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"
@@ -37,6 +38,7 @@
   const SIRegisterInfo *TRI = nullptr;
   const SIInstrInfo *TII = nullptr;
   LiveIntervals *LIS = 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.
@@ -50,13 +52,23 @@
 
   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
@@ -67,6 +79,7 @@
                       "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)
 
@@ -243,12 +256,66 @@
   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->getSGPRToVGPRSpills(FI);
+  Register PrevLaneVGPR;
+  bool SeenSpillInBlock = false;
+  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. Check if they dominate only for
+        // the first spill in case of tuple spills.
+        if (!SeenSpillInBlock) {
+          SeenSpillInBlock = true;
+          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();
   TRI = &TII->getRegisterInfo();
 
   LIS = getAnalysisIfAvailable<LiveIntervals>();
+  MDT = &getAnalysis<MachineDominatorTree>();
 
   assert(SaveBlocks.empty() && RestoreBlocks.empty());
 
@@ -258,7 +325,6 @@
   bool HasCSRs = spillCalleeSavedRegs(MF);
 
   MachineFrameInfo &MFI = MF.getFrameInfo();
-  MachineRegisterInfo &MRI = MF.getRegInfo();
   SIMachineFunctionInfo *FuncInfo = MF.getInfo<SIMachineFunctionInfo>();
 
   if (!MFI.hasStackObjects() && !HasCSRs) {
@@ -268,7 +334,6 @@
   }
 
   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.
@@ -284,6 +349,9 @@
     // 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))
@@ -291,23 +359,31 @@
 
         int FI = TII->getNamedOperand(MI, AMDGPU::OpName::addr)->getIndex();
         assert(MFI.getStackID(FI) == TargetStackID::SGPRSpill);
+        MachineInstrSpan MIS(&MI, &MBB);
         if (FuncInfo->allocateSGPRSpillToVGPR(MF, FI)) {
-          NewReservedRegs = true;
           bool Spilled = TRI->eliminateSGPRToVGPRSpillFrameIndex(MI, FI,
                                                                  nullptr, 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);
+      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
@@ -333,9 +409,5 @@
   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.cpp b/llvm/lib/Target/AMDGPU/SIMachineFunctionInfo.cpp
--- a/llvm/lib/Target/AMDGPU/SIMachineFunctionInfo.cpp
+++ b/llvm/lib/Target/AMDGPU/SIMachineFunctionInfo.cpp
@@ -287,33 +287,11 @@
 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.
-      SGPRToVGPRSpills.erase(FI);
-
-      // FIXME: We can run out of free registers with split allocation if
-      // IPRA is enabled and a called function already uses every VGPR.
-#if 0
-        DiagnosticInfoResourceLimit DiagOutOfRegs(MF.getFunction(),
-                                                  "VGPRs for SGPR spilling",
-                                                  0, DS_Error);
-        MF.getFunction().getContext().diagnose(DiagOutOfRegs);
-#endif
-      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();
   }
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
@@ -711,9 +711,6 @@
   for (MCPhysReg Reg : MFI->getVGPRSpillAGPRs())
     reserveRegisterTuples(Reserved, Reg);
 
-  for (auto Reg : MFI->getSGPRSpillVGPRs())
-    reserveRegisterTuples(Reserved, Reg);
-
   return Reserved;
 }
 
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
@@ -11,6 +11,7 @@
 ; CHECK-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; CHECK-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; CHECK-NEXT:    s_mov_b64 exec, s[16:17]
+; CHECK-NEXT:    ; implicit-def: $vgpr40
 ; CHECK-NEXT:    v_writelane_b32 v41, s33, 0
 ; CHECK-NEXT:    s_mov_b32 s33, s32
 ; CHECK-NEXT:    s_addk_i32 s32, 0x400
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
@@ -281,6 +282,7 @@
 ; 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
@@ -319,11 +321,12 @@
 ; MUBUF-NEXT:    v_writelane_b32 v41, s33, 0
 ; MUBUF-NEXT:    s_mov_b32 s33, s32
 ; MUBUF-NEXT:    s_addk_i32 s32, 0x400
-; MUBUF-NEXT:    v_writelane_b32 v40, s30, 0
-; MUBUF-NEXT:    v_writelane_b32 v40, s31, 1
+; MUBUF-NEXT:    ; implicit-def: $vgpr40
 ; MUBUF-NEXT:    s_getpc_b64 s[4:5]
 ; MUBUF-NEXT:    s_add_u32 s4, s4, external_void_func_byval@rel32@lo+4
 ; MUBUF-NEXT:    s_addc_u32 s5, s5, external_void_func_byval@rel32@hi+12
+; MUBUF-NEXT:    v_writelane_b32 v40, s30, 0
+; MUBUF-NEXT:    v_writelane_b32 v40, s31, 1
 ; MUBUF-NEXT:    s_waitcnt vmcnt(1)
 ; MUBUF-NEXT:    buffer_store_dword v1, off, s[0:3], s32
 ; MUBUF-NEXT:    s_waitcnt vmcnt(1)
@@ -400,11 +403,12 @@
 ; FLATSCR-NEXT:    v_writelane_b32 v41, s33, 0
 ; FLATSCR-NEXT:    s_mov_b32 s33, s32
 ; FLATSCR-NEXT:    s_add_i32 s32, s32, 16
-; FLATSCR-NEXT:    v_writelane_b32 v40, s30, 0
-; FLATSCR-NEXT:    v_writelane_b32 v40, s31, 1
+; FLATSCR-NEXT:    ; implicit-def: $vgpr40
 ; FLATSCR-NEXT:    s_getpc_b64 s[0:1]
 ; FLATSCR-NEXT:    s_add_u32 s0, s0, external_void_func_byval@rel32@lo+4
 ; FLATSCR-NEXT:    s_addc_u32 s1, s1, external_void_func_byval@rel32@hi+12
+; FLATSCR-NEXT:    v_writelane_b32 v40, s30, 0
+; FLATSCR-NEXT:    v_writelane_b32 v40, s31, 1
 ; FLATSCR-NEXT:    s_waitcnt vmcnt(0)
 ; FLATSCR-NEXT:    scratch_store_dwordx2 off, v[1:2], s32
 ; FLATSCR-NEXT:    scratch_load_dwordx2 v[1:2], v0, off offset:8
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
@@ -8,7 +8,7 @@
 ; CHECK-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; CHECK-NEXT:    s_waitcnt_vscnt null, 0x0
 ; CHECK-NEXT:    s_or_saveexec_b32 s4, -1
-; CHECK-NEXT:    buffer_store_dword v16, 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:    s_mov_b32 exec_lo, s4
 ; CHECK-NEXT:    v_mov_b32_e32 v14, v1
 ; CHECK-NEXT:    v_mov_b32_e32 v13, v2
@@ -38,10 +38,12 @@
 ; 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 v16, s4, 0
-; CHECK-NEXT:    v_writelane_b32 v16, s5, 1
-; CHECK-NEXT:    v_writelane_b32 v16, s6, 2
-; CHECK-NEXT:    v_writelane_b32 v16, 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:    buffer_store_dword v0, off, s[0:3], s32 offset:44 ; 4-byte Folded Spill
 ; CHECK-NEXT:    s_mov_b32 s6, 0
 ; CHECK-NEXT:    s_mov_b32 s4, s6
 ; CHECK-NEXT:    s_mov_b32 s5, s6
@@ -49,8 +51,11 @@
 ; 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:    buffer_store_dword v1, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
+; CHECK-NEXT:    buffer_load_dword v0, off, s[0:3], s32 offset:44 ; 4-byte Folded Reload
 ; CHECK-NEXT:    s_mov_b32 s4, exec_lo
-; CHECK-NEXT:    v_writelane_b32 v16, s4, 4
+; CHECK-NEXT:    s_waitcnt vmcnt(0)
+; CHECK-NEXT:    v_writelane_b32 v0, s4, 4
+; CHECK-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:44 ; 4-byte Folded Spill
 ; CHECK-NEXT:  .LBB0_1: ; =>This Inner Loop Header: Depth=1
 ; CHECK-NEXT:    buffer_load_dword v8, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
 ; CHECK-NEXT:    buffer_load_dword v9, off, s[0:3], s32 offset:12 ; 4-byte Folded Reload
@@ -77,6 +82,7 @@
 ; CHECK-NEXT:    v_readfirstlane_b32 s6, v2
 ; CHECK-NEXT:    v_readfirstlane_b32 s5, v1
 ; CHECK-NEXT:    v_readfirstlane_b32 s4, v0
