency stack slots if we're spilling
 # SGPRs to memory and potentially have an offset larger than fits in
@@ -29,7 +29,7 @@
     ; CHECK-NEXT: $sgpr6_sgpr7 = S_MOV_B64 $exec
     ; CHECK-NEXT: $exec = S_MOV_B64 1, implicit-def $vgpr2
     ; CHECK-NEXT: BUFFER_STORE_DWORD_OFFSET killed $vgpr2, $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr32, 0, 0, 0, implicit $exec :: (store (s32) into %stack.2, addrspace 5)
-    ; CHECK-NEXT: $vgpr2 = V_WRITELANE_B32 killed $sgpr10, 0, undef $vgpr2
+    ; CHECK-NEXT: $vgpr2 = V_WRITELANE_B32 $sgpr10, 0, undef $vgpr2
     ; CHECK-NEXT: BUFFER_STORE_DWORD_OFFSET killed $vgpr2, $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr32, 8, 0, 0, implicit $exec :: (store (s32) into %stack.0, addrspace 5)
     ; CHECK-NEXT: $vgpr2 = BUFFER_LOAD_DWORD_OFFSET $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr32, 0, 0, 0, implicit $exec :: (load (s32) from %stack.2, addrspace 5)
     ; CHECK-NEXT: $exec = S_MOV_B64 killed $sgpr6_sgpr7, implicit killed $vgpr2
diff --git a/llvm/test/CodeGen/AMDGPU/sgpr-spills-split-regalloc.ll b/llvm/test/CodeGen/AMDGPU/sgpr-spills-split-regalloc.ll
--- a/llvm/test/CodeGen/AMDGPU/sgpr-spills-split-regalloc.ll
+++ b/llvm/test/CodeGen/AMDGPU/sgpr-spills-split-regalloc.ll
@@ -16,10 +16,10 @@
 ; GCN-LABEL: spill_sgpr_with_no_lower_vgpr_available:
 ; GCN:       ; %bb.0:
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GCN-NEXT:    s_mov_b32 s6, s33
+; GCN-NEXT:    s_mov_b32 s14, s33
 ; GCN-NEXT:    s_mov_b32 s33, s32
-; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
-; GCN-NEXT:    buffer_store_dword v255, off, s[0:3], s33 offset:448 ; 4-byte Folded Spill
+; GCN-NEXT:    s_xor_saveexec_b64 s[4:5], -1
+; GCN-NEXT:    buffer_store_dword v0, off, s[0:3], s33 offset:452 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_add_i32 s32, s32, 0x7400
 ; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:440 ; 4-byte Folded Spill
@@ -133,13 +133,20 @@
 ; GCN-NEXT:    buffer_store_dword v252, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
 ; GCN-NEXT:    buffer_store_dword v253, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GCN-NEXT:    buffer_store_dword v254, off, s[0:3], s33 ; 4-byte Folded Spill
-; GCN-NEXT:    v_writelane_b32 v255, s30, 0
-; GCN-NEXT:    v_writelane_b32 v255, s31, 1
+; GCN-NEXT:    ; implicit-def: $vgpr0
+; GCN-NEXT:    v_writelane_b32 v0, s30, 0
+; GCN-NEXT:    v_writelane_b32 v0, s31, 1
+; GCN-NEXT:    s_or_saveexec_b64 s[12:13], -1
+; GCN-NEXT:    buffer_store_dword v0, off, s[0:3], s33 offset:448 ; 4-byte Folded Spill
+; GCN-NEXT:    s_mov_b64 exec, s[12:13]
 ; GCN-NEXT:    v_mov_b32_e32 v0, 0
 ; GCN-NEXT:    buffer_store_dword v0, off, s[0:3], s33 offset:444
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ;;#ASMEND
+; GCN-NEXT:    s_or_saveexec_b64 s[12:13], -1
+; GCN-NEXT:    buffer_load_dword v0, off, s[0:3], s33 offset:448 ; 4-byte Folded Reload
+; GCN-NEXT:    s_mov_b64 exec, s[12:13]
 ; GCN-NEXT:    s_getpc_b64 s[4:5]
 ; GCN-NEXT:    s_add_u32 s4, s4, child_function@gotpcrel32@lo+4
 ; GCN-NEXT:    s_addc_u32 s5, s5, child_function@gotpcrel32@hi+12
@@ -150,8 +157,8 @@
 ; GCN-NEXT:    s_mov_b64 s[2:3], s[10:11]
 ; GCN-NEXT:    s_waitcnt lgkmcnt(0)
 ; GCN-NEXT:    s_swappc_b64 s[30:31], s[4:5]
-; GCN-NEXT:    v_readlane_b32 s31, v255, 1
-; GCN-NEXT:    v_readlane_b32 s30, v255, 0
+; GCN-NEXT:    v_readlane_b32 s31, v0, 1
+; GCN-NEXT:    v_readlane_b32 s30, v0, 0
 ; GCN-NEXT:    buffer_load_dword v254, off, s[0:3], s33 ; 4-byte Folded Reload
 ; GCN-NEXT:    buffer_load_dword v253, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
 ; GCN-NEXT:    buffer_load_dword v252, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload
@@ -263,11 +270,11 @@
 ; GCN-NEXT:    buffer_load_dword v42, off, s[0:3], s33 offset:432 ; 4-byte Folded Reload
 ; GCN-NEXT:    buffer_load_dword v41, off, s[0:3], s33 offset:436 ; 4-byte Folded Reload
 ; GCN-NEXT:    buffer_load_dword v40, off, s[0:3], s33 offset:440 ; 4-byte Folded Reload
-; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
-; GCN-NEXT:    buffer_load_dword v255, off, s[0:3], s33 offset:448 ; 4-byte Folded Reload
+; GCN-NEXT:    s_xor_saveexec_b64 s[4:5], -1
+; GCN-NEXT:    buffer_load_dword v0, off, s[0:3], s33 offset:452 ; 4-byte Folded Reload
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_add_i32 s32, s32, 0xffff8c00
-; GCN-NEXT:    s_mov_b32 s33, s6
+; GCN-NEXT:    s_mov_b32 s33, s14
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
 ; GCN-NEXT:    s_setpc_b64 s[30:31]
   %alloca = alloca i32, align 4, addrspace(5)
@@ -308,10 +315,10 @@
 ; GCN-LABEL: spill_to_lowest_available_vgpr:
 ; GCN:       ; %bb.0:
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GCN-NEXT:    s_mov_b32 s6, s33
+; GCN-NEXT:    s_mov_b32 s14, s33
 ; GCN-NEXT:    s_mov_b32 s33, s32
-; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
-; GCN-NEXT:    buffer_store_dword v254, off, s[0:3], s33 offset:444 ; 4-byte Folded Spill
+; GCN-NEXT:    s_xor_saveexec_b64 s[4:5], -1
+; GCN-NEXT:    buffer_store_dword v0, off, s[0:3], s33 offset:448 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_add_i32 s32, s32, 0x7400
 ; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:436 ; 4-byte Folded Spill
@@ -424,13 +431,20 @@
 ; GCN-NEXT:    buffer_store_dword v251, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
 ; GCN-NEXT:    buffer_store_dword v252, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GCN-NEXT:    buffer_store_dword v253, off, s[0:3], s33 ; 4-byte Folded Spill
-; GCN-NEXT:    v_writelane_b32 v254, s30, 0
-; GCN-NEXT:    v_writelane_b32 v254, s31, 1
+; GCN-NEXT:    ; implicit-def: $vgpr0
+; GCN-NEXT:    v_writelane_b32 v0, s30, 0
+; GCN-NEXT:    v_writelane_b32 v0, s31, 1
+; GCN-NEXT:    s_or_saveexec_b64 s[12:13], -1
+; GCN-NEXT:    buffer_store_dword v0, off, s[0:3], s33 offset:444 ; 4-byte Folded Spill
+; GCN-NEXT:    s_mov_b64 exec, s[12:13]
 ; GCN-NEXT:    v_mov_b32_e32 v0, 0
 ; GCN-NEXT:    buffer_store_dword v0, off, s[0:3], s33 offset:440
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ;;#ASMEND
+; GCN-NEXT:    s_or_saveexec_b64 s[12:13], -1
+; GCN-NEXT:    buffer_load_dword v0, off, s[0:3], s33 offset:444 ; 4-byte Folded Reload
+; GCN-NEXT:    s_mov_b64 exec, s[12:13]
 ; GCN-NEXT:    s_getpc_b64 s[4:5]
 ; GCN-NEXT:    s_add_u32 s4, s4, child_function@gotpcrel32@lo+4
 ; GCN-NEXT:    s_addc_u32 s5, s5, child_function@gotpcrel32@hi+12
@@ -441,8 +455,8 @@
 ; GCN-NEXT:    s_mov_b64 s[2:3], s[10:11]
 ; GCN-NEXT:    s_waitcnt lgkmcnt(0)
 ; GCN-NEXT:    s_swappc_b64 s[30:31], s[4:5]
-; GCN-NEXT:    v_readlane_b32 s31, v254, 1
-; GCN-NEXT:    v_readlane_b32 s30, v254, 0
+; GCN-NEXT:    v_readlane_b32 s31, v0, 1
+; GCN-NEXT:    v_readlane_b32 s30, v0, 0
 ; GCN-NEXT:    buffer_load_dword v253, off, s[0:3], s33 ; 4-byte Folded Reload
 ; GCN-NEXT:    buffer_load_dword v252, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
 ; GCN-NEXT:    buffer_load_dword v251, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload
@@ -553,11 +567,11 @@
 ; GCN-NEXT:    buffer_load_dword v42, off, s[0:3], s33 offset:428 ; 4-byte Folded Reload
 ; GCN-NEXT:    buffer_load_dword v41, off, s[0:3], s33 offset:432 ; 4-byte Folded Reload
 ; GCN-NEXT:    buffer_load_dword v40, off, s[0:3], s33 offset:436 ; 4-byte Folded Reload
-; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
-; GCN-NEXT:    buffer_load_dword v254, off, s[0:3], s33 offset:444 ; 4-byte Folded Reload
+; GCN-NEXT:    s_xor_saveexec_b64 s[4:5], -1
+; GCN-NEXT:    buffer_load_dword v0, off, s[0:3], s33 offset:448 ; 4-byte Folded Reload
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_add_i32 s32, s32, 0xffff8c00
-; GCN-NEXT:    s_mov_b32 s33, s6
+; GCN-NEXT:    s_mov_b32 s33, s14
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
 ; GCN-NEXT:    s_setpc_b64 s[30:31]
   %alloca = alloca i32, align 4, addrspace(5)
@@ -598,8 +612,8 @@
 ; GCN-LABEL: spill_sgpr_with_sgpr_uses:
 ; GCN:       ; %bb.0:
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
-; GCN-NEXT:    buffer_store_dword v254, off, s[0:3], s32 offset:444 ; 4-byte Folded Spill
+; GCN-NEXT:    s_xor_saveexec_b64 s[4:5], -1
+; GCN-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:448 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s32 offset:436 ; 4-byte Folded Spill
 ; GCN-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:432 ; 4-byte Folded Spill
@@ -719,10 +733,18 @@
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; def s4
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    v_writelane_b32 v254, s4, 0
+; GCN-NEXT:    ; implicit-def: $vgpr0
+; GCN-NEXT:    v_writelane_b32 v0, s4, 0
+; GCN-NEXT:    s_or_saveexec_b64 s[8:9], -1
+; GCN-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:444 ; 4-byte Folded Spill
+; GCN-NEXT:    s_mov_b64 exec, s[8:9]
 ; GCN-NEXT:    s_cbranch_scc1 .LBB3_2
 ; GCN-NEXT:  ; %bb.1: ; %bb0
-; GCN-NEXT:    v_readlane_b32 s4, v254, 0
+; GCN-NEXT:    s_or_saveexec_b64 s[8:9], -1
+; GCN-NEXT:    buffer_load_dword v0, off, s[0:3], s32 offset:444 ; 4-byte Folded Reload
+; GCN-NEXT:    s_mov_b64 exec, s[8:9]
+; GCN-NEXT:    s_waitcnt vmcnt(0)
+; GCN-NEXT:    v_readlane_b32 s4, v0, 0
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ; use s4
 ; GCN-NEXT:    ;;#ASMEND
@@ -837,8 +859,8 @@
 ; GCN-NEXT:    buffer_load_dword v42, off, s[0:3], s32 offset:428 ; 4-byte Folded Reload
 ; GCN-NEXT:    buffer_load_dword v41, off, s[0:3], s32 offset:432 ; 4-byte Folded Reload
 ; GCN-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:436 ; 4-byte Folded Reload
-; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
-; GCN-NEXT:    buffer_load_dword v254, off, s[0:3], s32 offset:444 ; 4-byte Folded Reload
+; GCN-NEXT:    s_xor_saveexec_b64 s[4:5], -1
+; GCN-NEXT:    buffer_load_dword v0, off, s[0:3], s32 offset:448 ; 4-byte Folded Reload
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
 ; GCN-NEXT:    s_setpc_b64 s[30:31]
@@ -1161,7 +1183,8 @@
 ; GCN:       ; %bb.0:
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GCN-NEXT:    s_xor_saveexec_b64 s[4:5], -1
-; GCN-NEXT:    buffer_store_dword v4, off, s[0:3], s32 offset:464 ; 4-byte Folded Spill
+; GCN-NEXT:    buffer_store_dword v4, off, s[0:3], s32 offset:468 ; 4-byte Folded Spill
+; GCN-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:472 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s32 offset:444 ; 4-byte Folded Spill
 ; GCN-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:440 ; 4-byte Folded Spill
@@ -1275,45 +1298,54 @@
 ; GCN-NEXT:    buffer_store_dword v253, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
 ; GCN-NEXT:    buffer_store_dword v254, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GCN-NEXT:    buffer_store_dword v255, off, s[0:3], s32 ; 4-byte Folded Spill
+; GCN-NEXT:    ; implicit-def: $vgpr4
 ; GCN-NEXT:    v_writelane_b32 v4, s34, 0
 ; GCN-NEXT:    v_writelane_b32 v4, s35, 1
 ; GCN-NEXT:    v_writelane_b32 v4, s36, 2
 ; GCN-NEXT:    v_writelane_b32 v4, s37, 3
+; GCN-NEXT:    s_or_saveexec_b64 s[8:9], -1
+; GCN-NEXT:    buffer_store_dword v4, off, s[0:3], s32 offset:464 ; 4-byte Folded Spill
+; GCN-NEXT:    s_mov_b64 exec, s[8:9]
 ; GCN-NEXT:    v_mov_b32_e32 v5, v3
-; GCN-NEXT:    v_mov_b32_e32 v3, v1
+; GCN-NEXT:    v_mov_b32_e32 v3, v2
+; GCN-NEXT:    v_mov_b32_e32 v4, v1
+; GCN-NEXT:    v_mov_b32_e32 v1, v0
+; GCN-NEXT:    s_or_saveexec_b64 s[8:9], -1
+; GCN-NEXT:    buffer_load_dword v0, off, s[0:3], s32 offset:464 ; 4-byte Folded Reload
+; GCN-NEXT:    s_mov_b64 exec, s[8:9]
 ; GCN-NEXT:    ; implicit-def: $sgpr4
 ; GCN-NEXT:    ; implicit-def: $sgpr4
-; GCN-NEXT:    ; kill: def $vgpr3 killed $vgpr3 killed $exec
-; GCN-NEXT:    ; kill: def $vgpr0 killed $vgpr0 def $vgpr0_vgpr1 killed $exec
-; GCN-NEXT:    v_mov_b32_e32 v1, v3
+; GCN-NEXT:    ; kill: def $vgpr4 killed $vgpr4 killed $exec
+; GCN-NEXT:    ; kill: def $vgpr1 killed $vgpr1 def $vgpr1_vgpr2 killed $exec
+; GCN-NEXT:    v_mov_b32_e32 v2, v4
 ; GCN-NEXT:    ; implicit-def: $sgpr4
 ; GCN-NEXT:    ; implicit-def: $sgpr4
 ; GCN-NEXT:    ; kill: def $vgpr5 killed $vgpr5 killed $exec
-; GCN-NEXT:    ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
-; GCN-NEXT:    v_mov_b32_e32 v3, v5
+; GCN-NEXT:    ; kill: def $vgpr3 killed $vgpr3 def $vgpr3_vgpr4 killed $exec
+; GCN-NEXT:    v_mov_b32_e32 v4, v5
 ; GCN-NEXT:    ; implicit-def: $sgpr4_sgpr5
 ; GCN-NEXT:    ; implicit-def: $sgpr4_sgpr5
-; GCN-NEXT:    flat_load_dwordx4 v[5:8], v[2:3]
+; GCN-NEXT:    flat_load_dwordx4 v[3:6], v[3:4]
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
-; GCN-NEXT:    buffer_store_dword v5, off, s[0:3], s32 offset:448 ; 4-byte Folded Spill
+; GCN-NEXT:    buffer_store_dword v3, off, s[0:3], s32 offset:448 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
-; GCN-NEXT:    buffer_store_dword v6, off, s[0:3], s32 offset:452 ; 4-byte Folded Spill
-; GCN-NEXT:    buffer_store_dword v7, off, s[0:3], s32 offset:456 ; 4-byte Folded Spill
-; GCN-NEXT:    buffer_store_dword v8, off, s[0:3], s32 offset:460 ; 4-byte Folded Spill
+; GCN-NEXT:    buffer_store_dword v4, off, s[0:3], s32 offset:452 ; 4-byte Folded Spill
+; GCN-NEXT:    buffer_store_dword v5, off, s[0:3], s32 offset:456 ; 4-byte Folded Spill
+; GCN-NEXT:    buffer_store_dword v6, off, s[0:3], s32 offset:460 ; 4-byte Folded Spill
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ;;#ASMEND
-; GCN-NEXT:    buffer_load_dword v5, off, s[0:3], s32 offset:448 ; 4-byte Folded Reload
-; GCN-NEXT:    buffer_load_dword v6, off, s[0:3], s32 offset:452 ; 4-byte Folded Reload
-; GCN-NEXT:    buffer_load_dword v7, off, s[0:3], s32 offset:456 ; 4-byte Folded Reload
-; GCN-NEXT:    buffer_load_dword v8, off, s[0:3], s32 offset:460 ; 4-byte Folded Reload
+; GCN-NEXT:    buffer_load_dword v3, off, s[0:3], s32 offset:448 ; 4-byte Folded Reload
+; GCN-NEXT:    buffer_load_dword v4, off, s[0:3], s32 offset:452 ; 4-byte Folded Reload
+; GCN-NEXT:    buffer_load_dword v5, off, s[0:3], s32 offset:456 ; 4-byte Folded Reload
+; GCN-NEXT:    buffer_load_dword v6, off, s[0:3], s32 offset:460 ; 4-byte Folded Reload
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ;;#ASMEND
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
-; GCN-NEXT:    flat_store_dwordx4 v[0:1], v[5:8]
-; GCN-NEXT:    v_readlane_b32 s37, v4, 3
-; GCN-NEXT:    v_readlane_b32 s36, v4, 2
-; GCN-NEXT:    v_readlane_b32 s35, v4, 1
-; GCN-NEXT:    v_readlane_b32 s34, v4, 0
+; GCN-NEXT:    flat_store_dwordx4 v[1:2], v[3:6]
+; GCN-NEXT:    v_readlane_b32 s37, v0, 3
+; GCN-NEXT:    v_readlane_b32 s36, v0, 2
+; GCN-NEXT:    v_readlane_b32 s35, v0, 1
+; GCN-NEXT:    v_readlane_b32 s34, v0, 0
 ; GCN-NEXT:    buffer_load_dword v255, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GCN-NEXT:    buffer_load_dword v254, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
 ; GCN-NEXT:    buffer_load_dword v253, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
@@ -1427,7 +1459,8 @@
 ; GCN-NEXT:    buffer_load_dword v41, off, s[0:3], s32 offset:440 ; 4-byte Folded Reload
 ; GCN-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:444 ; 4-byte Folded Reload
 ; GCN-NEXT:    s_xor_saveexec_b64 s[4:5], -1
-; GCN-NEXT:    buffer_load_dword v4, off, s[0:3], s32 offset:464 ; 4-byte Folded Reload
+; GCN-NEXT:    buffer_load_dword v4, off, s[0:3], s32 offset:468 ; 4-byte Folded Reload
+; GCN-NEXT:    buffer_load_dword v0, off, s[0:3], s32 offset:472 ; 4-byte Folded Reload
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
 ; GCN-NEXT:    s_setpc_b64 s[30:31]
@@ -1508,8 +1541,11 @@
 ; GCN-LABEL: spill_sgpr_no_free_vgpr_ipra:
 ; GCN:       ; %bb.0:
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GCN-NEXT:    s_mov_b32 s6, s33
+; GCN-NEXT:    s_mov_b32 s14, s33
 ; GCN-NEXT:    s_mov_b32 s33, s32
+; GCN-NEXT:    s_xor_saveexec_b64 s[4:5], -1
+; GCN-NEXT:    buffer_store_dword v0, off, s[0:3], s33 offset:452 ; 4-byte Folded Spill
+; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_add_i32 s32, s32, 0x7400
 ; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:444 ; 4-byte Folded Spill
 ; GCN-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:440 ; 4-byte Folded Spill
@@ -1623,21 +1659,11 @@
 ; GCN-NEXT:    buffer_store_dword v253, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
 ; GCN-NEXT:    buffer_store_dword v254, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GCN-NEXT:    buffer_store_dword v255, off, s[0:3], s33 ; 4-byte Folded Spill
-; GCN-NEXT:    s_mov_b64 s[14:15], exec
-; GCN-NEXT:    s_mov_b64 exec, 1
-; GCN-NEXT:    buffer_store_dword v1, off, s[0:3], s33 offset:456
-; GCN-NEXT:    v_writelane_b32 v1, s30, 0
-; GCN-NEXT:    buffer_store_dword v1, off, s[0:3], s33 offset:448 ; 4-byte Folded Spill
-; GCN-NEXT:    buffer_load_dword v1, off, s[0:3], s33 offset:456
-; GCN-NEXT:    s_waitcnt vmcnt(0)
-; GCN-NEXT:    s_mov_b64 exec, s[14:15]
-; GCN-NEXT:    s_mov_b64 s[12:13], exec
-; GCN-NEXT:    s_mov_b64 exec, 1
-; GCN-NEXT:    buffer_store_dword v0, off, s[0:3], s33 offset:456
-; GCN-NEXT:    v_writelane_b32 v0, s31, 0
-; GCN-NEXT:    buffer_store_dword v0, off, s[0:3], s33 offset:452 ; 4-byte Folded Spill
-; GCN-NEXT:    buffer_load_dword v0, off, s[0:3], s33 offset:456
-; GCN-NEXT:    s_waitcnt vmcnt(0)
+; GCN-NEXT:    ; implicit-def: $vgpr0
+; GCN-NEXT:    v_writelane_b32 v0, s30, 0
+; GCN-NEXT:    v_writelane_b32 v0, s31, 1
+; GCN-NEXT:    s_or_saveexec_b64 s[12:13], -1
+; GCN-NEXT:    buffer_store_dword v0, off, s[0:3], s33 offset:448 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[12:13]
 ; GCN-NEXT:    s_getpc_b64 s[4:5]
 ; GCN-NEXT:    s_add_u32 s4, s4, child_function_ipra@rel32@lo+4
@@ -1647,24 +1673,12 @@
 ; GCN-NEXT:    s_mov_b64 s[0:1], s[8:9]
 ; GCN-NEXT:    s_mov_b64 s[2:3], s[10:11]
 ; GCN-NEXT:    s_swappc_b64 s[30:31], s[4:5]
-; GCN-NEXT:    s_mov_b64 s[8:9], exec
-; GCN-NEXT:    s_mov_b64 exec, 1
-; GCN-NEXT:    buffer_store_dword v1, off, s[0:3], s33 offset:456
-; GCN-NEXT:    buffer_load_dword v1, off, s[0:3], s33 offset:452 ; 4-byte Folded Reload
-; GCN-NEXT:    s_waitcnt vmcnt(0)
-; GCN-NEXT:    v_readlane_b32 s31, v1, 0
-; GCN-NEXT:    buffer_load_dword v1, off, s[0:3], s33 offset:456
-; GCN-NEXT:    s_waitcnt vmcnt(0)
-; GCN-NEXT:    s_mov_b64 exec, s[8:9]
-; GCN-NEXT:    s_mov_b64 s[4:5], exec
-; GCN-NEXT:    s_mov_b64 exec, 1
-; GCN-NEXT:    buffer_store_dword v0, off, s[0:3], s33 offset:456
+; GCN-NEXT:    s_or_saveexec_b64 s[12:13], -1
 ; GCN-NEXT:    buffer_load_dword v0, off, s[0:3], s33 offset:448 ; 4-byte Folded Reload
+; GCN-NEXT:    s_mov_b64 exec, s[12:13]
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
+; GCN-NEXT:    v_readlane_b32 s31, v0, 1
 ; GCN-NEXT:    v_readlane_b32 s30, v0, 0
-; GCN-NEXT:    buffer_load_dword v0, off, s[0:3], s33 offset:456
-; GCN-NEXT:    s_waitcnt vmcnt(0)
-; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    buffer_load_dword v255, off, s[0:3], s33 ; 4-byte Folded Reload
 ; GCN-NEXT:    buffer_load_dword v254, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
 ; GCN-NEXT:    buffer_load_dword v253, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload
@@ -1777,8 +1791,11 @@
 ; GCN-NEXT:    buffer_load_dword v42, off, s[0:3], s33 offset:436 ; 4-byte Folded Reload
 ; GCN-NEXT:    buffer_load_dword v41, off, s[0:3], s33 offset:440 ; 4-byte Folded Reload
 ; GCN-NEXT:    buffer_load_dword v40, off, s[0:3], s33 offset:444 ; 4-byte Folded Reload
+; GCN-NEXT:    s_xor_saveexec_b64 s[4:5], -1
+; GCN-NEXT:    buffer_load_dword v0, off, s[0:3], s33 offset:452 ; 4-byte Folded Reload
+; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_add_i32 s32, s32, 0xffff8c00
-; GCN-NEXT:    s_mov_b32 s33, s6
+; GCN-NEXT:    s_mov_b32 s33, s14
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
 ; GCN-NEXT:    s_setpc_b64 s[30:31]
   call void @child_function_ipra()
diff --git a/llvm/test/CodeGen/AMDGPU/si-spill-sgpr-stack.ll b/llvm/test/CodeGen/AMDGPU/si-spill-sgpr-stack.ll
--- a/llvm/test/CodeGen/AMDGPU/si-spill-sgpr-stack.ll
+++ b/llvm/test/CodeGen/AMDGPU/si-spill-sgpr-stack.ll
@@ -6,16 +6,16 @@
 ; ALL: s_mov_b32 s[[HI:[0-9]+]], 0xe80000
 
 ; Make sure we are handling hazards correctly.
-; SGPR: buffer_load_dword [[VHI:v[0-9]+]], off, s[{{[0-9]+:[0-9]+}}], 0 offset:4
+; SGPR: v_mov_b32_e32 v0, vcc_lo
+; SGPR-NEXT: s_or_saveexec_b64 [[EXEC_COPY:s\[[0-9]+:[0-9]+\]]], -1
+; SGPR-NEXT: buffer_load_dword [[VHI:v[0-9]+]], off, s[{{[0-9]+:[0-9]+}}], 0 offset:4 ; 4-byte Folded Reload
+; SGPR-NEXT: s_mov_b64 exec, [[EXEC_COPY]]
 ; SGPR-NEXT: s_waitcnt vmcnt(0)
 ; SGPR-NEXT: v_readlane_b32 s{{[0-9]+}}, [[VHI]], 0
 ; SGPR-NEXT: v_readlane_b32 s{{[0-9]+}}, [[VHI]], 1
 ; SGPR-NEXT: v_readlane_b32 s{{[0-9]+}}, [[VHI]], 2
 ; SGPR-NEXT: v_readlane_b32 s[[HI:[0-9]+]], [[VHI]], 3
-; SGPR-NEXT: buffer_load_dword [[VHI]], off, s[96:99], 0
-; SGPR-NEXT: s_waitcnt vmcnt(0)
-; SGPR-NEXT: s_mov_b64 exec, s[4:5]
-; SGPR-NEXT: s_nop 1
+; SGPR-NEXT: s_nop 4
 ; SGPR-NEXT: buffer_store_dword v0, off, s[0:3], 0
 
 ; ALL: s_endpgm
diff --git a/llvm/test/CodeGen/AMDGPU/sibling-call.ll b/llvm/test/CodeGen/AMDGPU/sibling-call.ll
--- a/llvm/test/CodeGen/AMDGPU/sibling-call.ll
+++ b/llvm/test/CodeGen/AMDGPU/sibling-call.ll
@@ -212,15 +212,15 @@
 ; GCN-DAG: s_addc_u32 s5, s5, i32_fastcc_i32_i32@gotpcrel32@hi+12
 
 ; GCN-DAG: v_writelane_b32 [[CSRV]], s30, 0
-; GCN-DAG: buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
-; GCN-DAG: buffer_store_dword v42, off, s[0:3], s33 ; 4-byte Folded Spill
+; GCN-DAG: buffer_store_dword v40, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
+; GCN-DAG: buffer_store_dword v41, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GCN-DAG: v_writelane_b32 [[CSRV]], s31, 1
 
 
 ; GCN: s_swappc_b64
 
-; GCN-DAG: buffer_load_dword v42, off, s[0:3], s33 ; 4-byte Folded Reload
-; GCN-DAG: buffer_load_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
+; GCN-DAG: buffer_load_dword v41, off, s[0:3], s33 ; 4-byte Folded Reload
+; GCN-DAG: buffer_load_dword v40, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
 
 ; GCN: s_getpc_b64 s[4:5]
 ; GCN-NEXT: s_add_u32 s4, s4, sibling_call_i32_fastcc_i32_i32@rel32@lo+4
diff --git a/llvm/test/CodeGen/AMDGPU/spill-csr-frame-ptr-reg-copy.ll b/llvm/test/CodeGen/AMDGPU/spill-csr-frame-ptr-reg-copy.ll
--- a/llvm/test/CodeGen/AMDGPU/spill-csr-frame-ptr-reg-copy.ll
+++ b/llvm/test/CodeGen/AMDGPU/spill-csr-frame-ptr-reg-copy.ll
@@ -2,20 +2,22 @@
 
 ; GCN-LABEL: {{^}}spill_csr_s5_copy:
 ; GCN: s_mov_b32 [[FP_SCRATCH_COPY:s[0-9]+]], s33
-; GCN: s_or_saveexec_b64
-; GCN-NEXT: buffer_store_dword v40, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
-; GCN-NEXT: buffer_store_dword v41, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
+; GCN: s_xor_saveexec_b64
+; GCN-NEXT: buffer_store_dword v0, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
+; GCN-NEXT: s_mov_b64 exec, -1
+; GCN-NEXT: buffer_store_dword v40, off, s[0:3], s33 offset:12 ; 4-byte Folded Spill
 ; GCN-NEXT: s_mov_b64 exec
-; GCN: v_writelane_b32 v41, [[FP_SCRATCH_COPY]], 0
+; GCN: v_writelane_b32 v40, [[FP_SCRATCH_COPY]], 2
 ; GCN: s_swappc_b64
 
 ; GCN: v_mov_b32_e32 [[K:v[0-9]+]], 9
 ; GCN: buffer_store_dword [[K]], off, s[0:3], s33{{$}}
 
-; GCN: v_readlane_b32 [[FP_SCRATCH_COPY:s[0-9]+]], v41, 0
-; GCN: s_or_saveexec_b64
-; GCN-NEXT: buffer_load_dword v40, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
-; GCN-NEXT: buffer_load_dword v41, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload
+; GCN: v_readlane_b32 [[FP_SCRATCH_COPY:s[0-9]+]], v40, 2
+; GCN: s_xor_saveexec_b64
+; GCN-NEXT: buffer_load_dword v0, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload
+; GCN-NEXT: s_mov_b64 exec, -1
+; GCN-NEXT: buffer_load_dword v40, off, s[0:3], s33 offset:12 ; 4-byte Folded Reload
 ; GCN: s_mov_b64 exec
 ; GCN: s_mov_b32 s33, [[FP_SCRATCH_COPY]]
 ; GCN: s_setpc_b64
diff --git a/llvm/test/CodeGen/AMDGPU/spill-offset-calculation.ll b/llvm/test/CodeGen/AMDGPU/spill-offset-calculation.ll
--- a/llvm/test/CodeGen/AMDGPU/spill-offset-calculation.ll
+++ b/llvm/test/CodeGen/AMDGPU/spill-offset-calculation.ll
@@ -122,16 +122,16 @@
 ; MUBUF-NEXT:    ;;#ASMEND
 ; MUBUF-NEXT:    buffer_load_dword v0, off, s[0:3], s32 offset:8 glc
 ; MUBUF-NEXT:    s_waitcnt vmcnt(0)
-; MUBUF-NEXT:    v_mov_b32_e32 v1, 0x1004
-; MUBUF-NEXT:    buffer_store_dword v0, v1, s[0:3], s32 offen ; 4-byte Folded Spill
+; MUBUF-NEXT:    s_add_i32 s10, s32, 0x40100
+; MUBUF-NEXT:    buffer_store_dword v0, off, s[0:3], s10 ; 4-byte Folded Spill
 ; MUBUF-NEXT:    ;;#ASMSTART
 ; MUBUF-NEXT:    ;;#ASMEND
 ; MUBUF-NEXT:    ;;#ASMSTART
 ; MUBUF-NEXT:    ;;#ASMEND
 ; MUBUF-NEXT:    ;;#ASMSTART
 ; MUBUF-NEXT:    ;;#ASMEND
-; MUBUF-NEXT:    v_mov_b32_e32 v1, 0x1004
-; MUBUF-NEXT:    buffer_load_dword v0, v1, s[0:3], s32 offen ; 4-byte Folded Reload
+; MUBUF-NEXT:    s_add_i32 s10, s32, 0x40100
+; MUBUF-NEXT:    buffer_load_dword v0, off, s[0:3], s10 ; 4-byte Folded Reload
 ; MUBUF-NEXT:    s_waitcnt vmcnt(0)
 ; MUBUF-NEXT:    ;;#ASMSTART
 ; MUBUF-NEXT:    ;;#ASMEND
@@ -205,16 +205,16 @@
 ; MUBUF-NEXT:    ;;#ASMEND
 ; MUBUF-NEXT:    buffer_load_dword v0, off, s[0:3], 0 offset:8 glc
 ; MUBUF-NEXT:    s_waitcnt vmcnt(0)
-; MUBUF-NEXT:    v_mov_b32_e32 v1, 0x1004
-; MUBUF-NEXT:    buffer_store_dword v0, v1, s[0:3], 0 offen ; 4-byte Folded Spill
+; MUBUF-NEXT:    s_mov_b32 s10, 0x40100
+; MUBUF-NEXT:    buffer_store_dword v0, off, s[0:3], s10 ; 4-byte Folded Spill
 ; MUBUF-NEXT:    ;;#ASMSTART
 ; MUBUF-NEXT:    ;;#ASMEND
 ; MUBUF-NEXT:    ;;#ASMSTART
 ; MUBUF-NEXT:    ;;#ASMEND
 ; MUBUF-NEXT:    ;;#ASMSTART
 ; MUBUF-NEXT:    ;;#ASMEND
-; MUBUF-NEXT:    v_mov_b32_e32 v1, 0x1004
-; MUBUF-NEXT:    buffer_load_dword v0, v1, s[0:3], 0 offen ; 4-byte Folded Reload
+; MUBUF-NEXT:    s_mov_b32 s10, 0x40100
+; MUBUF-NEXT:    buffer_load_dword v0, off, s[0:3], s10 ; 4-byte Folded Reload
 ; MUBUF-NEXT:    s_waitcnt vmcnt(0)
 ; MUBUF-NEXT:    ;;#ASMSTART
 ; MUBUF-NEXT:    ;;#ASMEND
@@ -660,5 +660,5 @@
 
 attributes #0 = { nounwind }
 attributes #1 = { nounwind "amdgpu-num-sgpr"="17" "amdgpu-num-vgpr"="8" }
-attributes #2 = { nounwind "amdgpu-num-sgpr"="14" "amdgpu-num-vgpr"="8" }
-attributes #3 = { nounwind "amdgpu-num-sgpr"="16" "amdgpu-num-vgpr"="8" }
+attributes #2 = { nounwind "amdgpu-num-sgpr"="16" "amdgpu-num-vgpr"="8" }
+attributes #3 = { nounwind "amdgpu-num-sgpr"="18" "amdgpu-num-vgpr"="8" }
diff --git a/llvm/test/CodeGen/AMDGPU/spill-reg-tuple-super-reg-use.mir b/llvm/test/CodeGen/AMDGPU/spill-reg-tuple-super-reg-use.mir
--- a/llvm/test/CodeGen/AMDGPU/spill-reg-tuple-super-reg-use.mir
+++ b/llvm/test/CodeGen/AMDGPU/spill-reg-tuple-super-reg-use.mir
@@ -1,5 +1,5 @@
 # NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
-# RUN: llc -march=amdgcn -mcpu=gfx900 -run-pass=si-lower-sgpr-spills,prologepilog,machine-cp -verify-machineinstrs %s -o - | FileCheck -check-prefix=GCN %s
+# RUN: llc -march=amdgcn -mcpu=gfx900 -start-before=si-lower-sgpr-spills -stop-after=prologepilog -verify-machineinstrs %s -o - | FileCheck -check-prefix=GCN %s
 