+; CHECK-NEXT:    buffer_load_dword v0, off, s[0:3], s32 offset:44 ; 4-byte Folded Reload
 ; 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
@@ -85,14 +91,16 @@
 ; 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 v16, s12, 5
-; CHECK-NEXT:    v_writelane_b32 v16, s13, 6
-; CHECK-NEXT:    v_writelane_b32 v16, s14, 7
-; CHECK-NEXT:    v_writelane_b32 v16, s15, 8
-; CHECK-NEXT:    v_writelane_b32 v16, s16, 9
-; CHECK-NEXT:    v_writelane_b32 v16, s17, 10
-; CHECK-NEXT:    v_writelane_b32 v16, s18, 11
-; CHECK-NEXT:    v_writelane_b32 v16, s19, 12
+; CHECK-NEXT:    s_waitcnt vmcnt(0)
+; 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:    buffer_store_dword v0, off, s[0:3], s32 offset:44 ; 4-byte Folded Spill
 ; CHECK-NEXT:    v_mov_b32_e32 v6, v8
 ; CHECK-NEXT:    v_mov_b32_e32 v7, v9
 ; CHECK-NEXT:    v_mov_b32_e32 v4, v10
@@ -111,25 +119,30 @@
 ; CHECK-NEXT:    v_cmp_eq_u64_e64 s5, s[8:9], v[2:3]
 ; CHECK-NEXT:    s_and_b32 s4, s4, s5
 ; CHECK-NEXT:    v_cmp_eq_u64_e64 s5, s[6:7], v[0:1]
+; CHECK-NEXT:    buffer_load_dword v0, off, s[0:3], s32 offset:44 ; 4-byte Folded Reload
 ; CHECK-NEXT:    s_and_b32 s4, s4, s5
 ; CHECK-NEXT:    s_and_saveexec_b32 s4, s4
-; CHECK-NEXT:    v_writelane_b32 v16, s4, 13
+; CHECK-NEXT:    s_waitcnt vmcnt(0)
+; CHECK-NEXT:    v_writelane_b32 v0, s4, 13
+; CHECK-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:44 ; 4-byte Folded Spill
 ; CHECK-NEXT:  ; %bb.2: ; in Loop: Header=BB0_1 Depth=1
-; CHECK-NEXT:    v_readlane_b32 s4, v16, 13
+; CHECK-NEXT:    buffer_load_dword v2, off, s[0:3], s32 offset:44 ; 4-byte Folded Reload
+; CHECK-NEXT:    s_waitcnt vmcnt(0)
+; CHECK-NEXT:    v_readlane_b32 s4, v2, 13
 ; 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 s8, v16, 5
-; CHECK-NEXT:    v_readlane_b32 s9, v16, 6
-; CHECK-NEXT:    v_readlane_b32 s10, v16, 7
-; CHECK-NEXT:    v_readlane_b32 s11, v16, 8
-; CHECK-NEXT:    v_readlane_b32 s12, v16, 9
-; CHECK-NEXT:    v_readlane_b32 s13, v16, 10
-; CHECK-NEXT:    v_readlane_b32 s14, v16, 11
-; CHECK-NEXT:    v_readlane_b32 s15, v16, 12
-; CHECK-NEXT:    v_readlane_b32 s16, v16, 0
-; CHECK-NEXT:    v_readlane_b32 s17, v16, 1
-; CHECK-NEXT:    v_readlane_b32 s18, v16, 2
-; CHECK-NEXT:    v_readlane_b32 s19, v16, 3
+; 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:    s_waitcnt vmcnt(0)
 ; 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)
@@ -137,7 +150,9 @@
 ; 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, v16, 4
+; CHECK-NEXT:    buffer_load_dword v0, off, s[0:3], s32 offset:44 ; 4-byte Folded Reload
+; 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
@@ -146,7 +161,7 @@
 ; CHECK-NEXT:    v_mov_b32_e32 v2, s4
 ; CHECK-NEXT:    v_mov_b32_e32 v3, s4
 ; CHECK-NEXT:    s_or_saveexec_b32 s4, -1
-; CHECK-NEXT:    buffer_load_dword v16, 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:    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
@@ -238,8 +238,9 @@
 ; 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, s33, 0
+; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
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
@@ -21,6 +21,7 @@
 ; FIXEDABI-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; FIXEDABI-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; FIXEDABI-NEXT:    s_mov_b64 exec, s[16:17]
+; FIXEDABI-NEXT:    ; implicit-def: $vgpr40
 ; FIXEDABI-NEXT:    v_writelane_b32 v41, s33, 0
 ; FIXEDABI-NEXT:    s_mov_b32 s33, s32
 ; FIXEDABI-NEXT:    s_addk_i32 s32, 0x400
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/call-preserved-registers.ll b/llvm/test/CodeGen/AMDGPU/call-preserved-registers.ll
--- a/llvm/test/CodeGen/AMDGPU/call-preserved-registers.ll
+++ b/llvm/test/CodeGen/AMDGPU/call-preserved-registers.ll
@@ -26,9 +26,9 @@
 ; MUBUF:   buffer_store_dword
 ; FLATSCR: scratch_store_dword
 ; FLATSCR: scratch_store_dword
+; GCN: v_writelane_b32 v41, s33, 0
 ; GCN: v_writelane_b32 v40, s30, 0
 ; GCN: v_writelane_b32 v40, s31, 1
-; GCN: v_writelane_b32 v41, s33, 0
 ; GCN: v_writelane_b32 v40, s34, 2
 ; GCN: v_writelane_b32 v40, s35, 3
 
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
@@ -222,6 +222,7 @@
 ; 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
@@ -277,8 +278,8 @@
 ; MUBUF-NEXT:   buffer_store_dword [[CSR_VGPR:v[0-9]+]], off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
 ; FLATSCR-NEXT: scratch_store_dword off, [[CSR_VGPR:v[0-9]+]], s32 offset:8 ; 4-byte Folded Spill
 ; GCN-NEXT: s_mov_b64 exec, [[COPY_EXEC0]]
-; GCN-COUNT-60: v_writelane_b32 v0
 ; GCN: s_mov_b32 [[TMP_SGPR:s[0-9]+]], s33
+; GCN-COUNT-60: v_writelane_b32 v0
 ; GCN: s_mov_b32 s33, s32
 ; GCN: v_writelane_b32 v0
 ; MUBUF:   buffer_store_dword v41, off, s[0:3], s33 ; 4-byte Folded Spill
@@ -323,10 +324,10 @@
 ; MUBUF-NEXT:   buffer_store_dword [[CSR_VGPR:v[0-9]+]], off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
 ; FLATSCR-NEXT: scratch_store_dword off, [[CSR_VGPR:v[0-9]+]], s32 offset:8 ; 4-byte Folded Spill
 ; GCN-NEXT: s_mov_b64 exec, [[COPY_EXEC0]]
-; GCN-COUNT-61: v_writelane_b32 v0,
+; MUBUF: s_mov_b32 [[FP_COPY:s[0-9]+]], s33
 ; FLATSCR: s_mov_b32 [[FP_COPY:s[0-9]+]], s33
+; GCN-COUNT-61: v_writelane_b32 v0,
 ; FLATSCR-NEXT: s_mov_b32 s33, s32
-; MUBUF: s_mov_b32 [[FP_COPY:s[0-9]+]], s33
 ; MUBUF-NEXT: s_mov_b32 s33, s32
 ; MUBUF:   buffer_store_dword v41, off, s[0:3], s33 ; 4-byte Folded Spill
 ; FLATSCR: scratch_store_dword off, v41, s33 ; 4-byte Folded Spill
@@ -397,6 +398,7 @@
 ; MUBUF-NEXT:   buffer_store_dword [[CSR_VGPR:v[0-9]+]], off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; FLATSCR-NEXT: scratch_store_dword off, [[CSR_VGPR:v[0-9]+]], s32 offset:4 ; 4-byte Folded Spill
 ; GCN-NEXT: s_mov_b64 exec, [[COPY_EXEC0]]
+; GCN-NEXT: ; implicit-def: $vgpr0
 ; GCN-NEXT: s_mov_b32 vcc_lo, s33
 ; GCN-NEXT: s_mov_b32 s33, s32
 ; GCN: v_writelane_b32 [[CSR_VGPR]], s30, 0
@@ -440,6 +442,7 @@
 ; MUBUF-NEXT:   buffer_store_dword [[CSR_VGPR:v[0-9]+]], off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; FLATSCR-NEXT: scratch_store_dword off, [[CSR_VGPR:v[0-9]+]], s32 offset:4 ; 4-byte Folded Spill
 ; GCN-NEXT: s_mov_b64 exec, [[COPY_EXEC0]]
+; GCN-NEXT: ; implicit-def: $vgpr48
 ; GCN-NEXT: s_mov_b32 vcc_lo, s33
 ; GCN-NEXT: s_mov_b32 s33, s32
 ; MUBUF:       s_addk_i32 s32, 0x300{{$}}
@@ -488,10 +491,12 @@
 ; FLATSCR-NEXT: s_add_i32 [[SCRATCH_SGPR:s[0-9]+]], s32, 0x1004
 ; FLATSCR-NEXT: scratch_store_dword off, [[CSR_VGPR:v[0-9]+]], [[SCRATCH_SGPR]] ; 4-byte Folded Spill
 ; GCN-NEXT: s_mov_b64 exec, [[COPY_EXEC0]]
+; MUBUF-NEXT: ; implicit-def: $vgpr48
 ; GCN-NEXT: s_mov_b32 vcc_lo, s33
 ; GCN-DAG:  s_mov_b32 s33, s32
 ; MUBUF-DAG:   s_add_i32 s32, s32, 0x40300{{$}}
 ; FLATSCR-DAG: s_addk_i32 s32, 0x100c{{$}}
+; FLATSCR-DAG: ; implicit-def: $vgpr48
 ; MUBUF-DAG:   buffer_store_dword
 ; FLATSCR-DAG: scratch_store_dword
 
@@ -626,14 +631,14 @@
 ; Make sure that the FP save happens after restoring exec from the same
 ; register.