 # Make sure the initial first $sgpr1 = COPY $sgpr2 copy is not deleted
 # by the copy propagation after lowering the spill.
@@ -26,11 +26,12 @@
     ; GCN-NEXT: BUFFER_STORE_DWORD_OFFSET $vgpr0, $sgpr100_sgpr101_sgpr102_sgpr103, $sgpr32, 0, 0, 0, implicit $exec :: (store (s32) into %stack.1, addrspace 5)
     ; GCN-NEXT: $exec = S_MOV_B64 killed $sgpr8_sgpr9
     ; GCN-NEXT: renamable $sgpr1 = COPY $sgpr2
-    ; GCN-NEXT: $vgpr0 = V_WRITELANE_B32 $sgpr0, 0, $vgpr0, implicit-def $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr0_sgpr1_sgpr2_sgpr3
-    ; GCN-NEXT: $vgpr0 = V_WRITELANE_B32 $sgpr1, 1, $vgpr0, implicit $sgpr0_sgpr1_sgpr2_sgpr3
-    ; GCN-NEXT: $vgpr0 = V_WRITELANE_B32 $sgpr2, 2, $vgpr0, implicit $sgpr0_sgpr1_sgpr2_sgpr3
-    ; GCN-NEXT: $vgpr0 = V_WRITELANE_B32 $sgpr3, 3, $vgpr0, implicit $sgpr0_sgpr1_sgpr2_sgpr3
-    ; GCN-NEXT: renamable $sgpr8 = COPY killed renamable $sgpr1
+    ; GCN-NEXT: renamable $vgpr0 = IMPLICIT_DEF
+    ; GCN-NEXT: renamable $vgpr0 = V_WRITELANE_B32 $sgpr0, 0, killed $vgpr0, implicit-def $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr0_sgpr1_sgpr2_sgpr3
+    ; GCN-NEXT: renamable $vgpr0 = V_WRITELANE_B32 $sgpr1, 1, killed $vgpr0, implicit $sgpr0_sgpr1_sgpr2_sgpr3
+    ; GCN-NEXT: renamable $vgpr0 = V_WRITELANE_B32 $sgpr2, 2, killed $vgpr0, implicit $sgpr0_sgpr1_sgpr2_sgpr3
+    ; GCN-NEXT: dead renamable $vgpr0 = V_WRITELANE_B32 $sgpr3, 3, killed $vgpr0, implicit $sgpr0_sgpr1_sgpr2_sgpr3
+    ; GCN-NEXT: renamable $sgpr8 = COPY renamable $sgpr1
     ; GCN-NEXT: $sgpr0_sgpr1 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
     ; GCN-NEXT: $vgpr0 = BUFFER_LOAD_DWORD_OFFSET $sgpr100_sgpr101_sgpr102_sgpr103, $sgpr32, 0, 0, 0, implicit $exec :: (load (s32) from %stack.1, addrspace 5)
     ; GCN-NEXT: $exec = S_MOV_B64 killed $sgpr0_sgpr1
@@ -63,10 +64,11 @@
     ; GCN-NEXT: BUFFER_STORE_DWORD_OFFSET $vgpr0, $sgpr100_sgpr101_sgpr102_sgpr103, $sgpr32, 0, 0, 0, implicit $exec :: (store (s32) into %stack.1, addrspace 5)
     ; GCN-NEXT: $exec = S_MOV_B64 killed $sgpr8_sgpr9
     ; GCN-NEXT: renamable $sgpr1 = COPY $sgpr2
-    ; GCN-NEXT: $vgpr0 = V_WRITELANE_B32 $sgpr0, 0, $vgpr0, implicit-def $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr0_sgpr1_sgpr2_sgpr3
-    ; GCN-NEXT: $vgpr0 = V_WRITELANE_B32 $sgpr1, 1, $vgpr0, implicit $sgpr0_sgpr1_sgpr2_sgpr3
-    ; GCN-NEXT: $vgpr0 = V_WRITELANE_B32 $sgpr2, 2, $vgpr0, implicit $sgpr0_sgpr1_sgpr2_sgpr3
-    ; GCN-NEXT: $vgpr0 = V_WRITELANE_B32 killed $sgpr3, 3, $vgpr0, implicit killed $sgpr0_sgpr1_sgpr2_sgpr3
+    ; GCN-NEXT: renamable $vgpr0 = IMPLICIT_DEF
+    ; GCN-NEXT: renamable $vgpr0 = V_WRITELANE_B32 $sgpr0, 0, killed $vgpr0, implicit-def $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr0_sgpr1_sgpr2_sgpr3
+    ; GCN-NEXT: renamable $vgpr0 = V_WRITELANE_B32 $sgpr1, 1, killed $vgpr0, implicit $sgpr0_sgpr1_sgpr2_sgpr3
+    ; GCN-NEXT: renamable $vgpr0 = V_WRITELANE_B32 $sgpr2, 2, killed $vgpr0, implicit $sgpr0_sgpr1_sgpr2_sgpr3
+    ; GCN-NEXT: dead renamable $vgpr0 = V_WRITELANE_B32 $sgpr3, 3, killed $vgpr0, implicit $sgpr0_sgpr1_sgpr2_sgpr3
     ; GCN-NEXT: $sgpr0_sgpr1 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
     ; GCN-NEXT: $vgpr0 = BUFFER_LOAD_DWORD_OFFSET $sgpr100_sgpr101_sgpr102_sgpr103, $sgpr32, 0, 0, 0, implicit $exec :: (load (s32) from %stack.1, addrspace 5)
     ; GCN-NEXT: $exec = S_MOV_B64 killed $sgpr0_sgpr1
@@ -93,12 +95,12 @@
     ; GCN-LABEL: name: spill_vgpr128_use_subreg
     ; GCN: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7
     ; GCN-NEXT: {{  $}}
-    ; GCN-NEXT: renamable $vgpr1 = COPY $vgpr2
+    ; GCN-NEXT: renamable $vgpr1 = COPY $vgpr2, implicit $exec
     ; GCN-NEXT: BUFFER_STORE_DWORD_OFFSET $vgpr0, $sgpr100_sgpr101_sgpr102_sgpr103, $sgpr32, 0, 0, 0, implicit $exec, implicit-def $vgpr0_vgpr1_vgpr2_vgpr3, implicit $vgpr0_vgpr1_vgpr2_vgpr3 :: (store (s32) into %stack.0, addrspace 5)
     ; GCN-NEXT: BUFFER_STORE_DWORD_OFFSET $vgpr1, $sgpr100_sgpr101_sgpr102_sgpr103, $sgpr32, 4, 0, 0, implicit $exec, implicit $vgpr0_vgpr1_vgpr2_vgpr3 :: (store (s32) into %stack.0 + 4, addrspace 5)
     ; GCN-NEXT: BUFFER_STORE_DWORD_OFFSET $vgpr2, $sgpr100_sgpr101_sgpr102_sgpr103, $sgpr32, 8, 0, 0, implicit $exec, implicit $vgpr0_vgpr1_vgpr2_vgpr3 :: (store (s32) into %stack.0 + 8, addrspace 5)
     ; GCN-NEXT: BUFFER_STORE_DWORD_OFFSET $vgpr3, $sgpr100_sgpr101_sgpr102_sgpr103, $sgpr32, 12, 0, 0, implicit $exec, implicit $vgpr0_vgpr1_vgpr2_vgpr3 :: (store (s32) into %stack.0 + 12, addrspace 5)
-    ; GCN-NEXT: renamable $vgpr8 = COPY killed renamable $vgpr1
+    ; GCN-NEXT: renamable $vgpr8 = COPY $vgpr2, implicit $exec
     ; GCN-NEXT: S_ENDPGM 0, implicit $vgpr8
     renamable $vgpr1 = COPY $vgpr2
     SI_SPILL_V128_SAVE renamable $vgpr0_vgpr1_vgpr2_vgpr3, %stack.0, $sgpr32, 0, implicit $exec :: (store (s128) into %stack.0, align 4, addrspace 5)
@@ -123,11 +125,11 @@
     ; GCN-LABEL: name: spill_vgpr128_use_kill
     ; GCN: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7
     ; GCN-NEXT: {{  $}}
-    ; GCN-NEXT: renamable $vgpr1 = COPY $vgpr2
-    ; GCN-NEXT: BUFFER_STORE_DWORD_OFFSET killed $vgpr0, $sgpr100_sgpr101_sgpr102_sgpr103, $sgpr32, 0, 0, 0, implicit $exec, implicit-def $vgpr0_vgpr1_vgpr2_vgpr3, implicit $vgpr0_vgpr1_vgpr2_vgpr3 :: (store (s32) into %stack.0, addrspace 5)
-    ; GCN-NEXT: BUFFER_STORE_DWORD_OFFSET killed $vgpr1, $sgpr100_sgpr101_sgpr102_sgpr103, $sgpr32, 4, 0, 0, implicit $exec, implicit $vgpr0_vgpr1_vgpr2_vgpr3 :: (store (s32) into %stack.0 + 4, addrspace 5)
-    ; GCN-NEXT: BUFFER_STORE_DWORD_OFFSET killed $vgpr2, $sgpr100_sgpr101_sgpr102_sgpr103, $sgpr32, 8, 0, 0, implicit $exec, implicit $vgpr0_vgpr1_vgpr2_vgpr3 :: (store (s32) into %stack.0 + 8, addrspace 5)
-    ; GCN-NEXT: BUFFER_STORE_DWORD_OFFSET killed $vgpr3, $sgpr100_sgpr101_sgpr102_sgpr103, $sgpr32, 12, 0, 0, implicit $exec, implicit killed $vgpr0_vgpr1_vgpr2_vgpr3 :: (store (s32) into %stack.0 + 12, addrspace 5)
+    ; GCN-NEXT: renamable $vgpr1 = COPY $vgpr2, implicit $exec
+    ; GCN-NEXT: BUFFER_STORE_DWORD_OFFSET $vgpr0, $sgpr100_sgpr101_sgpr102_sgpr103, $sgpr32, 0, 0, 0, implicit $exec, implicit-def $vgpr0_vgpr1_vgpr2_vgpr3, implicit $vgpr0_vgpr1_vgpr2_vgpr3 :: (store (s32) into %stack.0, addrspace 5)
+    ; GCN-NEXT: BUFFER_STORE_DWORD_OFFSET $vgpr1, $sgpr100_sgpr101_sgpr102_sgpr103, $sgpr32, 4, 0, 0, implicit $exec, implicit $vgpr0_vgpr1_vgpr2_vgpr3 :: (store (s32) into %stack.0 + 4, addrspace 5)
+    ; GCN-NEXT: BUFFER_STORE_DWORD_OFFSET $vgpr2, $sgpr100_sgpr101_sgpr102_sgpr103, $sgpr32, 8, 0, 0, implicit $exec, implicit $vgpr0_vgpr1_vgpr2_vgpr3 :: (store (s32) into %stack.0 + 8, addrspace 5)
+    ; GCN-NEXT: BUFFER_STORE_DWORD_OFFSET $vgpr3, $sgpr100_sgpr101_sgpr102_sgpr103, $sgpr32, 12, 0, 0, implicit $exec, implicit $vgpr0_vgpr1_vgpr2_vgpr3 :: (store (s32) into %stack.0 + 12, addrspace 5)
     ; GCN-NEXT: S_ENDPGM 0
     renamable $vgpr1 = COPY $vgpr2
     SI_SPILL_V128_SAVE renamable killed $vgpr0_vgpr1_vgpr2_vgpr3, %stack.0, $sgpr32, 0, implicit $exec :: (store (s128) into %stack.0, align 4, addrspace 5)
diff --git a/llvm/test/CodeGen/AMDGPU/spill-scavenge-offset.ll b/llvm/test/CodeGen/AMDGPU/spill-scavenge-offset.ll
--- a/llvm/test/CodeGen/AMDGPU/spill-scavenge-offset.ll
+++ b/llvm/test/CodeGen/AMDGPU/spill-scavenge-offset.ll
@@ -10085,16 +10085,25 @@
 ; GFX6-NEXT:    v_mbcnt_lo_u32_b32_e64 v0, -1, 0
 ; GFX6-NEXT:    v_mbcnt_hi_u32_b32_e32 v5, -1, v0
 ; GFX6-NEXT:    v_mov_b32_e32 v6, 0
-; GFX6-NEXT:    s_mov_b32 s38, 0
-; GFX6-NEXT:    s_mov_b32 s39, 0xf000
+; GFX6-NEXT:    s_mov_b32 s6, 0
+; GFX6-NEXT:    s_mov_b32 s7, 0xf000
 ; GFX6-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX6-NEXT:    s_mov_b64 s[36:37], s[2:3]
+; GFX6-NEXT:    s_mov_b64 s[4:5], s[2:3]
 ; GFX6-NEXT:    v_lshlrev_b32_e32 v7, 8, v5
 ; GFX6-NEXT:    v_mov_b32_e32 v8, v6
-; GFX6-NEXT:    buffer_load_dwordx4 v[0:3], v[7:8], s[36:39], 0 addr64 offset:240
+; GFX6-NEXT:    buffer_load_dwordx4 v[0:3], v[7:8], s[4:7], 0 addr64 offset:240
 ; GFX6-NEXT:    s_addc_u32 s41, s41, 0
+; GFX6-NEXT:    s_mov_b32 s2, 0x83800
+; GFX6-NEXT:    s_mov_b64 s[34:35], exec
+; GFX6-NEXT:    s_waitcnt vmcnt(0)
+; GFX6-NEXT:    buffer_store_dword v0, off, s[40:43], s2 ; 4-byte Folded Spill
+; GFX6-NEXT:    s_waitcnt vmcnt(0)
+; GFX6-NEXT:    buffer_store_dword v1, off, s[40:43], s2 offset:4 ; 4-byte Folded Spill
+; GFX6-NEXT:    buffer_store_dword v2, off, s[40:43], s2 offset:8 ; 4-byte Folded Spill
+; GFX6-NEXT:    buffer_store_dword v3, off, s[40:43], s2 offset:12 ; 4-byte Folded Spill
+; GFX6-NEXT:    s_waitcnt expcnt(0)
+; GFX6-NEXT:    buffer_load_dwordx4 v[0:3], v[7:8], s[4:7], 0 addr64 offset:224
 ; GFX6-NEXT:    s_mov_b32 s2, 0x83400
-; GFX6-NEXT:    s_mov_b64 s[44:45], exec
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    buffer_store_dword v0, off, s[40:43], s2 ; 4-byte Folded Spill
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
@@ -10102,7 +10111,7 @@
 ; GFX6-NEXT:    buffer_store_dword v2, off, s[40:43], s2 offset:8 ; 4-byte Folded Spill
 ; GFX6-NEXT:    buffer_store_dword v3, off, s[40:43], s2 offset:12 ; 4-byte Folded Spill
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    buffer_load_dwordx4 v[0:3], v[7:8], s[36:39], 0 addr64 offset:224
+; GFX6-NEXT:    buffer_load_dwordx4 v[0:3], v[7:8], s[4:7], 0 addr64 offset:208
 ; GFX6-NEXT:    s_mov_b32 s2, 0x83000
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    buffer_store_dword v0, off, s[40:43], s2 ; 4-byte Folded Spill
@@ -10111,7 +10120,7 @@
 ; GFX6-NEXT:    buffer_store_dword v2, off, s[40:43], s2 offset:8 ; 4-byte Folded Spill
 ; GFX6-NEXT:    buffer_store_dword v3, off, s[40:43], s2 offset:12 ; 4-byte Folded Spill
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    buffer_load_dwordx4 v[0:3], v[7:8], s[36:39], 0 addr64 offset:208
+; GFX6-NEXT:    buffer_load_dwordx4 v[0:3], v[7:8], s[4:7], 0 addr64 offset:192
 ; GFX6-NEXT:    s_mov_b32 s2, 0x82c00
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    buffer_store_dword v0, off, s[40:43], s2 ; 4-byte Folded Spill
@@ -10120,7 +10129,7 @@
 ; GFX6-NEXT:    buffer_store_dword v2, off, s[40:43], s2 offset:8 ; 4-byte Folded Spill
 ; GFX6-NEXT:    buffer_store_dword v3, off, s[40:43], s2 offset:12 ; 4-byte Folded Spill
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    buffer_load_dwordx4 v[0:3], v[7:8], s[36:39], 0 addr64 offset:192
+; GFX6-NEXT:    buffer_load_dwordx4 v[0:3], v[7:8], s[4:7], 0 addr64 offset:176
 ; GFX6-NEXT:    s_mov_b32 s2, 0x82800
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    buffer_store_dword v0, off, s[40:43], s2 ; 4-byte Folded Spill
@@ -10129,7 +10138,7 @@
 ; GFX6-NEXT:    buffer_store_dword v2, off, s[40:43], s2 offset:8 ; 4-byte Folded Spill
 ; GFX6-NEXT:    buffer_store_dword v3, off, s[40:43], s2 offset:12 ; 4-byte Folded Spill
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    buffer_load_dwordx4 v[0:3], v[7:8], s[36:39], 0 addr64 offset:176
+; GFX6-NEXT:    buffer_load_dwordx4 v[0:3], v[7:8], s[4:7], 0 addr64 offset:160
 ; GFX6-NEXT:    s_mov_b32 s2, 0x82400
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    buffer_store_dword v0, off, s[40:43], s2 ; 4-byte Folded Spill
@@ -10138,7 +10147,7 @@
 ; GFX6-NEXT:    buffer_store_dword v2, off, s[40:43], s2 offset:8 ; 4-byte Folded Spill
 ; GFX6-NEXT:    buffer_store_dword v3, off, s[40:43], s2 offset:12 ; 4-byte Folded Spill
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    buffer_load_dwordx4 v[0:3], v[7:8], s[36:39], 0 addr64 offset:160
+; GFX6-NEXT:    buffer_load_dwordx4 v[0:3], v[7:8], s[4:7], 0 addr64 offset:144
 ; GFX6-NEXT:    s_mov_b32 s2, 0x82000
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    buffer_store_dword v0, off, s[40:43], s2 ; 4-byte Folded Spill
@@ -10147,7 +10156,7 @@
 ; GFX6-NEXT:    buffer_store_dword v2, off, s[40:43], s2 offset:8 ; 4-byte Folded Spill
 ; GFX6-NEXT:    buffer_store_dword v3, off, s[40:43], s2 offset:12 ; 4-byte Folded Spill
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    buffer_load_dwordx4 v[0:3], v[7:8], s[36:39], 0 addr64 offset:144
+; GFX6-NEXT:    buffer_load_dwordx4 v[0:3], v[7:8], s[4:7], 0 addr64 offset:128
 ; GFX6-NEXT:    s_mov_b32 s2, 0x81c00
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    buffer_store_dword v0, off, s[40:43], s2 ; 4-byte Folded Spill
@@ -10156,7 +10165,7 @@
 ; GFX6-NEXT:    buffer_store_dword v2, off, s[40:43], s2 offset:8 ; 4-byte Folded Spill
 ; GFX6-NEXT:    buffer_store_dword v3, off, s[40:43], s2 offset:12 ; 4-byte Folded Spill
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    buffer_load_dwordx4 v[0:3], v[7:8], s[36:39], 0 addr64 offset:128
+; GFX6-NEXT:    buffer_load_dwordx4 v[0:3], v[7:8], s[4:7], 0 addr64 offset:112
 ; GFX6-NEXT:    s_mov_b32 s2, 0x81800
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    buffer_store_dword v0, off, s[40:43], s2 ; 4-byte Folded Spill
@@ -10165,7 +10174,7 @@
 ; GFX6-NEXT:    buffer_store_dword v2, off, s[40:43], s2 offset:8 ; 4-byte Folded Spill
 ; GFX6-NEXT:    buffer_store_dword v3, off, s[40:43], s2 offset:12 ; 4-byte Folded Spill
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    buffer_load_dwordx4 v[0:3], v[7:8], s[36:39], 0 addr64 offset:112
+; GFX6-NEXT:    buffer_load_dwordx4 v[0:3], v[7:8], s[4:7], 0 addr64 offset:96
 ; GFX6-NEXT:    s_mov_b32 s2, 0x81400
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    buffer_store_dword v0, off, s[40:43], s2 ; 4-byte Folded Spill
@@ -10174,7 +10183,7 @@
 ; GFX6-NEXT:    buffer_store_dword v2, off, s[40:43], s2 offset:8 ; 4-byte Folded Spill
 ; GFX6-NEXT:    buffer_store_dword v3, off, s[40:43], s2 offset:12 ; 4-byte Folded Spill
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    buffer_load_dwordx4 v[0:3], v[7:8], s[36:39], 0 addr64 offset:96
+; GFX6-NEXT:    buffer_load_dwordx4 v[0:3], v[7:8], s[4:7], 0 addr64 offset:80
 ; GFX6-NEXT:    s_mov_b32 s2, 0x81000
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    buffer_store_dword v0, off, s[40:43], s2 ; 4-byte Folded Spill
@@ -10183,17 +10192,8 @@
 ; GFX6-NEXT:    buffer_store_dword v2, off, s[40:43], s2 offset:8 ; 4-byte Folded Spill
 ; GFX6-NEXT:    buffer_store_dword v3, off, s[40:43], s2 offset:12 ; 4-byte Folded Spill
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    buffer_load_dwordx4 v[0:3], v[7:8], s[36:39], 0 addr64 offset:80
-; GFX6-NEXT:    s_mov_b32 s2, 0x80c00
-; GFX6-NEXT:    s_waitcnt vmcnt(0)
-; GFX6-NEXT:    buffer_store_dword v0, off, s[40:43], s2 ; 4-byte Folded Spill
-; GFX6-NEXT:    s_waitcnt vmcnt(0)
-; GFX6-NEXT:    buffer_store_dword v1, off, s[40:43], s2 offset:4 ; 4-byte Folded Spill
-; GFX6-NEXT:    buffer_store_dword v2, off, s[40:43], s2 offset:8 ; 4-byte Folded Spill
-; GFX6-NEXT:    buffer_store_dword v3, off, s[40:43], s2 offset:12 ; 4-byte Folded Spill
-; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    buffer_load_dwordx4 v[0:3], v[7:8], s[36:39], 0 addr64 offset:64
-; GFX6-NEXT:    s_mov_b32 s2, 0x80400
+; GFX6-NEXT:    buffer_load_dwordx4 v[0:3], v[7:8], s[4:7], 0 addr64 offset:64
+; GFX6-NEXT:    s_mov_b32 s2, 0x80800
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    buffer_store_dword v0, off, s[40:43], s2 ; 4-byte Folded Spill
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
@@ -10201,17 +10201,32 @@
 ; GFX6-NEXT:    buffer_store_dword v2, off, s[40:43], s2 offset:8 ; 4-byte Folded Spill
 ; GFX6-NEXT:    buffer_store_dword v3, off, s[40:43], s2 offset:12 ; 4-byte Folded Spill
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    buffer_load_dwordx4 v[0:3], v[7:8], s[36:39], 0 addr64
-; GFX6-NEXT:    buffer_load_dwordx4 v[9:12], v[7:8], s[36:39], 0 addr64 offset:16
-; GFX6-NEXT:    s_mov_b32 s2, 0x80800
+; GFX6-NEXT:    buffer_load_dwordx4 v[0:3], v[7:8], s[4:7], 0 addr64
+; GFX6-NEXT:    buffer_load_dwordx4 v[9:12], v[7:8], s[4:7], 0 addr64 offset:16
+; GFX6-NEXT:    s_mov_b32 s2, 0x80c00
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    buffer_store_dword v9, off, s[40:43], s2 ; 4-byte Folded Spill
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    buffer_store_dword v10, off, s[40:43], s2 offset:4 ; 4-byte Folded Spill
 ; GFX6-NEXT:    buffer_store_dword v11, off, s[40:43], s2 offset:8 ; 4-byte Folded Spill
 ; GFX6-NEXT:    buffer_store_dword v12, off, s[40:43], s2 offset:12 ; 4-byte Folded Spill
-; GFX6-NEXT:    buffer_load_dwordx4 v[13:16], v[7:8], s[36:39], 0 addr64 offset:32
-; GFX6-NEXT:    buffer_load_dwordx4 v[17:20], v[7:8], s[36:39], 0 addr64 offset:48
+; GFX6-NEXT:    buffer_load_dwordx4 v[13:16], v[7:8], s[4:7], 0 addr64 offset:32
+; GFX6-NEXT:    s_mov_b64 s[2:3], s[6:7]
+; GFX6-NEXT:    s_waitcnt expcnt(3)
+; GFX6-NEXT:    s_mov_b64 exec, 15
+; GFX6-NEXT:    buffer_store_dword v9, off, s[40:43], 0
+; GFX6-NEXT:    s_waitcnt expcnt(0)
+; GFX6-NEXT:    v_writelane_b32 v9, s0, 0
+; GFX6-NEXT:    v_writelane_b32 v9, s1, 1
+; GFX6-NEXT:    v_writelane_b32 v9, s2, 2
+; GFX6-NEXT:    v_writelane_b32 v9, s3, 3
+; GFX6-NEXT:    s_mov_b32 s8, 0x80400
+; GFX6-NEXT:    buffer_store_dword v9, off, s[40:43], s8 ; 4-byte Folded Spill
+; GFX6-NEXT:    s_waitcnt expcnt(0)
+; GFX6-NEXT:    buffer_load_dword v9, off, s[40:43], 0
+; GFX6-NEXT:    s_waitcnt vmcnt(0)
+; GFX6-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX6-NEXT:    buffer_load_dwordx4 v[17:20], v[7:8], s[4:7], 0 addr64 offset:48
 ; GFX6-NEXT:    v_lshlrev_b32_e32 v4, 13, v0
 ; GFX6-NEXT:    v_add_i32_e32 v4, vcc, 16, v4
 ; GFX6-NEXT:    v_mov_b32_e32 v7, 1
@@ -10219,6 +10234,7 @@
 ; GFX6-NEXT:    ;;#ASMSTART
 ; GFX6-NEXT:    ; def s[4:11]
 ; GFX6-NEXT:    ;;#ASMEND
+; GFX6-NEXT:    s_mov_b64 s[36:37], exec
 ; GFX6-NEXT:    s_mov_b64 exec, 0xff
 ; GFX6-NEXT:    buffer_store_dword v4, off, s[40:43], 0
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
@@ -10230,12 +10246,12 @@
 ; GFX6-NEXT:    v_writelane_b32 v4, s9, 5
 ; GFX6-NEXT:    v_writelane_b32 v4, s10, 6
 ; GFX6-NEXT:    v_writelane_b32 v4, s11, 7
-; GFX6-NEXT:    s_mov_b32 s2, 0x83800
+; GFX6-NEXT:    s_mov_b32 s2, 0x83c00
 ; GFX6-NEXT:    buffer_store_dword v4, off, s[40:43], s2 ; 4-byte Folded Spill
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
 ; GFX6-NEXT:    buffer_load_dword v4, off, s[40:43], 0
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
-; GFX6-NEXT:    s_mov_b64 exec, s[44:45]
+; GFX6-NEXT:    s_mov_b64 exec, s[36:37]
 ; GFX6-NEXT:    v_cmp_eq_u32_e32 vcc, 0, v0
 ; GFX6-NEXT:    ;;#ASMSTART
 ; GFX6-NEXT:    ; def s[8:15]
@@ -10253,272 +10269,211 @@
 ; GFX6-NEXT:    ; def s[2:3]
 ; GFX6-NEXT:    ;;#ASMEND
 ; GFX6-NEXT:    ;;#ASMSTART
-; GFX6-NEXT:    ; def s[36:37]
-; GFX6-NEXT:    ;;#ASMEND
-; GFX6-NEXT:    ;;#ASMSTART
 ; GFX6-NEXT:    ; def s33
 ; GFX6-NEXT:    ;;#ASMEND
 ; GFX6-NEXT:    s_and_saveexec_b64 s[34:35], vcc
 ; GFX6-NEXT:    s_cbranch_execz .LBB1_2
 ; GFX6-NEXT:  ; %bb.1: ; %bb0
-; GFX6-NEXT:    s_mov_b64 s[44:45], exec
-; GFX6-NEXT:    s_mov_b64 exec, 0xff
-; GFX6-NEXT:    buffer_store_dword v9, off, s[40:43], 0
-; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    v_writelane_b32 v9, s8, 0
-; GFX6-NEXT:    v_writelane_b32 v9, s9, 1
-; GFX6-NEXT:    v_writelane_b32 v9, s10, 2
-; GFX6-NEXT:    v_writelane_b32 v9, s11, 3
-; GFX6-NEXT:    v_writelane_b32 v9, s12, 4
-; GFX6-NEXT:    v_writelane_b32 v9, s13, 5
-; GFX6-NEXT:    v_writelane_b32 v9, s14, 6
-; GFX6-NEXT:    v_writelane_b32 v9, s15, 7
-; GFX6-NEXT:    v_mov_b32_e32 v4, 0x2100
-; GFX6-NEXT:    buffer_store_dword v9, v4, s[40:43], 0 offen ; 4-byte Folded Spill
-; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    buffer_load_dword v9, off, s[40:43], 0
-; GFX6-NEXT:    s_waitcnt vmcnt(0)
-; GFX6-NEXT:    s_mov_b64 exec, s[44:45]
-; GFX6-NEXT:    s_mov_b64 s[44:45], exec
-; GFX6-NEXT:    s_mov_b64 exec, 0xff
-; GFX6-NEXT:    v_mov_b32_e32 v4, 0x20e0
-; GFX6-NEXT:    buffer_store_dword v8, off, s[40:43], 0
-; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    buffer_load_dword v8, v4, s[40:43], 0 offen ; 4-byte Folded Reload
-; GFX6-NEXT:    s_waitcnt vmcnt(0)
-; GFX6-NEXT:    v_readlane_b32 s8, v8, 0
-; GFX6-NEXT:    v_readlane_b32 s9, v8, 1
-; GFX6-NEXT:    v_readlane_b32 s10, v8, 2
-; GFX6-NEXT:    v_readlane_b32 s11, v8, 3
-; GFX6-NEXT:    v_readlane_b32 s12, v8, 4
-; GFX6-NEXT:    v_readlane_b32 s13, v8, 5
-; GFX6-NEXT:    v_readlane_b32 s14, v8, 6
-; GFX6-NEXT:    v_readlane_b32 s15, v8, 7
-; GFX6-NEXT:    buffer_load_dword v8, off, s[40:43], 0
-; GFX6-NEXT:    s_waitcnt vmcnt(0)
-; GFX6-NEXT:    s_mov_b64 exec, s[44:45]
-; GFX6-NEXT:    s_mov_b64 s[44:45], exec
+; GFX6-NEXT:    s_mov_b64 s[38:39], exec
 ; GFX6-NEXT:    s_mov_b64 exec, 0xff
 ; GFX6-NEXT:    buffer_store_dword v7, off, s[40:43], 0
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    v_writelane_b32 v7, s16, 0
-; GFX6-NEXT:    v_writelane_b32 v7, s17, 1
-; GFX6-NEXT:    v_writelane_b32 v7, s18, 2
-; GFX6-NEXT:    v_writelane_b32 v7, s19, 3
-; GFX6-NEXT:    v_writelane_b32 v7, s20, 4
-; GFX6-NEXT:    v_writelane_b32 v7, s21, 5
-; GFX6-NEXT:    v_writelane_b32 v7, s22, 6
-; GFX6-NEXT:    v_writelane_b32 v7, s23, 7
-; GFX6-NEXT:    v_mov_b32_e32 v4, 0x2120
-; GFX6-NEXT:    buffer_store_dword v7, v4, s[40:43], 0 offen ; 4-byte Folded Spill
+; GFX6-NEXT:    v_writelane_b32 v7, s8, 0
+; GFX6-NEXT:    v_writelane_b32 v7, s9, 1
+; GFX6-NEXT:    v_writelane_b32 v7, s10, 2
+; GFX6-NEXT:    v_writelane_b32 v7, s11, 3
+; GFX6-NEXT:    v_writelane_b32 v7, s12, 4
+; GFX6-NEXT:    v_writelane_b32 v7, s13, 5
+; GFX6-NEXT:    v_writelane_b32 v7, s14, 6
+; GFX6-NEXT:    v_writelane_b32 v7, s15, 7
+; GFX6-NEXT:    s_mov_b32 s36, 0x84400
+; GFX6-NEXT:    buffer_store_dword v7, off, s[40:43], s36 ; 4-byte Folded Spill
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
 ; GFX6-NEXT:    buffer_load_dword v7, off, s[40:43], 0
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
-; GFX6-NEXT:    s_mov_b64 exec, s[44:45]
+; GFX6-NEXT:    s_mov_b64 exec, s[38:39]
 ; GFX6-NEXT:    s_mov_b64 s[44:45], exec
 ; GFX6-NEXT:    s_mov_b64 exec, 0xff
-; GFX6-NEXT:    v_mov_b32_e32 v4, 0x2100
-; GFX6-NEXT:    buffer_store_dword v9, off, s[40:43], 0
+; GFX6-NEXT:    s_mov_b32 s36, 0x83c00
+; GFX6-NEXT:    buffer_store_dword v4, off, s[40:43], 0
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    buffer_load_dword v9, v4, s[40:43], 0 offen ; 4-byte Folded Reload
+; GFX6-NEXT:    buffer_load_dword v4, off, s[40:43], s36 ; 4-byte Folded Reload
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
-; GFX6-NEXT:    v_readlane_b32 s16, v9, 0
-; GFX6-NEXT:    v_readlane_b32 s17, v9, 1
-; GFX6-NEXT:    v_readlane_b32 s18, v9, 2
-; GFX6-NEXT:    v_readlane_b32 s19, v9, 3
-; GFX6-NEXT:    v_readlane_b32 s20, v9, 4
-; GFX6-NEXT:    v_readlane_b32 s21, v9, 5
-; GFX6-NEXT:    v_readlane_b32 s22, v9, 6
-; GFX6-NEXT:    v_readlane_b32 s23, v9, 7
-; GFX6-NEXT:    buffer_load_dword v9, off, s[40:43], 0
+; GFX6-NEXT:    v_readlane_b32 s8, v4, 0
+; GFX6-NEXT:    v_readlane_b32 s9, v4, 1
+; GFX6-NEXT:    v_readlane_b32 s10, v4, 2
+; GFX6-NEXT:    v_readlane_b32 s11, v4, 3
+; GFX6-NEXT:    v_readlane_b32 s12, v4, 4
+; GFX6-NEXT:    v_readlane_b32 s13, v4, 5
+; GFX6-NEXT:    v_readlane_b32 s14, v4, 6
+; GFX6-NEXT:    v_readlane_b32 s15, v4, 7
+; GFX6-NEXT:    buffer_load_dword v4, off, s[40:43], 0
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    s_mov_b64 exec, s[44:45]
-; GFX6-NEXT:    s_mov_b64 s[44:45], exec
+; GFX6-NEXT:    s_mov_b64 s[38:39], exec
 ; GFX6-NEXT:    s_mov_b64 exec, 0xff
 ; GFX6-NEXT:    buffer_store_dword v8, off, s[40:43], 0
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    v_writelane_b32 v8, s24, 0
-; GFX6-NEXT:    v_writelane_b32 v8, s25, 1
-; GFX6-NEXT:    v_writelane_b32 v8, s26, 2
-; GFX6-NEXT:    v_writelane_b32 v8, s27, 3
-; GFX6-NEXT:    v_writelane_b32 v8, s28, 4
-; GFX6-NEXT:    v_writelane_b32 v8, s29, 5
-; GFX6-NEXT:    v_writelane_b32 v8, s30, 6
-; GFX6-NEXT:    v_writelane_b32 v8, s31, 7
-; GFX6-NEXT:    v_mov_b32_e32 v4, 0x2140
-; GFX6-NEXT:    buffer_store_dword v8, v4, s[40:43], 0 offen ; 4-byte Folded Spill
+; GFX6-NEXT:    v_writelane_b32 v8, s16, 0
+; GFX6-NEXT:    v_writelane_b32 v8, s17, 1
+; GFX6-NEXT:    v_writelane_b32 v8, s18, 2
+; GFX6-NEXT:    v_writelane_b32 v8, s19, 3
+; GFX6-NEXT:    v_writelane_b32 v8, s20, 4
+; GFX6-NEXT:    v_writelane_b32 v8, s21, 5
+; GFX6-NEXT:    v_writelane_b32 v8, s22, 6
+; GFX6-NEXT:    v_writelane_b32 v8, s23, 7
+; GFX6-NEXT:    s_mov_b32 s36, 0x84c00
+; GFX6-NEXT:    buffer_store_dword v8, off, s[40:43], s36 ; 4-byte Folded Spill
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
 ; GFX6-NEXT:    buffer_load_dword v8, off, s[40:43], 0
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
-; GFX6-NEXT:    s_mov_b64 exec, s[44:45]
+; GFX6-NEXT:    s_mov_b64 exec, s[38:39]
 ; GFX6-NEXT:    s_mov_b64 s[44:45], exec
 ; GFX6-NEXT:    s_mov_b64 exec, 0xff
-; GFX6-NEXT:    v_mov_b32_e32 v4, 0x2120
+; GFX6-NEXT:    s_mov_b32 s36, 0x84400
 ; GFX6-NEXT:    buffer_store_dword v7, off, s[40:43], 0
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    buffer_load_dword v7, v4, s[40:43], 0 offen ; 4-byte Folded Reload
+; GFX6-NEXT:    buffer_load_dword v7, off, s[40:43], s36 ; 4-byte Folded Reload
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
-; GFX6-NEXT:    v_readlane_b32 s24, v7, 0
-; GFX6-NEXT:    v_readlane_b32 s25, v7, 1
-; GFX6-NEXT:    v_readlane_b32 s26, v7, 2
-; GFX6-NEXT:    v_readlane_b32 s27, v7, 3
-; GFX6-NEXT:    v_readlane_b32 s28, v7, 4
-; GFX6-NEXT:    v_readlane_b32 s29, v7, 5
-; GFX6-NEXT:    v_readlane_b32 s30, v7, 6
-; GFX6-NEXT:    v_readlane_b32 s31, v7, 7
+; GFX6-NEXT:    v_readlane_b32 s16, v7, 0
+; GFX6-NEXT:    v_readlane_b32 s17, v7, 1
+; GFX6-NEXT:    v_readlane_b32 s18, v7, 2
+; GFX6-NEXT:    v_readlane_b32 s19, v7, 3
+; GFX6-NEXT:    v_readlane_b32 s20, v7, 4
+; GFX6-NEXT:    v_readlane_b32 s21, v7, 5
+; GFX6-NEXT:    v_readlane_b32 s22, v7, 6
+; GFX6-NEXT:    v_readlane_b32 s23, v7, 7
 ; GFX6-NEXT:    buffer_load_dword v7, off, s[40:43], 0
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    s_mov_b64 exec, s[44:45]
-; GFX6-NEXT:    s_mov_b64 s[44:45], exec
-; GFX6-NEXT:    s_mov_b64 exec, 15
-; GFX6-NEXT:    buffer_store_dword v10, off, s[40:43], 0
-; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    v_writelane_b32 v10, s0, 0
-; GFX6-NEXT:    v_writelane_b32 v10, s1, 1
-; GFX6-NEXT:    v_writelane_b32 v10, s2, 2
-; GFX6-NEXT:    v_writelane_b32 v10, s3, 3
-; GFX6-NEXT:    v_mov_b32_e32 v4, 0x2160
-; GFX6-NEXT:    buffer_store_dword v10, v4, s[40:43], 0 offen ; 4-byte Folded Spill
-; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    buffer_load_dword v10, off, s[40:43], 0
-; GFX6-NEXT:    s_waitcnt vmcnt(0)
-; GFX6-NEXT:    s_mov_b64 exec, s[44:45]
-; GFX6-NEXT:    s_mov_b64 s[44:45], exec
-; GFX6-NEXT:    s_mov_b64 exec, 15
-; GFX6-NEXT:    buffer_store_dword v8, off, s[40:43], 0
-; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    v_writelane_b32 v8, s4, 0
-; GFX6-NEXT:    v_writelane_b32 v8, s5, 1
-; GFX6-NEXT:    v_writelane_b32 v8, s6, 2
-; GFX6-NEXT:    v_writelane_b32 v8, s7, 3
-; GFX6-NEXT:    s_mov_b32 s0, 0x85c00
-; GFX6-NEXT:    buffer_store_dword v8, off, s[40:43], s0 ; 4-byte Folded Spill
-; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    buffer_load_dword v8, off, s[40:43], 0
-; GFX6-NEXT:    s_waitcnt vmcnt(0)
-; GFX6-NEXT:    s_mov_b64 exec, s[44:45]
-; GFX6-NEXT:    s_mov_b64 s[0:1], exec
-; GFX6-NEXT:    s_mov_b64 exec, 3
-; GFX6-NEXT:    buffer_store_dword v9, off, s[40:43], 0
+; GFX6-NEXT:    s_mov_b64 s[38:39], exec
+; GFX6-NEXT:    s_mov_b64 exec, 0xff
+; GFX6-NEXT:    buffer_store_dword v4, off, s[40:43], 0
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    v_writelane_b32 v9, s2, 0
-; GFX6-NEXT:    v_writelane_b32 v9, s3, 1
-; GFX6-NEXT:    s_mov_b32 s4, 0x86600
-; GFX6-NEXT:    buffer_store_dword v9, off, s[40:43], s4 ; 4-byte Folded Spill
+; GFX6-NEXT:    v_writelane_b32 v4, s24, 0
+; GFX6-NEXT:    v_writelane_b32 v4, s25, 1
+; GFX6-NEXT:    v_writelane_b32 v4, s26, 2
+; GFX6-NEXT:    v_writelane_b32 v4, s27, 3
+; GFX6-NEXT:    v_writelane_b32 v4, s28, 4
+; GFX6-NEXT:    v_writelane_b32 v4, s29, 5
+; GFX6-NEXT:    v_writelane_b32 v4, s30, 6
+; GFX6-NEXT:    v_writelane_b32 v4, s31, 7
+; GFX6-NEXT:    s_mov_b32 s36, 0x85400
+; GFX6-NEXT:    buffer_store_dword v4, off, s[40:43], s36 ; 4-byte Folded Spill
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    buffer_load_dword v9, off, s[40:43], 0
+; GFX6-NEXT:    buffer_load_dword v4, off, s[40:43], 0
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
-; GFX6-NEXT:    s_mov_b64 exec, s[0:1]
+; GFX6-NEXT:    s_mov_b64 exec, s[38:39]
 ; GFX6-NEXT:    s_mov_b64 s[44:45], exec
 ; GFX6-NEXT:    s_mov_b64 exec, 0xff
-; GFX6-NEXT:    v_mov_b32_e32 v4, 0x2140
-; GFX6-NEXT:    buffer_store_dword v7, off, s[40:43], 0
+; GFX6-NEXT:    s_mov_b32 s36, 0x84c00
+; GFX6-NEXT:    buffer_store_dword v9, off, s[40:43], 0
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    buffer_load_dword v7, v4, s[40:43], 0 offen ; 4-byte Folded Reload
+; GFX6-NEXT:    buffer_load_dword v9, off, s[40:43], s36 ; 4-byte Folded Reload
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
-; GFX6-NEXT:    v_readlane_b32 s0, v7, 0
-; GFX6-NEXT:    v_readlane_b32 s1, v7, 1
-; GFX6-NEXT:    v_readlane_b32 s2, v7, 2
-; GFX6-NEXT:    v_readlane_b32 s3, v7, 3
-; GFX6-NEXT:    v_readlane_b32 s4, v7, 4
-; GFX6-NEXT:    v_readlane_b32 s5, v7, 5
-; GFX6-NEXT:    v_readlane_b32 s6, v7, 6
-; GFX6-NEXT:    v_readlane_b32 s7, v7, 7
-; GFX6-NEXT:    buffer_load_dword v7, off, s[40:43], 0
+; GFX6-NEXT:    v_readlane_b32 s24, v9, 0
+; GFX6-NEXT:    v_readlane_b32 s25, v9, 1
+; GFX6-NEXT:    v_readlane_b32 s26, v9, 2
+; GFX6-NEXT:    v_readlane_b32 s27, v9, 3
+; GFX6-NEXT:    v_readlane_b32 s28, v9, 4
+; GFX6-NEXT:    v_readlane_b32 s29, v9, 5
+; GFX6-NEXT:    v_readlane_b32 s30, v9, 6
+; GFX6-NEXT:    v_readlane_b32 s31, v9, 7
+; GFX6-NEXT:    buffer_load_dword v9, off, s[40:43], 0
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    s_mov_b64 exec, s[44:45]
-; GFX6-NEXT:    s_mov_b64 s[44:45], exec
+; GFX6-NEXT:    s_mov_b64 s[36:37], exec
 ; GFX6-NEXT:    s_mov_b64 exec, 15
 ; GFX6-NEXT:    buffer_store_dword v8, off, s[40:43], 0
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    v_writelane_b32 v8, s36, 0
-; GFX6-NEXT:    v_writelane_b32 v8, s37, 1
-; GFX6-NEXT:    v_writelane_b32 v8, s38, 2
-; GFX6-NEXT:    v_writelane_b32 v8, s39, 3
-; GFX6-NEXT:    v_mov_b32_e32 v4, 0x2180
-; GFX6-NEXT:    buffer_store_dword v8, v4, s[40:43], 0 offen ; 4-byte Folded Spill
+; GFX6-NEXT:    v_writelane_b32 v8, s0, 0
+; GFX6-NEXT:    v_writelane_b32 v8, s1, 1
+; GFX6-NEXT:    v_writelane_b32 v8, s2, 2
+; GFX6-NEXT:    v_writelane_b32 v8, s3, 3
+; GFX6-NEXT:    s_mov_b32 s38, 0x85c00
+; GFX6-NEXT:    buffer_store_dword v8, off, s[40:43], s38 ; 4-byte Folded Spill
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
 ; GFX6-NEXT:    buffer_load_dword v8, off, s[40:43], 0
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
-; GFX6-NEXT:    s_mov_b64 exec, s[44:45]
+; GFX6-NEXT:    s_mov_b64 exec, s[36:37]
 ; GFX6-NEXT:    s_mov_b64 s[38:39], exec
-; GFX6-NEXT:    s_mov_b64 exec, 3
-; GFX6-NEXT:    buffer_store_dword v10, off, s[40:43], 0
+; GFX6-NEXT:    s_mov_b64 exec, 15
+; GFX6-NEXT:    buffer_store_dword v4, off, s[40:43], 0
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    v_writelane_b32 v10, s36, 0
-; GFX6-NEXT:    v_writelane_b32 v10, s37, 1
-; GFX6-NEXT:    s_mov_b32 s44, 0x86400
-; GFX6-NEXT:    buffer_store_dword v10, off, s[40:43], s44 ; 4-byte Folded Spill
+; GFX6-NEXT:    v_writelane_b32 v4, s4, 0
+; GFX6-NEXT:    v_writelane_b32 v4, s5, 1
+; GFX6-NEXT:    v_writelane_b32 v4, s6, 2
+; GFX6-NEXT:    v_writelane_b32 v4, s7, 3
+; GFX6-NEXT:    s_mov_b32 s0, 0x86000
+; GFX6-NEXT:    buffer_store_dword v4, off, s[40:43], s0 ; 4-byte Folded Spill
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    buffer_load_dword v10, off, s[40:43], 0
+; GFX6-NEXT:    buffer_load_dword v4, off, s[40:43], 0
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    s_mov_b64 exec, s[38:39]
 ; GFX6-NEXT:    s_mov_b64 s[44:45], exec
-; GFX6-NEXT:    s_mov_b64 exec, 15
-; GFX6-NEXT:    v_mov_b32_e32 v4, 0x2170
-; GFX6-NEXT:    buffer_store_dword v9, off, s[40:43], 0
+; GFX6-NEXT:    s_mov_b64 exec, 3
+; GFX6-NEXT:    buffer_store_dword v7, off, s[40:43], 0
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    buffer_load_dword v9, v4, s[40:43], 0 offen ; 4-byte Folded Reload
-; GFX6-NEXT:    s_waitcnt vmcnt(0)
-; GFX6-NEXT:    v_readlane_b32 s36, v9, 0
-; GFX6-NEXT:    v_readlane_b32 s37, v9, 1
-; GFX6-NEXT:    v_readlane_b32 s38, v9, 2
-; GFX6-NEXT:    v_readlane_b32 s39, v9, 3
-; GFX6-NEXT:    buffer_load_dword v9, off, s[40:43], 0
+; GFX6-NEXT:    v_writelane_b32 v7, s2, 0
+; GFX6-NEXT:    v_writelane_b32 v7, s3, 1
+; GFX6-NEXT:    s_mov_b32 s0, 0x86400
+; GFX6-NEXT:    buffer_store_dword v7, off, s[40:43], s0 ; 4-byte Folded Spill
+; GFX6-NEXT:    s_waitcnt expcnt(0)
+; GFX6-NEXT:    buffer_load_dword v7, off, s[40:43], 0
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    s_mov_b64 exec, s[44:45]
-; GFX6-NEXT:    s_not_b64 exec, exec
-; GFX6-NEXT:    v_mov_b32_e32 v4, 0x2190
-; GFX6-NEXT:    buffer_store_dword v7, off, s[40:43], 0
+; GFX6-NEXT:    s_mov_b64 s[36:37], exec
+; GFX6-NEXT:    s_mov_b64 exec, 0xff
+; GFX6-NEXT:    s_mov_b32 s38, 0x85400
+; GFX6-NEXT:    buffer_store_dword v9, off, s[40:43], 0
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    buffer_load_dword v7, v4, s[40:43], 0 offen ; 4-byte Folded Reload
-; GFX6-NEXT:    s_not_b64 exec, exec
-; GFX6-NEXT:    v_mov_b32_e32 v4, 0x2190
-; GFX6-NEXT:    buffer_load_dword v7, v4, s[40:43], 0 offen ; 4-byte Folded Reload
-; GFX6-NEXT:    s_not_b64 exec, exec
+; GFX6-NEXT:    buffer_load_dword v9, off, s[40:43], s38 ; 4-byte Folded Reload
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
-; GFX6-NEXT:    v_readlane_b32 s44, v7, 0
-; GFX6-NEXT:    v_readlane_b32 s45, v7, 1
-; GFX6-NEXT:    buffer_load_dword v7, off, s[40:43], 0
+; GFX6-NEXT:    v_readlane_b32 s0, v9, 0
+; GFX6-NEXT:    v_readlane_b32 s1, v9, 1
+; GFX6-NEXT:    v_readlane_b32 s2, v9, 2
+; GFX6-NEXT:    v_readlane_b32 s3, v9, 3
+; GFX6-NEXT:    v_readlane_b32 s4, v9, 4
+; GFX6-NEXT:    v_readlane_b32 s5, v9, 5
+; GFX6-NEXT:    v_readlane_b32 s6, v9, 6
+; GFX6-NEXT:    v_readlane_b32 s7, v9, 7
+; GFX6-NEXT:    buffer_load_dword v9, off, s[40:43], 0
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
-; GFX6-NEXT:    s_not_b64 exec, exec
-; GFX6-NEXT:    s_mov_b64 vcc, s[34:35]
-; GFX6-NEXT:    s_not_b64 exec, exec
-; GFX6-NEXT:    v_mov_b32_e32 v4, 0x2198
+; GFX6-NEXT:    s_mov_b64 exec, s[36:37]
+; GFX6-NEXT:    s_mov_b64 s[44:45], exec
+; GFX6-NEXT:    s_mov_b64 exec, 15
+; GFX6-NEXT:    v_mov_b32_e32 v4, 0x2180
 ; GFX6-NEXT:    buffer_store_dword v8, off, s[40:43], 0
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
 ; GFX6-NEXT:    buffer_load_dword v8, v4, s[40:43], 0 offen ; 4-byte Folded Reload
-; GFX6-NEXT:    s_not_b64 exec, exec
-; GFX6-NEXT:    v_mov_b32_e32 v4, 0x2198
-; GFX6-NEXT:    buffer_load_dword v8, v4, s[40:43], 0 offen ; 4-byte Folded Reload
-; GFX6-NEXT:    s_not_b64 exec, exec
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
-; GFX6-NEXT:    v_readlane_b32 s34, v8, 0
-; GFX6-NEXT:    v_readlane_b32 s35, v8, 1
+; GFX6-NEXT:    v_readlane_b32 s36, v8, 0
+; GFX6-NEXT:    v_readlane_b32 s37, v8, 1
+; GFX6-NEXT:    v_readlane_b32 s38, v8, 2
+; GFX6-NEXT:    v_readlane_b32 s39, v8, 3
 ; GFX6-NEXT:    buffer_load_dword v8, off, s[40:43], 0
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
-; GFX6-NEXT:    s_not_b64 exec, exec
-; GFX6-NEXT:    ;;#ASMSTART
-; GFX6-NEXT:    ; use s[8:15],s[16:23],s[24:31],s[0:7],s[36:39],s[34:35],s[44:45]
-; GFX6-NEXT:    ;;#ASMEND
-; GFX6-NEXT:    s_mov_b64 s[34:35], vcc
-; GFX6-NEXT:    s_mov_b64 s[8:9], exec
-; GFX6-NEXT:    s_mov_b64 exec, 15
-; GFX6-NEXT:    s_mov_b32 s0, 0x86000
+; GFX6-NEXT:    s_mov_b64 exec, s[44:45]
+; GFX6-NEXT:    s_mov_b64 vcc, s[34:35]
+; GFX6-NEXT:    s_mov_b64 s[44:45], exec
+; GFX6-NEXT:    s_mov_b64 exec, 3
+; GFX6-NEXT:    v_mov_b32_e32 v7, 0x2190
 ; GFX6-NEXT:    buffer_store_dword v4, off, s[40:43], 0
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
-; GFX6-NEXT:    buffer_load_dword v4, off, s[40:43], s0 ; 4-byte Folded Reload
+; GFX6-NEXT:    buffer_load_dword v4, v7, s[40:43], 0 offen ; 4-byte Folded Reload
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
-; GFX6-NEXT:    v_readlane_b32 s36, v4, 0
-; GFX6-NEXT:    v_readlane_b32 s37, v4, 1
-; GFX6-NEXT:    v_readlane_b32 s38, v4, 2
-; GFX6-NEXT:    v_readlane_b32 s39, v4, 3
+; GFX6-NEXT:    v_readlane_b32 s34, v4, 0
+; GFX6-NEXT:    v_readlane_b32 s35, v4, 1
 ; GFX6-NEXT:    buffer_load_dword v4, off, s[40:43], 0
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
-; GFX6-NEXT:    s_mov_b64 exec, s[8:9]
+; GFX6-NEXT:    s_mov_b64 exec, s[44:45]
+; GFX6-NEXT:    ;;#ASMSTART
+; GFX6-NEXT:    ; use s[8:15],s[16:23],s[24:31],s[0:7],s[36:39],s[34:35]
+; GFX6-NEXT:    ;;#ASMEND
+; GFX6-NEXT:    s_mov_b64 s[34:35], vcc
 ; GFX6-NEXT:    s_mov_b64 s[4:5], exec
 ; GFX6-NEXT:    s_mov_b64 exec, 15
-; GFX6-NEXT:    s_mov_b32 s6, 0x85800
+; GFX6-NEXT:    s_mov_b32 s6, 0x85c00
 ; GFX6-NEXT:    buffer_store_dword v7, off, s[40:43], 0
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
 ; GFX6-NEXT:    buffer_load_dword v7, off, s[40:43], s6 ; 4-byte Folded Reload
@@ -10530,19 +10485,19 @@
 ; GFX6-NEXT:    buffer_load_dword v7, off, s[40:43], 0
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    s_mov_b64 exec, s[4:5]
-; GFX6-NEXT:    s_mov_b32 s2, 0x83800
+; GFX6-NEXT:    s_mov_b32 s2, 0x83c00
 ; GFX6-NEXT:    buffer_store_dword v0, off, s[40:43], s2 ; 4-byte Folded Spill
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    buffer_store_dword v1, off, s[40:43], s2 offset:4 ; 4-byte Folded Spill
 ; GFX6-NEXT:    buffer_store_dword v2, off, s[40:43], s2 offset:8 ; 4-byte Folded Spill
 ; GFX6-NEXT:    buffer_store_dword v3, off, s[40:43], s2 offset:12 ; 4-byte Folded Spill
-; GFX6-NEXT:    s_mov_b32 s2, 0x84000
+; GFX6-NEXT:    s_mov_b32 s2, 0x84400
 ; GFX6-NEXT:    buffer_store_dword v13, off, s[40:43], s2 ; 4-byte Folded Spill
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    buffer_store_dword v14, off, s[40:43], s2 offset:4 ; 4-byte Folded Spill
 ; GFX6-NEXT:    buffer_store_dword v15, off, s[40:43], s2 offset:8 ; 4-byte Folded Spill
 ; GFX6-NEXT:    buffer_store_dword v16, off, s[40:43], s2 offset:12 ; 4-byte Folded Spill
-; GFX6-NEXT:    s_mov_b32 s2, 0x84800
+; GFX6-NEXT:    s_mov_b32 s2, 0x84c00
 ; GFX6-NEXT:    buffer_store_dword v17, off, s[40:43], s2 ; 4-byte Folded Spill
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    buffer_store_dword v18, off, s[40:43], s2 offset:4 ; 4-byte Folded Spill
@@ -10551,17 +10506,17 @@
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
 ; GFX6-NEXT:    ;;#ASMSTART
 ; GFX6-NEXT:    ;;#ASMEND
-; GFX6-NEXT:    s_mov_b32 s2, 0x84800
+; GFX6-NEXT:    s_mov_b32 s2, 0x84c00
 ; GFX6-NEXT:    buffer_load_dword v17, off, s[40:43], s2 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v18, off, s[40:43], s2 offset:4 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v19, off, s[40:43], s2 offset:8 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v20, off, s[40:43], s2 offset:12 ; 4-byte Folded Reload
-; GFX6-NEXT:    s_mov_b32 s2, 0x84000
+; GFX6-NEXT:    s_mov_b32 s2, 0x84400
 ; GFX6-NEXT:    buffer_load_dword v13, off, s[40:43], s2 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v14, off, s[40:43], s2 offset:4 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v15, off, s[40:43], s2 offset:8 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v16, off, s[40:43], s2 offset:12 ; 4-byte Folded Reload
-; GFX6-NEXT:    s_mov_b32 s2, 0x83800
+; GFX6-NEXT:    s_mov_b32 s2, 0x83c00
 ; GFX6-NEXT:    buffer_load_dword v0, off, s[40:43], s2 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v1, off, s[40:43], s2 offset:4 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v2, off, s[40:43], s2 offset:8 ; 4-byte Folded Reload
@@ -10580,14 +10535,28 @@
 ; GFX6-NEXT:    ;;#ASMEND
 ; GFX6-NEXT:  .LBB1_2: ; %ret
 ; GFX6-NEXT:    s_or_b64 exec, exec, s[34:35]
-; GFX6-NEXT:    s_mov_b32 s4, 0x83400
+; GFX6-NEXT:    s_mov_b64 s[8:9], exec
+; GFX6-NEXT:    s_mov_b64 exec, 15
+; GFX6-NEXT:    s_mov_b32 s2, 0x80400
+; GFX6-NEXT:    buffer_store_dword v10, off, s[40:43], 0
+; GFX6-NEXT:    s_waitcnt expcnt(0)
+; GFX6-NEXT:    buffer_load_dword v10, off, s[40:43], s2 ; 4-byte Folded Reload
+; GFX6-NEXT:    s_waitcnt vmcnt(0)
+; GFX6-NEXT:    v_readlane_b32 s4, v10, 0
+; GFX6-NEXT:    v_readlane_b32 s5, v10, 1
+; GFX6-NEXT:    v_readlane_b32 s6, v10, 2
+; GFX6-NEXT:    v_readlane_b32 s7, v10, 3
+; GFX6-NEXT:    buffer_load_dword v10, off, s[40:43], 0
+; GFX6-NEXT:    s_waitcnt vmcnt(0)
+; GFX6-NEXT:    s_mov_b64 exec, s[8:9]
+; GFX6-NEXT:    s_mov_b32 s4, 0x83800
 ; GFX6-NEXT:    v_lshl_b64 v[4:5], v[5:6], 8
 ; GFX6-NEXT:    buffer_load_dword v6, off, s[40:43], s4 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v7, off, s[40:43], s4 offset:4 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v8, off, s[40:43], s4 offset:8 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v9, off, s[40:43], s4 offset:12 ; 4-byte Folded Reload
-; GFX6-NEXT:    s_mov_b64 s[2:3], s[38:39]
-; GFX6-NEXT:    s_mov_b32 s4, 0x83000
+; GFX6-NEXT:    s_mov_b64 s[2:3], s[6:7]
+; GFX6-NEXT:    s_mov_b32 s4, 0x83400
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    buffer_store_dwordx4 v[6:9], v[4:5], s[0:3], 0 addr64 offset:240
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
@@ -10595,7 +10564,7 @@
 ; GFX6-NEXT:    buffer_load_dword v7, off, s[40:43], s4 offset:4 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v8, off, s[40:43], s4 offset:8 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v9, off, s[40:43], s4 offset:12 ; 4-byte Folded Reload
-; GFX6-NEXT:    s_mov_b32 s4, 0x82c00
+; GFX6-NEXT:    s_mov_b32 s4, 0x83000
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    buffer_store_dwordx4 v[6:9], v[4:5], s[0:3], 0 addr64 offset:224
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
@@ -10603,7 +10572,7 @@
 ; GFX6-NEXT:    buffer_load_dword v7, off, s[40:43], s4 offset:4 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v8, off, s[40:43], s4 offset:8 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v9, off, s[40:43], s4 offset:12 ; 4-byte Folded Reload
-; GFX6-NEXT:    s_mov_b32 s4, 0x82800
+; GFX6-NEXT:    s_mov_b32 s4, 0x82c00
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    buffer_store_dwordx4 v[6:9], v[4:5], s[0:3], 0 addr64 offset:208
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
@@ -10611,7 +10580,7 @@
 ; GFX6-NEXT:    buffer_load_dword v7, off, s[40:43], s4 offset:4 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v8, off, s[40:43], s4 offset:8 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v9, off, s[40:43], s4 offset:12 ; 4-byte Folded Reload
-; GFX6-NEXT:    s_mov_b32 s4, 0x82400
+; GFX6-NEXT:    s_mov_b32 s4, 0x82800
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    buffer_store_dwordx4 v[6:9], v[4:5], s[0:3], 0 addr64 offset:192
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
@@ -10619,7 +10588,7 @@
 ; GFX6-NEXT:    buffer_load_dword v7, off, s[40:43], s4 offset:4 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v8, off, s[40:43], s4 offset:8 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v9, off, s[40:43], s4 offset:12 ; 4-byte Folded Reload
-; GFX6-NEXT:    s_mov_b32 s4, 0x82000
+; GFX6-NEXT:    s_mov_b32 s4, 0x82400
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    buffer_store_dwordx4 v[6:9], v[4:5], s[0:3], 0 addr64 offset:176
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
@@ -10627,7 +10596,7 @@
 ; GFX6-NEXT:    buffer_load_dword v7, off, s[40:43], s4 offset:4 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v8, off, s[40:43], s4 offset:8 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v9, off, s[40:43], s4 offset:12 ; 4-byte Folded Reload
-; GFX6-NEXT:    s_mov_b32 s4, 0x81c00
+; GFX6-NEXT:    s_mov_b32 s4, 0x82000
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    buffer_store_dwordx4 v[6:9], v[4:5], s[0:3], 0 addr64 offset:160
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
@@ -10635,7 +10604,7 @@
 ; GFX6-NEXT:    buffer_load_dword v7, off, s[40:43], s4 offset:4 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v8, off, s[40:43], s4 offset:8 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v9, off, s[40:43], s4 offset:12 ; 4-byte Folded Reload
-; GFX6-NEXT:    s_mov_b32 s4, 0x81800
+; GFX6-NEXT:    s_mov_b32 s4, 0x81c00
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    buffer_store_dwordx4 v[6:9], v[4:5], s[0:3], 0 addr64 offset:144
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
@@ -10643,7 +10612,7 @@
 ; GFX6-NEXT:    buffer_load_dword v7, off, s[40:43], s4 offset:4 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v8, off, s[40:43], s4 offset:8 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v9, off, s[40:43], s4 offset:12 ; 4-byte Folded Reload
-; GFX6-NEXT:    s_mov_b32 s4, 0x81400
+; GFX6-NEXT:    s_mov_b32 s4, 0x81800
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    buffer_store_dwordx4 v[6:9], v[4:5], s[0:3], 0 addr64 offset:128
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
@@ -10651,7 +10620,7 @@
 ; GFX6-NEXT:    buffer_load_dword v7, off, s[40:43], s4 offset:4 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v8, off, s[40:43], s4 offset:8 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v9, off, s[40:43], s4 offset:12 ; 4-byte Folded Reload
-; GFX6-NEXT:    s_mov_b32 s4, 0x81000
+; GFX6-NEXT:    s_mov_b32 s4, 0x81400
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    buffer_store_dwordx4 v[6:9], v[4:5], s[0:3], 0 addr64 offset:112
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
@@ -10659,7 +10628,7 @@
 ; GFX6-NEXT:    buffer_load_dword v7, off, s[40:43], s4 offset:4 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v8, off, s[40:43], s4 offset:8 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v9, off, s[40:43], s4 offset:12 ; 4-byte Folded Reload
-; GFX6-NEXT:    s_mov_b32 s4, 0x80c00
+; GFX6-NEXT:    s_mov_b32 s4, 0x81000
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    buffer_store_dwordx4 v[6:9], v[4:5], s[0:3], 0 addr64 offset:96
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
@@ -10667,7 +10636,7 @@
 ; GFX6-NEXT:    buffer_load_dword v7, off, s[40:43], s4 offset:4 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v8, off, s[40:43], s4 offset:8 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v9, off, s[40:43], s4 offset:12 ; 4-byte Folded Reload
-; GFX6-NEXT:    s_mov_b32 s4, 0x80400
+; GFX6-NEXT:    s_mov_b32 s4, 0x80800
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    buffer_store_dwordx4 v[6:9], v[4:5], s[0:3], 0 addr64 offset:80
 ; GFX6-NEXT:    s_waitcnt expcnt(0)
@@ -10675,7 +10644,7 @@
 ; GFX6-NEXT:    buffer_load_dword v7, off, s[40:43], s4 offset:4 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v8, off, s[40:43], s4 offset:8 ; 4-byte Folded Reload
 ; GFX6-NEXT:    buffer_load_dword v9, off, s[40:43], s4 offset:12 ; 4-byte Folded Reload
-; GFX6-NEXT:    s_mov_b32 s4, 0x80800
+; GFX6-NEXT:    s_mov_b32 s4, 0x80c00
 ; GFX6-NEXT:    s_waitcnt vmcnt(0)
 ; GFX6-NEXT:    buffer_store_dwordx4 v[6:9], v[4:5], s[0:3], 0 addr64 offset:64
 ; GFX6-NEXT:    buffer_store_dwordx4 v[17:20], v[4:5], s[0:3], 0 addr64 offset:48
@@ -10777,16 +10746,13 @@
 ; GFX9-FLATSCR-NEXT:    ; def s[38:39]
 ; GFX9-FLATSCR-NEXT:    ;;#ASMEND
 ; GFX9-FLATSCR-NEXT:    ;;#ASMSTART
-; GFX9-FLATSCR-NEXT:    ; def s[44:45]
-; GFX9-FLATSCR-NEXT:    ;;#ASMEND
-; GFX9-FLATSCR-NEXT:    ;;#ASMSTART
 ; GFX9-FLATSCR-NEXT:    ; def s33
 ; GFX9-FLATSCR-NEXT:    ;;#ASMEND
 ; GFX9-FLATSCR-NEXT:    s_and_saveexec_b64 s[34:35], vcc
 ; GFX9-FLATSCR-NEXT:    s_cbranch_execz .LBB1_2
 ; GFX9-FLATSCR-NEXT:  ; %bb.1: ; %bb0
 ; GFX9-FLATSCR-NEXT:    ;;#ASMSTART
-; GFX9-FLATSCR-NEXT:    ; use s[0:7],s[8:15],s[16:23],s[24:31],s[40:43],s[38:39],s[44:45]
+; GFX9-FLATSCR-NEXT:    ; use s[0:7],s[8:15],s[16:23],s[24:31],s[40:43],s[38:39]
 ; GFX9-FLATSCR-NEXT:    ;;#ASMEND
 ; GFX9-FLATSCR-NEXT:    s_movk_i32 s0, 0x20d0
 ; GFX9-FLATSCR-NEXT:    scratch_store_dwordx4 off, v[0:3], s0 ; 16-byte Folded Spill
@@ -10931,16 +10897,13 @@
 ; GFX10-FLATSCR-NEXT:    ; def s[34:35]
 ; GFX10-FLATSCR-NEXT:    ;;#ASMEND
 ; GFX10-FLATSCR-NEXT:    ;;#ASMSTART
-; GFX10-FLATSCR-NEXT:    ; def s[38:39]
-; GFX10-FLATSCR-NEXT:    ;;#ASMEND
-; GFX10-FLATSCR-NEXT:    ;;#ASMSTART
-; GFX10-FLATSCR-NEXT:    ; def s44
+; GFX10-FLATSCR-NEXT:    ; def s38
 ; GFX10-FLATSCR-NEXT:    ;;#ASMEND
 ; GFX10-FLATSCR-NEXT:    v_cmpx_eq_u32_e32 0, v0
 ; GFX10-FLATSCR-NEXT:    s_cbranch_execz .LBB1_2
 ; GFX10-FLATSCR-NEXT:  ; %bb.1: ; %bb0
 ; GFX10-FLATSCR-NEXT:    ;;#ASMSTART
-; GFX10-FLATSCR-NEXT:    ; use s[0:7],s[8:15],s[16:23],s[24:31],s[40:43],s[34:35],s[38:39]
+; GFX10-FLATSCR-NEXT:    ; use s[0:7],s[8:15],s[16:23],s[24:31],s[40:43],s[34:35]
 ; GFX10-FLATSCR-NEXT:    ;;#ASMEND
 ; GFX10-FLATSCR-NEXT:    s_movk_i32 s0, 0x2010
 ; GFX10-FLATSCR-NEXT:    v_mov_b32_e32 v88, v59
@@ -11118,15 +11081,14 @@
   %sgpr3 = call <8 x i32> asm sideeffect "; def $0", "=s" ()
   %sgpr4 = call <4 x i32> asm sideeffect "; def $0", "=s" ()
   %sgpr5 = call <2 x i32> asm sideeffect "; def $0", "=s" ()
-  %sgpr6 = call <2 x i32> asm sideeffect "; def $0", "=s" ()
-  %sgpr7 = call i32 asm sideeffect "; def $0", "=s" ()
+  %sgpr6 = call i32 asm sideeffect "; def $0", "=s" ()
 
   %cmp = icmp eq i32 %x, 0
   br i1 %cmp, label %bb0, label %ret
 
 bb0:
 ; create SGPR pressure
-  call void asm sideeffect "; use $0,$1,$2,$3,$4,$5,$6", "s,s,s,s,s,s,s,s"(<8 x i32> %sgpr0, <8 x i32> %sgpr1, <8 x i32> %sgpr2, <8 x i32> %sgpr3, <4 x i32> %sgpr4, <2 x i32> %sgpr5, <2 x i32> %sgpr6, i32 %sgpr7)
+  call void asm sideeffect "; use $0,$1,$2,$3,$4,$5", "s,s,s,s,s,s,s"(<8 x i32> %sgpr0, <8 x i32> %sgpr1, <8 x i32> %sgpr2, <8 x i32> %sgpr3, <4 x i32> %sgpr4, <2 x i32> %sgpr5, i32 %sgpr6)
 
 ; mark most VGPR registers as used to increase register pressure
   call void asm sideeffect "", "~{v4},~{v8},~{v12},~{v16},~{v20},~{v24},~{v28},~{v32}" ()
diff --git a/llvm/test/CodeGen/AMDGPU/spill-sgpr-csr-live-ins.mir b/llvm/test/CodeGen/AMDGPU/spill-sgpr-csr-live-ins.mir
--- a/llvm/test/CodeGen/AMDGPU/spill-sgpr-csr-live-ins.mir
+++ b/llvm/test/CodeGen/AMDGPU/spill-sgpr-csr-live-ins.mir
@@ -10,9 +10,10 @@
   bb.0:
     liveins: $sgpr50
     ; CHECK-LABEL: name: spill_csr_sgpr_argument
-    ; CHECK: liveins: $sgpr50, $vgpr0
+    ; CHECK: liveins: $sgpr50
     ; CHECK-NEXT: {{  $}}
-    ; CHECK-NEXT: $vgpr0 = V_WRITELANE_B32 $sgpr50, 0, $vgpr0
+    ; CHECK-NEXT: [[DEF:%[0-9]+]]:vgpr_32 = IMPLICIT_DEF
+    ; CHECK-NEXT: [[V_WRITELANE_B32_:%[0-9]+]]:vgpr_32 = V_WRITELANE_B32 $sgpr50, 0, [[V_WRITELANE_B32_]]
     ; CHECK-NEXT: S_NOP 0, implicit $sgpr50
     ; CHECK-NEXT: $sgpr50 = S_MOV_B32 0
     S_NOP 0, implicit $sgpr50
diff --git a/llvm/test/CodeGen/AMDGPU/spill-sgpr-stack-no-sgpr.ll b/llvm/test/CodeGen/AMDGPU/spill-sgpr-stack-no-sgpr.ll
--- a/llvm/test/CodeGen/AMDGPU/spill-sgpr-stack-no-sgpr.ll
+++ b/llvm/test/CodeGen/AMDGPU/spill-sgpr-stack-no-sgpr.ll
@@ -1,59 +1,50 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc -march=amdgcn -mcpu=gfx1010 -mattr=-wavefrontsize32,+wavefrontsize64 -verify-machineinstrs < %s | FileCheck -check-prefix=GFX10 %s
 
-; Spill an SGPR to scratch without having spare SGPRs available to save exec
+; The test was originally written to spill an SGPR to scratch without having spare SGPRs available to save exec.
+; This scenario no longer exists when we enabled SGPR spill into virtual VGPRs.
 
 define amdgpu_kernel void @test() #1 {
 ; GFX10-LABEL: test:
 ; GFX10:       ; %bb.0:
-; GFX10-NEXT:    s_mov_b32 s8, SCRATCH_RSRC_DWORD0
-; GFX10-NEXT:    s_mov_b32 s9, SCRATCH_RSRC_DWORD1
-; GFX10-NEXT:    s_mov_b32 s10, -1
-; GFX10-NEXT:    s_mov_b32 s11, 0x31e16000
-; GFX10-NEXT:    s_add_u32 s8, s8, s1
-; GFX10-NEXT:    s_addc_u32 s9, s9, 0
+; GFX10-NEXT:    s_mov_b32 s12, SCRATCH_RSRC_DWORD0
+; GFX10-NEXT:    s_mov_b32 s13, SCRATCH_RSRC_DWORD1
+; GFX10-NEXT:    s_mov_b32 s14, -1
+; GFX10-NEXT:    s_mov_b32 s15, 0x31e16000
+; GFX10-NEXT:    s_add_u32 s12, s12, s1
+; GFX10-NEXT:    s_addc_u32 s13, s13, 0
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; def s[0:7]
 ; GFX10-NEXT:    ;;#ASMEND
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; def s[8:12]
 ; GFX10-NEXT:    ;;#ASMEND
-; GFX10-NEXT:    s_not_b64 exec, exec
-; GFX10-NEXT:    buffer_store_dword v0, off, s[8:11], 0
+; GFX10-NOT:    s_not_b64 exec, exec
+; GFX10-NEXT:    ; implicit-def: $vgpr0
 ; GFX10-NEXT:    v_writelane_b32 v0, s8, 0
 ; GFX10-NEXT:    v_writelane_b32 v0, s9, 1
 ; GFX10-NEXT:    v_writelane_b32 v0, s10, 2
 ; GFX10-NEXT:    v_writelane_b32 v0, s11, 3
 ; GFX10-NEXT:    v_writelane_b32 v0, s12, 4
-; GFX10-NEXT:    buffer_store_dword v0, off, s[8:11], 0 offset:4 ; 4-byte Folded Spill
+; GFX10-NEXT:    s_or_saveexec_b64 s[14:15], -1
+; GFX10-NEXT:    buffer_store_dword v0, off, s[12:15], 0 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
-; GFX10-NEXT:    s_not_b64 exec, exec
-; GFX10-NEXT:    buffer_store_dword v0, off, s[8:11], 0 offset:4 ; 4-byte Folded Spill
-; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
-; GFX10-NEXT:    s_not_b64 exec, exec
-; GFX10-NEXT:    buffer_load_dword v0, off, s[8:11], 0
-; GFX10-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
-; GFX10-NEXT:    s_not_b64 exec, exec
+; GFX10-NEXT:    s_mov_b64 exec, s[14:15]
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ;;#ASMEND
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; use s[0:7]
 ; GFX10-NEXT:    ;;#ASMEND
-; GFX10-NEXT:    s_mov_b64 s[6:7], exec
-; GFX10-NEXT:    s_mov_b64 exec, 31
-; GFX10-NEXT:    buffer_store_dword v0, off, s[8:11], 0
-; GFX10-NEXT:    buffer_load_dword v0, off, s[8:11], 0 offset:4 ; 4-byte Folded Reload
+; GFX10-NEXT:    s_or_saveexec_b64 s[14:15], -1
+; GFX10-NEXT:    buffer_load_dword v0, off, s[12:15], 0 offset:4 ; 4-byte Folded Reload
+; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
+; GFX10-NEXT:    s_mov_b64 exec, s[14:15]
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    v_readlane_b32 s0, v0, 0
 ; GFX10-NEXT:    v_readlane_b32 s1, v0, 1
 ; GFX10-NEXT:    v_readlane_b32 s2, v0, 2
 ; GFX10-NEXT:    v_readlane_b32 s3, v0, 3
 ; GFX10-NEXT:    v_readlane_b32 s4, v0, 4
-; GFX10-NEXT:    buffer_load_dword v0, off, s[8:11], 0
-; GFX10-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
-; GFX10-NEXT:    s_mov_b64 exec, s[6:7]
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; use s[0:4]
 ; GFX10-NEXT:    ;;#ASMEND
@@ -67,4 +58,4 @@
 }
 