 ; GCN-LABEL: {{^}}callee_need_to_spill_fp_to_reg:
-; GCN-NOT: v_writelane_b32 v40, s33
+; GCN-NOT: v_writelane_b32 v48, s33
 ; FLATSCR: s_or_saveexec_b64 s[0:1], -1
 ; FLATSCR: s_mov_b64 exec, s[0:1]
 ; FLATSCR: s_mov_b32 s0, s33
 ; FLATSCR: s_mov_b32 s33, s32
 ; FLATSCR: s_mov_b32 s33, s0
 ; FLATSCR: s_or_saveexec_b64 s[0:1], -1
-; GCN-NOT: v_readlane_b32 s33, v40
+; GCN-NOT: v_readlane_b32 s33, v48
 ; GCN:     s_setpc_b64
 define void @callee_need_to_spill_fp_to_reg() #1 {
   call void asm sideeffect "; clobber nonpreserved SGPRs and 64 CSRs",
@@ -664,7 +669,7 @@
 ; MUBUF-NEXT: buffer_store_dword v39, off, s[0:3], [[SCRATCH_SGPR]] ; 4-byte Folded Spill
 ; MUBUF: v_mov_b32_e32 v0, s33
 ; GCN-NOT: v_mov_b32_e32 v0, 0x100c
-; MUBUF-NEXT: s_add_i32 [[SCRATCH_SGPR:s[0-9]+]], s32, 0x40200
+; MUBUF-DAG: s_add_i32 [[SCRATCH_SGPR:s[0-9]+]], s32, 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/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]+}}:
@@ -162,7 +162,7 @@
 ; 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]+}}:
@@ -171,7 +171,7 @@
 ; 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_mov_b64 exec, s[{{[0-9:]+}}]
+; GCN-O0-DAG: 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_]+]]
 ; GCN-O0-NEXT: {{^}}[[THEN_INNER]]:
@@ -181,7 +181,7 @@
 ; 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
@@ -262,7 +262,7 @@
 ; 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_mov_b64 exec, s[{{[0-9:]+}}]
+; GCN-O0-DAG: 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_]+]]
 ; GCN-O0-NEXT: {{^}}[[THEN_OUTER]]:
@@ -272,14 +272,14 @@
 ; 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]+}}:
@@ -289,7 +289,7 @@
 ; 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]+}}:
@@ -359,7 +359,7 @@
 ; 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]+}}:
@@ -421,61 +421,73 @@
 ; 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: buffer_store_dword [[VGPR]], off, s{{\[[0-9]+:[0-9]+\]}}, s32 offset:68
 ; GCN-O0: [[INNER_LOOP:.LBB[0-9]+_[0-9]+]]:
-; 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: buffer_load_dword
-; 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 offset:68
+; 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:      buffer_load_dword
+; 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-NEXT: buffer_store_dword [[RESTORED_VGPR]], off, s{{\[[0-9]+:[0-9]+\]}}, s32 offset:68
+; GCN-O0:      buffer_load_dword [[RESTORED_1_VGPR:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, s32 offset:68
 ; 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_1_VGPR]], s{{[0-9]+}}, [[INNER_LOOP_OUT_EXEC_SPILL_LANE_0:[0-9]+]]
+; GCN-O0-DAG:  v_writelane_b32 [[RESTORED_1_VGPR]], s{{[0-9]+}}, [[INNER_LOOP_OUT_EXEC_SPILL_LANE_1:[0-9]+]]
+; GCN-O0-DAG:  v_writelane_b32 [[RESTORED_1_VGPR]], s{{[0-9]+}}, [[INNER_LOOP_IN_EXEC_SPILL_LANE_0]]
+; GCN-O0-DAG:  v_writelane_b32 [[RESTORED_1_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_1_VGPR]], s{{[0-9]+}}, [[INNER_LOOP_BACK_EDGE_EXEC_SPILL_LANE_0]]
+; GCN-O0-DAG:  v_writelane_b32 [[RESTORED_1_VGPR]], s{{[0-9]+}}, [[INNER_LOOP_BACK_EDGE_EXEC_SPILL_LANE_1]]
+; GCN-O0-NEXT: buffer_store_dword [[RESTORED_1_VGPR]], off, s{{\[[0-9]+:[0-9]+\]}}, s32 offset:68
 ; 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_2_VGPR:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, s32 offset:68
+; GCN-O0-DAG:  v_readlane_b32 s{{[0-9]+}}, [[RESTORED_2_VGPR]], [[INNER_LOOP_OUT_EXEC_SPILL_LANE_0]]
+; GCN-O0-DAG:  v_readlane_b32 s{{[0-9]+}}, [[RESTORED_2_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_2_VGPR]], s{{[0-9]+}}, [[FLOW2_IN_EXEC_SPILL_LANE_0:[0-9]+]]
+; GCN-O0-DAG:  v_writelane_b32 [[RESTORED_2_VGPR]], s{{[0-9]+}}, [[FLOW2_IN_EXEC_SPILL_LANE_1:[0-9]+]]
+; GCN-O0-NEXT: buffer_store_dword [[RESTORED_2_VGPR]], off, s{{\[[0-9]+:[0-9]+\]}}, s32 offset:68
 ; 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_3_VGPR:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, s32 offset:68
+; GCN-O0-DAG:  v_readlane_b32 s{{[0-9]+}}, [[RESTORED_3_VGPR]], [[FLOW2_IN_EXEC_SPILL_LANE_0]]
+; GCN-O0-DAG:  v_readlane_b32 s{{[0-9]+}}, [[RESTORED_3_VGPR]], [[FLOW2_IN_EXEC_SPILL_LANE_1]]
 ; GCN-O0:      s_branch [[FLOW:.LBB[0-9_]+]]
 ; GCN-O0: {{^}}[[FLOW]]:
 ; 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: buffer_store_dword [[RESTORED_3_VGPR]], off, s{{\[[0-9]+:[0-9]+\]}}, s32 offset:68
 ; 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 offset:68
+; 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: buffer_store_dword [[RESTORED_4_VGPR]], off, s{{\[[0-9]+:[0-9]+\]}}, s32 offset:68
 ; GCN-O0: {{^}}[[FLOW3]]:
+; GCN-O0:      buffer_load_dword [[RESTORED_5_VGPR:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, s32 offset:68
 ; 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-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:
@@ -82,7 +82,7 @@
 }
 
 ; 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
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
@@ -31,6 +31,7 @@
 ; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GCN-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[16:17]
+; GCN-NEXT:    ; implicit-def: $vgpr40
 ; GCN-NEXT:    v_writelane_b32 v41, s33, 0
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
@@ -67,6 +68,7 @@
 ; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GCN-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[16:17]
+; GCN-NEXT:    ; implicit-def: $vgpr40
 ; GCN-NEXT:    v_writelane_b32 v41, s33, 0
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
@@ -103,6 +105,7 @@
 ; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GCN-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[16:17]
+; GCN-NEXT:    ; implicit-def: $vgpr40
 ; GCN-NEXT:    v_writelane_b32 v41, s33, 0
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
@@ -139,6 +142,7 @@
 ; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GCN-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[16:17]
+; GCN-NEXT:    ; implicit-def: $vgpr40
 ; GCN-NEXT:    v_writelane_b32 v41, s33, 0
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
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_OR_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, 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
@@ -17,26 +17,27 @@
 ; CHECK-NEXT:  ; %bb.0:
 ; CHECK-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; CHECK-NEXT:    s_or_saveexec_b64 s[16:17], -1
-; CHECK-NEXT:    buffer_store_dword v40, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
+; CHECK-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; CHECK-NEXT:    buffer_store_dword v42, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
 ; CHECK-NEXT:    s_mov_b64 exec, s[16:17]
-; 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:    ; implicit-def: $vgpr41
 ; CHECK-NEXT:    v_writelane_b32 v42, s33, 0
+; 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:    s_mov_b32 s33, s32
 ; CHECK-NEXT:    s_addk_i32 s32, 0x400
-; CHECK-NEXT:    v_writelane_b32 v40, s44, 12
-; CHECK-NEXT:    v_writelane_b32 v40, s46, 13
+; CHECK-NEXT:    v_writelane_b32 v41, s44, 12
+; CHECK-NEXT:    v_writelane_b32 v41, s46, 13
 ; CHECK-NEXT:    s_mov_b64 s[40:41], s[4:5]
 ; CHECK-NEXT:    ;DEBUG_VALUE: dummy:dummy <- undef
 ; CHECK-NEXT:  .Ltmp0:
@@ -44,11 +45,11 @@
 ; 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, 14
+; CHECK-NEXT:    v_writelane_b32 v41, s47, 14
 ; 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:    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, s14
 ; CHECK-NEXT:    s_mov_b32 s43, s13
 ; CHECK-NEXT:    s_mov_b32 s44, s12
@@ -64,33 +65,33 @@
 ; CHECK-NEXT:    s_mov_b32 s12, s44
 ; CHECK-NEXT:    s_mov_b32 s13, s43
 ; CHECK-NEXT:    s_mov_b32 s14, 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, 14
-; CHECK-NEXT:    v_readlane_b32 s46, 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, 14
+; CHECK-NEXT:    v_readlane_b32 s46, 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:    s_addk_i32 s32, 0xfc00
 ; CHECK-NEXT:    v_readlane_b32 s33, v42, 0
 ; CHECK-NEXT:    s_or_saveexec_b64 s[4:5], -1
-; CHECK-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
+; CHECK-NEXT:    buffer_load_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
 ; CHECK-NEXT:    buffer_load_dword v42, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
 ; CHECK-NEXT:    s_mov_b64 exec, s[4:5]
 ; CHECK-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(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
@@ -118,26 +118,19 @@
 ; FLAT_SCR_OPT-NEXT:    s_addc_u32 s3, s3, 0
 ; 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_clause 0x1
 ; 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 s105, 0
-; FLAT_SCR_OPT-NEXT:    scratch_store_dword off, v72, s105
-; 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 s105, 4
-; FLAT_SCR_OPT-NEXT:    v_writelane_b32 v72, s3, 1
-; FLAT_SCR_OPT-NEXT:    scratch_store_dword off, v72, s105 ; 4-byte Folded Spill
-; FLAT_SCR_OPT-NEXT:    s_waitcnt_depctr 0xffe3
-; FLAT_SCR_OPT-NEXT:    s_mov_b32 s105, 0
-; FLAT_SCR_OPT-NEXT:    scratch_load_dword v72, off, s105
-; 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_load_dword vcc_lo, s[0:1], 0x8
+; FLAT_SCR_OPT-NEXT:    ; implicit-def: $vgpr0
+; FLAT_SCR_OPT-NEXT:    s_mov_b32 s104, 4
 ; FLAT_SCR_OPT-NEXT:    ; kill: killed $sgpr0_sgpr1
+; FLAT_SCR_OPT-NEXT:    s_waitcnt lgkmcnt(0)
+; FLAT_SCR_OPT-NEXT:    v_writelane_b32 v0, s2, 0
+; FLAT_SCR_OPT-NEXT:    v_writelane_b32 v0, s3, 1
+; FLAT_SCR_OPT-NEXT:    scratch_store_dword off, v0, s104 ; 4-byte Folded Spill
 ; FLAT_SCR_OPT-NEXT:    ;;#ASMSTART
 ; FLAT_SCR_OPT-NEXT:    ;;#ASMEND
+; FLAT_SCR_OPT-NEXT:    s_mov_b32 s2, 4
 ; FLAT_SCR_OPT-NEXT:    ;;#ASMSTART
 ; FLAT_SCR_OPT-NEXT:    ;;#ASMEND
 ; FLAT_SCR_OPT-NEXT:    ;;#ASMSTART
@@ -220,8 +213,6 @@
 ; FLAT_SCR_OPT-NEXT:    ;;#ASMEND
 ; FLAT_SCR_OPT-NEXT:    ;;#ASMSTART
 ; FLAT_SCR_OPT-NEXT:    ;;#ASMEND
-; FLAT_SCR_OPT-NEXT:    s_waitcnt lgkmcnt(0)
-; FLAT_SCR_OPT-NEXT:    v_mov_b32_e32 v0, vcc_lo
 ; FLAT_SCR_OPT-NEXT:    ;;#ASMSTART
 ; FLAT_SCR_OPT-NEXT:    ;;#ASMEND
 ; FLAT_SCR_OPT-NEXT:    ;;#ASMSTART
@@ -230,48 +221,30 @@
 ; 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_waitcnt_depctr 0xffe3
-; FLAT_SCR_OPT-NEXT:    s_mov_b32 s3, 0
-; 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:    scratch_load_dword v1, off, s2 ; 4-byte Folded Reload
+; FLAT_SCR_OPT-NEXT:    v_mov_b32_e32 v0, vcc_lo
 ; 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
 ;
 ; FLAT_SCR_ARCH-LABEL: test:
 ; FLAT_SCR_ARCH:       ; %bb.0:
+; FLAT_SCR_ARCH-NEXT:    s_clause 0x1
 ; 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 s105, 0
-; FLAT_SCR_ARCH-NEXT:    scratch_store_dword off, v72, s105
-; 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 s105, 4
-; FLAT_SCR_ARCH-NEXT:    v_writelane_b32 v72, s3, 1
-; FLAT_SCR_ARCH-NEXT:    scratch_store_dword off, v72, s105 ; 4-byte Folded Spill
-; FLAT_SCR_ARCH-NEXT:    s_waitcnt_depctr 0xffe3
-; FLAT_SCR_ARCH-NEXT:    s_mov_b32 s105, 0
-; FLAT_SCR_ARCH-NEXT:    scratch_load_dword v72, off, s105
-; 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_load_dword vcc_lo, s[0:1], 0x8
+; FLAT_SCR_ARCH-NEXT:    ; implicit-def: $vgpr0
+; FLAT_SCR_ARCH-NEXT:    s_mov_b32 s104, 4
 ; FLAT_SCR_ARCH-NEXT:    ; kill: killed $sgpr0_sgpr1
+; FLAT_SCR_ARCH-NEXT:    s_waitcnt lgkmcnt(0)
+; FLAT_SCR_ARCH-NEXT:    v_writelane_b32 v0, s2, 0
+; FLAT_SCR_ARCH-NEXT:    v_writelane_b32 v0, s3, 1
+; FLAT_SCR_ARCH-NEXT:    scratch_store_dword off, v0, s104 ; 4-byte Folded Spill
 ; FLAT_SCR_ARCH-NEXT:    ;;#ASMSTART
 ; FLAT_SCR_ARCH-NEXT:    ;;#ASMEND
+; FLAT_SCR_ARCH-NEXT:    s_mov_b32 s2, 4
 ; FLAT_SCR_ARCH-NEXT:    ;;#ASMSTART
 ; FLAT_SCR_ARCH-NEXT:    ;;#ASMEND
 ; FLAT_SCR_ARCH-NEXT:    ;;#ASMSTART
@@ -354,8 +327,6 @@
 ; FLAT_SCR_ARCH-NEXT:    ;;#ASMEND
 ; FLAT_SCR_ARCH-NEXT:    ;;#ASMSTART
 ; FLAT_SCR_ARCH-NEXT:    ;;#ASMEND
-; FLAT_SCR_ARCH-NEXT:    s_waitcnt lgkmcnt(0)
-; FLAT_SCR_ARCH-NEXT:    v_mov_b32_e32 v0, vcc_lo
 ; FLAT_SCR_ARCH-NEXT:    ;;#ASMSTART
 ; FLAT_SCR_ARCH-NEXT:    ;;#ASMEND
 ; FLAT_SCR_ARCH-NEXT:    ;;#ASMSTART
@@ -364,22 +335,11 @@
 ; 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_waitcnt_depctr 0xffe3
-; FLAT_SCR_ARCH-NEXT:    s_mov_b32 s3, 0
-; 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:    scratch_load_dword v1, off, s2 ; 4-byte Folded Reload
+; FLAT_SCR_ARCH-NEXT:    v_mov_b32_e32 v0, vcc_lo
 ; 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
@@ -14,6 +14,7 @@
 ; SPILL-TO-VGPR-NEXT:    buffer_store_dword v40, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; SPILL-TO-VGPR-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
 ; SPILL-TO-VGPR-NEXT:    s_mov_b64 exec, s[4:5]
+; SPILL-TO-VGPR-NEXT:    ; implicit-def: $vgpr40
 ; SPILL-TO-VGPR-NEXT:    v_writelane_b32 v41, s33, 0
 ; SPILL-TO-VGPR-NEXT:    s_mov_b32 s33, s32
 ; SPILL-TO-VGPR-NEXT:    s_addk_i32 s32, 0x400
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
@@ -11,6 +11,8 @@
 ; SDAG-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; SDAG-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; SDAG-NEXT:    s_mov_b64 exec, s[34:35]
+; SDAG-NEXT:    ; implicit-def: $vgpr40
+; SDAG-NEXT:    s_mov_b32 s36, s33
 ; SDAG-NEXT:    v_writelane_b32 v40, s4, 0
 ; SDAG-NEXT:    v_writelane_b32 v40, s5, 1
 ; SDAG-NEXT:    v_writelane_b32 v40, s6, 2
@@ -31,7 +33,6 @@
 ; 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_mov_b32 s36, s33
 ; SDAG-NEXT:    s_mov_b32 s33, s32
 ; SDAG-NEXT:    s_addk_i32 s32, 0x400
 ; SDAG-NEXT:    v_writelane_b32 v40, s24, 20
@@ -91,6 +92,8 @@
 ; GISEL-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GISEL-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GISEL-NEXT:    s_mov_b64 exec, s[34:35]
+; GISEL-NEXT:    ; implicit-def: $vgpr40
+; GISEL-NEXT:    s_mov_b32 s36, s33
 ; GISEL-NEXT:    v_writelane_b32 v40, s4, 0
 ; GISEL-NEXT:    v_writelane_b32 v40, s5, 1
 ; GISEL-NEXT:    v_writelane_b32 v40, s6, 2
@@ -111,7 +114,6 @@
 ; 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_mov_b32 s36, s33
 ; GISEL-NEXT:    s_mov_b32 s33, s32
 ; GISEL-NEXT:    s_addk_i32 s32, 0x400
 ; GISEL-NEXT:    v_writelane_b32 v40, s24, 20
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
@@ -100,6 +100,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -131,15 +132,16 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; 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_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; 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_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:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
@@ -164,15 +166,16 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; 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_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; 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_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:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s31, v40, 1
@@ -201,6 +204,7 @@
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    global_load_ubyte v0, v[0:1], off glc
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -209,6 +213,7 @@
 ; 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]
@@ -234,14 +239,16 @@
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; 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]
@@ -269,14 +276,16 @@
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; 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]
@@ -307,6 +316,7 @@
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    global_load_ubyte v0, v[0:1], off glc
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -315,6 +325,7 @@
 ; 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]
@@ -340,14 +351,16 @@
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; 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]
@@ -375,14 +388,16 @@
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; 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]
@@ -411,6 +426,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -441,15 +457,16 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; 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_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; 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_i8@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_i8@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
@@ -473,15 +490,16 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; 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_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; 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_i8@rel32@lo+4
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_i8@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
@@ -509,6 +527,7 @@
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    global_load_sbyte v0, v[0:1], off glc
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -540,8 +559,9 @@
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
@@ -573,8 +593,9 @@
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
@@ -609,6 +630,7 @@
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    global_load_ubyte v0, v[0:1], off glc
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -640,8 +662,9 @@
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
@@ -673,8 +696,9 @@
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
@@ -707,6 +731,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -737,15 +762,16 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; 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_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; 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_i16@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_i16@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
@@ -769,15 +795,16 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; 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_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; 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_i16@rel32@lo+4
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_i16@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
@@ -805,6 +832,7 @@
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    global_load_ushort v0, v[0:1], off glc
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -836,8 +864,9 @@
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
@@ -869,8 +898,9 @@
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
@@ -905,6 +935,7 @@
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    global_load_ushort v0, v[0:1], off glc
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -936,8 +967,9 @@
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
@@ -969,8 +1001,9 @@
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
@@ -1003,6 +1036,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -1033,15 +1067,16 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; 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:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; 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_i32@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_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
@@ -1065,15 +1100,16 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; 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:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; 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_i32@rel32@lo+4
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_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
@@ -1099,6 +1135,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -1130,8 +1167,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; 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:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
@@ -1163,8 +1201,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; 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:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
@@ -1201,6 +1240,7 @@
 ; 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:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -1232,15 +1272,16 @@
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; 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_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    global_load_dwordx4 v[0:3], v[0:1], off
-; 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_v2i64@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v2i64@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
@@ -1266,15 +1307,16 @@
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
 ; 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_writelane_b32 v41, s33, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    global_load_dwordx4 v[0:3], v[0:1], off
-; 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_v2i64@rel32@lo+4
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_v2i64@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
@@ -1301,6 +1343,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -1334,8 +1377,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; 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
@@ -1369,8 +1413,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; 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
@@ -1409,6 +1454,7 @@
 ; 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:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -1442,8 +1488,9 @@
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; 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_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    global_load_dwordx4 v[0:3], v[0:1], off
@@ -1478,8 +1525,9 @@
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
 ; 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_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    global_load_dwordx4 v[0:3], v[0:1], off
@@ -1520,6 +1568,7 @@
 ; 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:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -1555,8 +1604,9 @@
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; 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
@@ -1593,8 +1643,9 @@
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
 ; 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
@@ -1633,6 +1684,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -1663,15 +1715,16 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; 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:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; 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_f16@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_f16@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
@@ -1695,15 +1748,16 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; 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:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; 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_f16@rel32@lo+4
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_f16@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
@@ -1729,6 +1783,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -1759,15 +1814,16 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; 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:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; 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_f32@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_f32@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
@@ -1791,15 +1847,16 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; 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:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; 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_f32@rel32@lo+4
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_f32@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
@@ -1825,6 +1882,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -1856,8 +1914,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; 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_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
@@ -1889,8 +1948,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; 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_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
@@ -1924,6 +1984,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -1956,8 +2017,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; 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_writelane_b32 v41, s33, 0
@@ -1990,8 +2052,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; 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_writelane_b32 v41, s33, 0
@@ -2026,6 +2089,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -2060,8 +2124,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; 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
@@ -2096,8 +2161,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; 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
@@ -2134,6 +2200,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -2165,8 +2232,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; 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:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
@@ -2198,8 +2266,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; 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:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
@@ -2233,6 +2302,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -2266,8 +2336,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; 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
@@ -2301,8 +2372,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; 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
@@ -2338,6 +2410,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -2373,8 +2446,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; 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
@@ -2410,8 +2484,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; 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
@@ -2450,6 +2525,7 @@
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    global_load_dword v0, v[0:1], off
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -2480,8 +2556,9 @@
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
@@ -2512,8 +2589,9 @@
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
@@ -2547,6 +2625,7 @@
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    global_load_dwordx2 v[0:1], v[0:1], off
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -2577,8 +2656,9 @@
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
@@ -2609,8 +2689,9 @@
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
@@ -2644,6 +2725,7 @@
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    global_load_dwordx2 v[0:1], v[0:1], off
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -2674,8 +2756,9 @@
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
@@ -2706,8 +2789,9 @@
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
@@ -2740,6 +2824,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -2771,8 +2856,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; 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:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
@@ -2804,8 +2890,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; 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:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
@@ -2839,6 +2926,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -2870,8 +2958,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; 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:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
@@ -2903,8 +2992,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; 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:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
@@ -2939,6 +3029,7 @@
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    global_load_dwordx2 v[0:1], v[0:1], off
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -2969,8 +3060,9 @@
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
@@ -3001,8 +3093,9 @@
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
@@ -3035,6 +3128,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -3066,8 +3160,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; 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:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
@@ -3099,8 +3194,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; 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:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
@@ -3135,6 +3231,7 @@
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    global_load_dword v0, v[0:1], off
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -3165,8 +3262,9 @@
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
@@ -3197,8 +3295,9 @@
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
@@ -3232,6 +3331,7 @@
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    global_load_dwordx2 v[0:1], v[0:1], off
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -3262,8 +3362,9 @@
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
@@ -3294,8 +3395,9 @@
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
@@ -3328,6 +3430,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -3359,8 +3462,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; 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_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
@@ -3392,8 +3496,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; 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_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
@@ -3427,6 +3532,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -3459,8 +3565,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; 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_writelane_b32 v41, s33, 0
@@ -3493,8 +3600,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; 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_writelane_b32 v41, s33, 0
@@ -3529,6 +3637,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -3562,8 +3671,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; 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
@@ -3597,8 +3707,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; 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
@@ -3635,6 +3746,7 @@
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    global_load_dwordx4 v[0:3], v[0:1], off
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -3665,8 +3777,9 @@
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
@@ -3697,8 +3810,9 @@
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
@@ -3731,6 +3845,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -3764,8 +3879,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; 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
@@ -3799,8 +3915,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; 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
@@ -3836,6 +3953,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -3870,8 +3988,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; 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