 attributes #0 = { nounwind }
-attributes #1 = { nounwind "amdgpu-num-sgpr"="16" "amdgpu-num-vgpr"="8" }
+attributes #1 = { nounwind "amdgpu-num-sgpr"="18" "amdgpu-num-vgpr"="8" }
diff --git a/llvm/test/CodeGen/AMDGPU/spill-sgpr-to-virtual-vgpr.mir b/llvm/test/CodeGen/AMDGPU/spill-sgpr-to-virtual-vgpr.mir
new file mode 100644
--- /dev/null
+++ b/llvm/test/CodeGen/AMDGPU/spill-sgpr-to-virtual-vgpr.mir
@@ -0,0 +1,320 @@
+# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
+# RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx906 -run-pass=si-lower-sgpr-spills -verify-machineinstrs %s -o - | FileCheck -check-prefix=GCN %s
+
+# A simple SGPR spill. Implicit def for lane VGPR should be inserted just before the spill instruction.
+---
+name:            sgpr32_spill
+tracksRegLiveness: true
+frameInfo:
+  maxAlignment:    4
+stack:
+  - { id: 0, type: spill-slot, size: 4, alignment: 4, stack-id: sgpr-spill }
+machineFunctionInfo:
+  isEntryFunction: false
+  scratchRSrcReg:  '$sgpr0_sgpr1_sgpr2_sgpr3'
+  stackPtrOffsetReg: '$sgpr32'
+  frameOffsetReg: '$sgpr33'
+  hasSpilledSGPRs: true
+body:             |
+  bb.0:
+    liveins: $sgpr30_sgpr31, $sgpr10
+    ; GCN-LABEL: name: sgpr32_spill
+    ; GCN: liveins: $sgpr30_sgpr31, $sgpr10
+    ; GCN-NEXT: {{  $}}
+    ; GCN-NEXT: S_NOP 0
+    ; GCN-NEXT: [[V_WRITELANE_B32_:%[0-9]+]]:vgpr_32 = IMPLICIT_DEF
+    ; GCN-NEXT: [[V_WRITELANE_B32_]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr10, 0, [[V_WRITELANE_B32_]]
+    ; GCN-NEXT: $sgpr10 = V_READLANE_B32 [[V_WRITELANE_B32_]], 0
+    ; GCN-NEXT: S_SETPC_B64 $sgpr30_sgpr31
+    S_NOP 0
+    SI_SPILL_S32_SAVE killed $sgpr10, %stack.0, implicit $exec, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr32
+    renamable $sgpr10 = SI_SPILL_S32_RESTORE %stack.0, implicit $exec, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr32
+    S_SETPC_B64 $sgpr30_sgpr31
+...
+
+# Needed an additional virtual lane register as the lanes of current register are fully occupied while spilling a wide SGPR tuple.
+# There must be two implicit def for the two lane VGPRs.
+
+---
+name:            sgpr_spill_lane_crossover
+tracksRegLiveness: true
+frameInfo:
+  maxAlignment:    4
+stack:
+  - { id: 0, type: spill-slot, size: 4, alignment: 4, stack-id: sgpr-spill }
+  - { id: 1, type: spill-slot, size: 128, alignment: 4, stack-id: sgpr-spill }
+machineFunctionInfo:
+  isEntryFunction: false
+  scratchRSrcReg:  '$sgpr0_sgpr1_sgpr2_sgpr3'
+  stackPtrOffsetReg: '$sgpr32'
+  frameOffsetReg: '$sgpr33'
+  hasSpilledSGPRs: true
+body:             |
+  bb.0:
+    liveins: $sgpr30_sgpr31, $sgpr10, $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71, $sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79, $sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87, $sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-LABEL: name: sgpr_spill_lane_crossover
+    ; GCN: liveins: $sgpr10, $sgpr64, $sgpr65, $sgpr66, $sgpr67, $sgpr68, $sgpr69, $sgpr70, $sgpr71, $sgpr72, $sgpr73, $sgpr74, $sgpr75, $sgpr76, $sgpr77, $sgpr78, $sgpr79, $sgpr80, $sgpr81, $sgpr82, $sgpr83, $sgpr84, $sgpr85, $sgpr86, $sgpr87, $sgpr88, $sgpr89, $sgpr90, $sgpr91, $sgpr92, $sgpr93, $sgpr94, $sgpr95, $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71, $sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79, $sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87, $sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95, $sgpr30_sgpr31
+    ; GCN-NEXT: {{  $}}
+    ; GCN-NEXT: [[V_WRITELANE_B32_1:%[0-9]+]]:vgpr_32 = IMPLICIT_DEF
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr64, 0, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr65, 1, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr66, 2, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr67, 3, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr68, 4, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr69, 5, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr70, 6, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr71, 7, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr72, 8, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr73, 9, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr74, 10, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr75, 11, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr76, 12, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr77, 13, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr78, 14, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr79, 15, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr80, 16, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr81, 17, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr82, 18, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr83, 19, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr84, 20, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr85, 21, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr86, 22, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr87, 23, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr88, 24, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr89, 25, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr90, 26, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr91, 27, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr92, 28, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr93, 29, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr94, 30, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr95, 31, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: S_NOP 0
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr10, 32, [[V_WRITELANE_B32_1]]
+    ; GCN-NEXT: [[V_WRITELANE_B32_2:%[0-9]+]]:vgpr_32 = IMPLICIT_DEF
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr64, 33, [[V_WRITELANE_B32_1]], implicit-def $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95, implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr65, 34, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr66, 35, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr67, 36, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr68, 37, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr69, 38, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr70, 39, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr71, 40, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr72, 41, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr73, 42, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr74, 43, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr75, 44, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr76, 45, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr77, 46, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr78, 47, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr79, 48, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr80, 49, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr81, 50, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr82, 51, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr83, 52, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr84, 53, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr85, 54, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr86, 55, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr87, 56, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr88, 57, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr89, 58, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr90, 59, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr91, 60, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr92, 61, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr93, 62, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_1]]:vgpr_32 = V_WRITELANE_B32 $sgpr94, 63, [[V_WRITELANE_B32_1]], implicit $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: [[V_WRITELANE_B32_2]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr95, 0, [[V_WRITELANE_B32_2]], implicit killed $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: S_NOP 0
+    ; GCN-NEXT: $sgpr64 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 33, implicit-def $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
+    ; GCN-NEXT: $sgpr65 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 34
+    ; GCN-NEXT: $sgpr66 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 35
+    ; GCN-NEXT: $sgpr67 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 36
+    ; GCN-NEXT: $sgpr68 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 37
+    ; GCN-NEXT: $sgpr69 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 38
+    ; GCN-NEXT: $sgpr70 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 39
+    ; GCN-NEXT: $sgpr71 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 40
+    ; GCN-NEXT: $sgpr72 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 41
+    ; GCN-NEXT: $sgpr73 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 42
+    ; GCN-NEXT: $sgpr74 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 43
+    ; GCN-NEXT: $sgpr75 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 44
+    ; GCN-NEXT: $sgpr76 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 45
+    ; GCN-NEXT: $sgpr77 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 46
+    ; GCN-NEXT: $sgpr78 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 47
+    ; GCN-NEXT: $sgpr79 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 48
+    ; GCN-NEXT: $sgpr80 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 49
+    ; GCN-NEXT: $sgpr81 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 50
+    ; GCN-NEXT: $sgpr82 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 51
+    ; GCN-NEXT: $sgpr83 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 52
+    ; GCN-NEXT: $sgpr84 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 53
+    ; GCN-NEXT: $sgpr85 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 54
+    ; GCN-NEXT: $sgpr86 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 55
+    ; GCN-NEXT: $sgpr87 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 56
+    ; GCN-NEXT: $sgpr88 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 57
+    ; GCN-NEXT: $sgpr89 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 58
+    ; GCN-NEXT: $sgpr90 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 59
+    ; GCN-NEXT: $sgpr91 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 60
+    ; GCN-NEXT: $sgpr92 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 61
+    ; GCN-NEXT: $sgpr93 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 62
+    ; GCN-NEXT: $sgpr94 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 63
+    ; GCN-NEXT: $sgpr95 = V_READLANE_B32 [[V_WRITELANE_B32_2]], 0
+    ; GCN-NEXT: $sgpr10 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 32
+    ; GCN-NEXT: S_SETPC_B64 $sgpr30_sgpr31
+    S_NOP 0
+    SI_SPILL_S32_SAVE killed $sgpr10, %stack.0, implicit $exec, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr32
+    SI_SPILL_S1024_SAVE killed $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95, %stack.1, implicit $exec, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr32
+    S_NOP 0
+    renamable $sgpr64_sgpr65_sgpr66_sgpr67_sgpr68_sgpr69_sgpr70_sgpr71_sgpr72_sgpr73_sgpr74_sgpr75_sgpr76_sgpr77_sgpr78_sgpr79_sgpr80_sgpr81_sgpr82_sgpr83_sgpr84_sgpr85_sgpr86_sgpr87_sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95 = SI_SPILL_S1024_RESTORE %stack.1, implicit $exec, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr32
+    renamable $sgpr10 = SI_SPILL_S32_RESTORE %stack.0, implicit $exec, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr32
+    S_SETPC_B64 $sgpr30_sgpr31
+...
+
+# The implicit def for the lane VGPR should be inserted at the common dominator block (the entry block here).
+
+---
+name:            lane_vgpr_implicit_def_at_common_dominator_block
+tracksRegLiveness: true
+frameInfo:
+  maxAlignment:    4
+stack:
+  - { id: 0, type: spill-slot, size: 4, alignment: 4, stack-id: sgpr-spill }
+machineFunctionInfo:
+  isEntryFunction: false
+  scratchRSrcReg:  '$sgpr0_sgpr1_sgpr2_sgpr3'
+  stackPtrOffsetReg: '$sgpr32'
+  frameOffsetReg: '$sgpr33'
+  hasSpilledSGPRs: true
+body:             |
+  ; GCN-LABEL: name: lane_vgpr_implicit_def_at_common_dominator_block
+  ; GCN: bb.0:
+  ; GCN-NEXT:   successors: %bb.2(0x40000000), %bb.1(0x40000000)
+  ; GCN-NEXT:   liveins: $sgpr10, $sgpr11, $sgpr30_sgpr31
+  ; GCN-NEXT: {{  $}}
+  ; GCN-NEXT:   S_NOP 0
+  ; GCN-NEXT:   S_CMP_EQ_U32 $sgpr11, 0, implicit-def $scc
+  ; GCN-NEXT:   [[DEF:%[0-9]+]]:vgpr_32 = IMPLICIT_DEF
+  ; GCN-NEXT:   S_CBRANCH_SCC1 %bb.2, implicit killed $scc
+  ; GCN-NEXT: {{  $}}
+  ; GCN-NEXT: bb.1:
+  ; GCN-NEXT:   successors: %bb.3(0x80000000)
+  ; GCN-NEXT:   liveins: $sgpr10, $sgpr30_sgpr31
+  ; GCN-NEXT: {{  $}}
+  ; GCN-NEXT:   $sgpr10 = S_MOV_B32 10
+  ; GCN-NEXT:   [[V_WRITELANE_B32_:%[0-9]+]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr10, 0, [[V_WRITELANE_B32_]]
+  ; GCN-NEXT:   S_BRANCH %bb.3
+  ; GCN-NEXT: {{  $}}
+  ; GCN-NEXT: bb.2:
+  ; GCN-NEXT:   successors: %bb.3(0x80000000)
+  ; GCN-NEXT:   liveins: $sgpr10, $sgpr30_sgpr31
+  ; GCN-NEXT: {{  $}}
+  ; GCN-NEXT:   $sgpr10 = S_MOV_B32 20
+  ; GCN-NEXT:   [[V_WRITELANE_B32_1:%[0-9]+]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr10, 0, [[V_WRITELANE_B32_1]]
+  ; GCN-NEXT:   S_BRANCH %bb.3
+  ; GCN-NEXT: {{  $}}
+  ; GCN-NEXT: bb.3:
+  ; GCN-NEXT:   liveins: $sgpr10, $sgpr30_sgpr31
+  ; GCN-NEXT: {{  $}}
+  ; GCN-NEXT:   $sgpr10 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 0
+  ; GCN-NEXT:   S_SETPC_B64 $sgpr30_sgpr31, implicit $sgpr10
+  bb.0:
+    liveins: $sgpr10, $sgpr11, $sgpr30_sgpr31
+    S_NOP 0
+    S_CMP_EQ_U32 $sgpr11, 0, implicit-def $scc
+    S_CBRANCH_SCC1 %bb.2, implicit killed $scc
+  bb.1:
+    liveins: $sgpr10, $sgpr30_sgpr31
+    $sgpr10 = S_MOV_B32 10
+    SI_SPILL_S32_SAVE killed $sgpr10, %stack.0, implicit $exec, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr32
+    S_BRANCH %bb.3
+  bb.2:
+    liveins: $sgpr10, $sgpr30_sgpr31
+    $sgpr10 = S_MOV_B32 20
+    SI_SPILL_S32_SAVE killed $sgpr10, %stack.0, implicit $exec, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr32
+    S_BRANCH %bb.3
+  bb.3:
+    liveins: $sgpr10, $sgpr30_sgpr31
+    renamable $sgpr10 = SI_SPILL_S32_RESTORE %stack.0, implicit $exec, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr32
+    S_SETPC_B64 $sgpr30_sgpr31, implicit $sgpr10
+...
+
+# The common dominator block is visited only at the end. The insertion point was initially identified to the
+# terminator instruction in the dominator block which later becomes the point where a spill get inserted in the same block.
+
+---
+name:            dominator_block_follows_the_successors_bbs
+tracksRegLiveness: true
+frameInfo:
+  maxAlignment:    4
+stack:
+  - { id: 0, type: spill-slot, size: 4, alignment: 4, stack-id: sgpr-spill }
+machineFunctionInfo:
+  isEntryFunction: false
+  scratchRSrcReg:  '$sgpr0_sgpr1_sgpr2_sgpr3'
+  stackPtrOffsetReg: '$sgpr32'
+  frameOffsetReg: '$sgpr33'
+  hasSpilledSGPRs: true
+body:             |
+  ; GCN-LABEL: name: dominator_block_follows_the_successors_bbs
+  ; GCN: bb.0:
+  ; GCN-NEXT:   successors: %bb.3(0x80000000)
+  ; GCN-NEXT:   liveins: $sgpr10, $sgpr11, $sgpr30_sgpr31
+  ; GCN-NEXT: {{  $}}
+  ; GCN-NEXT:   S_NOP 0
+  ; GCN-NEXT:   S_BRANCH %bb.3
+  ; GCN-NEXT: {{  $}}
+  ; GCN-NEXT: bb.1:
+  ; GCN-NEXT:   successors: %bb.2(0x80000000)
+  ; GCN-NEXT:   liveins: $sgpr10, $sgpr30_sgpr31
+  ; GCN-NEXT: {{  $}}
+  ; GCN-NEXT:   $sgpr10 = V_READLANE_B32 %0, 0
+  ; GCN-NEXT:   $sgpr10 = S_ADD_I32 $sgpr10, 15, implicit-def dead $scc
+  ; GCN-NEXT:   S_BRANCH %bb.2
+  ; GCN-NEXT: {{  $}}
+  ; GCN-NEXT: bb.2:
+  ; GCN-NEXT:   successors: %bb.3(0x80000000)
+  ; GCN-NEXT:   liveins: $sgpr10, $sgpr30_sgpr31
+  ; GCN-NEXT: {{  $}}
+  ; GCN-NEXT:   $sgpr10 = V_READLANE_B32 %0, 0
+  ; GCN-NEXT:   $sgpr10 = S_ADD_I32 $sgpr10, 20, implicit-def dead $scc
+  ; GCN-NEXT:   S_BRANCH %bb.3
+  ; GCN-NEXT: {{  $}}
+  ; GCN-NEXT: bb.3:
+  ; GCN-NEXT:   successors: %bb.2(0x40000000), %bb.1(0x40000000)
+  ; GCN-NEXT:   liveins: $sgpr10, $sgpr11, $sgpr30_sgpr31
+  ; GCN-NEXT: {{  $}}
+  ; GCN-NEXT:   $sgpr10 = S_MOV_B32 10
+  ; GCN-NEXT:   [[DEF:%[0-9]+]]:vgpr_32 = IMPLICIT_DEF
+  ; GCN-NEXT:   [[V_WRITELANE_B32_:%[0-9]+]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr10, 0, [[V_WRITELANE_B32_]]
+  ; GCN-NEXT:   S_CMP_EQ_U32 $sgpr11, 0, implicit-def $scc
+  ; GCN-NEXT:   S_CBRANCH_SCC1 %bb.2, implicit killed $scc
+  ; GCN-NEXT:   S_BRANCH %bb.1
+  ; GCN-NEXT: {{  $}}
+  ; GCN-NEXT: bb.4:
+  ; GCN-NEXT:   liveins: $sgpr10, $sgpr30_sgpr31
+  ; GCN-NEXT: {{  $}}
+  ; GCN-NEXT:   S_NOP 0
+  ; GCN-NEXT:   S_SETPC_B64 $sgpr30_sgpr31, implicit $sgpr10
+  bb.0:
+    liveins: $sgpr10, $sgpr11, $sgpr30_sgpr31
+    S_NOP 0
+    S_BRANCH %bb.3
+  bb.1:
+    liveins: $sgpr10, $sgpr30_sgpr31
+    renamable $sgpr10 = SI_SPILL_S32_RESTORE %stack.0, implicit $exec, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr32
+    $sgpr10 = S_ADD_I32 $sgpr10, 15, implicit-def dead $scc
+    S_BRANCH %bb.2
+  bb.2:
+    liveins: $sgpr10, $sgpr30_sgpr31
+    renamable $sgpr10 = SI_SPILL_S32_RESTORE %stack.0, implicit $exec, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr32
+    $sgpr10 = S_ADD_I32 $sgpr10, 20, implicit-def dead $scc
+    S_BRANCH %bb.3
+  bb.3:
+    liveins: $sgpr10, $sgpr11, $sgpr30_sgpr31
+    $sgpr10 = S_MOV_B32 10
+    SI_SPILL_S32_SAVE killed $sgpr10, %stack.0, implicit $exec, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr32
+    S_CMP_EQ_U32 $sgpr11, 0, implicit-def $scc
+    S_CBRANCH_SCC1 %bb.2, implicit killed $scc
+    S_BRANCH %bb.1
+  bb.4:
+    liveins: $sgpr10, $sgpr30_sgpr31
+    S_NOP 0
+    S_SETPC_B64 $sgpr30_sgpr31, implicit $sgpr10
+...
diff --git a/llvm/test/CodeGen/AMDGPU/spill-vgpr-to-agpr-update-regscavenger.ll b/llvm/test/CodeGen/AMDGPU/spill-vgpr-to-agpr-update-regscavenger.ll
new file mode 100644
--- /dev/null
+++ b/llvm/test/CodeGen/AMDGPU/spill-vgpr-to-agpr-update-regscavenger.ll
@@ -0,0 +1,85 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx908 -O0 -verify-machineinstrs -o - %s | FileCheck %s
+
+; Regression test for `processFunctionBeforeFrameFinalized`:
+; Check that it correctly updates RegisterScavenger so we
+; don't end up with bad machine code due to using undefined
+; physical registers.
+
+define void @test() {
+; CHECK-LABEL: test:
+; CHECK:       ; %bb.0: ; %bb.0
+; CHECK-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; CHECK-NEXT:    s_xor_saveexec_b64 s[4:5], -1
+; CHECK-NEXT:    buffer_store_dword v0, off, s[0:3], s32 ; 4-byte Folded Spill
+; CHECK-NEXT:    s_mov_b64 exec, s[4:5]
+; CHECK-NEXT:  .LBB0_1: ; %bb.1
+; CHECK-NEXT:    ; =>This Inner Loop Header: Depth=1
+; CHECK-NEXT:    s_cbranch_scc1 .LBB0_3
+; CHECK-NEXT:  ; %bb.2: ; %bb.2
+; CHECK-NEXT:    ; in Loop: Header=BB0_1 Depth=1
+; CHECK-NEXT:  .LBB0_3: ; %bb.3
+; CHECK-NEXT:    ; in Loop: Header=BB0_1 Depth=1
+; CHECK-NEXT:    ; implicit-def: $sgpr4
+; CHECK-NEXT:    v_mov_b32_e32 v0, s4
+; CHECK-NEXT:    v_readfirstlane_b32 s6, v0
+; CHECK-NEXT:    s_mov_b64 s[4:5], -1
+; CHECK-NEXT:    s_mov_b32 s7, 0
+; CHECK-NEXT:    s_cmp_eq_u32 s6, s7
+; CHECK-NEXT:    ; implicit-def: $vgpr0
+; CHECK-NEXT:    v_writelane_b32 v0, s4, 0
+; CHECK-NEXT:    v_writelane_b32 v0, s5, 1
+; CHECK-NEXT:    s_mov_b64 s[10:11], exec
+; CHECK-NEXT:    s_mov_b64 exec, -1
+; CHECK-NEXT:    v_accvgpr_write_b32 a0, v0 ; Reload Reuse
+; CHECK-NEXT:    s_mov_b64 exec, s[10:11]
+; CHECK-NEXT:    s_cbranch_scc1 .LBB0_5
+; CHECK-NEXT:  ; %bb.4: ; %bb.4
+; CHECK-NEXT:    ; in Loop: Header=BB0_1 Depth=1
+; CHECK-NEXT:    s_or_saveexec_b64 s[10:11], -1
+; CHECK-NEXT:    v_accvgpr_read_b32 v0, a0 ; Reload Reuse
+; CHECK-NEXT:    s_mov_b64 exec, s[10:11]
+; CHECK-NEXT:    s_mov_b64 s[4:5], 0
+; CHECK-NEXT:    v_writelane_b32 v0, s4, 0
+; CHECK-NEXT:    v_writelane_b32 v0, s5, 1
+; CHECK-NEXT:    s_or_saveexec_b64 s[10:11], -1
+; CHECK-NEXT:    s_nop 0
+; CHECK-NEXT:    v_accvgpr_write_b32 a0, v0 ; Reload Reuse
+; CHECK-NEXT:    s_mov_b64 exec, s[10:11]
+; CHECK-NEXT:  .LBB0_5: ; %Flow
+; CHECK-NEXT:    ; in Loop: Header=BB0_1 Depth=1
+; CHECK-NEXT:    s_or_saveexec_b64 s[10:11], -1
+; CHECK-NEXT:    s_nop 0
+; CHECK-NEXT:    v_accvgpr_read_b32 v0, a0 ; Reload Reuse
+; CHECK-NEXT:    s_mov_b64 exec, s[10:11]
+; CHECK-NEXT:    v_readlane_b32 s4, v0, 0
+; CHECK-NEXT:    v_readlane_b32 s5, v0, 1
+; CHECK-NEXT:    v_cndmask_b32_e64 v0, 0, 1, s[4:5]
+; CHECK-NEXT:    s_mov_b32 s4, 1
+; CHECK-NEXT:    ; implicit-def: $sgpr5
+; CHECK-NEXT:    v_cmp_ne_u32_e64 s[4:5], v0, s4
+; CHECK-NEXT:    s_and_b64 vcc, exec, s[4:5]
+; CHECK-NEXT:    s_cbranch_vccnz .LBB0_1
+; CHECK-NEXT:  ; %bb.6: ; %bb.5
+; CHECK-NEXT:    s_xor_saveexec_b64 s[4:5], -1
+; CHECK-NEXT:    buffer_load_dword v0, off, s[0:3], s32 ; 4-byte Folded Reload
+; CHECK-NEXT:    s_mov_b64 exec, s[4:5]
+; CHECK-NEXT:    s_waitcnt vmcnt(0)
+; CHECK-NEXT:    s_setpc_b64 s[30:31]
+bb.0:
+  br label %bb.1
+bb.1:                                                ; preds = %bb.4, %bb.0
+  br i1 poison, label %bb.2, label %bb.3
+bb.2:                                                ; preds = %bb.1
+  br label %bb.3
+bb.3:                                                ; preds = %bb.2, %bb.1
+  %call = tail call i32 @llvm.amdgcn.readfirstlane(i32 poison)
+  %cmp = icmp eq i32 %call, 0
+  br i1 %cmp, label %bb.5, label %bb.4
+bb.4:                                                ; preds = %bb.3
+  br label %bb.1
+bb.5:                                                ; preds = %bb.3
+  ret void
+}
+
+declare i32 @llvm.amdgcn.readfirstlane(i32)
diff --git a/llvm/test/CodeGen/AMDGPU/spill-writelane-vgprs.ll b/llvm/test/CodeGen/AMDGPU/spill-writelane-vgprs.ll
--- a/llvm/test/CodeGen/AMDGPU/spill-writelane-vgprs.ll
+++ b/llvm/test/CodeGen/AMDGPU/spill-writelane-vgprs.ll
@@ -10,16 +10,17 @@
 ; GCN-LABEL: sgpr_spill_writelane:
 ; GCN:       ; %bb.0:
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GCN-NEXT:    s_xor_saveexec_b64 s[4:5], -1
+; GCN-NEXT:    s_xor_saveexec_b64 s[6:7], -1
 ; GCN-NEXT:    buffer_store_dword v0, off, s[0:3], s32 ; 4-byte Folded Spill
-; GCN-NEXT:    s_mov_b64 exec, s[4:5]
+; GCN-NEXT:    s_mov_b64 exec, s[6:7]
+; GCN-NEXT:    ; implicit-def: $vgpr0
 ; GCN-NEXT:    v_writelane_b32 v0, s35, 0
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ;;#ASMEND
 ; GCN-NEXT:    v_readlane_b32 s35, v0, 0
-; GCN-NEXT:    s_xor_saveexec_b64 s[4:5], -1
+; GCN-NEXT:    s_xor_saveexec_b64 s[6:7], -1
 ; GCN-NEXT:    buffer_load_dword v0, off, s[0:3], s32 ; 4-byte Folded Reload
-; GCN-NEXT:    s_mov_b64 exec, s[4:5]
+; GCN-NEXT:    s_mov_b64 exec, s[6:7]
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
 ; GCN-NEXT:    s_setpc_b64 s[30:31]
   call void asm sideeffect "", "~{s35}"()
diff --git a/llvm/test/CodeGen/AMDGPU/spill192.mir b/llvm/test/CodeGen/AMDGPU/spill192.mir
--- a/llvm/test/CodeGen/AMDGPU/spill192.mir
+++ b/llvm/test/CodeGen/AMDGPU/spill192.mir
@@ -32,32 +32,29 @@
   ; EXPANDED-LABEL: name: spill_restore_sgpr192
   ; EXPANDED: bb.0:
   ; EXPANDED-NEXT:   successors: %bb.1(0x80000000)
-  ; EXPANDED-NEXT:   liveins: $vgpr0
   ; EXPANDED-NEXT: {{  $}}
   ; EXPANDED-NEXT:   S_NOP 0, implicit-def renamable $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9
-  ; EXPANDED-NEXT:   $vgpr0 = V_WRITELANE_B32 $sgpr4, 0, $vgpr0, implicit-def $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9, implicit $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9
-  ; EXPANDED-NEXT:   $vgpr0 = V_WRITELANE_B32 $sgpr5, 1, $vgpr0, implicit $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9
-  ; EXPANDED-NEXT:   $vgpr0 = V_WRITELANE_B32 $sgpr6, 2, $vgpr0, implicit $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9
-  ; EXPANDED-NEXT:   $vgpr0 = V_WRITELANE_B32 $sgpr7, 3, $vgpr0, implicit $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9
-  ; EXPANDED-NEXT:   $vgpr0 = V_WRITELANE_B32 $sgpr8, 4, $vgpr0, implicit $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9
-  ; EXPANDED-NEXT:   $vgpr0 = V_WRITELANE_B32 killed $sgpr9, 5, $vgpr0, implicit killed $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9
+  ; EXPANDED-NEXT:   [[DEF:%[0-9]+]]:vgpr_32 = IMPLICIT_DEF
+  ; EXPANDED-NEXT:   [[V_WRITELANE_B32_:%[0-9]+]]:vgpr_32 = V_WRITELANE_B32 $sgpr4, 0, [[V_WRITELANE_B32_]], implicit-def $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9, implicit $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9
+  ; EXPANDED-NEXT:   [[V_WRITELANE_B32_1:%[0-9]+]]:vgpr_32 = V_WRITELANE_B32 $sgpr5, 1, [[V_WRITELANE_B32_1]], implicit $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9
+  ; EXPANDED-NEXT:   [[V_WRITELANE_B32_1:%[0-9]+]]:vgpr_32 = V_WRITELANE_B32 $sgpr6, 2, [[V_WRITELANE_B32_1]], implicit $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9
+  ; EXPANDED-NEXT:   [[V_WRITELANE_B32_1:%[0-9]+]]:vgpr_32 = V_WRITELANE_B32 $sgpr7, 3, [[V_WRITELANE_B32_1]], implicit $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9
+  ; EXPANDED-NEXT:   [[V_WRITELANE_B32_1:%[0-9]+]]:vgpr_32 = V_WRITELANE_B32 $sgpr8, 4, [[V_WRITELANE_B32_1]], implicit $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9
+  ; EXPANDED-NEXT:   [[V_WRITELANE_B32_1:%[0-9]+]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr9, 5, [[V_WRITELANE_B32_1]], implicit killed $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9
   ; EXPANDED-NEXT:   S_CBRANCH_SCC1 %bb.1, implicit undef $scc
   ; EXPANDED-NEXT: {{  $}}
   ; EXPANDED-NEXT: bb.1:
   ; EXPANDED-NEXT:   successors: %bb.2(0x80000000)
-  ; EXPANDED-NEXT:   liveins: $vgpr0
   ; EXPANDED-NEXT: {{  $}}
   ; EXPANDED-NEXT:   S_NOP 1
   ; EXPANDED-NEXT: {{  $}}
   ; EXPANDED-NEXT: bb.2:
-  ; EXPANDED-NEXT:   liveins: $vgpr0
-  ; EXPANDED-NEXT: {{  $}}
-  ; EXPANDED-NEXT:   $sgpr4 = V_READLANE_B32 $vgpr0, 0, implicit-def $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9
-  ; EXPANDED-NEXT:   $sgpr5 = V_READLANE_B32 $vgpr0, 1
-  ; EXPANDED-NEXT:   $sgpr6 = V_READLANE_B32 $vgpr0, 2
-  ; EXPANDED-NEXT:   $sgpr7 = V_READLANE_B32 $vgpr0, 3
-  ; EXPANDED-NEXT:   $sgpr8 = V_READLANE_B32 $vgpr0, 4
-  ; EXPANDED-NEXT:   $sgpr9 = V_READLANE_B32 $vgpr0, 5
+  ; EXPANDED-NEXT:   $sgpr4 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 0, implicit-def $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9
+  ; EXPANDED-NEXT:   $sgpr5 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 1
+  ; EXPANDED-NEXT:   $sgpr6 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 2
+  ; EXPANDED-NEXT:   $sgpr7 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 3
+  ; EXPANDED-NEXT:   $sgpr8 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 4
+  ; EXPANDED-NEXT:   $sgpr9 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 5
   ; EXPANDED-NEXT:   S_NOP 0, implicit killed renamable $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9
   bb.0:
     S_NOP 0, implicit-def %0:sgpr_192
diff --git a/llvm/test/CodeGen/AMDGPU/spill224.mir b/llvm/test/CodeGen/AMDGPU/spill224.mir
--- a/llvm/test/CodeGen/AMDGPU/spill224.mir
+++ b/llvm/test/CodeGen/AMDGPU/spill224.mir
@@ -30,34 +30,31 @@
   ; EXPANDED-LABEL: name: spill_restore_sgpr224
   ; EXPANDED: bb.0:
   ; EXPANDED-NEXT:   successors: %bb.1(0x80000000)
-  ; EXPANDED-NEXT:   liveins: $vgpr0
   ; EXPANDED-NEXT: {{  $}}
   ; EXPANDED-NEXT:   S_NOP 0, implicit-def renamable $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10
-  ; EXPANDED-NEXT:   $vgpr0 = V_WRITELANE_B32 $sgpr4, 0, $vgpr0, implicit-def $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10, implicit $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10
-  ; EXPANDED-NEXT:   $vgpr0 = V_WRITELANE_B32 $sgpr5, 1, $vgpr0, implicit $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10
-  ; EXPANDED-NEXT:   $vgpr0 = V_WRITELANE_B32 $sgpr6, 2, $vgpr0, implicit $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10
-  ; EXPANDED-NEXT:   $vgpr0 = V_WRITELANE_B32 $sgpr7, 3, $vgpr0, implicit $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10
-  ; EXPANDED-NEXT:   $vgpr0 = V_WRITELANE_B32 $sgpr8, 4, $vgpr0, implicit $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10
-  ; EXPANDED-NEXT:   $vgpr0 = V_WRITELANE_B32 $sgpr9, 5, $vgpr0, implicit $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10
-  ; EXPANDED-NEXT:   $vgpr0 = V_WRITELANE_B32 killed $sgpr10, 6, $vgpr0, implicit killed $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10
+  ; EXPANDED-NEXT:   [[DEF:%[0-9]+]]:vgpr_32 = IMPLICIT_DEF
+  ; EXPANDED-NEXT:   [[V_WRITELANE_B32_:%[0-9]+]]:vgpr_32 = V_WRITELANE_B32 $sgpr4, 0, [[V_WRITELANE_B32_]], implicit-def $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10, implicit $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10
+  ; EXPANDED-NEXT:   [[V_WRITELANE_B32_1:%[0-9]+]]:vgpr_32 = V_WRITELANE_B32 $sgpr5, 1, [[V_WRITELANE_B32_1]], implicit $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10
+  ; EXPANDED-NEXT:   [[V_WRITELANE_B32_1:%[0-9]+]]:vgpr_32 = V_WRITELANE_B32 $sgpr6, 2, [[V_WRITELANE_B32_1]], implicit $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10
+  ; EXPANDED-NEXT:   [[V_WRITELANE_B32_1:%[0-9]+]]:vgpr_32 = V_WRITELANE_B32 $sgpr7, 3, [[V_WRITELANE_B32_1]], implicit $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10
+  ; EXPANDED-NEXT:   [[V_WRITELANE_B32_1:%[0-9]+]]:vgpr_32 = V_WRITELANE_B32 $sgpr8, 4, [[V_WRITELANE_B32_1]], implicit $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10
+  ; EXPANDED-NEXT:   [[V_WRITELANE_B32_1:%[0-9]+]]:vgpr_32 = V_WRITELANE_B32 $sgpr9, 5, [[V_WRITELANE_B32_1]], implicit $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10
+  ; EXPANDED-NEXT:   [[V_WRITELANE_B32_1:%[0-9]+]]:vgpr_32 = V_WRITELANE_B32 killed $sgpr10, 6, [[V_WRITELANE_B32_1]], implicit killed $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10
   ; EXPANDED-NEXT:   S_CBRANCH_SCC1 %bb.1, implicit undef $scc
   ; EXPANDED-NEXT: {{  $}}
   ; EXPANDED-NEXT: bb.1:
   ; EXPANDED-NEXT:   successors: %bb.2(0x80000000)
-  ; EXPANDED-NEXT:   liveins: $vgpr0
   ; EXPANDED-NEXT: {{  $}}
   ; EXPANDED-NEXT:   S_NOP 1
   ; EXPANDED-NEXT: {{  $}}
   ; EXPANDED-NEXT: bb.2:
-  ; EXPANDED-NEXT:   liveins: $vgpr0
-  ; EXPANDED-NEXT: {{  $}}
-  ; EXPANDED-NEXT:   $sgpr4 = V_READLANE_B32 $vgpr0, 0, implicit-def $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10
-  ; EXPANDED-NEXT:   $sgpr5 = V_READLANE_B32 $vgpr0, 1
-  ; EXPANDED-NEXT:   $sgpr6 = V_READLANE_B32 $vgpr0, 2
-  ; EXPANDED-NEXT:   $sgpr7 = V_READLANE_B32 $vgpr0, 3
-  ; EXPANDED-NEXT:   $sgpr8 = V_READLANE_B32 $vgpr0, 4
-  ; EXPANDED-NEXT:   $sgpr9 = V_READLANE_B32 $vgpr0, 5
-  ; EXPANDED-NEXT:   $sgpr10 = V_READLANE_B32 $vgpr0, 6
+  ; EXPANDED-NEXT:   $sgpr4 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 0, implicit-def $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10