@@ -3906,8 +4025,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; 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
@@ -3946,6 +4066,7 @@
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; 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:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -3980,8 +4101,9 @@
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
@@ -4017,8 +4139,9 @@
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    s_waitcnt lgkmcnt(0)
@@ -4056,6 +4179,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -4093,8 +4217,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; 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
@@ -4132,8 +4257,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; 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
@@ -4175,6 +4301,7 @@
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; 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:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -4211,8 +4338,9 @@
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
@@ -4250,8 +4378,9 @@
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    s_waitcnt lgkmcnt(0)
@@ -4293,6 +4422,7 @@
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; 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:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -4334,8 +4464,9 @@
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
@@ -4377,8 +4508,9 @@
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    s_waitcnt lgkmcnt(0)
@@ -4425,6 +4557,7 @@
 ; 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:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
@@ -4468,8 +4601,9 @@
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    global_load_dword v33, v[0:1], off
@@ -4514,8 +4648,9 @@
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    global_load_dword v33, v[0:1], off
@@ -4560,33 +4695,34 @@
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
-; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
+; GFX9-NEXT:    buffer_store_dword v42, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v43, off, s[0:3], s32 offset:12 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v43, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
+; 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 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:    s_addk_i32 s32, 0xf800
 ; GFX9-NEXT:    v_readlane_b32 s33, v43, 0
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
-; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
+; GFX9-NEXT:    buffer_load_dword v42, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
 ; GFX9-NEXT:    buffer_load_dword v43, off, s[0:3], s32 offset:12 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
@@ -4597,36 +4733,37 @@
 ; GFX10-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
-; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
+; GFX10-NEXT:    buffer_store_dword v42, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
 ; GFX10-NEXT:    buffer_store_dword v43, off, s[0:3], s32 offset:12 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; GFX10-NEXT:    v_writelane_b32 v43, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
-; 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_mov_b32_e32 v42, v1
-; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
+; GFX10-NEXT:    v_mov_b32_e32 v41, v1
+; GFX10-NEXT:    v_writelane_b32 v42, s31, 1
 ; 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:    s_addk_i32 s32, 0xfc00
 ; GFX10-NEXT:    v_readlane_b32 s33, v43, 0
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    s_clause 0x1
-; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:8
+; GFX10-NEXT:    buffer_load_dword v42, off, s[0:3], s32 offset:8
 ; GFX10-NEXT:    buffer_load_dword v43, off, s[0:3], s32 offset:12
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
@@ -4638,36 +4775,37 @@
 ; GFX10-SCRATCH-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-SCRATCH-NEXT:    s_or_saveexec_b32 s0, -1
-; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v40, s32 offset:8 ; 4-byte Folded Spill
+; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v42, s32 offset:8 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v43, s32 offset:12 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v43, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
-; 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_mov_b32_e32 v42, v1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 1
+; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v41, v1
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v42, s31, 1
 ; 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:    s_addk_i32 s32, 0xffe0
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s33, v43, 0
 ; GFX10-SCRATCH-NEXT:    s_or_saveexec_b32 s0, -1
 ; GFX10-SCRATCH-NEXT:    s_clause 0x1
-; GFX10-SCRATCH-NEXT:    scratch_load_dword v40, off, s32 offset:8
+; GFX10-SCRATCH-NEXT:    scratch_load_dword v42, off, s32 offset:8
 ; GFX10-SCRATCH-NEXT:    scratch_load_dword v43, off, s32 offset:12
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
@@ -4688,6 +4826,7 @@
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; 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:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -4722,8 +4861,9 @@
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
@@ -4759,8 +4899,9 @@
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    s_waitcnt lgkmcnt(0)
@@ -4801,6 +4942,7 @@
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; 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
@@ -4836,11 +4978,12 @@
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 8
 ; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    s_addk_i32 s32, 0x400
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 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:    s_addk_i32 s32, 0x400
 ; 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
@@ -4872,11 +5015,12 @@
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v1, 8
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
+; 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:    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:    s_add_i32 s32, s32, 32
 ; 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
@@ -4916,6 +5060,7 @@
 ; 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
@@ -4958,17 +5103,18 @@
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 8
 ; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    s_addk_i32 s32, 0x400
 ; 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, s30, 0
 ; GFX10-NEXT:    v_lshrrev_b32_e64 v1, 5, s33
-; GFX10-NEXT:    s_addk_i32 s32, 0x400
-; 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_sret_struct_i8_i32_byval_struct_i8_i32@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_sret_struct_i8_i32_byval_struct_i8_i32@rel32@hi+12
 ; GFX10-NEXT:    v_add_nc_u32_e32 v0, 8, v0
+; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    s_clause 0x1
 ; GFX10-NEXT:    buffer_load_ubyte v0, off, s[0:3], s33 offset:8
@@ -5009,9 +5155,10 @@
 ; 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
@@ -5063,6 +5210,7 @@
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; 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:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -5115,8 +5263,9 @@
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
@@ -5170,8 +5319,9 @@
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
 ; 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:    v_writelane_b32 v41, s33, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 1
@@ -5231,6 +5381,8 @@
 ; 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
@@ -5261,7 +5413,6 @@
 ; 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]
@@ -5326,8 +5477,9 @@
 ; 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
@@ -5421,8 +5573,9 @@
 ; 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
@@ -5515,6 +5668,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -5546,15 +5700,16 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; 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_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; 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_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:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
@@ -5579,15 +5734,16 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; 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_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; 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_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:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s31, v40, 1
@@ -5614,8 +5770,9 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
-; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
+; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
@@ -5646,15 +5803,16 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_movk_i32 s4, 0x7b
-; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i8_inreg@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_i8_inreg@rel32@hi+12
+; 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:    v_readlane_b32 s31, v40, 2
@@ -5680,15 +5838,16 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_movk_i32 s4, 0x7b
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 1
 ; 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
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_i8_inreg@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 2
 ; GFX10-SCRATCH-NEXT:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s31, v40, 2
@@ -5716,8 +5875,9 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
-; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
+; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
@@ -5748,15 +5908,16 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_movk_i32 s4, 0x7b
-; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i16_inreg@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_i16_inreg@rel32@hi+12
+; 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:    v_readlane_b32 s31, v40, 2
@@ -5782,15 +5943,16 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_movk_i32 s4, 0x7b
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 1
 ; 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
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_i16_inreg@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 2
 ; GFX10-SCRATCH-NEXT:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s31, v40, 2
@@ -5818,8 +5980,9 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
-; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
+; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
@@ -5850,15 +6013,16 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 42
-; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i32_inreg@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_i32_inreg@rel32@hi+12
+; 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:    v_readlane_b32 s31, v40, 2
@@ -5884,15 +6048,16 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 42
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 1
 ; 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
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_i32_inreg@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 2
 ; GFX10-SCRATCH-NEXT:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s31, v40, 2
@@ -5920,9 +6085,10 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
@@ -5955,16 +6121,17 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_movk_i32 s4, 0x7b
-; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-NEXT:    s_mov_b32 s5, 0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i64_inreg@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_i64_inreg@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-NEXT:    s_mov_b32 s5, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 3
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
@@ -5992,16 +6159,17 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_movk_i32 s4, 0x7b
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 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
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_i64_inreg@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 3
 ; GFX10-SCRATCH-NEXT:    s_swappc_b64 s[30:31], s[0:1]
@@ -6031,10 +6199,11 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_mov_b64 s[34:35], 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:    s_mov_b64 s[34:35], 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s7, 3
 ; GFX9-NEXT:    s_load_dwordx4 s[4:7], s[34:35], 0x0
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
@@ -6070,8 +6239,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_mov_b64 s[34:35], 0
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
@@ -6111,8 +6281,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_mov_b64 s[0:1], 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
@@ -6155,11 +6326,12 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s33, 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 v40, s7, 3
-; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 4
@@ -6196,16 +6368,17 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 1
-; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-NEXT:    s_mov_b32 s5, 2
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2i64_inreg@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v2i64_inreg@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-NEXT:    s_mov_b32 s5, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-NEXT:    s_mov_b32 s6, 3
 ; GFX10-NEXT:    v_writelane_b32 v40, s7, 3
@@ -6239,16 +6412,17 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 2
 ; 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
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_v2i64_inreg@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 2
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s6, 3
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s7, 3
@@ -6284,10 +6458,11 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_mov_b64 s[34:35], 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:    s_mov_b64 s[34:35], 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s7, 3
 ; GFX9-NEXT:    s_load_dwordx4 s[4:7], s[34:35], 0x0
 ; GFX9-NEXT:    v_writelane_b32 v40, s8, 4
@@ -6329,8 +6504,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_mov_b64 s[34:35], 0
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
@@ -6376,8 +6552,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_mov_b64 s[0:1], 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
@@ -6428,11 +6605,12 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_mov_b64 s[34:35], 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 v40, s7, 3
-; GFX9-NEXT:    s_mov_b64 s[34:35], 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s8, 4
 ; GFX9-NEXT:    s_load_dwordx4 s[4:7], s[34:35], 0x0
 ; GFX9-NEXT:    v_writelane_b32 v40, s9, 5
@@ -6479,8 +6657,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_mov_b64 s[34:35], 0
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
@@ -6532,8 +6711,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_mov_b64 s[0:1], 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
@@ -6589,8 +6769,9 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
-; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
+; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
@@ -6621,15 +6802,16 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_movk_i32 s4, 0x4400
-; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_f16_inreg@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_f16_inreg@rel32@hi+12
+; 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:    v_readlane_b32 s31, v40, 2
@@ -6655,15 +6837,16 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_movk_i32 s4, 0x4400
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 1
 ; 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
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_f16_inreg@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 2
 ; GFX10-SCRATCH-NEXT:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s31, v40, 2
@@ -6691,8 +6874,9 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
-; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
+; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
@@ -6723,15 +6907,16 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 4.0
-; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_f32_inreg@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_f32_inreg@rel32@hi+12
+; 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:    v_readlane_b32 s31, v40, 2
@@ -6757,15 +6942,16 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 4.0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 1
 ; 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
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_f32_inreg@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 2
 ; GFX10-SCRATCH-NEXT:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s31, v40, 2
@@ -6793,9 +6979,10 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
@@ -6828,16 +7015,17 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 1.0
-; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-NEXT:    s_mov_b32 s5, 2.0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2f32_inreg@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v2f32_inreg@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-NEXT:    s_mov_b32 s5, 2.0
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 3
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
@@ -6865,16 +7053,17 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 1.0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 2.0
 ; 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
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_v2f32_inreg@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 2.0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 3
 ; GFX10-SCRATCH-NEXT:    s_swappc_b64 s[30:31], s[0:1]
@@ -6904,10 +7093,11 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s33, 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, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 3
@@ -6942,16 +7132,17 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 1.0
-; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-NEXT:    s_mov_b32 s5, 2.0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3f32_inreg@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v3f32_inreg@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-NEXT:    s_mov_b32 s5, 2.0
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-NEXT:    s_mov_b32 s6, 4.0
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 3
@@ -6982,16 +7173,17 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 1.0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 2.0
 ; 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
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_v3f32_inreg@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 2.0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s6, 4.0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 3
@@ -7024,12 +7216,13 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s33, 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 v40, s7, 3
 ; GFX9-NEXT:    v_writelane_b32 v40, s8, 4
-; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 5
@@ -7068,16 +7261,17 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 1.0
-; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-NEXT:    s_mov_b32 s5, 2.0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v5f32_inreg@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v5f32_inreg@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-NEXT:    s_mov_b32 s5, 2.0
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-NEXT:    s_mov_b32 s6, 4.0
 ; GFX10-NEXT:    v_writelane_b32 v40, s7, 3
@@ -7114,16 +7308,17 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 1.0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 2.0
 ; 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
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_v5f32_inreg@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 2.0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s6, 4.0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s7, 3
@@ -7162,9 +7357,10 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
@@ -7197,16 +7393,17 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 0
-; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-NEXT:    s_mov_b32 s5, 0x40100000
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_f64_inreg@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_f64_inreg@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-NEXT:    s_mov_b32 s5, 0x40100000
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 3
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
@@ -7234,16 +7431,17 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 0x40100000
 ; 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
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_f64_inreg@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 0x40100000
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 3
 ; GFX10-SCRATCH-NEXT:    s_swappc_b64 s[30:31], s[0:1]
@@ -7273,11 +7471,12 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s33, 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 v40, s7, 3
-; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 4
@@ -7314,16 +7513,17 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 0
-; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-NEXT:    s_mov_b32 s5, 2.0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2f64_inreg@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v2f64_inreg@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-NEXT:    s_mov_b32 s5, 2.0
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-NEXT:    s_mov_b32 s6, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s7, 3
@@ -7357,16 +7557,17 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 2.0
 ; 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
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_v2f64_inreg@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 2.0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s6, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s7, 3
@@ -7402,13 +7603,14 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s33, 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 v40, s7, 3
 ; GFX9-NEXT:    v_writelane_b32 v40, s8, 4
 ; GFX9-NEXT:    v_writelane_b32 v40, s9, 5
-; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 6
@@ -7449,16 +7651,17 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 0
-; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-NEXT:    s_mov_b32 s5, 2.0
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3f64_inreg@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v3f64_inreg@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-NEXT:    s_mov_b32 s5, 2.0
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-NEXT:    s_mov_b32 s6, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s7, 3
@@ -7498,16 +7701,17 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 2.0
 ; 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
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_v3f64_inreg@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 2.0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s6, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s7, 3
@@ -7549,9 +7753,10 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    s_load_dword s4, s[34:35], 0x0
-; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
@@ -7581,15 +7786,16 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_load_dword s4, s[34:35], 0x0
-; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2i16_inreg@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v2i16_inreg@rel32@hi+12
+; 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:    v_readlane_b32 s31, v40, 2
@@ -7615,15 +7821,16 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_load_dword s4, s[0:1], 0x0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 1
 ; 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
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_v2i16_inreg@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 2
 ; GFX10-SCRATCH-NEXT:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s31, v40, 2
@@ -7652,10 +7859,11 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; 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:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
@@ -7686,8 +7894,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
@@ -7722,8 +7931,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
@@ -7761,10 +7971,11 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; 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:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
@@ -7795,8 +8006,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
@@ -7831,8 +8043,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
@@ -7870,9 +8083,10 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
@@ -7905,16 +8119,17 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 0x20001
-; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-NEXT:    s_mov_b32 s5, 3
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3i16_inreg@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v3i16_inreg@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-NEXT:    s_mov_b32 s5, 3
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 3
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
@@ -7942,16 +8157,17 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 0x20001
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 3
 ; 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
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_v3i16_inreg@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 3
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 3
 ; GFX10-SCRATCH-NEXT:    s_swappc_b64 s[30:31], s[0:1]
@@ -7981,9 +8197,10 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
@@ -8016,16 +8233,17 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 0x40003c00
-; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-NEXT:    s_movk_i32 s5, 0x4400
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3f16_inreg@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v3f16_inreg@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-NEXT:    s_movk_i32 s5, 0x4400
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 3
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
@@ -8053,16 +8271,17 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 0x40003c00
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-SCRATCH-NEXT:    s_movk_i32 s5, 0x4400
 ; 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
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_v3f16_inreg@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-SCRATCH-NEXT:    s_movk_i32 s5, 0x4400
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 3
 ; GFX10-SCRATCH-NEXT:    s_swappc_b64 s[30:31], s[0:1]
@@ -8092,10 +8311,11 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; 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:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
@@ -8126,8 +8346,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
@@ -8162,8 +8383,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
@@ -8201,9 +8423,10 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
@@ -8236,16 +8459,17 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 0x20001
-; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-NEXT:    s_mov_b32 s5, 0x40003
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v4i16_inreg@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v4i16_inreg@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-NEXT:    s_mov_b32 s5, 0x40003
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 3
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
@@ -8273,16 +8497,17 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 0x20001
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 0x40003
 ; 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
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_v4i16_inreg@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 0x40003
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 3
 ; GFX10-SCRATCH-NEXT:    s_swappc_b64 s[30:31], s[0:1]
@@ -8312,9 +8537,10 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    s_load_dword s4, s[34:35], 0x0
-; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
@@ -8344,15 +8570,16 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_load_dword s4, s[34:35], 0x0
-; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2f16_inreg@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v2f16_inreg@rel32@hi+12
+; 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:    v_readlane_b32 s31, v40, 2
@@ -8378,15 +8605,16 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_load_dword s4, s[0:1], 0x0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 1
 ; 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
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_v2f16_inreg@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 2
 ; GFX10-SCRATCH-NEXT:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s31, v40, 2
@@ -8415,10 +8643,11 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; 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:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
@@ -8449,8 +8678,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
@@ -8485,8 +8715,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
@@ -8524,9 +8755,10 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
@@ -8559,16 +8791,17 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 1
-; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-NEXT:    s_mov_b32 s5, 2
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2i32_inreg@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v2i32_inreg@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-NEXT:    s_mov_b32 s5, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 3
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
@@ -8596,16 +8829,17 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 2
 ; 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
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_v2i32_inreg@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 2
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 3
 ; GFX10-SCRATCH-NEXT:    s_swappc_b64 s[30:31], s[0:1]
@@ -8635,10 +8869,11 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s33, 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, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 3
@@ -8673,16 +8908,17 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 3
-; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-NEXT:    s_mov_b32 s5, 4
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3i32_inreg@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v3i32_inreg@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-NEXT:    s_mov_b32 s5, 4
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-NEXT:    s_mov_b32 s6, 5
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 3
@@ -8713,16 +8949,17 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 3
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 4
 ; 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
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_v3i32_inreg@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 4
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s6, 5
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 3
@@ -8755,11 +8992,12 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s33, 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 v40, s7, 3
-; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 4
@@ -8796,16 +9034,17 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 3
-; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-NEXT:    s_mov_b32 s5, 4
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3i32_i32_inreg@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v3i32_i32_inreg@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-NEXT:    s_mov_b32 s5, 4
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-NEXT:    s_mov_b32 s6, 5
 ; GFX10-NEXT:    v_writelane_b32 v40, s7, 3
@@ -8839,16 +9078,17 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 3
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 4
 ; 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
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_v3i32_i32_inreg@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 4
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s6, 5
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s7, 3
@@ -8884,12 +9124,13 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s33, 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 v40, s7, 3
 ; GFX9-NEXT:    s_load_dwordx4 s[4:7], s[34:35], 0x0
-; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 4
@@ -8922,8 +9163,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
@@ -8962,8 +9204,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
@@ -9005,11 +9248,12 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s33, 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 v40, s7, 3
-; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 4
@@ -9046,16 +9290,17 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 1
-; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-NEXT:    s_mov_b32 s5, 2
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v4i32_inreg@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v4i32_inreg@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-NEXT:    s_mov_b32 s5, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-NEXT:    s_mov_b32 s6, 3
 ; GFX10-NEXT:    v_writelane_b32 v40, s7, 3
@@ -9089,16 +9334,17 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 2
 ; 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
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_v4i32_inreg@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 2
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s6, 3
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s7, 3
@@ -9134,12 +9380,13 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s33, 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 v40, s7, 3
 ; GFX9-NEXT:    v_writelane_b32 v40, s8, 4
-; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 5
@@ -9178,16 +9425,17 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 1
-; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-NEXT:    s_mov_b32 s5, 2
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v5i32_inreg@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v5i32_inreg@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-NEXT:    s_mov_b32 s5, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-NEXT:    s_mov_b32 s6, 3
 ; GFX10-NEXT:    v_writelane_b32 v40, s7, 3
@@ -9224,16 +9472,17 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 2
 ; 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
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_v5i32_inreg@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 2
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s6, 3
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s7, 3
@@ -9272,10 +9521,11 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; 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_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX9-NEXT:    v_writelane_b32 v40, s7, 3
 ; GFX9-NEXT:    v_writelane_b32 v40, s8, 4
 ; GFX9-NEXT:    v_writelane_b32 v40, s9, 5
@@ -9320,8 +9570,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
@@ -9370,8 +9621,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
@@ -9424,6 +9676,8 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s33, 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
@@ -9432,7 +9686,6 @@
 ; 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:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 8
@@ -9477,16 +9730,17 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
+; GFX10-NEXT:    ; implicit-def: $vgpr40
+; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 1
-; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-NEXT:    s_mov_b32 s5, 2
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v8i32_inreg@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v8i32_inreg@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-NEXT:    s_mov_b32 s5, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-NEXT:    s_mov_b32 s6, 3
 ; GFX10-NEXT:    v_writelane_b32 v40, s7, 3
@@ -9532,16 +9786,17 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, 1
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 2
 ; 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
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, external_void_func_v8i32_inreg@rel32@hi+12
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-SCRATCH-NEXT:    s_mov_b32 s5, 2
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s6, 3
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s7, 3
@@ -9589,6 +9844,8 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s6, 2
@@ -9600,7 +9857,6 @@
 ; GFX9-NEXT:    v_writelane_b32 v40, s12, 8
 ; GFX9-NEXT:    v_writelane_b32 v40, s13, 9
 ; 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
 ; GFX9-NEXT:    v_writelane_b32 v40, s16, 12
 ; GFX9-NEXT:    v_writelane_b32 v40, s17, 13
@@ -9653,8 +9909,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
@@ -9719,8 +9976,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
@@ -9789,6 +10047,8 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s6, 2
@@ -9805,7 +10065,6 @@
 ; GFX9-NEXT:    v_writelane_b32 v40, s17, 13
 ; GFX9-NEXT:    v_writelane_b32 v40, s18, 14
 ; 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
 ; GFX9-NEXT:    v_writelane_b32 v40, s21, 17
 ; GFX9-NEXT:    v_writelane_b32 v40, s22, 18
@@ -9897,8 +10156,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
@@ -10008,8 +10268,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
@@ -10119,6 +10380,8 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s6, 2
@@ -10134,7 +10397,6 @@
 ; 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:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX9-NEXT:    v_writelane_b32 v40, s19, 15
 ; GFX9-NEXT:    v_writelane_b32 v40, s20, 16
 ; GFX9-NEXT:    v_writelane_b32 v40, s21, 17
@@ -10232,8 +10494,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
@@ -10348,8 +10611,9 @@
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v41, s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x0
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
@@ -10469,6 +10733,7 @@
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; 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 v40, s31, 1
@@ -10505,8 +10770,9 @@
 ; 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:    s_getpc_b64 s[34:35]
 ; GFX10-NEXT:    s_add_u32 s34, s34, stack_passed_f64_arg@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s35, s35, stack_passed_f64_arg@rel32@hi+12
@@ -10540,9 +10806,10 @@
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 0
-; GFX10-SCRATCH-NEXT:    scratch_load_dwordx2 v[32:33], off, s33
+; GFX10-SCRATCH-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 32
+; GFX10-SCRATCH-NEXT:    scratch_load_dwordx2 v[32:33], off, s33
+; 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, stack_passed_f64_arg@rel32@lo+4
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, stack_passed_f64_arg@rel32@hi+12
@@ -10583,6 +10850,7 @@
 ; 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
@@ -10651,11 +10919,12 @@
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; 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
@@ -10722,13 +10991,14 @@
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 16
+; 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:    scratch_store_dwordx4 off, v[0:3], s32
 ; 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
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v3, 1
-; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v4, 1
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v5, 1
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v6, 2
 ; GFX10-SCRATCH-NEXT:    v_mov_b32_e32 v7, 2
@@ -10814,6 +11084,7 @@
 ; 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
@@ -10888,6 +11159,7 @@
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 12
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 13
 ; GFX10-NEXT:    v_mov_b32_e32 v3, 14
+; 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
@@ -10965,9 +11237,10 @@
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; 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
@@ -11054,6 +11327,7 @@
 ; 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
@@ -11128,6 +11402,7 @@
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 0x41400000
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 0x41500000
 ; GFX10-NEXT:    v_mov_b32_e32 v3, 0x41600000
+; 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
@@ -11205,9 +11480,10 @@
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
 ; 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
@@ -12,9 +12,10 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
+; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
@@ -48,8 +49,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
@@ -89,6 +91,7 @@
 ; GFX9-NEXT:    s_or_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:    ;;#ASMSTART
@@ -110,6 +113,7 @@
 ; 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:    ;;#ASMSTART
@@ -141,8 +145,9 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
-; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
+; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
@@ -180,8 +185,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
@@ -222,34 +228,35 @@
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
-; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
+; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v42, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr41
 ; GFX9-NEXT:    v_writelane_b32 v42, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 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:    v_writelane_b32 v41, s30, 0
+; 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:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v42, 0
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
-; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
+; GFX9-NEXT:    buffer_load_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
 ; GFX9-NEXT:    buffer_load_dword v42, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
@@ -260,36 +267,37 @@
 ; GFX10-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
-; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
+; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    buffer_store_dword v42, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr41
 ; GFX10-NEXT:    v_writelane_b32 v42, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v41, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 ; 4-byte Folded Spill
-; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
+; 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_mov_b32_e32 v41, v31
+; GFX10-NEXT:    v_writelane_b32 v41, s31, 1
+; GFX10-NEXT:    v_mov_b32_e32 v40, v31
 ; 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:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v42, 0
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    s_clause 0x1
-; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:4
+; GFX10-NEXT:    buffer_load_dword v41, off, s[0:3], s32 offset:4
 ; GFX10-NEXT:    buffer_load_dword v42, off, s[0:3], s32 offset:8
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
@@ -310,8 +318,9 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
-; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
+; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
@@ -349,17 +358,18 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; 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, s30, 1
 ; 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, s31, 2
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
@@ -394,8 +404,9 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
-; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
+; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
@@ -433,17 +444,18 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    s_getpc_b64 s[36:37]
 ; GFX10-NEXT:    s_add_u32 s36, s36, external_void_func_void@rel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s37, s37, external_void_func_void@rel32@hi+12
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 1
 ; 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, s31, 2
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[36:37]
@@ -478,6 +490,7 @@
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v42, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr41
 ; GFX9-NEXT:    v_writelane_b32 v42, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -515,15 +528,16 @@
 ; GFX10-NEXT:    buffer_store_dword v42, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; GFX10-NEXT:    v_writelane_b32 v41, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr41
 ; GFX10-NEXT:    v_writelane_b32 v42, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v41, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
-; GFX10-NEXT:    v_writelane_b32 v41, s31, 1
 ; 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
@@ -557,6 +571,7 @@
 ; GFX9-NEXT:    s_or_saveexec_b64 s[4:5], -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:    ; implicit-def: $vgpr0
 ; GFX9-NEXT:    v_writelane_b32 v0, s33, 0
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; clobber
@@ -576,6 +591,7 @@
 ; 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, s4
+; GFX10-NEXT:    ; implicit-def: $vgpr0
 ; GFX10-NEXT:    v_writelane_b32 v0, s33, 0
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; clobber
@@ -599,6 +615,7 @@
 ; GFX9-NEXT:    s_or_saveexec_b64 s[4:5], -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:    ; implicit-def: $vgpr0
 ; GFX9-NEXT:    v_writelane_b32 v0, s34, 0
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; clobber
@@ -618,6 +635,7 @@