+  ; EXPANDED-NEXT:   $sgpr5 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 1
+  ; EXPANDED-NEXT:   $sgpr6 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 2
+  ; EXPANDED-NEXT:   $sgpr7 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 3
+  ; EXPANDED-NEXT:   $sgpr8 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 4
+  ; EXPANDED-NEXT:   $sgpr9 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 5
+  ; EXPANDED-NEXT:   $sgpr10 = V_READLANE_B32 [[V_WRITELANE_B32_1]], 6
   ; EXPANDED-NEXT:   S_NOP 0, implicit killed renamable $sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10
   bb.0:
     S_NOP 0, implicit-def %0:sgpr_224
diff --git a/llvm/test/CodeGen/AMDGPU/tail-call-amdgpu-gfx.ll b/llvm/test/CodeGen/AMDGPU/tail-call-amdgpu-gfx.ll
--- a/llvm/test/CodeGen/AMDGPU/tail-call-amdgpu-gfx.ll
+++ b/llvm/test/CodeGen/AMDGPU/tail-call-amdgpu-gfx.ll
@@ -22,8 +22,9 @@
 ; GCN-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GCN-NEXT:    buffer_store_dword v1, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[34:35]
-; GCN-NEXT:    v_writelane_b32 v1, s4, 0
+; GCN-NEXT:    ; implicit-def: $vgpr1
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
+; GCN-NEXT:    v_writelane_b32 v1, s4, 0
 ; GCN-NEXT:    v_writelane_b32 v1, s30, 1
 ; GCN-NEXT:    v_add_f32_e32 v0, 1.0, v0
 ; GCN-NEXT:    s_mov_b32 s4, 2.0
diff --git a/llvm/test/CodeGen/AMDGPU/tuple-allocation-failure.ll b/llvm/test/CodeGen/AMDGPU/tuple-allocation-failure.ll
--- a/llvm/test/CodeGen/AMDGPU/tuple-allocation-failure.ll
+++ b/llvm/test/CodeGen/AMDGPU/tuple-allocation-failure.ll
@@ -10,18 +10,20 @@
 ; GFX90A-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX90A-NEXT:    s_mov_b32 s16, s33
 ; GFX90A-NEXT:    s_mov_b32 s33, s32
-; GFX90A-NEXT:    s_or_saveexec_b64 s[18:19], -1
-; GFX90A-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
-; GFX90A-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
+; GFX90A-NEXT:    s_xor_saveexec_b64 s[18:19], -1
+; GFX90A-NEXT:    buffer_store_dword v0, off, s[0:3], s33 ; 4-byte Folded Spill
+; GFX90A-NEXT:    s_mov_b64 exec, -1
+; GFX90A-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX90A-NEXT:    s_mov_b64 exec, s[18:19]
 ; GFX90A-NEXT:    s_addk_i32 s32, 0x400
-; GFX90A-NEXT:    v_writelane_b32 v41, s16, 0
+; GFX90A-NEXT:    v_writelane_b32 v40, s16, 0
 ; GFX90A-NEXT:    s_getpc_b64 s[16:17]
 ; GFX90A-NEXT:    s_add_u32 s16, s16, wobble@gotpcrel32@lo+4
 ; GFX90A-NEXT:    s_addc_u32 s17, s17, wobble@gotpcrel32@hi+12
 ; GFX90A-NEXT:    s_load_dwordx2 s[16:17], s[16:17], 0x0
-; GFX90A-NEXT:    v_writelane_b32 v40, s30, 0
-; GFX90A-NEXT:    v_writelane_b32 v40, s31, 1
+; GFX90A-NEXT:    ; implicit-def: $vgpr0
+; GFX90A-NEXT:    v_writelane_b32 v0, s30, 0
+; GFX90A-NEXT:    v_writelane_b32 v0, s31, 1
 ; GFX90A-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX90A-NEXT:    s_swappc_b64 s[30:31], s[16:17]
 bb:
@@ -35,12 +37,12 @@
 ; GLOBALNESS1-NEXT:    s_mov_b64 s[54:55], s[6:7]
 ; GLOBALNESS1-NEXT:    s_load_dwordx4 s[36:39], s[8:9], 0x0
 ; GLOBALNESS1-NEXT:    s_load_dword s6, s[8:9], 0x14
-; GLOBALNESS1-NEXT:    v_mov_b32_e32 v43, v0
-; GLOBALNESS1-NEXT:    v_mov_b32_e32 v40, 0
+; GLOBALNESS1-NEXT:    v_mov_b32_e32 v41, v0
+; GLOBALNESS1-NEXT:    v_mov_b32_e32 v42, 0
 ; GLOBALNESS1-NEXT:    v_pk_mov_b32 v[0:1], 0, 0
-; GLOBALNESS1-NEXT:    global_store_dword v[0:1], v40, off
+; GLOBALNESS1-NEXT:    global_store_dword v[0:1], v42, off
 ; GLOBALNESS1-NEXT:    s_waitcnt lgkmcnt(0)
-; GLOBALNESS1-NEXT:    global_load_dword v0, v40, s[36:37]
+; GLOBALNESS1-NEXT:    global_load_dword v0, v42, s[36:37]
 ; GLOBALNESS1-NEXT:    s_add_u32 flat_scratch_lo, s12, s17
 ; GLOBALNESS1-NEXT:    s_mov_b64 s[64:65], s[4:5]
 ; GLOBALNESS1-NEXT:    s_load_dwordx2 s[4:5], s[8:9], 0x18
@@ -48,11 +50,11 @@
 ; GLOBALNESS1-NEXT:    s_addc_u32 flat_scratch_hi, s13, 0
 ; GLOBALNESS1-NEXT:    s_add_u32 s0, s0, s17
 ; GLOBALNESS1-NEXT:    s_addc_u32 s1, s1, 0
-; GLOBALNESS1-NEXT:    v_mov_b32_e32 v41, 0x40994400
+; GLOBALNESS1-NEXT:    v_mov_b32_e32 v43, 0x40994400
 ; GLOBALNESS1-NEXT:    s_bitcmp1_b32 s38, 0
 ; GLOBALNESS1-NEXT:    s_waitcnt lgkmcnt(0)
-; GLOBALNESS1-NEXT:    v_cmp_ngt_f64_e64 s[36:37], s[4:5], v[40:41]
-; GLOBALNESS1-NEXT:    v_cmp_ngt_f64_e64 s[40:41], s[4:5], 0
+; GLOBALNESS1-NEXT:    v_cmp_ngt_f64_e64 s[40:41], s[4:5], v[42:43]
+; GLOBALNESS1-NEXT:    v_cmp_ngt_f64_e64 s[42:43], s[4:5], 0
 ; GLOBALNESS1-NEXT:    s_cselect_b64 s[4:5], -1, 0
 ; GLOBALNESS1-NEXT:    s_xor_b64 s[94:95], s[4:5], -1
 ; GLOBALNESS1-NEXT:    s_bitcmp1_b32 s6, 0
@@ -65,33 +67,34 @@
 ; GLOBALNESS1-NEXT:    s_add_u32 s6, s6, wobble@gotpcrel32@lo+4
 ; GLOBALNESS1-NEXT:    s_addc_u32 s7, s7, wobble@gotpcrel32@hi+12
 ; GLOBALNESS1-NEXT:    s_xor_b64 s[86:87], s[4:5], -1
+; GLOBALNESS1-NEXT:    ; implicit-def: $vgpr40
 ; GLOBALNESS1-NEXT:    s_load_dwordx2 s[66:67], s[6:7], 0x0
 ; GLOBALNESS1-NEXT:    s_mov_b32 s98, s16
 ; GLOBALNESS1-NEXT:    s_mov_b64 s[62:63], s[8:9]
 ; GLOBALNESS1-NEXT:    s_mov_b32 s99, s15
-; GLOBALNESS1-NEXT:    s_mov_b32 s100, s14
+; GLOBALNESS1-NEXT:    s_mov_b32 s56, s14
 ; GLOBALNESS1-NEXT:    s_mov_b64 s[34:35], s[10:11]
 ; GLOBALNESS1-NEXT:    s_mov_b64 s[92:93], 0x80
-; GLOBALNESS1-NEXT:    v_cmp_ne_u32_e64 s[42:43], 1, v1
+; GLOBALNESS1-NEXT:    v_cmp_ne_u32_e64 s[36:37], 1, v1
 ; GLOBALNESS1-NEXT:    s_mov_b32 s69, 0x3ff00000
 ; GLOBALNESS1-NEXT:    s_mov_b32 s32, 0
 ; GLOBALNESS1-NEXT:    ; implicit-def: $agpr32_agpr33_agpr34_agpr35_agpr36_agpr37_agpr38_agpr39_agpr40_agpr41_agpr42_agpr43_agpr44_agpr45_agpr46_agpr47_agpr48_agpr49_agpr50_agpr51_agpr52_agpr53_agpr54_agpr55_agpr56_agpr57_agpr58_agpr59_agpr60_agpr61_agpr62_agpr63
 ; GLOBALNESS1-NEXT:    s_waitcnt vmcnt(0)
 ; GLOBALNESS1-NEXT:    v_cmp_gt_i32_e64 s[4:5], 0, v0
-; GLOBALNESS1-NEXT:    v_writelane_b32 v42, s4, 0
-; GLOBALNESS1-NEXT:    v_writelane_b32 v42, s5, 1
+; GLOBALNESS1-NEXT:    v_writelane_b32 v40, s4, 0
+; GLOBALNESS1-NEXT:    v_writelane_b32 v40, s5, 1
 ; GLOBALNESS1-NEXT:    v_cmp_eq_u32_e64 s[4:5], 1, v0
-; GLOBALNESS1-NEXT:    v_writelane_b32 v42, s4, 2
-; GLOBALNESS1-NEXT:    v_writelane_b32 v42, s5, 3
+; GLOBALNESS1-NEXT:    v_writelane_b32 v40, s4, 2
+; GLOBALNESS1-NEXT:    v_writelane_b32 v40, s5, 3
 ; GLOBALNESS1-NEXT:    v_cmp_eq_u32_e64 s[4:5], 0, v0
-; GLOBALNESS1-NEXT:    v_writelane_b32 v42, s4, 4
+; GLOBALNESS1-NEXT:    v_writelane_b32 v40, s4, 4
 ; GLOBALNESS1-NEXT:    v_cmp_gt_i32_e64 s[90:91], 1, v0
-; GLOBALNESS1-NEXT:    v_writelane_b32 v42, s5, 5
+; GLOBALNESS1-NEXT:    v_writelane_b32 v40, s5, 5
 ; GLOBALNESS1-NEXT:    s_branch .LBB1_4
 ; GLOBALNESS1-NEXT:  .LBB1_1: ; %bb70.i
 ; GLOBALNESS1-NEXT:    ; in Loop: Header=BB1_4 Depth=1
-; GLOBALNESS1-NEXT:    v_readlane_b32 s6, v42, 4
-; GLOBALNESS1-NEXT:    v_readlane_b32 s7, v42, 5
+; GLOBALNESS1-NEXT:    v_readlane_b32 s6, v40, 4
+; GLOBALNESS1-NEXT:    v_readlane_b32 s7, v40, 5
 ; GLOBALNESS1-NEXT:    s_andn2_b64 vcc, exec, s[6:7]
 ; GLOBALNESS1-NEXT:    s_cbranch_vccz .LBB1_29
 ; GLOBALNESS1-NEXT:  .LBB1_2: ; %Flow6
@@ -141,19 +144,19 @@
 ; GLOBALNESS1-NEXT:    v_pk_mov_b32 v[0:1], s[92:93], s[92:93] op_sel:[0,1]
 ; GLOBALNESS1-NEXT:    flat_load_dword v44, v[0:1]
 ; GLOBALNESS1-NEXT:    s_add_u32 s8, s62, 40
-; GLOBALNESS1-NEXT:    buffer_store_dword v40, off, s[0:3], 0
+; GLOBALNESS1-NEXT:    buffer_store_dword v42, off, s[0:3], 0
 ; GLOBALNESS1-NEXT:    flat_load_dword v45, v[0:1]
 ; GLOBALNESS1-NEXT:    s_addc_u32 s9, s63, 0
 ; GLOBALNESS1-NEXT:    s_mov_b64 s[4:5], s[64:65]
 ; GLOBALNESS1-NEXT:    s_mov_b64 s[6:7], s[54:55]
 ; GLOBALNESS1-NEXT:    s_mov_b64 s[10:11], s[34:35]
-; GLOBALNESS1-NEXT:    s_mov_b32 s12, s100
+; GLOBALNESS1-NEXT:    s_mov_b32 s12, s56
 ; GLOBALNESS1-NEXT:    s_mov_b32 s13, s99
 ; GLOBALNESS1-NEXT:    s_mov_b32 s14, s98
-; GLOBALNESS1-NEXT:    v_mov_b32_e32 v31, v43
+; GLOBALNESS1-NEXT:    v_mov_b32_e32 v31, v41
 ; GLOBALNESS1-NEXT:    s_waitcnt lgkmcnt(0)
 ; GLOBALNESS1-NEXT:    s_swappc_b64 s[30:31], s[66:67]
-; GLOBALNESS1-NEXT:    s_and_b64 vcc, exec, s[42:43]
+; GLOBALNESS1-NEXT:    s_and_b64 vcc, exec, s[36:37]
 ; GLOBALNESS1-NEXT:    s_mov_b64 s[6:7], -1
 ; GLOBALNESS1-NEXT:    ; implicit-def: $sgpr4_sgpr5
 ; GLOBALNESS1-NEXT:    s_cbranch_vccnz .LBB1_8
@@ -220,23 +223,23 @@
 ; GLOBALNESS1-NEXT:  ; %bb.10: ; %bb33.i
 ; GLOBALNESS1-NEXT:    ; in Loop: Header=BB1_4 Depth=1
 ; GLOBALNESS1-NEXT:    global_load_dwordx2 v[0:1], v[32:33], off
-; GLOBALNESS1-NEXT:    v_readlane_b32 s4, v42, 0
-; GLOBALNESS1-NEXT:    v_readlane_b32 s5, v42, 1
+; GLOBALNESS1-NEXT:    v_readlane_b32 s4, v40, 0
+; GLOBALNESS1-NEXT:    v_readlane_b32 s5, v40, 1
+; GLOBALNESS1-NEXT:    s_mov_b64 s[72:73], s[36:37]
+; GLOBALNESS1-NEXT:    s_mov_b32 s75, s39
 ; GLOBALNESS1-NEXT:    s_andn2_b64 vcc, exec, s[4:5]
 ; GLOBALNESS1-NEXT:    s_cbranch_vccnz .LBB1_12
 ; GLOBALNESS1-NEXT:  ; %bb.11: ; %bb39.i
 ; GLOBALNESS1-NEXT:    ; in Loop: Header=BB1_4 Depth=1
-; GLOBALNESS1-NEXT:    v_mov_b32_e32 v41, v40
+; GLOBALNESS1-NEXT:    v_mov_b32_e32 v43, v42
 ; GLOBALNESS1-NEXT:    v_pk_mov_b32 v[2:3], 0, 0
-; GLOBALNESS1-NEXT:    global_store_dwordx2 v[2:3], v[40:41], off
+; GLOBALNESS1-NEXT:    global_store_dwordx2 v[2:3], v[42:43], off
 ; GLOBALNESS1-NEXT:  .LBB1_12: ; %bb44.lr.ph.i
 ; GLOBALNESS1-NEXT:    ; in Loop: Header=BB1_4 Depth=1
 ; GLOBALNESS1-NEXT:    v_cmp_ne_u32_e32 vcc, 0, v45
 ; GLOBALNESS1-NEXT:    v_cndmask_b32_e32 v2, 0, v44, vcc
-; GLOBALNESS1-NEXT:    s_mov_b64 s[72:73], s[42:43]
-; GLOBALNESS1-NEXT:    s_mov_b32 s75, s39
 ; GLOBALNESS1-NEXT:    s_waitcnt vmcnt(0)
-; GLOBALNESS1-NEXT:    v_cmp_nlt_f64_e64 s[56:57], 0, v[0:1]
+; GLOBALNESS1-NEXT:    v_cmp_nlt_f64_e64 s[36:37], 0, v[0:1]
 ; GLOBALNESS1-NEXT:    v_cmp_eq_u32_e64 s[58:59], 0, v2
 ; GLOBALNESS1-NEXT:    s_branch .LBB1_15
 ; GLOBALNESS1-NEXT:  .LBB1_13: ; %Flow7
@@ -257,15 +260,15 @@
 ; GLOBALNESS1-NEXT:    s_cbranch_vccnz .LBB1_14
 ; GLOBALNESS1-NEXT:  ; %bb.17: ; %bb50.i
 ; GLOBALNESS1-NEXT:    ; in Loop: Header=BB1_15 Depth=2
-; GLOBALNESS1-NEXT:    s_andn2_b64 vcc, exec, s[36:37]
+; GLOBALNESS1-NEXT:    s_andn2_b64 vcc, exec, s[40:41]
 ; GLOBALNESS1-NEXT:    s_cbranch_vccnz .LBB1_20
 ; GLOBALNESS1-NEXT:  ; %bb.18: ; %bb3.i.i
 ; GLOBALNESS1-NEXT:    ; in Loop: Header=BB1_15 Depth=2
-; GLOBALNESS1-NEXT:    s_andn2_b64 vcc, exec, s[40:41]
+; GLOBALNESS1-NEXT:    s_andn2_b64 vcc, exec, s[42:43]
 ; GLOBALNESS1-NEXT:    s_cbranch_vccnz .LBB1_20
 ; GLOBALNESS1-NEXT:  ; %bb.19: ; %bb6.i.i
 ; GLOBALNESS1-NEXT:    ; in Loop: Header=BB1_15 Depth=2
-; GLOBALNESS1-NEXT:    s_andn2_b64 vcc, exec, s[56:57]
+; GLOBALNESS1-NEXT:    s_andn2_b64 vcc, exec, s[36:37]
 ; GLOBALNESS1-NEXT:  .LBB1_20: ; %spam.exit.i
 ; GLOBALNESS1-NEXT:    ; in Loop: Header=BB1_15 Depth=2
 ; GLOBALNESS1-NEXT:    s_andn2_b64 vcc, exec, s[90:91]
@@ -278,28 +281,28 @@
 ; GLOBALNESS1-NEXT:    s_mov_b64 s[6:7], s[54:55]
 ; GLOBALNESS1-NEXT:    s_mov_b64 s[8:9], s[60:61]
 ; GLOBALNESS1-NEXT:    s_mov_b64 s[10:11], s[34:35]
-; GLOBALNESS1-NEXT:    s_mov_b32 s12, s100
+; GLOBALNESS1-NEXT:    s_mov_b32 s12, s56
 ; GLOBALNESS1-NEXT:    s_mov_b32 s13, s99
 ; GLOBALNESS1-NEXT:    s_mov_b32 s14, s98
-; GLOBALNESS1-NEXT:    v_mov_b32_e32 v31, v43
+; GLOBALNESS1-NEXT:    v_mov_b32_e32 v31, v41
 ; GLOBALNESS1-NEXT:    s_swappc_b64 s[30:31], s[66:67]
 ; GLOBALNESS1-NEXT:    v_pk_mov_b32 v[44:45], 0, 0
 ; GLOBALNESS1-NEXT:    s_mov_b64 s[4:5], s[64:65]
 ; GLOBALNESS1-NEXT:    s_mov_b64 s[6:7], s[54:55]
 ; GLOBALNESS1-NEXT:    s_mov_b64 s[8:9], s[60:61]
 ; GLOBALNESS1-NEXT:    s_mov_b64 s[10:11], s[34:35]
-; GLOBALNESS1-NEXT:    s_mov_b32 s12, s100
+; GLOBALNESS1-NEXT:    s_mov_b32 s12, s56
 ; GLOBALNESS1-NEXT:    s_mov_b32 s13, s99
 ; GLOBALNESS1-NEXT:    s_mov_b32 s14, s98
-; GLOBALNESS1-NEXT:    v_mov_b32_e32 v31, v43
+; GLOBALNESS1-NEXT:    v_mov_b32_e32 v31, v41
 ; GLOBALNESS1-NEXT:    global_store_dwordx2 v[44:45], a[32:33], off
 ; GLOBALNESS1-NEXT:    s_swappc_b64 s[30:31], s[66:67]
 ; GLOBALNESS1-NEXT:    s_and_saveexec_b64 s[4:5], s[58:59]
 ; GLOBALNESS1-NEXT:    s_cbranch_execz .LBB1_13
 ; GLOBALNESS1-NEXT:  ; %bb.22: ; %bb62.i
 ; GLOBALNESS1-NEXT:    ; in Loop: Header=BB1_15 Depth=2
-; GLOBALNESS1-NEXT:    v_mov_b32_e32 v41, v40
-; GLOBALNESS1-NEXT:    global_store_dwordx2 v[44:45], v[40:41], off
+; GLOBALNESS1-NEXT:    v_mov_b32_e32 v43, v42
+; GLOBALNESS1-NEXT:    global_store_dwordx2 v[44:45], v[42:43], off
 ; GLOBALNESS1-NEXT:    s_branch .LBB1_13
 ; GLOBALNESS1-NEXT:  .LBB1_23: ; %LeafBlock
 ; GLOBALNESS1-NEXT:    ; in Loop: Header=BB1_4 Depth=1
@@ -314,14 +317,14 @@
 ; GLOBALNESS1-NEXT:    s_branch .LBB1_3
 ; GLOBALNESS1-NEXT:  .LBB1_25: ; %Flow14
 ; GLOBALNESS1-NEXT:    ; in Loop: Header=BB1_4 Depth=1
-; GLOBALNESS1-NEXT:    s_mov_b64 s[4:5], s[36:37]
 ; GLOBALNESS1-NEXT:    s_mov_b32 s36, s93
 ; GLOBALNESS1-NEXT:    s_mov_b32 s37, s93
 ; GLOBALNESS1-NEXT:    s_mov_b32 s38, s93
 ; GLOBALNESS1-NEXT:    s_mov_b32 s39, s93
-; GLOBALNESS1-NEXT:    s_mov_b64 s[6:7], s[40:41]
+; GLOBALNESS1-NEXT:    s_mov_b64 s[4:5], s[40:41]
 ; GLOBALNESS1-NEXT:    s_mov_b32 s40, s93
 ; GLOBALNESS1-NEXT:    s_mov_b32 s41, s93
+; GLOBALNESS1-NEXT:    s_mov_b64 s[6:7], s[42:43]
 ; GLOBALNESS1-NEXT:    s_mov_b32 s42, s93
 ; GLOBALNESS1-NEXT:    s_mov_b32 s43, s93
 ; GLOBALNESS1-NEXT:    s_mov_b32 s44, s93
@@ -350,10 +353,10 @@
 ; GLOBALNESS1-NEXT:    v_pk_mov_b32 v[26:27], s[62:63], s[62:63] op_sel:[0,1]
 ; GLOBALNESS1-NEXT:    v_pk_mov_b32 v[28:29], s[64:65], s[64:65] op_sel:[0,1]
 ; GLOBALNESS1-NEXT:    v_pk_mov_b32 v[30:31], s[66:67], s[66:67] op_sel:[0,1]
-; GLOBALNESS1-NEXT:    s_mov_b64 s[40:41], s[6:7]
-; GLOBALNESS1-NEXT:    s_mov_b64 s[36:37], s[4:5]
+; GLOBALNESS1-NEXT:    s_mov_b64 s[42:43], s[6:7]
+; GLOBALNESS1-NEXT:    s_mov_b64 s[40:41], s[4:5]
 ; GLOBALNESS1-NEXT:    s_mov_b32 s39, s75
-; GLOBALNESS1-NEXT:    s_mov_b64 s[42:43], s[72:73]
+; GLOBALNESS1-NEXT:    s_mov_b64 s[36:37], s[72:73]
 ; GLOBALNESS1-NEXT:  .LBB1_26: ; %Flow15
 ; GLOBALNESS1-NEXT:    ; in Loop: Header=BB1_4 Depth=1
 ; GLOBALNESS1-NEXT:    s_or_b64 exec, exec, s[70:71]
@@ -361,21 +364,21 @@
 ; GLOBALNESS1-NEXT:    s_cbranch_execz .LBB1_2
 ; GLOBALNESS1-NEXT:  ; %bb.27: ; %bb67.i
 ; GLOBALNESS1-NEXT:    ; in Loop: Header=BB1_4 Depth=1
-; GLOBALNESS1-NEXT:    v_readlane_b32 s6, v42, 2
-; GLOBALNESS1-NEXT:    v_readlane_b32 s7, v42, 3
+; GLOBALNESS1-NEXT:    v_readlane_b32 s6, v40, 2
+; GLOBALNESS1-NEXT:    v_readlane_b32 s7, v40, 3
 ; GLOBALNESS1-NEXT:    s_andn2_b64 vcc, exec, s[6:7]
 ; GLOBALNESS1-NEXT:    s_cbranch_vccnz .LBB1_1
 ; GLOBALNESS1-NEXT:  ; %bb.28: ; %bb69.i
 ; GLOBALNESS1-NEXT:    ; in Loop: Header=BB1_4 Depth=1
-; GLOBALNESS1-NEXT:    v_mov_b32_e32 v41, v40
+; GLOBALNESS1-NEXT:    v_mov_b32_e32 v43, v42
 ; GLOBALNESS1-NEXT:    v_pk_mov_b32 v[32:33], 0, 0
-; GLOBALNESS1-NEXT:    global_store_dwordx2 v[32:33], v[40:41], off
+; GLOBALNESS1-NEXT:    global_store_dwordx2 v[32:33], v[42:43], off
 ; GLOBALNESS1-NEXT:    s_branch .LBB1_1
 ; GLOBALNESS1-NEXT:  .LBB1_29: ; %bb73.i
 ; GLOBALNESS1-NEXT:    ; in Loop: Header=BB1_4 Depth=1
-; GLOBALNESS1-NEXT:    v_mov_b32_e32 v41, v40
+; GLOBALNESS1-NEXT:    v_mov_b32_e32 v43, v42
 ; GLOBALNESS1-NEXT:    v_pk_mov_b32 v[32:33], 0, 0
-; GLOBALNESS1-NEXT:    global_store_dwordx2 v[32:33], v[40:41], off
+; GLOBALNESS1-NEXT:    global_store_dwordx2 v[32:33], v[42:43], off
 ; GLOBALNESS1-NEXT:    s_branch .LBB1_2
 ; GLOBALNESS1-NEXT:  .LBB1_30: ; %loop.exit.guard
 ; GLOBALNESS1-NEXT:    s_andn2_b64 vcc, exec, s[4:5]
@@ -387,10 +390,10 @@
 ; GLOBALNESS1-NEXT:    s_mov_b64 s[4:5], s[64:65]
 ; GLOBALNESS1-NEXT:    s_mov_b64 s[6:7], s[54:55]
 ; GLOBALNESS1-NEXT:    s_mov_b64 s[10:11], s[34:35]
-; GLOBALNESS1-NEXT:    s_mov_b32 s12, s100
+; GLOBALNESS1-NEXT:    s_mov_b32 s12, s56
 ; GLOBALNESS1-NEXT:    s_mov_b32 s13, s99
 ; GLOBALNESS1-NEXT:    s_mov_b32 s14, s98
-; GLOBALNESS1-NEXT:    v_mov_b32_e32 v31, v43
+; GLOBALNESS1-NEXT:    v_mov_b32_e32 v31, v41
 ; GLOBALNESS1-NEXT:    s_getpc_b64 s[16:17]
 ; GLOBALNESS1-NEXT:    s_add_u32 s16, s16, widget@rel32@lo+4
 ; GLOBALNESS1-NEXT:    s_addc_u32 s17, s17, widget@rel32@hi+12
@@ -405,10 +408,10 @@
 ; GLOBALNESS1-NEXT:    s_mov_b64 s[4:5], s[64:65]
 ; GLOBALNESS1-NEXT:    s_mov_b64 s[6:7], s[54:55]
 ; GLOBALNESS1-NEXT:    s_mov_b64 s[10:11], s[34:35]
-; GLOBALNESS1-NEXT:    s_mov_b32 s12, s100
+; GLOBALNESS1-NEXT:    s_mov_b32 s12, s56
 ; GLOBALNESS1-NEXT:    s_mov_b32 s13, s99
 ; GLOBALNESS1-NEXT:    s_mov_b32 s14, s98
-; GLOBALNESS1-NEXT:    v_mov_b32_e32 v31, v43
+; GLOBALNESS1-NEXT:    v_mov_b32_e32 v31, v41
 ; GLOBALNESS1-NEXT:    s_getpc_b64 s[16:17]
 ; GLOBALNESS1-NEXT:    s_add_u32 s16, s16, widget@rel32@lo+4
 ; GLOBALNESS1-NEXT:    s_addc_u32 s17, s17, widget@rel32@hi+12
@@ -420,12 +423,12 @@
 ; GLOBALNESS0-NEXT:    s_mov_b64 s[54:55], s[6:7]
 ; GLOBALNESS0-NEXT:    s_load_dwordx4 s[36:39], s[8:9], 0x0
 ; GLOBALNESS0-NEXT:    s_load_dword s6, s[8:9], 0x14
-; GLOBALNESS0-NEXT:    v_mov_b32_e32 v43, v0
-; GLOBALNESS0-NEXT:    v_mov_b32_e32 v40, 0
+; GLOBALNESS0-NEXT:    v_mov_b32_e32 v41, v0
+; GLOBALNESS0-NEXT:    v_mov_b32_e32 v42, 0
 ; GLOBALNESS0-NEXT:    v_pk_mov_b32 v[0:1], 0, 0
-; GLOBALNESS0-NEXT:    global_store_dword v[0:1], v40, off
+; GLOBALNESS0-NEXT:    global_store_dword v[0:1], v42, off
 ; GLOBALNESS0-NEXT:    s_waitcnt lgkmcnt(0)
-; GLOBALNESS0-NEXT:    global_load_dword v0, v40, s[36:37]
+; GLOBALNESS0-NEXT:    global_load_dword v0, v42, s[36:37]
 ; GLOBALNESS0-NEXT:    s_add_u32 flat_scratch_lo, s12, s17
 ; GLOBALNESS0-NEXT:    s_mov_b64 s[62:63], s[4:5]
 ; GLOBALNESS0-NEXT:    s_load_dwordx2 s[4:5], s[8:9], 0x18
@@ -433,11 +436,11 @@
 ; GLOBALNESS0-NEXT:    s_addc_u32 flat_scratch_hi, s13, 0
 ; GLOBALNESS0-NEXT:    s_add_u32 s0, s0, s17
 ; GLOBALNESS0-NEXT:    s_addc_u32 s1, s1, 0
-; GLOBALNESS0-NEXT:    v_mov_b32_e32 v41, 0x40994400
+; GLOBALNESS0-NEXT:    v_mov_b32_e32 v43, 0x40994400
 ; GLOBALNESS0-NEXT:    s_bitcmp1_b32 s38, 0
 ; GLOBALNESS0-NEXT:    s_waitcnt lgkmcnt(0)
-; GLOBALNESS0-NEXT:    v_cmp_ngt_f64_e64 s[36:37], s[4:5], v[40:41]
-; GLOBALNESS0-NEXT:    v_cmp_ngt_f64_e64 s[40:41], s[4:5], 0
+; GLOBALNESS0-NEXT:    v_cmp_ngt_f64_e64 s[40:41], s[4:5], v[42:43]
+; GLOBALNESS0-NEXT:    v_cmp_ngt_f64_e64 s[42:43], s[4:5], 0
 ; GLOBALNESS0-NEXT:    s_cselect_b64 s[4:5], -1, 0
 ; GLOBALNESS0-NEXT:    s_xor_b64 s[94:95], s[4:5], -1
 ; GLOBALNESS0-NEXT:    s_bitcmp1_b32 s6, 0
@@ -450,33 +453,34 @@
 ; GLOBALNESS0-NEXT:    s_add_u32 s6, s6, wobble@gotpcrel32@lo+4
 ; GLOBALNESS0-NEXT:    s_addc_u32 s7, s7, wobble@gotpcrel32@hi+12
 ; GLOBALNESS0-NEXT:    s_xor_b64 s[86:87], s[4:5], -1
+; GLOBALNESS0-NEXT:    ; implicit-def: $vgpr40
 ; GLOBALNESS0-NEXT:    s_load_dwordx2 s[66:67], s[6:7], 0x0
 ; GLOBALNESS0-NEXT:    s_mov_b32 s98, s16
 ; GLOBALNESS0-NEXT:    s_mov_b64 s[60:61], s[8:9]
 ; GLOBALNESS0-NEXT:    s_mov_b32 s99, s15
-; GLOBALNESS0-NEXT:    s_mov_b32 s100, s14
+; GLOBALNESS0-NEXT:    s_mov_b32 s56, s14
 ; GLOBALNESS0-NEXT:    s_mov_b64 s[34:35], s[10:11]
 ; GLOBALNESS0-NEXT:    s_mov_b64 s[92:93], 0x80
-; GLOBALNESS0-NEXT:    v_cmp_ne_u32_e64 s[42:43], 1, v1
+; GLOBALNESS0-NEXT:    v_cmp_ne_u32_e64 s[36:37], 1, v1
 ; GLOBALNESS0-NEXT:    s_mov_b32 s69, 0x3ff00000
 ; GLOBALNESS0-NEXT:    s_mov_b32 s32, 0
 ; GLOBALNESS0-NEXT:    ; implicit-def: $agpr32_agpr33_agpr34_agpr35_agpr36_agpr37_agpr38_agpr39_agpr40_agpr41_agpr42_agpr43_agpr44_agpr45_agpr46_agpr47_agpr48_agpr49_agpr50_agpr51_agpr52_agpr53_agpr54_agpr55_agpr56_agpr57_agpr58_agpr59_agpr60_agpr61_agpr62_agpr63
 ; GLOBALNESS0-NEXT:    s_waitcnt vmcnt(0)
 ; GLOBALNESS0-NEXT:    v_cmp_gt_i32_e64 s[4:5], 0, v0
-; GLOBALNESS0-NEXT:    v_writelane_b32 v42, s4, 0
-; GLOBALNESS0-NEXT:    v_writelane_b32 v42, s5, 1
+; GLOBALNESS0-NEXT:    v_writelane_b32 v40, s4, 0
+; GLOBALNESS0-NEXT:    v_writelane_b32 v40, s5, 1
 ; GLOBALNESS0-NEXT:    v_cmp_eq_u32_e64 s[4:5], 1, v0
-; GLOBALNESS0-NEXT:    v_writelane_b32 v42, s4, 2
-; GLOBALNESS0-NEXT:    v_writelane_b32 v42, s5, 3
+; GLOBALNESS0-NEXT:    v_writelane_b32 v40, s4, 2
+; GLOBALNESS0-NEXT:    v_writelane_b32 v40, s5, 3
 ; GLOBALNESS0-NEXT:    v_cmp_eq_u32_e64 s[4:5], 0, v0
-; GLOBALNESS0-NEXT:    v_writelane_b32 v42, s4, 4
+; GLOBALNESS0-NEXT:    v_writelane_b32 v40, s4, 4
 ; GLOBALNESS0-NEXT:    v_cmp_gt_i32_e64 s[90:91], 1, v0
-; GLOBALNESS0-NEXT:    v_writelane_b32 v42, s5, 5
+; GLOBALNESS0-NEXT:    v_writelane_b32 v40, s5, 5
 ; GLOBALNESS0-NEXT:    s_branch .LBB1_4
 ; GLOBALNESS0-NEXT:  .LBB1_1: ; %bb70.i
 ; GLOBALNESS0-NEXT:    ; in Loop: Header=BB1_4 Depth=1
-; GLOBALNESS0-NEXT:    v_readlane_b32 s6, v42, 4
-; GLOBALNESS0-NEXT:    v_readlane_b32 s7, v42, 5
+; GLOBALNESS0-NEXT:    v_readlane_b32 s6, v40, 4
+; GLOBALNESS0-NEXT:    v_readlane_b32 s7, v40, 5
 ; GLOBALNESS0-NEXT:    s_andn2_b64 vcc, exec, s[6:7]
 ; GLOBALNESS0-NEXT:    s_cbranch_vccz .LBB1_29
 ; GLOBALNESS0-NEXT:  .LBB1_2: ; %Flow6
@@ -526,19 +530,19 @@
 ; GLOBALNESS0-NEXT:    v_pk_mov_b32 v[0:1], s[92:93], s[92:93] op_sel:[0,1]
 ; GLOBALNESS0-NEXT:    flat_load_dword v44, v[0:1]
 ; GLOBALNESS0-NEXT:    s_add_u32 s8, s60, 40
-; GLOBALNESS0-NEXT:    buffer_store_dword v40, off, s[0:3], 0
+; GLOBALNESS0-NEXT:    buffer_store_dword v42, off, s[0:3], 0
 ; GLOBALNESS0-NEXT:    flat_load_dword v45, v[0:1]
 ; GLOBALNESS0-NEXT:    s_addc_u32 s9, s61, 0
 ; GLOBALNESS0-NEXT:    s_mov_b64 s[4:5], s[62:63]
 ; GLOBALNESS0-NEXT:    s_mov_b64 s[6:7], s[54:55]
 ; GLOBALNESS0-NEXT:    s_mov_b64 s[10:11], s[34:35]
-; GLOBALNESS0-NEXT:    s_mov_b32 s12, s100
+; GLOBALNESS0-NEXT:    s_mov_b32 s12, s56
 ; GLOBALNESS0-NEXT:    s_mov_b32 s13, s99
 ; GLOBALNESS0-NEXT:    s_mov_b32 s14, s98
-; GLOBALNESS0-NEXT:    v_mov_b32_e32 v31, v43
+; GLOBALNESS0-NEXT:    v_mov_b32_e32 v31, v41
 ; GLOBALNESS0-NEXT:    s_waitcnt lgkmcnt(0)
 ; GLOBALNESS0-NEXT:    s_swappc_b64 s[30:31], s[66:67]
-; GLOBALNESS0-NEXT:    s_and_b64 vcc, exec, s[42:43]
+; GLOBALNESS0-NEXT:    s_and_b64 vcc, exec, s[36:37]
 ; GLOBALNESS0-NEXT:    s_mov_b64 s[6:7], -1
 ; GLOBALNESS0-NEXT:    ; implicit-def: $sgpr4_sgpr5
 ; GLOBALNESS0-NEXT:    s_cbranch_vccnz .LBB1_8
@@ -605,23 +609,23 @@
 ; GLOBALNESS0-NEXT:  ; %bb.10: ; %bb33.i
 ; GLOBALNESS0-NEXT:    ; in Loop: Header=BB1_4 Depth=1
 ; GLOBALNESS0-NEXT:    global_load_dwordx2 v[0:1], v[32:33], off
-; GLOBALNESS0-NEXT:    v_readlane_b32 s4, v42, 0
-; GLOBALNESS0-NEXT:    v_readlane_b32 s5, v42, 1
+; GLOBALNESS0-NEXT:    v_readlane_b32 s4, v40, 0
+; GLOBALNESS0-NEXT:    v_readlane_b32 s5, v40, 1
+; GLOBALNESS0-NEXT:    s_mov_b64 s[72:73], s[36:37]
+; GLOBALNESS0-NEXT:    s_mov_b32 s75, s39
 ; GLOBALNESS0-NEXT:    s_andn2_b64 vcc, exec, s[4:5]
 ; GLOBALNESS0-NEXT:    s_cbranch_vccnz .LBB1_12
 ; GLOBALNESS0-NEXT:  ; %bb.11: ; %bb39.i
 ; GLOBALNESS0-NEXT:    ; in Loop: Header=BB1_4 Depth=1
-; GLOBALNESS0-NEXT:    v_mov_b32_e32 v41, v40
+; GLOBALNESS0-NEXT:    v_mov_b32_e32 v43, v42
 ; GLOBALNESS0-NEXT:    v_pk_mov_b32 v[2:3], 0, 0
-; GLOBALNESS0-NEXT:    global_store_dwordx2 v[2:3], v[40:41], off
+; GLOBALNESS0-NEXT:    global_store_dwordx2 v[2:3], v[42:43], off
 ; GLOBALNESS0-NEXT:  .LBB1_12: ; %bb44.lr.ph.i
 ; GLOBALNESS0-NEXT:    ; in Loop: Header=BB1_4 Depth=1
 ; GLOBALNESS0-NEXT:    v_cmp_ne_u32_e32 vcc, 0, v45
 ; GLOBALNESS0-NEXT:    v_cndmask_b32_e32 v2, 0, v44, vcc
-; GLOBALNESS0-NEXT:    s_mov_b64 s[72:73], s[42:43]
-; GLOBALNESS0-NEXT:    s_mov_b32 s75, s39
 ; GLOBALNESS0-NEXT:    s_waitcnt vmcnt(0)
-; GLOBALNESS0-NEXT:    v_cmp_nlt_f64_e64 s[56:57], 0, v[0:1]
+; GLOBALNESS0-NEXT:    v_cmp_nlt_f64_e64 s[36:37], 0, v[0:1]
 ; GLOBALNESS0-NEXT:    v_cmp_eq_u32_e64 s[58:59], 0, v2
 ; GLOBALNESS0-NEXT:    s_branch .LBB1_15
 ; GLOBALNESS0-NEXT:  .LBB1_13: ; %Flow7
@@ -642,15 +646,15 @@
 ; GLOBALNESS0-NEXT:    s_cbranch_vccnz .LBB1_14
 ; GLOBALNESS0-NEXT:  ; %bb.17: ; %bb50.i
 ; GLOBALNESS0-NEXT:    ; in Loop: Header=BB1_15 Depth=2
-; GLOBALNESS0-NEXT:    s_andn2_b64 vcc, exec, s[36:37]
+; GLOBALNESS0-NEXT:    s_andn2_b64 vcc, exec, s[40:41]
 ; GLOBALNESS0-NEXT:    s_cbranch_vccnz .LBB1_20
 ; GLOBALNESS0-NEXT:  ; %bb.18: ; %bb3.i.i
 ; GLOBALNESS0-NEXT:    ; in Loop: Header=BB1_15 Depth=2
-; GLOBALNESS0-NEXT:    s_andn2_b64 vcc, exec, s[40:41]
+; GLOBALNESS0-NEXT:    s_andn2_b64 vcc, exec, s[42:43]
 ; GLOBALNESS0-NEXT:    s_cbranch_vccnz .LBB1_20
 ; GLOBALNESS0-NEXT:  ; %bb.19: ; %bb6.i.i
 ; GLOBALNESS0-NEXT:    ; in Loop: Header=BB1_15 Depth=2
-; GLOBALNESS0-NEXT:    s_andn2_b64 vcc, exec, s[56:57]
+; GLOBALNESS0-NEXT:    s_andn2_b64 vcc, exec, s[36:37]
 ; GLOBALNESS0-NEXT:  .LBB1_20: ; %spam.exit.i
 ; GLOBALNESS0-NEXT:    ; in Loop: Header=BB1_15 Depth=2
 ; GLOBALNESS0-NEXT:    s_andn2_b64 vcc, exec, s[90:91]
@@ -663,28 +667,28 @@
 ; GLOBALNESS0-NEXT:    s_mov_b64 s[6:7], s[54:55]
 ; GLOBALNESS0-NEXT:    s_mov_b64 s[8:9], s[64:65]
 ; GLOBALNESS0-NEXT:    s_mov_b64 s[10:11], s[34:35]
-; GLOBALNESS0-NEXT:    s_mov_b32 s12, s100
+; GLOBALNESS0-NEXT:    s_mov_b32 s12, s56
 ; GLOBALNESS0-NEXT:    s_mov_b32 s13, s99
 ; GLOBALNESS0-NEXT:    s_mov_b32 s14, s98
-; GLOBALNESS0-NEXT:    v_mov_b32_e32 v31, v43
+; GLOBALNESS0-NEXT:    v_mov_b32_e32 v31, v41
 ; GLOBALNESS0-NEXT:    s_swappc_b64 s[30:31], s[66:67]
 ; GLOBALNESS0-NEXT:    v_pk_mov_b32 v[44:45], 0, 0
 ; GLOBALNESS0-NEXT:    s_mov_b64 s[4:5], s[62:63]
 ; GLOBALNESS0-NEXT:    s_mov_b64 s[6:7], s[54:55]
 ; GLOBALNESS0-NEXT:    s_mov_b64 s[8:9], s[64:65]
 ; GLOBALNESS0-NEXT:    s_mov_b64 s[10:11], s[34:35]
-; GLOBALNESS0-NEXT:    s_mov_b32 s12, s100
+; GLOBALNESS0-NEXT:    s_mov_b32 s12, s56
 ; GLOBALNESS0-NEXT:    s_mov_b32 s13, s99
 ; GLOBALNESS0-NEXT:    s_mov_b32 s14, s98
-; GLOBALNESS0-NEXT:    v_mov_b32_e32 v31, v43
+; GLOBALNESS0-NEXT:    v_mov_b32_e32 v31, v41
 ; GLOBALNESS0-NEXT:    global_store_dwordx2 v[44:45], a[32:33], off
 ; GLOBALNESS0-NEXT:    s_swappc_b64 s[30:31], s[66:67]
 ; GLOBALNESS0-NEXT:    s_and_saveexec_b64 s[4:5], s[58:59]
 ; GLOBALNESS0-NEXT:    s_cbranch_execz .LBB1_13
 ; GLOBALNESS0-NEXT:  ; %bb.22: ; %bb62.i
 ; GLOBALNESS0-NEXT:    ; in Loop: Header=BB1_15 Depth=2
-; GLOBALNESS0-NEXT:    v_mov_b32_e32 v41, v40
-; GLOBALNESS0-NEXT:    global_store_dwordx2 v[44:45], v[40:41], off
+; GLOBALNESS0-NEXT:    v_mov_b32_e32 v43, v42
+; GLOBALNESS0-NEXT:    global_store_dwordx2 v[44:45], v[42:43], off
 ; GLOBALNESS0-NEXT:    s_branch .LBB1_13
 ; GLOBALNESS0-NEXT:  .LBB1_23: ; %LeafBlock
 ; GLOBALNESS0-NEXT:    ; in Loop: Header=BB1_4 Depth=1
@@ -699,14 +703,14 @@
 ; GLOBALNESS0-NEXT:    s_branch .LBB1_3
 ; GLOBALNESS0-NEXT:  .LBB1_25: ; %Flow14
 ; GLOBALNESS0-NEXT:    ; in Loop: Header=BB1_4 Depth=1
-; GLOBALNESS0-NEXT:    s_mov_b64 s[4:5], s[36:37]
 ; GLOBALNESS0-NEXT:    s_mov_b32 s36, s93
 ; GLOBALNESS0-NEXT:    s_mov_b32 s37, s93
 ; GLOBALNESS0-NEXT:    s_mov_b32 s38, s93
 ; GLOBALNESS0-NEXT:    s_mov_b32 s39, s93
-; GLOBALNESS0-NEXT:    s_mov_b64 s[6:7], s[40:41]
+; GLOBALNESS0-NEXT:    s_mov_b64 s[4:5], s[40:41]
 ; GLOBALNESS0-NEXT:    s_mov_b32 s40, s93
 ; GLOBALNESS0-NEXT:    s_mov_b32 s41, s93
+; GLOBALNESS0-NEXT:    s_mov_b64 s[6:7], s[42:43]
 ; GLOBALNESS0-NEXT:    s_mov_b32 s42, s93
 ; GLOBALNESS0-NEXT:    s_mov_b32 s43, s93
 ; GLOBALNESS0-NEXT:    s_mov_b32 s44, s93
@@ -735,10 +739,10 @@
 ; GLOBALNESS0-NEXT:    v_pk_mov_b32 v[26:27], s[62:63], s[62:63] op_sel:[0,1]
 ; GLOBALNESS0-NEXT:    v_pk_mov_b32 v[28:29], s[64:65], s[64:65] op_sel:[0,1]
 ; GLOBALNESS0-NEXT:    v_pk_mov_b32 v[30:31], s[66:67], s[66:67] op_sel:[0,1]
-; GLOBALNESS0-NEXT:    s_mov_b64 s[40:41], s[6:7]
-; GLOBALNESS0-NEXT:    s_mov_b64 s[36:37], s[4:5]
+; GLOBALNESS0-NEXT:    s_mov_b64 s[42:43], s[6:7]
+; GLOBALNESS0-NEXT:    s_mov_b64 s[40:41], s[4:5]
 ; GLOBALNESS0-NEXT:    s_mov_b32 s39, s75
-; GLOBALNESS0-NEXT:    s_mov_b64 s[42:43], s[72:73]
+; GLOBALNESS0-NEXT:    s_mov_b64 s[36:37], s[72:73]
 ; GLOBALNESS0-NEXT:  .LBB1_26: ; %Flow15
 ; GLOBALNESS0-NEXT:    ; in Loop: Header=BB1_4 Depth=1
 ; GLOBALNESS0-NEXT:    s_or_b64 exec, exec, s[70:71]
@@ -746,21 +750,21 @@
 ; GLOBALNESS0-NEXT:    s_cbranch_execz .LBB1_2
 ; GLOBALNESS0-NEXT:  ; %bb.27: ; %bb67.i
 ; GLOBALNESS0-NEXT:    ; in Loop: Header=BB1_4 Depth=1
-; GLOBALNESS0-NEXT:    v_readlane_b32 s6, v42, 2
-; GLOBALNESS0-NEXT:    v_readlane_b32 s7, v42, 3
+; GLOBALNESS0-NEXT:    v_readlane_b32 s6, v40, 2
+; GLOBALNESS0-NEXT:    v_readlane_b32 s7, v40, 3
 ; GLOBALNESS0-NEXT:    s_andn2_b64 vcc, exec, s[6:7]
 ; GLOBALNESS0-NEXT:    s_cbranch_vccnz .LBB1_1
 ; GLOBALNESS0-NEXT:  ; %bb.28: ; %bb69.i
 ; GLOBALNESS0-NEXT:    ; in Loop: Header=BB1_4 Depth=1
-; GLOBALNESS0-NEXT:    v_mov_b32_e32 v41, v40
+; GLOBALNESS0-NEXT:    v_mov_b32_e32 v43, v42
 ; GLOBALNESS0-NEXT:    v_pk_mov_b32 v[32:33], 0, 0
-; GLOBALNESS0-NEXT:    global_store_dwordx2 v[32:33], v[40:41], off
+; GLOBALNESS0-NEXT:    global_store_dwordx2 v[32:33], v[42:43], off
 ; GLOBALNESS0-NEXT:    s_branch .LBB1_1
 ; GLOBALNESS0-NEXT:  .LBB1_29: ; %bb73.i
 ; GLOBALNESS0-NEXT:    ; in Loop: Header=BB1_4 Depth=1
-; GLOBALNESS0-NEXT:    v_mov_b32_e32 v41, v40
+; GLOBALNESS0-NEXT:    v_mov_b32_e32 v43, v42
 ; GLOBALNESS0-NEXT:    v_pk_mov_b32 v[32:33], 0, 0
-; GLOBALNESS0-NEXT:    global_store_dwordx2 v[32:33], v[40:41], off
+; GLOBALNESS0-NEXT:    global_store_dwordx2 v[32:33], v[42:43], off
 ; GLOBALNESS0-NEXT:    s_branch .LBB1_2
 ; GLOBALNESS0-NEXT:  .LBB1_30: ; %loop.exit.guard
 ; GLOBALNESS0-NEXT:    s_andn2_b64 vcc, exec, s[4:5]
@@ -772,10 +776,10 @@
 ; GLOBALNESS0-NEXT:    s_mov_b64 s[4:5], s[62:63]
 ; GLOBALNESS0-NEXT:    s_mov_b64 s[6:7], s[54:55]
 ; GLOBALNESS0-NEXT:    s_mov_b64 s[10:11], s[34:35]
-; GLOBALNESS0-NEXT:    s_mov_b32 s12, s100
+; GLOBALNESS0-NEXT:    s_mov_b32 s12, s56
 ; GLOBALNESS0-NEXT:    s_mov_b32 s13, s99
 ; GLOBALNESS0-NEXT:    s_mov_b32 s14, s98
-; GLOBALNESS0-NEXT:    v_mov_b32_e32 v31, v43
+; GLOBALNESS0-NEXT:    v_mov_b32_e32 v31, v41
 ; GLOBALNESS0-NEXT:    s_getpc_b64 s[16:17]
 ; GLOBALNESS0-NEXT:    s_add_u32 s16, s16, widget@rel32@lo+4
 ; GLOBALNESS0-NEXT:    s_addc_u32 s17, s17, widget@rel32@hi+12
@@ -790,10 +794,10 @@
 ; GLOBALNESS0-NEXT:    s_mov_b64 s[4:5], s[62:63]
 ; GLOBALNESS0-NEXT:    s_mov_b64 s[6:7], s[54:55]
 ; GLOBALNESS0-NEXT:    s_mov_b64 s[10:11], s[34:35]
-; GLOBALNESS0-NEXT:    s_mov_b32 s12, s100
+; GLOBALNESS0-NEXT:    s_mov_b32 s12, s56
 ; GLOBALNESS0-NEXT:    s_mov_b32 s13, s99
 ; GLOBALNESS0-NEXT:    s_mov_b32 s14, s98
-; GLOBALNESS0-NEXT:    v_mov_b32_e32 v31, v43
+; GLOBALNESS0-NEXT:    v_mov_b32_e32 v31, v41
 ; GLOBALNESS0-NEXT:    s_getpc_b64 s[16:17]
 ; GLOBALNESS0-NEXT:    s_add_u32 s16, s16, widget@rel32@lo+4
 ; GLOBALNESS0-NEXT:    s_addc_u32 s17, s17, widget@rel32@hi+12
diff --git a/llvm/test/CodeGen/AMDGPU/unstructured-cfg-def-use-issue.ll b/llvm/test/CodeGen/AMDGPU/unstructured-cfg-def-use-issue.ll
--- a/llvm/test/CodeGen/AMDGPU/unstructured-cfg-def-use-issue.ll
+++ b/llvm/test/CodeGen/AMDGPU/unstructured-cfg-def-use-issue.ll
@@ -14,6 +14,7 @@
 ; GCN-NEXT:    s_mov_b64 exec, s[18:19]
 ; GCN-NEXT:    v_writelane_b32 v41, s16, 0
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
+; GCN-NEXT:    ; implicit-def: $vgpr40
 ; GCN-NEXT:    v_writelane_b32 v40, s30, 0
 ; GCN-NEXT:    v_writelane_b32 v40, s31, 1
 ; GCN-NEXT:    v_mov_b32_e32 v0, 0
@@ -189,36 +190,38 @@
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GCN-NEXT:    s_mov_b32 s16, s33
 ; GCN-NEXT:    s_mov_b32 s33, s32
-; GCN-NEXT:    s_or_saveexec_b64 s[18:19], -1
-; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:20 ; 4-byte Folded Spill
-; GCN-NEXT:    buffer_store_dword v46, off, s[0:3], s33 offset:24 ; 4-byte Folded Spill
+; GCN-NEXT:    s_xor_saveexec_b64 s[18:19], -1
+; GCN-NEXT:    buffer_store_dword v0, off, s[0:3], s33 offset:20 ; 4-byte Folded Spill
+; GCN-NEXT:    s_mov_b64 exec, -1
+; GCN-NEXT:    buffer_store_dword v45, off, s[0:3], s33 offset:24 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[18:19]
-; GCN-NEXT:    v_writelane_b32 v46, s16, 0
+; GCN-NEXT:    v_writelane_b32 v45, s16, 0
 ; GCN-NEXT:    s_addk_i32 s32, 0x800
-; GCN-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:16 ; 4-byte Folded Spill
-; GCN-NEXT:    buffer_store_dword v42, off, s[0:3], s33 offset:12 ; 4-byte Folded Spill
-; GCN-NEXT:    buffer_store_dword v43, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
-; GCN-NEXT:    buffer_store_dword v44, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
-; GCN-NEXT:    buffer_store_dword v45, off, s[0:3], s33 ; 4-byte Folded Spill
-; GCN-NEXT:    v_writelane_b32 v40, s30, 0
-; GCN-NEXT:    v_writelane_b32 v40, s31, 1
-; GCN-NEXT:    v_writelane_b32 v40, s34, 2
-; GCN-NEXT:    v_writelane_b32 v40, s35, 3
-; GCN-NEXT:    v_writelane_b32 v40, s36, 4
-; GCN-NEXT:    v_writelane_b32 v40, s37, 5
-; GCN-NEXT:    v_writelane_b32 v40, s38, 6
-; GCN-NEXT:    v_writelane_b32 v40, s39, 7
-; GCN-NEXT:    v_writelane_b32 v40, s40, 8
-; GCN-NEXT:    v_writelane_b32 v40, s41, 9
-; GCN-NEXT:    v_writelane_b32 v40, s42, 10
-; GCN-NEXT:    v_writelane_b32 v40, s43, 11
-; GCN-NEXT:    v_writelane_b32 v40, s44, 12
-; GCN-NEXT:    v_writelane_b32 v40, s45, 13
-; GCN-NEXT:    v_writelane_b32 v40, s46, 14
-; GCN-NEXT:    v_writelane_b32 v40, s47, 15
-; GCN-NEXT:    v_writelane_b32 v40, s48, 16
-; GCN-NEXT:    v_writelane_b32 v40, s49, 17
-; GCN-NEXT:    v_mov_b32_e32 v41, v31
+; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:16 ; 4-byte Folded Spill
+; GCN-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:12 ; 4-byte Folded Spill
+; GCN-NEXT:    buffer_store_dword v42, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
+; GCN-NEXT:    buffer_store_dword v43, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
+; GCN-NEXT:    buffer_store_dword v44, off, s[0:3], s33 ; 4-byte Folded Spill
+; GCN-NEXT:    ; implicit-def: $vgpr0
+; GCN-NEXT:    v_writelane_b32 v0, s30, 0
+; GCN-NEXT:    v_writelane_b32 v0, s31, 1
+; GCN-NEXT:    v_writelane_b32 v0, s34, 2
+; GCN-NEXT:    v_writelane_b32 v0, s35, 3
+; GCN-NEXT:    v_writelane_b32 v0, s36, 4
+; GCN-NEXT:    v_writelane_b32 v0, s37, 5
+; GCN-NEXT:    v_writelane_b32 v0, s38, 6
+; GCN-NEXT:    v_writelane_b32 v0, s39, 7
+; GCN-NEXT:    v_writelane_b32 v0, s40, 8
+; GCN-NEXT:    v_writelane_b32 v0, s41, 9
+; GCN-NEXT:    v_writelane_b32 v0, s42, 10
+; GCN-NEXT:    v_writelane_b32 v0, s43, 11
+; GCN-NEXT:    v_writelane_b32 v0, s44, 12
+; GCN-NEXT:    v_writelane_b32 v0, s45, 13
+; GCN-NEXT:    v_writelane_b32 v0, s46, 14
+; GCN-NEXT:    v_writelane_b32 v0, s47, 15
+; GCN-NEXT:    v_writelane_b32 v0, s48, 16
+; GCN-NEXT:    v_writelane_b32 v0, s49, 17
+; GCN-NEXT:    v_mov_b32_e32 v40, v31
 ; GCN-NEXT:    s_mov_b32 s44, s15
 ; GCN-NEXT:    s_mov_b32 s45, s14
 ; GCN-NEXT:    s_mov_b32 s46, s13
@@ -230,24 +233,24 @@
 ; GCN-NEXT:    s_mov_b64 s[4:5], 0
 ; GCN-NEXT:    v_mov_b32_e32 v0, 0
 ; GCN-NEXT:    v_mov_b32_e32 v1, 0
-; GCN-NEXT:    v_and_b32_e32 v2, 0x3ff, v41
-; GCN-NEXT:    v_mov_b32_e32 v43, 0
-; GCN-NEXT:    flat_load_dword v44, v[0:1]
-; GCN-NEXT:    v_mov_b32_e32 v45, 0x7fc00000
+; GCN-NEXT:    v_and_b32_e32 v2, 0x3ff, v40
+; GCN-NEXT:    v_mov_b32_e32 v42, 0
+; GCN-NEXT:    flat_load_dword v43, v[0:1]
+; GCN-NEXT:    v_mov_b32_e32 v44, 0x7fc00000
 ; GCN-NEXT:    s_getpc_b64 s[48:49]
 ; GCN-NEXT:    s_add_u32 s48, s48, spam@rel32@lo+4
 ; GCN-NEXT:    s_addc_u32 s49, s49, spam@rel32@hi+12
-; GCN-NEXT:    v_lshlrev_b32_e32 v42, 2, v2
+; GCN-NEXT:    v_lshlrev_b32_e32 v41, 2, v2
 ; GCN-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GCN-NEXT:    v_cmp_eq_f32_e64 s[42:43], 0, v44
+; GCN-NEXT:    v_cmp_eq_f32_e64 s[42:43], 0, v43
 ; GCN-NEXT:    s_branch .LBB1_3
 ; GCN-NEXT:  .LBB1_1: ; %bb10
 ; GCN-NEXT:    ; in Loop: Header=BB1_3 Depth=1
 ; GCN-NEXT:    s_or_b64 exec, exec, s[6:7]
-; GCN-NEXT:    buffer_store_dword v45, off, s[0:3], 0
+; GCN-NEXT:    buffer_store_dword v44, off, s[0:3], 0
 ; GCN-NEXT:  .LBB1_2: ; %bb18
 ; GCN-NEXT:    ; in Loop: Header=BB1_3 Depth=1
-; GCN-NEXT:    buffer_store_dword v45, off, s[0:3], 0
+; GCN-NEXT:    buffer_store_dword v44, off, s[0:3], 0
 ; GCN-NEXT:    s_mov_b64 s[4:5], 0
 ; GCN-NEXT:  .LBB1_3: ; %bb2
 ; GCN-NEXT:    ; =>This Loop Header: Depth=1
@@ -256,8 +259,8 @@
 ; GCN-NEXT:  .LBB1_4: ; %bb2
 ; GCN-NEXT:    ; Parent Loop BB1_3 Depth=1
 ; GCN-NEXT:    ; => This Inner Loop Header: Depth=2
-; GCN-NEXT:    flat_load_dword v0, v[42:43]
-; GCN-NEXT:    buffer_store_dword v43, off, s[0:3], 0
+; GCN-NEXT:    flat_load_dword v0, v[41:42]
+; GCN-NEXT:    buffer_store_dword v42, off, s[0:3], 0
 ; GCN-NEXT:    s_waitcnt vmcnt(1)
 ; GCN-NEXT:    v_cmp_gt_i32_e32 vcc, 3, v0
 ; GCN-NEXT:    s_and_saveexec_b64 s[8:9], vcc
@@ -289,7 +292,7 @@
 ; GCN-NEXT:    s_mov_b32 s13, s46
 ; GCN-NEXT:    s_mov_b32 s14, s45
 ; GCN-NEXT:    s_mov_b32 s15, s44
-; GCN-NEXT:    v_mov_b32_e32 v31, v41
+; GCN-NEXT:    v_mov_b32_e32 v31, v40
 ; GCN-NEXT:    s_swappc_b64 s[30:31], s[48:49]
 ; GCN-NEXT:    v_cmp_eq_f32_e32 vcc, 0, v0
 ; GCN-NEXT:    s_mov_b64 s[4:5], 0
@@ -304,10 +307,10 @@
 ; GCN-NEXT:  ; %bb.9: ; %bb16
 ; GCN-NEXT:    ; in Loop: Header=BB1_3 Depth=1
 ; GCN-NEXT:    s_or_b64 exec, exec, s[4:5]
-; GCN-NEXT:    buffer_store_dword v45, off, s[0:3], 0
+; GCN-NEXT:    buffer_store_dword v44, off, s[0:3], 0
 ; GCN-NEXT:  .LBB1_10: ; %bb17
 ; GCN-NEXT:    ; in Loop: Header=BB1_3 Depth=1
-; GCN-NEXT:    buffer_store_dword v44, off, s[0:3], 0
+; GCN-NEXT:    buffer_store_dword v43, off, s[0:3], 0
 ; GCN-NEXT:    s_branch .LBB1_2
 bb:
   %tmp = load float, float* null, align 16
diff --git a/llvm/test/CodeGen/AMDGPU/vgpr-tuple-allocation.ll b/llvm/test/CodeGen/AMDGPU/vgpr-tuple-allocation.ll
--- a/llvm/test/CodeGen/AMDGPU/vgpr-tuple-allocation.ll
+++ b/llvm/test/CodeGen/AMDGPU/vgpr-tuple-allocation.ll
@@ -14,7 +14,7 @@
 ; GFX9-NEXT:    s_mov_b32 s4, s33
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_or_saveexec_b64 s[6:7], -1
-; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:16 ; 4-byte Folded Spill
+; GFX9-NEXT:    buffer_store_dword v44, off, s[0:3], s33 offset:16 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v45, off, s[0:3], s33 offset:20 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[6:7]
 ; GFX9-NEXT:    v_mov_b32_e32 v36, v16
@@ -22,10 +22,10 @@
 ; GFX9-NEXT:    v_mov_b32_e32 v34, v14
 ; GFX9-NEXT:    v_mov_b32_e32 v33, v13
 ; GFX9-NEXT:    v_mov_b32_e32 v32, v12
-; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:12 ; 4-byte Folded Spill
-; GFX9-NEXT:    buffer_store_dword v42, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
-; GFX9-NEXT:    buffer_store_dword v43, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
-; GFX9-NEXT:    buffer_store_dword v44, off, s[0:3], s33 ; 4-byte Folded Spill
+; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:12 ; 4-byte Folded Spill
+; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
+; GFX9-NEXT:    buffer_store_dword v42, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
+; GFX9-NEXT:    buffer_store_dword v43, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ;;#ASMEND
 ; GFX9-NEXT:    ;;#ASMSTART
@@ -34,30 +34,31 @@
 ; GFX9-NEXT:    ;;#ASMEND
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ;;#ASMEND
-; GFX9-NEXT:    image_gather4_c_b_cl v[41:44], v[32:36], s[4:11], s[4:7] dmask:0x1
+; GFX9-NEXT:    image_gather4_c_b_cl v[40:43], v[32:36], s[4:11], s[4:7] dmask:0x1
 ; GFX9-NEXT:    s_addk_i32 s32, 0x800
 ; GFX9-NEXT:    v_writelane_b32 v45, s4, 0
 ; GFX9-NEXT:    s_getpc_b64 s[4:5]
 ; GFX9-NEXT:    s_add_u32 s4, s4, extern_func@gotpcrel32@lo+4
 ; GFX9-NEXT:    s_addc_u32 s5, s5, extern_func@gotpcrel32@hi+12
 ; GFX9-NEXT:    s_load_dwordx2 s[4:5], s[4:5], 0x0
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
-; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
+; GFX9-NEXT:    ; implicit-def: $vgpr44
+; GFX9-NEXT:    v_writelane_b32 v44, s30, 0
+; GFX9-NEXT:    v_writelane_b32 v44, s31, 1
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX9-NEXT:    s_swappc_b64 s[30:31], s[4:5]
-; GFX9-NEXT:    v_mov_b32_e32 v0, v41
-; GFX9-NEXT:    v_mov_b32_e32 v1, v42
-; GFX9-NEXT:    v_mov_b32_e32 v2, v43
-; GFX9-NEXT:    v_mov_b32_e32 v3, v44
-; GFX9-NEXT:    buffer_load_dword v44, off, s[0:3], s33 ; 4-byte Folded Reload
-; GFX9-NEXT:    buffer_load_dword v43, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
-; GFX9-NEXT:    buffer_load_dword v42, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload
-; GFX9-NEXT:    buffer_load_dword v41, off, s[0:3], s33 offset:12 ; 4-byte Folded Reload
-; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
-; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
+; GFX9-NEXT:    v_mov_b32_e32 v0, v40
+; GFX9-NEXT:    v_mov_b32_e32 v1, v41
+; GFX9-NEXT:    v_mov_b32_e32 v2, v42
+; GFX9-NEXT:    v_mov_b32_e32 v3, v43
+; GFX9-NEXT:    buffer_load_dword v43, off, s[0:3], s33 ; 4-byte Folded Reload
+; GFX9-NEXT:    buffer_load_dword v42, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
+; GFX9-NEXT:    buffer_load_dword v41, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload
+; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s33 offset:12 ; 4-byte Folded Reload
+; GFX9-NEXT:    v_readlane_b32 s31, v44, 1
+; GFX9-NEXT:    v_readlane_b32 s30, v44, 0
 ; GFX9-NEXT:    v_readlane_b32 s4, v45, 0
 ; GFX9-NEXT:    s_or_saveexec_b64 s[6:7], -1
-; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s33 offset:16 ; 4-byte Folded Reload
+; GFX9-NEXT:    buffer_load_dword v44, off, s[0:3], s33 offset:16 ; 4-byte Folded Reload
 ; GFX9-NEXT:    buffer_load_dword v45, off, s[0:3], s33 offset:20 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[6:7]
 ; GFX9-NEXT:    s_addk_i32 s32, 0xf800
@@ -72,7 +73,7 @@
 ; GFX10-NEXT:    s_mov_b32 s4, s33
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_or_saveexec_b32 s5, -1
-; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:16 ; 4-byte Folded Spill
+; GFX10-NEXT:    buffer_store_dword v44, off, s[0:3], s33 offset:16 ; 4-byte Folded Spill
 ; GFX10-NEXT:    buffer_store_dword v45, off, s[0:3], s33 offset:20 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s5
@@ -81,10 +82,10 @@
 ; GFX10-NEXT:    v_mov_b32_e32 v34, v14
 ; GFX10-NEXT:    v_mov_b32_e32 v33, v13
 ; GFX10-NEXT:    v_mov_b32_e32 v32, v12
-; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:12 ; 4-byte Folded Spill
-; GFX10-NEXT:    buffer_store_dword v42, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
-; GFX10-NEXT:    buffer_store_dword v43, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
-; GFX10-NEXT:    buffer_store_dword v44, off, s[0:3], s33 ; 4-byte Folded Spill
+; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:12 ; 4-byte Folded Spill
+; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
+; GFX10-NEXT:    buffer_store_dword v42, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
+; GFX10-NEXT:    buffer_store_dword v43, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ;;#ASMEND
 ; GFX10-NEXT:    ;;#ASMSTART
@@ -93,32 +94,33 @@
 ; GFX10-NEXT:    ;;#ASMEND
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ;;#ASMEND
-; GFX10-NEXT:    image_gather4_c_b_cl v[41:44], v[32:36], s[4:11], s[4:7] dmask:0x1 dim:SQ_RSRC_IMG_2D
+; GFX10-NEXT:    image_gather4_c_b_cl v[40:43], v[32:36], s[4:11], s[4:7] dmask:0x1 dim:SQ_RSRC_IMG_2D
 ; GFX10-NEXT:    s_addk_i32 s32, 0x400
 ; GFX10-NEXT:    v_writelane_b32 v45, s4, 0
 ; GFX10-NEXT:    s_getpc_b64 s[4:5]
 ; GFX10-NEXT:    s_add_u32 s4, s4, extern_func@gotpcrel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s5, s5, extern_func@gotpcrel32@hi+12
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr44
 ; GFX10-NEXT:    s_load_dwordx2 s[4:5], s[4:5], 0x0
-; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
+; GFX10-NEXT:    v_writelane_b32 v44, s30, 0
+; GFX10-NEXT:    v_writelane_b32 v44, s31, 1
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[4:5]
-; GFX10-NEXT:    v_mov_b32_e32 v0, v41
-; GFX10-NEXT:    v_mov_b32_e32 v1, v42
-; GFX10-NEXT:    v_mov_b32_e32 v2, v43
-; GFX10-NEXT:    v_mov_b32_e32 v3, v44
+; GFX10-NEXT:    v_mov_b32_e32 v0, v40
+; GFX10-NEXT:    v_mov_b32_e32 v1, v41
+; GFX10-NEXT:    v_mov_b32_e32 v2, v42
+; GFX10-NEXT:    v_mov_b32_e32 v3, v43
 ; GFX10-NEXT:    s_clause 0x3
-; GFX10-NEXT:    buffer_load_dword v44, off, s[0:3], s33
-; GFX10-NEXT:    buffer_load_dword v43, off, s[0:3], s33 offset:4
-; GFX10-NEXT:    buffer_load_dword v42, off, s[0:3], s33 offset:8
-; GFX10-NEXT:    buffer_load_dword v41, off, s[0:3], s33 offset:12
-; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
-; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
+; GFX10-NEXT:    buffer_load_dword v43, off, s[0:3], s33
+; GFX10-NEXT:    buffer_load_dword v42, off, s[0:3], s33 offset:4
+; GFX10-NEXT:    buffer_load_dword v41, off, s[0:3], s33 offset:8
+; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s33 offset:12
+; GFX10-NEXT:    v_readlane_b32 s31, v44, 1
+; GFX10-NEXT:    v_readlane_b32 s30, v44, 0
 ; GFX10-NEXT:    v_readlane_b32 s4, v45, 0
 ; GFX10-NEXT:    s_or_saveexec_b32 s5, -1
 ; GFX10-NEXT:    s_clause 0x1
-; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s33 offset:16
+; GFX10-NEXT:    buffer_load_dword v44, off, s[0:3], s33 offset:16
 ; GFX10-NEXT:    buffer_load_dword v45, off, s[0:3], s33 offset:20
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s5
@@ -135,17 +137,17 @@
 ; GFX11-NEXT:    s_mov_b32 s33, s32
 ; GFX11-NEXT:    s_or_saveexec_b32 s1, -1
 ; GFX11-NEXT:    s_clause 0x1
-; GFX11-NEXT:    scratch_store_b32 off, v40, s33 offset:16
+; GFX11-NEXT:    scratch_store_b32 off, v44, s33 offset:16
 ; GFX11-NEXT:    scratch_store_b32 off, v45, s33 offset:20
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX11-NEXT:    v_dual_mov_b32 v36, v16 :: v_dual_mov_b32 v35, v15
 ; GFX11-NEXT:    v_dual_mov_b32 v34, v14 :: v_dual_mov_b32 v33, v13
 ; GFX11-NEXT:    v_mov_b32_e32 v32, v12
 ; GFX11-NEXT:    s_clause 0x3
-; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:12
-; GFX11-NEXT:    scratch_store_b32 off, v42, s33 offset:8
-; GFX11-NEXT:    scratch_store_b32 off, v43, s33 offset:4
-; GFX11-NEXT:    scratch_store_b32 off, v44, s33
+; GFX11-NEXT:    scratch_store_b32 off, v40, s33 offset:12
+; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:8
+; GFX11-NEXT:    scratch_store_b32 off, v42, s33 offset:4
+; GFX11-NEXT:    scratch_store_b32 off, v43, s33
 ; GFX11-NEXT:    ;;#ASMSTART
 ; GFX11-NEXT:    ;;#ASMEND
 ; GFX11-NEXT:    ;;#ASMSTART
@@ -154,30 +156,31 @@
 ; GFX11-NEXT:    ;;#ASMEND
 ; GFX11-NEXT:    ;;#ASMSTART
 ; GFX11-NEXT:    ;;#ASMEND
-; GFX11-NEXT:    image_gather4_c_b_cl v[41:44], v[32:36], s[0:7], s[0:3] dmask:0x1 dim:SQ_RSRC_IMG_2D
+; GFX11-NEXT:    image_gather4_c_b_cl v[40:43], v[32:36], s[0:7], s[0:3] dmask:0x1 dim:SQ_RSRC_IMG_2D
 ; GFX11-NEXT:    s_add_i32 s32, s32, 32
 ; GFX11-NEXT:    v_writelane_b32 v45, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, extern_func@gotpcrel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, extern_func@gotpcrel32@hi+12
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr44
 ; GFX11-NEXT:    s_load_b64 s[0:1], s[0:1], 0x0
-; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
+; GFX11-NEXT:    v_writelane_b32 v44, s30, 0
+; GFX11-NEXT:    v_writelane_b32 v44, s31, 1
 ; GFX11-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
-; GFX11-NEXT:    v_dual_mov_b32 v0, v41 :: v_dual_mov_b32 v1, v42
-; GFX11-NEXT:    v_dual_mov_b32 v2, v43 :: v_dual_mov_b32 v3, v44
+; GFX11-NEXT:    v_dual_mov_b32 v0, v40 :: v_dual_mov_b32 v1, v41
+; GFX11-NEXT:    v_dual_mov_b32 v2, v42 :: v_dual_mov_b32 v3, v43
 ; GFX11-NEXT:    s_clause 0x3
-; GFX11-NEXT:    scratch_load_b32 v44, off, s33
-; GFX11-NEXT:    scratch_load_b32 v43, off, s33 offset:4
-; GFX11-NEXT:    scratch_load_b32 v42, off, s33 offset:8
-; GFX11-NEXT:    scratch_load_b32 v41, off, s33 offset:12
-; GFX11-NEXT:    v_readlane_b32 s31, v40, 1
-; GFX11-NEXT:    v_readlane_b32 s30, v40, 0
+; GFX11-NEXT:    scratch_load_b32 v43, off, s33
+; GFX11-NEXT:    scratch_load_b32 v42, off, s33 offset:4
+; GFX11-NEXT:    scratch_load_b32 v41, off, s33 offset:8
+; GFX11-NEXT:    scratch_load_b32 v40, off, s33 offset:12
+; GFX11-NEXT:    v_readlane_b32 s31, v44, 1
+; GFX11-NEXT:    v_readlane_b32 s30, v44, 0
 ; GFX11-NEXT:    v_readlane_b32 s0, v45, 0
 ; GFX11-NEXT:    s_or_saveexec_b32 s1, -1
 ; GFX11-NEXT:    s_clause 0x1
-; GFX11-NEXT:    scratch_load_b32 v40, off, s33 offset:16
+; GFX11-NEXT:    scratch_load_b32 v44, off, s33 offset:16
 ; GFX11-NEXT:    scratch_load_b32 v45, off, s33 offset:20
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX11-NEXT:    s_addk_i32 s32, 0xffe0
@@ -215,44 +218,45 @@
 ; GFX9-NEXT:    s_mov_b32 s4, s33
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_or_saveexec_b64 s[6:7], -1
-; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:20 ; 4-byte Folded Spill
+; GFX9-NEXT:    buffer_store_dword v45, off, s[0:3], s33 offset:20 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v46, off, s[0:3], s33 offset:24 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[6:7]
-; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:16 ; 4-byte Folded Spill
-; GFX9-NEXT:    buffer_store_dword v42, off, s[0:3], s33 offset:12 ; 4-byte Folded Spill
-; GFX9-NEXT:    buffer_store_dword v43, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
-; GFX9-NEXT:    buffer_store_dword v44, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
-; GFX9-NEXT:    buffer_store_dword v45, off, s[0:3], s33 ; 4-byte Folded Spill
-; GFX9-NEXT:    v_mov_b32_e32 v45, v16
-; GFX9-NEXT:    v_mov_b32_e32 v44, v15
-; GFX9-NEXT:    v_mov_b32_e32 v43, v14
-; GFX9-NEXT:    v_mov_b32_e32 v42, v13
-; GFX9-NEXT:    v_mov_b32_e32 v41, v12
-; GFX9-NEXT:    image_gather4_c_b_cl v[0:3], v[41:45], s[4:11], s[4:7] dmask:0x1
+; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:16 ; 4-byte Folded Spill
+; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:12 ; 4-byte Folded Spill
+; GFX9-NEXT:    buffer_store_dword v42, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
+; GFX9-NEXT:    buffer_store_dword v43, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
+; GFX9-NEXT:    buffer_store_dword v44, off, s[0:3], s33 ; 4-byte Folded Spill
+; GFX9-NEXT:    v_mov_b32_e32 v44, v16
+; GFX9-NEXT:    v_mov_b32_e32 v43, v15
+; GFX9-NEXT:    v_mov_b32_e32 v42, v14
+; GFX9-NEXT:    v_mov_b32_e32 v41, v13
+; GFX9-NEXT:    v_mov_b32_e32 v40, v12
+; GFX9-NEXT:    image_gather4_c_b_cl v[0:3], v[40:44], s[4:11], s[4:7] dmask:0x1
 ; GFX9-NEXT:    s_addk_i32 s32, 0x800
 ; GFX9-NEXT:    v_writelane_b32 v46, s4, 0
 ; GFX9-NEXT:    s_getpc_b64 s[4:5]
 ; GFX9-NEXT:    s_add_u32 s4, s4, extern_func@gotpcrel32@lo+4
 ; GFX9-NEXT:    s_addc_u32 s5, s5, extern_func@gotpcrel32@hi+12
 ; GFX9-NEXT:    s_load_dwordx2 s[4:5], s[4:5], 0x0
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
-; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
+; GFX9-NEXT:    ; implicit-def: $vgpr45
+; GFX9-NEXT:    v_writelane_b32 v45, s30, 0
+; GFX9-NEXT:    v_writelane_b32 v45, s31, 1
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    global_store_dwordx4 v[0:1], v[0:3], off
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX9-NEXT:    s_swappc_b64 s[30:31], s[4:5]
-; GFX9-NEXT:    image_gather4_c_b_cl v[0:3], v[41:45], s[4:11], s[4:7] dmask:0x1
+; GFX9-NEXT:    image_gather4_c_b_cl v[0:3], v[40:44], s[4:11], s[4:7] dmask:0x1
 ; GFX9-NEXT:    s_nop 0
-; GFX9-NEXT:    buffer_load_dword v45, off, s[0:3], s33 ; 4-byte Folded Reload
-; GFX9-NEXT:    buffer_load_dword v44, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
-; GFX9-NEXT:    buffer_load_dword v43, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload
-; GFX9-NEXT:    buffer_load_dword v42, off, s[0:3], s33 offset:12 ; 4-byte Folded Reload
-; GFX9-NEXT:    buffer_load_dword v41, off, s[0:3], s33 offset:16 ; 4-byte Folded Reload
-; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
-; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
+; GFX9-NEXT:    buffer_load_dword v44, off, s[0:3], s33 ; 4-byte Folded Reload
+; GFX9-NEXT:    buffer_load_dword v43, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
+; GFX9-NEXT:    buffer_load_dword v42, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload
+; GFX9-NEXT:    buffer_load_dword v41, off, s[0:3], s33 offset:12 ; 4-byte Folded Reload
+; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s33 offset:16 ; 4-byte Folded Reload
+; GFX9-NEXT:    v_readlane_b32 s31, v45, 1
+; GFX9-NEXT:    v_readlane_b32 s30, v45, 0
 ; GFX9-NEXT:    v_readlane_b32 s4, v46, 0
 ; GFX9-NEXT:    s_or_saveexec_b64 s[6:7], -1
-; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s33 offset:20 ; 4-byte Folded Reload
+; GFX9-NEXT:    buffer_load_dword v45, off, s[0:3], s33 offset:20 ; 4-byte Folded Reload
 ; GFX9-NEXT:    buffer_load_dword v46, off, s[0:3], s33 offset:24 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[6:7]
 ; GFX9-NEXT:    s_addk_i32 s32, 0xf800
@@ -267,46 +271,47 @@
 ; GFX10-NEXT:    s_mov_b32 s4, s33
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_or_saveexec_b32 s5, -1
-; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:20 ; 4-byte Folded Spill
+; GFX10-NEXT:    buffer_store_dword v45, off, s[0:3], s33 offset:20 ; 4-byte Folded Spill
 ; GFX10-NEXT:    buffer_store_dword v46, off, s[0:3], s33 offset:24 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s5
-; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:16 ; 4-byte Folded Spill
-; GFX10-NEXT:    buffer_store_dword v42, off, s[0:3], s33 offset:12 ; 4-byte Folded Spill
-; GFX10-NEXT:    buffer_store_dword v43, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
-; GFX10-NEXT:    buffer_store_dword v44, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
-; GFX10-NEXT:    buffer_store_dword v45, off, s[0:3], s33 ; 4-byte Folded Spill
+; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s33 offset:16 ; 4-byte Folded Spill
+; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:12 ; 4-byte Folded Spill
+; GFX10-NEXT:    buffer_store_dword v42, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
+; GFX10-NEXT:    buffer_store_dword v43, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
+; GFX10-NEXT:    buffer_store_dword v44, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX10-NEXT:    image_gather4_c_b_cl v[0:3], v[12:16], s[4:11], s[4:7] dmask:0x1 dim:SQ_RSRC_IMG_2D
 ; GFX10-NEXT:    s_addk_i32 s32, 0x400
 ; GFX10-NEXT:    v_writelane_b32 v46, s4, 0
 ; GFX10-NEXT:    s_getpc_b64 s[4:5]
 ; GFX10-NEXT:    s_add_u32 s4, s4, extern_func@gotpcrel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s5, s5, extern_func@gotpcrel32@hi+12
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr45
+; GFX10-NEXT:    v_mov_b32_e32 v40, v16
 ; GFX10-NEXT:    s_load_dwordx2 s[4:5], s[4:5], 0x0
-; GFX10-NEXT:    v_mov_b32_e32 v41, v16
-; GFX10-NEXT:    v_mov_b32_e32 v42, v15
-; GFX10-NEXT:    v_mov_b32_e32 v43, v14
-; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    v_mov_b32_e32 v44, v13
-; GFX10-NEXT:    v_mov_b32_e32 v45, v12
+; GFX10-NEXT:    v_writelane_b32 v45, s30, 0
+; GFX10-NEXT:    v_mov_b32_e32 v41, v15
+; GFX10-NEXT:    v_mov_b32_e32 v42, v14
+; GFX10-NEXT:    v_mov_b32_e32 v43, v13
+; GFX10-NEXT:    v_mov_b32_e32 v44, v12
+; GFX10-NEXT:    v_writelane_b32 v45, s31, 1
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    global_store_dwordx4 v[0:1], v[0:3], off
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[4:5]
-; GFX10-NEXT:    image_gather4_c_b_cl v[0:3], [v45, v44, v43, v42, v41], s[4:11], s[4:7] dmask:0x1 dim:SQ_RSRC_IMG_2D
+; GFX10-NEXT:    image_gather4_c_b_cl v[0:3], [v44, v43, v42, v41, v40], s[4:11], s[4:7] dmask:0x1 dim:SQ_RSRC_IMG_2D
 ; GFX10-NEXT:    s_clause 0x4
-; GFX10-NEXT:    buffer_load_dword v45, off, s[0:3], s33
-; GFX10-NEXT:    buffer_load_dword v44, off, s[0:3], s33 offset:4
-; GFX10-NEXT:    buffer_load_dword v43, off, s[0:3], s33 offset:8
-; GFX10-NEXT:    buffer_load_dword v42, off, s[0:3], s33 offset:12
-; GFX10-NEXT:    buffer_load_dword v41, off, s[0:3], s33 offset:16
-; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
-; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
+; GFX10-NEXT:    buffer_load_dword v44, off, s[0:3], s33
+; GFX10-NEXT:    buffer_load_dword v43, off, s[0:3], s33 offset:4
+; GFX10-NEXT:    buffer_load_dword v42, off, s[0:3], s33 offset:8
+; GFX10-NEXT:    buffer_load_dword v41, off, s[0:3], s33 offset:12
+; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s33 offset:16
+; GFX10-NEXT:    v_readlane_b32 s31, v45, 1
+; GFX10-NEXT:    v_readlane_b32 s30, v45, 0
 ; GFX10-NEXT:    v_readlane_b32 s4, v46, 0
 ; GFX10-NEXT:    s_or_saveexec_b32 s5, -1
 ; GFX10-NEXT:    s_clause 0x1
-; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s33 offset:20
+; GFX10-NEXT:    buffer_load_dword v45, off, s[0:3], s33 offset:20
 ; GFX10-NEXT:    buffer_load_dword v46, off, s[0:3], s33 offset:24
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s5
@@ -323,44 +328,45 @@
 ; GFX11-NEXT:    s_mov_b32 s33, s32
 ; GFX11-NEXT:    s_or_saveexec_b32 s1, -1
 ; GFX11-NEXT:    s_clause 0x1
-; GFX11-NEXT:    scratch_store_b32 off, v40, s33 offset:20
+; GFX11-NEXT:    scratch_store_b32 off, v45, s33 offset:20
 ; GFX11-NEXT:    scratch_store_b32 off, v46, s33 offset:24
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX11-NEXT:    s_clause 0x4
-; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:16
-; GFX11-NEXT:    scratch_store_b32 off, v42, s33 offset:12
-; GFX11-NEXT:    scratch_store_b32 off, v43, s33 offset:8
-; GFX11-NEXT:    scratch_store_b32 off, v44, s33 offset:4
-; GFX11-NEXT:    scratch_store_b32 off, v45, s33
+; GFX11-NEXT:    scratch_store_b32 off, v40, s33 offset:16
+; GFX11-NEXT:    scratch_store_b32 off, v41, s33 offset:12
+; GFX11-NEXT:    scratch_store_b32 off, v42, s33 offset:8
+; GFX11-NEXT:    scratch_store_b32 off, v43, s33 offset:4
+; GFX11-NEXT:    scratch_store_b32 off, v44, s33
 ; GFX11-NEXT:    image_gather4_c_b_cl v[0:3], v[12:16], s[0:7], s[0:3] dmask:0x1 dim:SQ_RSRC_IMG_2D
 ; GFX11-NEXT:    s_add_i32 s32, s32, 32
 ; GFX11-NEXT:    v_writelane_b32 v46, s0, 0
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, extern_func@gotpcrel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, extern_func@gotpcrel32@hi+12
-; GFX11-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX11-NEXT:    ; implicit-def: $vgpr45
+; GFX11-NEXT:    v_dual_mov_b32 v40, v16 :: v_dual_mov_b32 v41, v15
 ; GFX11-NEXT:    s_load_b64 s[0:1], s[0:1], 0x0
-; GFX11-NEXT:    v_dual_mov_b32 v41, v16 :: v_dual_mov_b32 v42, v15
-; GFX11-NEXT:    v_dual_mov_b32 v43, v14 :: v_dual_mov_b32 v44, v13
-; GFX11-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX11-NEXT:    v_mov_b32_e32 v45, v12
+; GFX11-NEXT:    v_writelane_b32 v45, s30, 0
+; GFX11-NEXT:    v_dual_mov_b32 v42, v14 :: v_dual_mov_b32 v43, v13
+; GFX11-NEXT:    v_mov_b32_e32 v44, v12
+; GFX11-NEXT:    v_writelane_b32 v45, s31, 1
 ; GFX11-NEXT:    s_waitcnt vmcnt(0)
 ; GFX11-NEXT:    global_store_b128 v[0:1], v[0:3], off
 ; GFX11-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
-; GFX11-NEXT:    image_gather4_c_b_cl v[0:3], [v45, v44, v43, v42, v41], s[0:7], s[0:3] dmask:0x1 dim:SQ_RSRC_IMG_2D
+; GFX11-NEXT:    image_gather4_c_b_cl v[0:3], [v44, v43, v42, v41, v40], s[0:7], s[0:3] dmask:0x1 dim:SQ_RSRC_IMG_2D
 ; GFX11-NEXT:    s_clause 0x4
-; GFX11-NEXT:    scratch_load_b32 v45, off, s33
-; GFX11-NEXT:    scratch_load_b32 v44, off, s33 offset:4
-; GFX11-NEXT:    scratch_load_b32 v43, off, s33 offset:8
-; GFX11-NEXT:    scratch_load_b32 v42, off, s33 offset:12
-; GFX11-NEXT:    scratch_load_b32 v41, off, s33 offset:16
-; GFX11-NEXT:    v_readlane_b32 s31, v40, 1
-; GFX11-NEXT:    v_readlane_b32 s30, v40, 0
+; GFX11-NEXT:    scratch_load_b32 v44, off, s33
+; GFX11-NEXT:    scratch_load_b32 v43, off, s33 offset:4
+; GFX11-NEXT:    scratch_load_b32 v42, off, s33 offset:8
+; GFX11-NEXT:    scratch_load_b32 v41, off, s33 offset:12
+; GFX11-NEXT:    scratch_load_b32 v40, off, s33 offset:16
+; GFX11-NEXT:    v_readlane_b32 s31, v45, 1
+; GFX11-NEXT:    v_readlane_b32 s30, v45, 0
 ; GFX11-NEXT:    v_readlane_b32 s0, v46, 0
 ; GFX11-NEXT:    s_or_saveexec_b32 s1, -1
 ; GFX11-NEXT:    s_clause 0x1
-; GFX11-NEXT:    scratch_load_b32 v40, off, s33 offset:20
+; GFX11-NEXT:    scratch_load_b32 v45, off, s33 offset:20
 ; GFX11-NEXT:    scratch_load_b32 v46, off, s33 offset:24
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s1
 ; GFX11-NEXT:    s_addk_i32 s32, 0xffe0
diff --git a/llvm/test/CodeGen/AMDGPU/wwm-register-spill-during-regalloc.ll b/llvm/test/CodeGen/AMDGPU/wwm-register-spill-during-regalloc.ll
new file mode 100644
--- /dev/null
+++ b/llvm/test/CodeGen/AMDGPU/wwm-register-spill-during-regalloc.ll
@@ -0,0 +1,166 @@
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx906 -stop-after=virtregrewriter,1 --verify-machineinstrs -o - %s | FileCheck -check-prefix=WWM-SPILL %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx906 -O0 -stop-after=regallocfast,1 --verify-machineinstrs -o - %s | FileCheck -check-prefix=WWM-SPILL-O0 %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx906 --verify-machineinstrs -o - %s | FileCheck -check-prefix=GCN %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx906 -O0 --verify-machineinstrs -o - %s | FileCheck -check-prefix=GCN-O0 %s
+
+; Test whole-wave register spilling.
+
+; In the testcase, the return address registers (SGPR30_SGPR31) should be preserved across the call.
+; Since the test limits the VGPR numbers, they are all in the call-clobber (scratch) range and RA should
+; spill any VGPR borrowed for spilling SGPRs. The writelane/readlane instructions that spill/restore
+; SGPRs into/from VGPR are whole-wave operations and hence the VGPRs involved in such operations require
+; whole-wave spilling.
+
+define void @test() #0 {
+; WWM-SPILL-LABEL: name: test
+; WWM-SPILL: bb.0 (%ir-block.0):
+; WWM-SPILL-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $sgpr15, $sgpr30, $sgpr31, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11
+; WWM-SPILL-NEXT: {{  $}}
+; WWM-SPILL-NEXT:   renamable $vgpr0 = IMPLICIT_DEF
+; WWM-SPILL-NEXT:   renamable $vgpr0 = V_WRITELANE_B32 killed $sgpr30, 0, killed $vgpr0
+; WWM-SPILL-NEXT:   renamable $vgpr0 = V_WRITELANE_B32 killed $sgpr31, 1, killed $vgpr0
+; WWM-SPILL-NEXT:   SI_SPILL_WWM_V32_SAVE killed $vgpr0, %stack.2, $sgpr32, 0, implicit $exec :: (store (s32) into %stack.2, addrspace 5)
+; WWM-SPILL-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def dead $scc, implicit-def $sgpr32, implicit $sgpr32
+; WWM-SPILL-NEXT:   renamable $sgpr16_sgpr17 = SI_PC_ADD_REL_OFFSET target-flags(amdgpu-gotprel32-lo) @ext_func + 4, target-flags(amdgpu-gotprel32-hi) @ext_func + 12, implicit-def dead $scc
+; WWM-SPILL-NEXT:   renamable $sgpr16_sgpr17 = S_LOAD_DWORDX2_IMM killed renamable $sgpr16_sgpr17, 0, 0 :: (dereferenceable invariant load (s64) from got, addrspace 4)
+; WWM-SPILL-NEXT:   dead $sgpr30_sgpr31 = SI_CALL killed renamable $sgpr16_sgpr17, @ext_func, csr_amdgpu, implicit $sgpr4_sgpr5, implicit $sgpr6_sgpr7, implicit $sgpr8_sgpr9, implicit $sgpr10_sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $sgpr15, implicit $vgpr31, implicit $sgpr0_sgpr1_sgpr2_sgpr3
+; WWM-SPILL-NEXT:   ADJCALLSTACKDOWN 0, 0, implicit-def dead $scc, implicit-def $sgpr32, implicit $sgpr32
+; WWM-SPILL-NEXT:   renamable $vgpr0 = SI_SPILL_WWM_V32_RESTORE %stack.2, $sgpr32, 0, implicit $exec :: (load (s32) from %stack.2, addrspace 5)
+; WWM-SPILL-NEXT:   $sgpr31 = V_READLANE_B32 $vgpr0, 1
+; WWM-SPILL-NEXT:   $sgpr30 = V_READLANE_B32 killed $vgpr0, 0
+; WWM-SPILL-NEXT:   SI_RETURN
+;
+; WWM-SPILL-O0-LABEL: name: test
+; WWM-SPILL-O0: bb.0 (%ir-block.0):
+; WWM-SPILL-O0-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $sgpr15, $sgpr30, $sgpr31, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11
+; WWM-SPILL-O0-NEXT: {{  $}}
+; WWM-SPILL-O0-NEXT:   renamable $vgpr0 = IMPLICIT_DEF
+; WWM-SPILL-O0-NEXT:   renamable $vgpr0 = V_WRITELANE_B32 killed $sgpr30, 0, $vgpr0
+; WWM-SPILL-O0-NEXT:   renamable $vgpr0 = V_WRITELANE_B32 killed $sgpr31, 1, $vgpr0
+; WWM-SPILL-O0-NEXT:   SI_SPILL_WWM_V32_SAVE $vgpr0, %stack.2, $sgpr32, 0, implicit $exec :: (store (s32) into %stack.2, addrspace 5)
+; WWM-SPILL-O0-NEXT:   renamable $vgpr0 = COPY $vgpr31
+; WWM-SPILL-O0-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def dead $scc, implicit-def $sgpr32, implicit $sgpr32
+; WWM-SPILL-O0-NEXT:   renamable $sgpr16_sgpr17 = SI_PC_ADD_REL_OFFSET target-flags(amdgpu-gotprel32-lo) @ext_func + 4, target-flags(amdgpu-gotprel32-hi) @ext_func + 12, implicit-def dead $scc
+; WWM-SPILL-O0-NEXT:   renamable $sgpr16_sgpr17 = S_LOAD_DWORDX2_IMM killed renamable $sgpr16_sgpr17, 0, 0 :: (dereferenceable invariant load (s64) from got, addrspace 4)
+; WWM-SPILL-O0-NEXT:   renamable $sgpr20_sgpr21_sgpr22_sgpr23 = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+; WWM-SPILL-O0-NEXT:   $vgpr31 = COPY killed renamable $vgpr0
+; WWM-SPILL-O0-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY killed renamable $sgpr20_sgpr21_sgpr22_sgpr23
+; WWM-SPILL-O0-NEXT:   dead $sgpr30_sgpr31 = SI_CALL killed renamable $sgpr16_sgpr17, @ext_func, csr_amdgpu, implicit killed $sgpr4_sgpr5, implicit killed $sgpr6_sgpr7, implicit killed $sgpr8_sgpr9, implicit killed $sgpr10_sgpr11, implicit killed $sgpr12, implicit killed $sgpr13, implicit killed $sgpr14, implicit killed $sgpr15, implicit $vgpr31, implicit $sgpr0_sgpr1_sgpr2_sgpr3
+; WWM-SPILL-O0-NEXT:   $vgpr0 = SI_SPILL_WWM_V32_RESTORE %stack.2, $sgpr32, 0, implicit $exec :: (load (s32) from %stack.2, addrspace 5)
+; WWM-SPILL-O0-NEXT:   ADJCALLSTACKDOWN 0, 0, implicit-def dead $scc, implicit-def $sgpr32, implicit $sgpr32
+; WWM-SPILL-O0-NEXT:   dead $sgpr31 = V_READLANE_B32 $vgpr0, 1
+; WWM-SPILL-O0-NEXT:   dead $sgpr30 = V_READLANE_B32 killed $vgpr0, 0
+; WWM-SPILL-O0-NEXT:   SI_RETURN
+;
+; GCN-LABEL: test:
+; GCN:       ; %bb.0:
+; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GCN-NEXT:    s_mov_b32 s16, s33
+; GCN-NEXT:    s_mov_b32 s33, s32
+; GCN-NEXT:    s_xor_saveexec_b64 s[18:19], -1
+; GCN-NEXT:    buffer_store_dword v0, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
+; GCN-NEXT:    s_mov_b64 exec, s[18:19]
+; GCN-NEXT:    v_mov_b32_e32 v1, s34
+; GCN-NEXT:    buffer_store_dword v1, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
+; GCN-NEXT:    v_mov_b32_e32 v1, s35
+; GCN-NEXT:    ; implicit-def: $vgpr0
+; GCN-NEXT:    buffer_store_dword v1, off, s[0:3], s33 offset:12 ; 4-byte Folded Spill
+; GCN-NEXT:    v_mov_b32_e32 v1, s16
+; GCN-NEXT:    v_writelane_b32 v0, s30, 0
+; GCN-NEXT:    buffer_store_dword v1, off, s[0:3], s33 offset:16 ; 4-byte Folded Spill
+; GCN-NEXT:    s_addk_i32 s32, 0x800
+; GCN-NEXT:    v_writelane_b32 v0, s31, 1
+; GCN-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GCN-NEXT:    buffer_store_dword v0, off, s[0:3], s33 ; 4-byte Folded Spill
+; GCN-NEXT:    s_mov_b64 exec, s[34:35]
+; GCN-NEXT:    s_getpc_b64 s[16:17]
+; GCN-NEXT:    s_add_u32 s16, s16, ext_func@gotpcrel32@lo+4
+; GCN-NEXT:    s_addc_u32 s17, s17, ext_func@gotpcrel32@hi+12
+; GCN-NEXT:    s_load_dwordx2 s[16:17], s[16:17], 0x0
+; GCN-NEXT:    s_waitcnt lgkmcnt(0)
+; GCN-NEXT:    s_swappc_b64 s[30:31], s[16:17]
+; GCN-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GCN-NEXT:    buffer_load_dword v0, off, s[0:3], s33 ; 4-byte Folded Reload
+; GCN-NEXT:    s_mov_b64 exec, s[34:35]
+; GCN-NEXT:    s_waitcnt vmcnt(0)
+; GCN-NEXT:    v_readlane_b32 s31, v0, 1
+; GCN-NEXT:    v_readlane_b32 s30, v0, 0
+; GCN-NEXT:    buffer_load_dword v0, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload
+; GCN-NEXT:    s_waitcnt vmcnt(0)
+; GCN-NEXT:    v_readfirstlane_b32 s34, v0
+; GCN-NEXT:    buffer_load_dword v0, off, s[0:3], s33 offset:12 ; 4-byte Folded Reload
+; GCN-NEXT:    s_waitcnt vmcnt(0)
+; GCN-NEXT:    v_readfirstlane_b32 s35, v0
+; GCN-NEXT:    buffer_load_dword v0, off, s[0:3], s33 offset:16 ; 4-byte Folded Reload
+; GCN-NEXT:    s_waitcnt vmcnt(0)
+; GCN-NEXT:    v_readfirstlane_b32 s4, v0
+; GCN-NEXT:    s_xor_saveexec_b64 s[6:7], -1
+; GCN-NEXT:    buffer_load_dword v0, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
+; GCN-NEXT:    s_mov_b64 exec, s[6:7]
+; GCN-NEXT:    s_addk_i32 s32, 0xf800
+; GCN-NEXT:    s_mov_b32 s33, s4
+; GCN-NEXT:    s_waitcnt vmcnt(0)
+; GCN-NEXT:    s_setpc_b64 s[30:31]
+;
+; GCN-O0-LABEL: test:
+; GCN-O0:       ; %bb.0:
+; GCN-O0-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GCN-O0-NEXT:    s_mov_b32 s16, s33
+; GCN-O0-NEXT:    s_mov_b32 s33, s32
+; GCN-O0-NEXT:    s_xor_saveexec_b64 s[18:19], -1
+; GCN-O0-NEXT:    buffer_store_dword v0, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
+; GCN-O0-NEXT:    s_mov_b64 exec, s[18:19]
+; GCN-O0-NEXT:    v_mov_b32_e32 v1, s34
+; GCN-O0-NEXT:    buffer_store_dword v1, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
+; GCN-O0-NEXT:    v_mov_b32_e32 v1, s35
+; GCN-O0-NEXT:    buffer_store_dword v1, off, s[0:3], s33 offset:12 ; 4-byte Folded Spill
+; GCN-O0-NEXT:    v_mov_b32_e32 v1, s16
+; GCN-O0-NEXT:    buffer_store_dword v1, off, s[0:3], s33 offset:16 ; 4-byte Folded Spill
+; GCN-O0-NEXT:    s_add_i32 s32, s32, 0x800
+; GCN-O0-NEXT:    ; implicit-def: $vgpr0
+; GCN-O0-NEXT:    v_writelane_b32 v0, s30, 0
+; GCN-O0-NEXT:    v_writelane_b32 v0, s31, 1
+; GCN-O0-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GCN-O0-NEXT:    buffer_store_dword v0, off, s[0:3], s33 ; 4-byte Folded Spill
+; GCN-O0-NEXT:    s_mov_b64 exec, s[34:35]
+; GCN-O0-NEXT:    v_mov_b32_e32 v0, v31
+; GCN-O0-NEXT:    s_getpc_b64 s[16:17]
+; GCN-O0-NEXT:    s_add_u32 s16, s16, ext_func@gotpcrel32@lo+4
+; GCN-O0-NEXT:    s_addc_u32 s17, s17, ext_func@gotpcrel32@hi+12
+; GCN-O0-NEXT:    s_load_dwordx2 s[16:17], s[16:17], 0x0
+; GCN-O0-NEXT:    s_mov_b64 s[22:23], s[2:3]
+; GCN-O0-NEXT:    s_mov_b64 s[20:21], s[0:1]
+; GCN-O0-NEXT:    v_mov_b32_e32 v31, v0
+; GCN-O0-NEXT:    s_mov_b64 s[0:1], s[20:21]
+; GCN-O0-NEXT:    s_mov_b64 s[2:3], s[22:23]
+; GCN-O0-NEXT:    s_waitcnt lgkmcnt(0)
+; GCN-O0-NEXT:    s_swappc_b64 s[30:31], s[16:17]
+; GCN-O0-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GCN-O0-NEXT:    buffer_load_dword v0, off, s[0:3], s33 ; 4-byte Folded Reload
+; GCN-O0-NEXT:    s_mov_b64 exec, s[34:35]
+; GCN-O0-NEXT:    s_waitcnt vmcnt(0)
+; GCN-O0-NEXT:    v_readlane_b32 s31, v0, 1
+; GCN-O0-NEXT:    v_readlane_b32 s30, v0, 0
+; GCN-O0-NEXT:    buffer_load_dword v0, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload
+; GCN-O0-NEXT:    s_waitcnt vmcnt(0)
+; GCN-O0-NEXT:    v_readfirstlane_b32 s34, v0
+; GCN-O0-NEXT:    buffer_load_dword v0, off, s[0:3], s33 offset:12 ; 4-byte Folded Reload
+; GCN-O0-NEXT:    s_waitcnt vmcnt(0)
+; GCN-O0-NEXT:    v_readfirstlane_b32 s35, v0
+; GCN-O0-NEXT:    buffer_load_dword v0, off, s[0:3], s33 offset:16 ; 4-byte Folded Reload
+; GCN-O0-NEXT:    s_waitcnt vmcnt(0)
+; GCN-O0-NEXT:    v_readfirstlane_b32 s4, v0
+; GCN-O0-NEXT:    s_xor_saveexec_b64 s[6:7], -1
+; GCN-O0-NEXT:    buffer_load_dword v0, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
+; GCN-O0-NEXT:    s_mov_b64 exec, s[6:7]
+; GCN-O0-NEXT:    s_add_i32 s32, s32, 0xfffff800
+; GCN-O0-NEXT:    s_mov_b32 s33, s4
+; GCN-O0-NEXT:    s_waitcnt vmcnt(0)
+; GCN-O0-NEXT:    s_setpc_b64 s[30:31]
+  call void @ext_func()
+  ret void
+}
+
+declare void @ext_func();
+
+attributes #0 = { nounwind "amdgpu-num-vgpr"="4" }
diff --git a/llvm/test/CodeGen/AMDGPU/wwm-reserved-spill.ll b/llvm/test/CodeGen/AMDGPU/wwm-reserved-spill.ll
--- a/llvm/test/CodeGen/AMDGPU/wwm-reserved-spill.ll
+++ b/llvm/test/CodeGen/AMDGPU/wwm-reserved-spill.ll
@@ -133,10 +133,12 @@
 ; GFX9-O0:       ; %bb.0: ; %entry
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-O0-NEXT:    s_xor_saveexec_b64 s[34:35], -1
-; GFX9-O0-NEXT:    buffer_store_dword v3, off, s[0:3], s32 offset:16 ; 4-byte Folded Spill
-; GFX9-O0-NEXT:    buffer_store_dword v1, off, s[0:3], s32 offset:20 ; 4-byte Folded Spill
-; GFX9-O0-NEXT:    buffer_store_dword v2, off, s[0:3], s32 offset:24 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:20 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    buffer_store_dword v4, off, s[0:3], s32 offset:24 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    buffer_store_dword v1, off, s[0:3], s32 offset:28 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    buffer_store_dword v2, off, s[0:3], s32 offset:32 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-O0-NEXT:    v_mov_b32_e32 v3, v0
 ; GFX9-O0-NEXT:    s_mov_b32 s36, s4
 ; GFX9-O0-NEXT:    ; kill: def $sgpr36 killed $sgpr36 def $sgpr36_sgpr37_sgpr38_sgpr39
 ; GFX9-O0-NEXT:    s_mov_b32 s37, s5
@@ -144,16 +146,17 @@
 ; GFX9-O0-NEXT:    s_mov_b32 s39, s7
 ; GFX9-O0-NEXT:    s_mov_b64 s[42:43], s[38:39]
 ; GFX9-O0-NEXT:    s_mov_b64 s[40:41], s[36:37]
-; GFX9-O0-NEXT:    v_writelane_b32 v3, s40, 0
-; GFX9-O0-NEXT:    v_writelane_b32 v3, s41, 1
-; GFX9-O0-NEXT:    v_writelane_b32 v3, s42, 2
-; GFX9-O0-NEXT:    v_writelane_b32 v3, s43, 3
+; GFX9-O0-NEXT:    ; implicit-def: $vgpr0
+; GFX9-O0-NEXT:    v_writelane_b32 v0, s40, 0
+; GFX9-O0-NEXT:    v_writelane_b32 v0, s41, 1
+; GFX9-O0-NEXT:    v_writelane_b32 v0, s42, 2
+; GFX9-O0-NEXT:    v_writelane_b32 v0, s43, 3
 ; GFX9-O0-NEXT:    s_mov_b32 s34, 0
 ; GFX9-O0-NEXT:    buffer_load_dwordx2 v[4:5], off, s[36:39], s34
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
-; GFX9-O0-NEXT:    buffer_store_dword v4, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    buffer_store_dword v4, off, s[0:3], s32 offset:12 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
-; GFX9-O0-NEXT:    buffer_store_dword v5, off, s[0:3], s32 offset:12 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    buffer_store_dword v5, off, s[0:3], s32 offset:16 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    ; implicit-def: $sgpr36_sgpr37
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v1, v4
 ; GFX9-O0-NEXT:    s_not_b64 exec, exec
@@ -166,22 +169,25 @@
 ; GFX9-O0-NEXT:    v_add_u32_e64 v1, v1, v2
 ; GFX9-O0-NEXT:    s_mov_b64 exec, s[36:37]
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v4, v1
-; GFX9-O0-NEXT:    buffer_store_dword v4, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
-; GFX9-O0-NEXT:    v_cmp_eq_u32_e64 s[36:37], v0, s34
-; GFX9-O0-NEXT:    v_mov_b32_e32 v0, s34
-; GFX9-O0-NEXT:    buffer_store_dword v0, off, s[0:3], s32 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    buffer_store_dword v4, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    v_cmp_eq_u32_e64 s[36:37], v3, s34
+; GFX9-O0-NEXT:    v_mov_b32_e32 v3, s34
+; GFX9-O0-NEXT:    buffer_store_dword v3, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    s_mov_b64 s[34:35], exec
-; GFX9-O0-NEXT:    v_writelane_b32 v3, s34, 4
-; GFX9-O0-NEXT:    v_writelane_b32 v3, s35, 5
+; GFX9-O0-NEXT:    v_writelane_b32 v0, s34, 4
+; GFX9-O0-NEXT:    v_writelane_b32 v0, s35, 5
+; GFX9-O0-NEXT:    s_or_saveexec_b64 s[44:45], -1
+; GFX9-O0-NEXT:    buffer_store_dword v0, off, s[0:3], s32 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    s_mov_b64 exec, s[44:45]
 ; GFX9-O0-NEXT:    s_and_b64 s[34:35], s[34:35], s[36:37]
 ; GFX9-O0-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-O0-NEXT:    s_cbranch_execz .LBB1_2
 ; GFX9-O0-NEXT:  ; %bb.1: ; %if
-; GFX9-O0-NEXT:    buffer_load_dword v4, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v3, off, s[0:3], s32 offset:12 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
-; GFX9-O0-NEXT:    buffer_load_dword v5, off, s[0:3], s32 offset:12 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v4, off, s[0:3], s32 offset:16 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
-; GFX9-O0-NEXT:    v_mov_b32_e32 v0, v5
+; GFX9-O0-NEXT:    v_mov_b32_e32 v0, v4
 ; GFX9-O0-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX9-O0-NEXT:    s_mov_b64 exec, s[34:35]
@@ -194,20 +200,23 @@
 ; GFX9-O0-NEXT:    v_add_u32_e64 v1, v2, v1
 ; GFX9-O0-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v0, v1
-; GFX9-O0-NEXT:    buffer_store_dword v0, off, s[0:3], s32 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:  .LBB1_2: ; %merge
-; GFX9-O0-NEXT:    buffer_load_dword v0, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v0, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
+; GFX9-O0-NEXT:    buffer_load_dword v3, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    s_or_saveexec_b64 s[44:45], -1
 ; GFX9-O0-NEXT:    buffer_load_dword v4, off, s[0:3], s32 ; 4-byte Folded Reload
-; GFX9-O0-NEXT:    v_readlane_b32 s34, v3, 4
-; GFX9-O0-NEXT:    v_readlane_b32 s35, v3, 5
-; GFX9-O0-NEXT:    s_or_b64 exec, exec, s[34:35]
-; GFX9-O0-NEXT:    v_readlane_b32 s36, v3, 0
-; GFX9-O0-NEXT:    v_readlane_b32 s37, v3, 1
-; GFX9-O0-NEXT:    v_readlane_b32 s38, v3, 2
-; GFX9-O0-NEXT:    v_readlane_b32 s39, v3, 3
+; GFX9-O0-NEXT:    s_mov_b64 exec, s[44:45]
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
-; GFX9-O0-NEXT:    v_cmp_eq_u32_e64 s[34:35], v0, v4
+; GFX9-O0-NEXT:    v_readlane_b32 s34, v4, 4
+; GFX9-O0-NEXT:    v_readlane_b32 s35, v4, 5
+; GFX9-O0-NEXT:    s_or_b64 exec, exec, s[34:35]
+; GFX9-O0-NEXT:    v_readlane_b32 s36, v4, 0
+; GFX9-O0-NEXT:    v_readlane_b32 s37, v4, 1
+; GFX9-O0-NEXT:    v_readlane_b32 s38, v4, 2
+; GFX9-O0-NEXT:    v_readlane_b32 s39, v4, 3
+; GFX9-O0-NEXT:    v_cmp_eq_u32_e64 s[34:35], v0, v3
 ; GFX9-O0-NEXT:    v_cndmask_b32_e64 v0, 0, 1, s[34:35]
 ; GFX9-O0-NEXT:    s_mov_b32 s34, 1
 ; GFX9-O0-NEXT:    v_lshlrev_b32_e64 v0, s34, v0
@@ -216,9 +225,10 @@
 ; GFX9-O0-NEXT:    s_mov_b32 s34, 0
 ; GFX9-O0-NEXT:    buffer_store_dword v0, off, s[36:39], s34 offset:4
 ; GFX9-O0-NEXT:    s_xor_saveexec_b64 s[34:35], -1
-; GFX9-O0-NEXT:    buffer_load_dword v3, off, s[0:3], s32 offset:16 ; 4-byte Folded Reload
-; GFX9-O0-NEXT:    buffer_load_dword v1, off, s[0:3], s32 offset:20 ; 4-byte Folded Reload
-; GFX9-O0-NEXT:    buffer_load_dword v2, off, s[0:3], s32 offset:24 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v0, off, s[0:3], s32 offset:20 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v4, off, s[0:3], s32 offset:24 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v1, off, s[0:3], s32 offset:28 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v2, off, s[0:3], s32 offset:32 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-O0-NEXT:    s_setpc_b64 s[30:31]
@@ -335,13 +345,17 @@
 ; GFX9-O0-NEXT:    s_mov_b32 s35, s33
 ; GFX9-O0-NEXT:    s_mov_b32 s33, s32
 ; GFX9-O0-NEXT:    s_xor_saveexec_b64 s[36:37], -1
-; GFX9-O0-NEXT:    buffer_store_dword v3, off, s[0:3], s33 ; 4-byte Folded Spill
-; GFX9-O0-NEXT:    buffer_store_dword v2, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
-; GFX9-O0-NEXT:    buffer_store_dword v1, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    buffer_store_dword v0, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    buffer_store_dword v2, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    buffer_store_dword v1, off, s[0:3], s33 offset:12 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    s_mov_b64 exec, s[36:37]
-; GFX9-O0-NEXT:    s_add_i32 s32, s32, 0x400
-; GFX9-O0-NEXT:    v_writelane_b32 v3, s30, 0
-; GFX9-O0-NEXT:    v_writelane_b32 v3, s31, 1
+; GFX9-O0-NEXT:    s_add_i32 s32, s32, 0x800
+; GFX9-O0-NEXT:    ; implicit-def: $vgpr0
+; GFX9-O0-NEXT:    v_writelane_b32 v0, s30, 0
+; GFX9-O0-NEXT:    v_writelane_b32 v0, s31, 1
+; GFX9-O0-NEXT:    s_or_saveexec_b64 s[48:49], -1
+; GFX9-O0-NEXT:    buffer_store_dword v0, off, s[0:3], s33 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    s_mov_b64 exec, s[48:49]
 ; GFX9-O0-NEXT:    s_mov_b32 s36, s4
 ; GFX9-O0-NEXT:    ; kill: def $sgpr36 killed $sgpr36 def $sgpr36_sgpr37_sgpr38_sgpr39
 ; GFX9-O0-NEXT:    s_mov_b32 s37, s5
@@ -364,18 +378,22 @@
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v0, v2
 ; GFX9-O0-NEXT:    s_swappc_b64 s[30:31], s[42:43]
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v1, v0
+; GFX9-O0-NEXT:    s_or_saveexec_b64 s[48:49], -1
+; GFX9-O0-NEXT:    buffer_load_dword v0, off, s[0:3], s33 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    s_mov_b64 exec, s[48:49]
 ; GFX9-O0-NEXT:    v_add_u32_e64 v1, v1, v2
 ; GFX9-O0-NEXT:    s_mov_b64 exec, s[40:41]
-; GFX9-O0-NEXT:    v_mov_b32_e32 v0, v1
-; GFX9-O0-NEXT:    buffer_store_dword v0, off, s[36:39], s34 offset:4
-; GFX9-O0-NEXT:    v_readlane_b32 s31, v3, 1
-; GFX9-O0-NEXT:    v_readlane_b32 s30, v3, 0
+; GFX9-O0-NEXT:    v_mov_b32_e32 v3, v1
+; GFX9-O0-NEXT:    buffer_store_dword v3, off, s[36:39], s34 offset:4
+; GFX9-O0-NEXT:    s_waitcnt vmcnt(1)
+; GFX9-O0-NEXT:    v_readlane_b32 s31, v0, 1
+; GFX9-O0-NEXT:    v_readlane_b32 s30, v0, 0
 ; GFX9-O0-NEXT:    s_xor_saveexec_b64 s[36:37], -1
-; GFX9-O0-NEXT:    buffer_load_dword v3, off, s[0:3], s33 ; 4-byte Folded Reload
-; GFX9-O0-NEXT:    buffer_load_dword v2, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
-; GFX9-O0-NEXT:    buffer_load_dword v1, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v0, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v2, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v1, off, s[0:3], s33 offset:12 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_mov_b64 exec, s[36:37]
-; GFX9-O0-NEXT:    s_add_i32 s32, s32, 0xfffffc00
+; GFX9-O0-NEXT:    s_add_i32 s32, s32, 0xfffff800
 ; GFX9-O0-NEXT:    s_mov_b32 s33, s35
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-O0-NEXT:    s_setpc_b64 s[30:31]
@@ -390,8 +408,9 @@
 ; GFX9-O3-NEXT:    buffer_store_dword v2, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    buffer_store_dword v1, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    s_mov_b64 exec, s[34:35]
-; GFX9-O3-NEXT:    v_writelane_b32 v3, s30, 0
+; GFX9-O3-NEXT:    ; implicit-def: $vgpr3
 ; GFX9-O3-NEXT:    s_addk_i32 s32, 0x400
+; GFX9-O3-NEXT:    v_writelane_b32 v3, s30, 0
 ; GFX9-O3-NEXT:    v_writelane_b32 v3, s31, 1
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v2, s8
 ; GFX9-O3-NEXT:    s_not_b64 exec, exec
@@ -516,37 +535,39 @@
 ; GFX9-O0-LABEL: strict_wwm_call_i64:
 ; GFX9-O0:       ; %bb.0:
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX9-O0-NEXT:    s_mov_b32 s42, s33
+; GFX9-O0-NEXT:    s_mov_b32 s44, s33
 ; GFX9-O0-NEXT:    s_mov_b32 s33, s32
 ; GFX9-O0-NEXT:    s_xor_saveexec_b64 s[34:35], -1
-; GFX9-O0-NEXT:    buffer_store_dword v10, off, s[0:3], s33 ; 4-byte Folded Spill
-; GFX9-O0-NEXT:    buffer_store_dword v8, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    buffer_store_dword v0, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    buffer_store_dword v6, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    buffer_store_dword v8, off, s[0:3], s33 offset:12 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
-; GFX9-O0-NEXT:    buffer_store_dword v9, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
-; GFX9-O0-NEXT:    buffer_store_dword v2, off, s[0:3], s33 offset:12 ; 4-byte Folded Spill
-; GFX9-O0-NEXT:    buffer_store_dword v3, off, s[0:3], s33 offset:16 ; 4-byte Folded Spill
-; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
-; GFX9-O0-NEXT:    buffer_store_dword v4, off, s[0:3], s33 offset:20 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    buffer_store_dword v9, off, s[0:3], s33 offset:16 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    buffer_store_dword v2, off, s[0:3], s33 offset:20 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    buffer_store_dword v3, off, s[0:3], s33 offset:24 ; 4-byte Folded Spill
-; GFX9-O0-NEXT:    buffer_store_dword v2, off, s[0:3], s33 offset:28 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
+; GFX9-O0-NEXT:    buffer_store_dword v4, off, s[0:3], s33 offset:28 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    buffer_store_dword v3, off, s[0:3], s33 offset:32 ; 4-byte Folded Spill
-; GFX9-O0-NEXT:    buffer_store_dword v4, off, s[0:3], s33 offset:36 ; 4-byte Folded Spill
-; GFX9-O0-NEXT:    buffer_store_dword v5, off, s[0:3], s33 offset:40 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    buffer_store_dword v2, off, s[0:3], s33 offset:36 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
+; GFX9-O0-NEXT:    buffer_store_dword v3, off, s[0:3], s33 offset:40 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    buffer_store_dword v4, off, s[0:3], s33 offset:44 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    buffer_store_dword v5, off, s[0:3], s33 offset:48 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    s_mov_b64 exec, s[34:35]
-; GFX9-O0-NEXT:    s_add_i32 s32, s32, 0xc00
-; GFX9-O0-NEXT:    v_writelane_b32 v10, s30, 0
-; GFX9-O0-NEXT:    v_writelane_b32 v10, s31, 1
+; GFX9-O0-NEXT:    s_add_i32 s32, s32, 0x1000
+; GFX9-O0-NEXT:    ; implicit-def: $vgpr0
+; GFX9-O0-NEXT:    v_writelane_b32 v0, s30, 0
+; GFX9-O0-NEXT:    v_writelane_b32 v0, s31, 1
 ; GFX9-O0-NEXT:    s_mov_b32 s34, s8
 ; GFX9-O0-NEXT:    s_mov_b32 s36, s4
 ; GFX9-O0-NEXT:    ; kill: def $sgpr36 killed $sgpr36 def $sgpr36_sgpr37_sgpr38_sgpr39
 ; GFX9-O0-NEXT:    s_mov_b32 s37, s5
 ; GFX9-O0-NEXT:    s_mov_b32 s38, s6
 ; GFX9-O0-NEXT:    s_mov_b32 s39, s7
-; GFX9-O0-NEXT:    v_writelane_b32 v10, s36, 2
-; GFX9-O0-NEXT:    v_writelane_b32 v10, s37, 3
-; GFX9-O0-NEXT:    v_writelane_b32 v10, s38, 4
-; GFX9-O0-NEXT:    v_writelane_b32 v10, s39, 5
+; GFX9-O0-NEXT:    v_writelane_b32 v0, s36, 2
+; GFX9-O0-NEXT:    v_writelane_b32 v0, s37, 3
+; GFX9-O0-NEXT:    v_writelane_b32 v0, s38, 4
+; GFX9-O0-NEXT:    v_writelane_b32 v0, s39, 5
 ; GFX9-O0-NEXT:    ; kill: def $sgpr34 killed $sgpr34 def $sgpr34_sgpr35
 ; GFX9-O0-NEXT:    s_mov_b32 s35, s9
 ; GFX9-O0-NEXT:    ; kill: def $sgpr40_sgpr41 killed $sgpr34_sgpr35
@@ -558,8 +579,11 @@
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v9, s37
 ; GFX9-O0-NEXT:    s_not_b64 exec, exec
 ; GFX9-O0-NEXT:    s_or_saveexec_b64 s[34:35], -1
-; GFX9-O0-NEXT:    v_writelane_b32 v10, s34, 6
-; GFX9-O0-NEXT:    v_writelane_b32 v10, s35, 7
+; GFX9-O0-NEXT:    v_writelane_b32 v0, s34, 6
+; GFX9-O0-NEXT:    v_writelane_b32 v0, s35, 7
+; GFX9-O0-NEXT:    s_or_saveexec_b64 s[42:43], -1
+; GFX9-O0-NEXT:    buffer_store_dword v0, off, s[0:3], s33 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    s_mov_b64 exec, s[42:43]
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v2, v8
 ; GFX9-O0-NEXT:    s_mov_b32 s34, 32
 ; GFX9-O0-NEXT:    ; implicit-def: $sgpr36_sgpr37
@@ -576,13 +600,20 @@
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v1, v3
 ; GFX9-O0-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX9-O0-NEXT:    s_swappc_b64 s[30:31], s[34:35]
-; GFX9-O0-NEXT:    v_readlane_b32 s34, v10, 6
-; GFX9-O0-NEXT:    v_readlane_b32 s35, v10, 7
-; GFX9-O0-NEXT:    v_readlane_b32 s36, v10, 2
-; GFX9-O0-NEXT:    v_readlane_b32 s37, v10, 3
-; GFX9-O0-NEXT:    v_readlane_b32 s38, v10, 4
-; GFX9-O0-NEXT:    v_readlane_b32 s39, v10, 5
+; GFX9-O0-NEXT:    s_or_saveexec_b64 s[42:43], -1
+; GFX9-O0-NEXT:    buffer_load_dword v6, off, s[0:3], s33 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    s_mov_b64 exec, s[42:43]
+; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
+; GFX9-O0-NEXT:    v_readlane_b32 s34, v6, 6
+; GFX9-O0-NEXT:    v_readlane_b32 s35, v6, 7
+; GFX9-O0-NEXT:    v_readlane_b32 s36, v6, 2
+; GFX9-O0-NEXT:    v_readlane_b32 s37, v6, 3
+; GFX9-O0-NEXT:    v_readlane_b32 s38, v6, 4
+; GFX9-O0-NEXT:    v_readlane_b32 s39, v6, 5
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v2, v0
+; GFX9-O0-NEXT:    s_or_saveexec_b64 s[42:43], -1
+; GFX9-O0-NEXT:    buffer_load_dword v0, off, s[0:3], s33 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    s_mov_b64 exec, s[42:43]
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v3, v1
 ; GFX9-O0-NEXT:    ; implicit-def: $sgpr40
 ; GFX9-O0-NEXT:    ; implicit-def: $sgpr40
@@ -592,36 +623,38 @@
 ; GFX9-O0-NEXT:    v_add_co_u32_e64 v2, s[40:41], v2, v4
 ; GFX9-O0-NEXT:    v_addc_co_u32_e64 v3, s[40:41], v3, v5, s[40:41]
 ; GFX9-O0-NEXT:    s_mov_b64 exec, s[34:35]
-; GFX9-O0-NEXT:    v_mov_b32_e32 v0, v2
-; GFX9-O0-NEXT:    v_mov_b32_e32 v1, v3
+; GFX9-O0-NEXT:    v_mov_b32_e32 v6, v2
+; GFX9-O0-NEXT:    v_mov_b32_e32 v7, v3
 ; GFX9-O0-NEXT:    s_mov_b32 s34, 0
-; GFX9-O0-NEXT:    buffer_store_dwordx2 v[0:1], off, s[36:39], s34 offset:4
-; GFX9-O0-NEXT:    v_readlane_b32 s31, v10, 1
-; GFX9-O0-NEXT:    v_readlane_b32 s30, v10, 0
+; GFX9-O0-NEXT:    buffer_store_dwordx2 v[6:7], off, s[36:39], s34 offset:4
+; GFX9-O0-NEXT:    s_waitcnt vmcnt(1)
+; GFX9-O0-NEXT:    v_readlane_b32 s31, v0, 1
+; GFX9-O0-NEXT:    v_readlane_b32 s30, v0, 0
 ; GFX9-O0-NEXT:    s_xor_saveexec_b64 s[34:35], -1
-; GFX9-O0-NEXT:    buffer_load_dword v10, off, s[0:3], s33 ; 4-byte Folded Reload
-; GFX9-O0-NEXT:    s_nop 0
-; GFX9-O0-NEXT:    buffer_load_dword v8, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v0, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v6, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
-; GFX9-O0-NEXT:    buffer_load_dword v9, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v8, off, s[0:3], s33 offset:12 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
-; GFX9-O0-NEXT:    buffer_load_dword v2, off, s[0:3], s33 offset:12 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v9, off, s[0:3], s33 offset:16 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
-; GFX9-O0-NEXT:    buffer_load_dword v3, off, s[0:3], s33 offset:16 ; 4-byte Folded Reload
-; GFX9-O0-NEXT:    s_nop 0
-; GFX9-O0-NEXT:    buffer_load_dword v4, off, s[0:3], s33 offset:20 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v2, off, s[0:3], s33 offset:20 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
 ; GFX9-O0-NEXT:    buffer_load_dword v3, off, s[0:3], s33 offset:24 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
-; GFX9-O0-NEXT:    buffer_load_dword v2, off, s[0:3], s33 offset:28 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v4, off, s[0:3], s33 offset:28 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
 ; GFX9-O0-NEXT:    buffer_load_dword v3, off, s[0:3], s33 offset:32 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
-; GFX9-O0-NEXT:    buffer_load_dword v4, off, s[0:3], s33 offset:36 ; 4-byte Folded Reload
-; GFX9-O0-NEXT:    buffer_load_dword v5, off, s[0:3], s33 offset:40 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v2, off, s[0:3], s33 offset:36 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    s_nop 0
+; GFX9-O0-NEXT:    buffer_load_dword v3, off, s[0:3], s33 offset:40 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    s_nop 0
+; GFX9-O0-NEXT:    buffer_load_dword v4, off, s[0:3], s33 offset:44 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v5, off, s[0:3], s33 offset:48 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_mov_b64 exec, s[34:35]
-; GFX9-O0-NEXT:    s_add_i32 s32, s32, 0xfffff400
-; GFX9-O0-NEXT:    s_mov_b32 s33, s42
+; GFX9-O0-NEXT:    s_add_i32 s32, s32, 0xfffff000
+; GFX9-O0-NEXT:    s_mov_b32 s33, s44
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-O0-NEXT:    s_setpc_b64 s[30:31]
 ;
@@ -641,8 +674,9 @@
 ; GFX9-O3-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-O3-NEXT:    buffer_store_dword v3, off, s[0:3], s33 offset:24 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    s_mov_b64 exec, s[34:35]
-; GFX9-O3-NEXT:    v_writelane_b32 v8, s30, 0
+; GFX9-O3-NEXT:    ; implicit-def: $vgpr8
 ; GFX9-O3-NEXT:    s_addk_i32 s32, 0x800
+; GFX9-O3-NEXT:    v_writelane_b32 v8, s30, 0
 ; GFX9-O3-NEXT:    v_writelane_b32 v8, s31, 1
 ; GFX9-O3-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-O3-NEXT:    s_getpc_b64 s[36:37]
diff --git a/llvm/test/CodeGen/MIR/AMDGPU/machine-function-info-after-pei.ll b/llvm/test/CodeGen/MIR/AMDGPU/machine-function-info-after-pei.ll
--- a/llvm/test/CodeGen/MIR/AMDGPU/machine-function-info-after-pei.ll
+++ b/llvm/test/CodeGen/MIR/AMDGPU/machine-function-info-after-pei.ll
@@ -37,6 +37,7 @@
 ; AFTER-PEI-NEXT: occupancy: 5
 ; AFTER-PEI-NEXT: scavengeFI: '%fixed-stack.0'
 ; AFTER-PEI-NEXT: vgprForAGPRCopy: ''
+; AFTER-PEI-NEXT: sgprForEXECCopy: ''
 ; AFTER-PEI-NEXT: body:
 define amdgpu_kernel void @scavenge_fi(i32 addrspace(1)* %out, i32 %in) #0 {
   %wide.sgpr0 = call <32 x i32>  asm sideeffect "; def $0", "=s" () #0
diff --git a/llvm/test/CodeGen/MIR/AMDGPU/machine-function-info-no-ir.mir b/llvm/test/CodeGen/MIR/AMDGPU/machine-function-info-no-ir.mir
--- a/llvm/test/CodeGen/MIR/AMDGPU/machine-function-info-no-ir.mir
+++ b/llvm/test/CodeGen/MIR/AMDGPU/machine-function-info-no-ir.mir
@@ -46,6 +46,7 @@
 # FULL-NEXT:  highBitsOf32BitAddress: 0
 # FULL-NEXT:  occupancy: 10
 # FULL-NEXT:  vgprForAGPRCopy: ''
+# FULL-NEXT:  sgprForEXECCopy: ''
 # FULL-NEXT: body:
 
 # SIMPLE: machineFunctionInfo:
@@ -144,6 +145,7 @@
 # FULL-NEXT:  highBitsOf32BitAddress: 0
 # FULL-NEXT:  occupancy: 10
 # FULL-NEXT: vgprForAGPRCopy: ''
+# FULL-NEXT: sgprForEXECCopy: ''
 # FULL-NEXT: body:
 
 # SIMPLE: machineFunctionInfo:
@@ -213,6 +215,7 @@
 # FULL-NEXT:  highBitsOf32BitAddress: 0
 # FULL-NEXT:  occupancy: 10
 # FULL-NEXT: vgprForAGPRCopy: ''
+# FULL-NEXT: sgprForEXECCopy: ''
 # FULL-NEXT: body:
 
 # SIMPLE: machineFunctionInfo:
@@ -283,6 +286,7 @@
 # FULL-NEXT:  highBitsOf32BitAddress: 0
 # FULL-NEXT:  occupancy: 10
 # FULL-NEXT: vgprForAGPRCopy: ''
+# FULL-NEXT: sgprForEXECCopy: ''
 # FULL-NEXT: body:
 
 # SIMPLE: machineFunctionInfo:
@@ -529,3 +533,28 @@
     SI_RETURN
 
 ...
+
+---
+# ALL-LABEL: name: sgpr_for_exec_copy
+# ALL: sgprForEXECCopy: '$sgpr2_sgpr3'
+name: sgpr_for_exec_copy
+machineFunctionInfo:
+  sgprForEXECCopy: '$sgpr2_sgpr3'
+body:             |
+  bb.0:
+    SI_RETURN
+
+...
+
+---
+# ALL-LABEL: name: sgpr_for_exec_copy_noreg
+# FULL: sgprForEXECCopy: ''
+# SIMPLE-NOT: sgprForEXECCopy
+name: sgpr_for_exec_copy_noreg
+machineFunctionInfo:
+  sgprForEXECCopy: '$noreg'
+body:             |
+  bb.0:
+    SI_RETURN
+
+...
diff --git a/llvm/test/CodeGen/MIR/AMDGPU/machine-function-info.ll b/llvm/test/CodeGen/MIR/AMDGPU/machine-function-info.ll
--- a/llvm/test/CodeGen/MIR/AMDGPU/machine-function-info.ll
+++ b/llvm/test/CodeGen/MIR/AMDGPU/machine-function-info.ll
@@ -40,6 +40,7 @@
 ; CHECK-NEXT: highBitsOf32BitAddress: 0
 ; CHECK-NEXT: occupancy: 10
 ; CHECK-NEXT: vgprForAGPRCopy: ''
+; CHECK-NEXT: sgprForEXECCopy: '$sgpr100_sgpr101'
 ; CHECK-NEXT: body:
 define amdgpu_kernel void @kernel(i32 %arg0, i64 %arg1, <16 x i32> %arg2) {
   %gep = getelementptr inbounds [512 x float], [512 x float] addrspace(3)* @lds, i32 0, i32 %arg0
@@ -80,6 +81,7 @@
 ; CHECK-NEXT: highBitsOf32BitAddress: 0
 ; CHECK-NEXT: occupancy: 10
 ; CHECK-NEXT: vgprForAGPRCopy: ''
+; CHECK-NEXT: sgprForEXECCopy: '$sgpr100_sgpr101'
 ; CHECK-NEXT: body:
 define amdgpu_ps void @ps_shader(i32 %arg0, i32 inreg %arg1) {
   %gep = getelementptr inbounds [128 x i32], [128 x i32] addrspace(2)* @gds, i32 0, i32 %arg0
@@ -134,6 +136,7 @@
 ; CHECK-NEXT: highBitsOf32BitAddress: 0
 ; CHECK-NEXT: occupancy: 10
 ; CHECK-NEXT: vgprForAGPRCopy: ''
+; CHECK-NEXT: sgprForEXECCopy: '$sgpr100_sgpr101'
 ; CHECK-NEXT: body:
 define void @function() {
   ret void
@@ -180,6 +183,7 @@
 ; CHECK-NEXT: highBitsOf32BitAddress: 0
 ; CHECK-NEXT: occupancy: 10
 ; CHECK-NEXT: vgprForAGPRCopy: ''
+; CHECK-NEXT: sgprForEXECCopy: '$sgpr100_sgpr101'
 ; CHECK-NEXT: body:
 define void @function_nsz() #0 {
   ret void
diff --git a/llvm/test/CodeGen/MIR/AMDGPU/sgpr-for-exec-copy-invalid-reg.mir b/llvm/test/CodeGen/MIR/AMDGPU/sgpr-for-exec-copy-invalid-reg.mir
new file mode 100644
--- /dev/null
+++ b/llvm/test/CodeGen/MIR/AMDGPU/sgpr-for-exec-copy-invalid-reg.mir
@@ -0,0 +1,12 @@
+# RUN: not llc -mtriple=amdgcn-amd-amdhsa -run-pass=none -verify-machineinstrs %s -o /dev/null 2>&1 | FileCheck -check-prefix=ERR %s
+
+---
+name: invalid_reg
+machineFunctionInfo:
+# ERR: [[@LINE+1]]:21: unknown register name 'srst'
+  sgprForEXECCopy: '$srst'
+body:             |
+  bb.0:
+    S_ENDPGM 0
+
+...
diff --git a/llvm/test/CodeGen/MIR/AMDGPU/stack-id-assert.mir b/llvm/test/CodeGen/MIR/AMDGPU/stack-id-assert.mir
--- a/llvm/test/CodeGen/MIR/AMDGPU/stack-id-assert.mir
+++ b/llvm/test/CodeGen/MIR/AMDGPU/stack-id-assert.mir
@@ -3,7 +3,7 @@
 # contains not dead objects only. So using objects IDs as offset in the storage
 # caused out of bounds access.
 
-# RUN: llc -march=amdgcn -run-pass=si-lower-sgpr-spills,prologepilog -verify-machineinstrs -o - %s | FileCheck %s
+# RUN: llc -march=amdgcn -start-before=si-lower-sgpr-spills -stop-after=prologepilog -verify-machineinstrs -o - %s | FileCheck %s
 
 # CHECK-LABEL: name:            foo
 # CHECK: {{^}}fixedStack:      []