 ; 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, s4
+; GFX10-NEXT:    ; implicit-def: $vgpr0
 ; GFX10-NEXT:    v_writelane_b32 v0, s34, 0
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; clobber
@@ -642,6 +660,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -671,8 +690,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
@@ -704,6 +724,7 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
@@ -733,8 +754,9 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    s_getpc_b64 s[34:35]
@@ -766,8 +788,9 @@
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
-; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX9-NEXT:    ; implicit-def: $vgpr40
 ; GFX9-NEXT:    v_writelane_b32 v41, s33, 0
+; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
@@ -804,17 +827,18 @@
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr40
 ; GFX10-NEXT:    v_writelane_b32 v41, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; 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, s30, 1
 ; 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, s31, 2
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
@@ -845,16 +869,17 @@
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
-; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
+; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v42, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
-; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX9-NEXT:    ; implicit-def: $vgpr41
 ; GFX9-NEXT:    v_writelane_b32 v42, s33, 0
+; GFX9-NEXT:    v_writelane_b32 v41, s4, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
-; 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:    v_writelane_b32 v41, s30, 1
+; 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
@@ -862,7 +887,7 @@
 ; 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
@@ -871,16 +896,16 @@
 ; 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:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v42, 0
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
-; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
+; GFX9-NEXT:    buffer_load_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
 ; GFX9-NEXT:    buffer_load_dword v42, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
@@ -891,44 +916,45 @@
 ; GFX10-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
-; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
+; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    buffer_store_dword v42, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
-; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr41
 ; GFX10-NEXT:    v_writelane_b32 v42, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v41, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 ; 4-byte Folded Spill
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 1
+; 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 v41, s30, 1
 ; GFX10-NEXT:    s_mov_b32 s4, s40
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; def v32
 ; GFX10-NEXT:    ;;#ASMEND
-; GFX10-NEXT:    v_mov_b32_e32 v41, v32
-; GFX10-NEXT:    v_writelane_b32 v40, s31, 2
+; 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 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:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v42, 0
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    s_clause 0x1
-; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:4
+; GFX10-NEXT:    buffer_load_dword v41, off, s[0:3], s32 offset:4
 ; GFX10-NEXT:    buffer_load_dword v42, off, s[0:3], s32 offset:8
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
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
@@ -33,6 +33,7 @@
 ; 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)
@@ -61,8 +62,9 @@
 ; 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:    v_writelane_b32 v1, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr1
 ; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
+; GFX10-NEXT:    v_writelane_b32 v1, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v1, s31, 1
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
@@ -112,6 +114,7 @@
 ; 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)
@@ -140,8 +143,9 @@
 ; 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:    v_writelane_b32 v1, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr1
 ; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
+; GFX10-NEXT:    v_writelane_b32 v1, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v1, s31, 1
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
@@ -191,6 +195,7 @@
 ; 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)
@@ -219,8 +224,9 @@
 ; 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:    v_writelane_b32 v1, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr1
 ; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
+; GFX10-NEXT:    v_writelane_b32 v1, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v1, s31, 1
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
@@ -272,6 +278,7 @@
 ; 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)
@@ -300,8 +307,9 @@
 ; 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:    v_writelane_b32 v2, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr2
 ; 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]
@@ -1380,8 +1388,9 @@
 ; 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]
@@ -1410,9 +1419,10 @@
 ; 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]
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
@@ -400,6 +400,7 @@
 ; GCN-NEXT:    v_writelane_b32 v41, s33, 0
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; 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
@@ -480,6 +481,7 @@
 ; GISEL-NEXT:    v_writelane_b32 v41, s33, 0
 ; GISEL-NEXT:    s_mov_b32 s33, s32
 ; 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
@@ -564,6 +566,7 @@
 ; GCN-NEXT:    v_writelane_b32 v41, s33, 0
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; 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
@@ -647,6 +650,7 @@
 ; GISEL-NEXT:    v_writelane_b32 v41, s33, 0
 ; GISEL-NEXT:    s_mov_b32 s33, s32
 ; 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
@@ -732,6 +736,7 @@
 ; GCN-NEXT:    v_writelane_b32 v41, s33, 0
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; 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
@@ -814,6 +819,7 @@
 ; GISEL-NEXT:    v_writelane_b32 v41, s33, 0
 ; GISEL-NEXT:    s_mov_b32 s33, s32
 ; 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
@@ -901,6 +907,7 @@
 ; GCN-NEXT:    v_writelane_b32 v41, s33, 0
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; 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
@@ -992,6 +999,7 @@
 ; GISEL-NEXT:    v_writelane_b32 v41, s33, 0
 ; GISEL-NEXT:    s_mov_b32 s33, s32
 ; 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
@@ -1093,6 +1101,7 @@
 ; GCN-NEXT:    s_mov_b32 s5, s33
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; 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
@@ -1187,6 +1196,7 @@
 ; GISEL-NEXT:    s_mov_b32 s5, s33
 ; GISEL-NEXT:    s_mov_b32 s33, s32
 ; 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
@@ -1280,96 +1290,97 @@
 ; 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 v40, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
+; GCN-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_mov_b32 s10, s33
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; 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 s8, v1
 ; GCN-NEXT:    v_readfirstlane_b32 s9, v2
 ; GCN-NEXT:    v_cmp_eq_u64_e32 vcc, s[8:9], v[1:2]
 ; GCN-NEXT:    s_and_saveexec_b64 s[6:7], 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[8:9]
 ; GCN-NEXT:    ; implicit-def: $vgpr1_vgpr2
 ; GCN-NEXT:    s_xor_b64 exec, exec, s[6:7]
 ; 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_addk_i32 s32, 0xfc00
 ; GCN-NEXT:    s_mov_b32 s33, s10
 ; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
-; GCN-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
+; GCN-NEXT:    buffer_load_dword v41, off, s[0:3], s32 offset:4 ; 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]
@@ -1378,96 +1389,97 @@
 ; GISEL:       ; %bb.0:
 ; GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GISEL-NEXT:    s_or_saveexec_b64 s[4:5], -1
-; GISEL-NEXT:    buffer_store_dword v40, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
+; GISEL-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GISEL-NEXT:    s_mov_b64 exec, s[4:5]
 ; GISEL-NEXT:    s_mov_b32 s10, s33
 ; GISEL-NEXT:    s_mov_b32 s33, s32
 ; 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_addk_i32 s32, 0xfc00
 ; GISEL-NEXT:    s_mov_b32 s33, s10
 ; GISEL-NEXT:    s_or_saveexec_b64 s[4:5], -1
-; GISEL-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
+; GISEL-NEXT:    buffer_load_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
 ; GISEL-NEXT:    s_mov_b64 exec, s[4:5]
 ; GISEL-NEXT:    s_waitcnt vmcnt(0)
 ; GISEL-NEXT:    s_setpc_b64 s[30:31]
@@ -1489,6 +1501,7 @@
 ; GCN-NEXT:    s_mov_b32 s10, s33
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; 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
@@ -1585,6 +1598,7 @@
 ; GISEL-NEXT:    s_mov_b32 s10, s33
 ; GISEL-NEXT:    s_mov_b32 s33, s32
 ; 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
@@ -1686,6 +1700,7 @@
 ; GCN-NEXT:    s_mov_b32 s10, s33
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; 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
@@ -1779,6 +1794,7 @@
 ; GISEL-NEXT:    s_mov_b32 s10, s33
 ; GISEL-NEXT:    s_mov_b32 s33, s32
 ; 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
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
@@ -262,7 +262,10 @@
 ; 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
+; 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: buffer_load_dword
 ; 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]+]]
@@ -288,6 +291,8 @@
 ; 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
 ; 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_format_x [[RES:v[0-9]+]], [[IDX]], s[[[S0]]:[[S3]]], {{.*}} idxen
 ; W64-O0: s_waitcnt vmcnt(0)
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
@@ -189,45 +189,46 @@
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    s_or_saveexec_b64 s[4:5], -1
-; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 offset:12 ; 4-byte Folded Spill
+; GFX9-NEXT:    buffer_store_dword v42, off, s[0:3], s32 offset:12 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v44, off, s[0:3], s32 offset:16 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[4:5]
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX9-NEXT:    ; implicit-def: $vgpr42
 ; GFX9-NEXT:    v_writelane_b32 v44, s33, 0
+; GFX9-NEXT:    v_writelane_b32 v42, s30, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x800
-; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    v_writelane_b32 v40, s34, 2
+; GFX9-NEXT:    v_writelane_b32 v42, s31, 1
+; GFX9-NEXT:    v_writelane_b32 v42, s34, 2
 ; 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, s35, 3
+; GFX9-NEXT:    v_writelane_b32 v42, s35, 3
 ; GFX9-NEXT:    s_load_dwordx2 s[34:35], 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_and_b32_e32 v43, 0xffffff, 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:    v_and_b32_e32 v43, 0xffffff, v40
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
-; GFX9-NEXT:    v_mad_u32_u24 v41, v42, v41, v43
-; GFX9-NEXT:    v_mov_b32_e32 v0, v41
+; GFX9-NEXT:    v_mad_u32_u24 v40, v41, v40, v43
+; GFX9-NEXT:    v_mov_b32_e32 v0, v40
 ; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
-; GFX9-NEXT:    v_add_u32_e32 v0, v41, v43
+; GFX9-NEXT:    v_add_u32_e32 v0, v40, v43
 ; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; 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 s35, 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 s35, 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:    s_addk_i32 s32, 0xf800
 ; GFX9-NEXT:    v_readlane_b32 s33, v44, 0
 ; GFX9-NEXT:    s_or_saveexec_b64 s[4:5], -1
-; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:12 ; 4-byte Folded Reload
+; GFX9-NEXT:    buffer_load_dword v42, off, s[0:3], s32 offset:12 ; 4-byte Folded Reload
 ; GFX9-NEXT:    buffer_load_dword v44, off, s[0:3], s32 offset:16 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
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
@@ -28,14 +28,16 @@
 ; CHECK:       ; %bb.0: ; %bb
 ; CHECK-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; CHECK-NEXT:    s_or_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:8 ; 4-byte Folded Spill
 ; CHECK-NEXT:    s_mov_b64 exec, s[4:5]
 ; CHECK-NEXT:    s_mov_b32 s6, s33
 ; CHECK-NEXT:    s_mov_b32 s33, s32
 ; 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:    buffer_store_dword v0, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; 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
@@ -44,16 +46,18 @@
 ; 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:    buffer_load_dword v0, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
 ; 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_add_i32 s32, s32, 0xfffffc00
 ; CHECK-NEXT:    s_mov_b32 s33, s6
 ; CHECK-NEXT:    s_or_saveexec_b64 s[4:5], -1
-; CHECK-NEXT:    buffer_load_dword v1, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
+; CHECK-NEXT:    buffer_load_dword v0, off, s[0:3], s32 offset:8 ; 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]
@@ -91,20 +95,21 @@
 ; CHECK:       ; %bb.0: ; %bb
 ; CHECK-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; CHECK-NEXT:    s_or_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_or_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:    buffer_load_dword v0, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
 ; CHECK-NEXT:    s_mov_b64 exec, s[6:7]
 ; CHECK-NEXT:    s_setpc_b64 s[4:5]
 bb:
@@ -158,6 +163,7 @@
 ; CHECK-NEXT:    s_mov_b32 s6, s33
 ; CHECK-NEXT:    s_mov_b32 s33, s32
 ; 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]
@@ -187,13 +193,15 @@
 ; CHECK:       ; %bb.0: ; %entry
 ; CHECK-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; CHECK-NEXT:    s_or_saveexec_b64 s[4:5], -1
-; CHECK-NEXT:    buffer_store_dword v2, off, s[0:3], s32 ; 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:    s_mov_b32 s7, s33
 ; CHECK-NEXT:    s_mov_b32 s33, s32
 ; 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:    buffer_store_dword v0, off, s[0:3], s33 ; 4-byte Folded Spill
 ; 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
@@ -202,12 +210,14 @@
 ; 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:    buffer_load_dword v1, off, s[0:3], s33 ; 4-byte Folded Reload
+; 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_add_i32 s32, s32, 0xfffffc00
 ; CHECK-NEXT:    s_mov_b32 s33, s7
 ; CHECK-NEXT:    s_or_saveexec_b64 s[4:5], -1
-; CHECK-NEXT:    buffer_load_dword v2, off, s[0:3], s32 ; 4-byte Folded Reload
+; CHECK-NEXT:    buffer_load_dword v0, off, s[0:3], s32 offset:4 ; 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]
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
@@ -14,16 +14,18 @@
 ; CHECK-NEXT:  ; %bb.0: ; %entry
 ; CHECK-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; CHECK-NEXT:    s_or_saveexec_b64 s[16:17], -1
-; CHECK-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
-; CHECK-NEXT:    buffer_store_dword v41, 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:    buffer_store_dword v40, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
 ; CHECK-NEXT:    s_mov_b64 exec, s[16:17]
-; CHECK-NEXT:    v_writelane_b32 v41, s33, 0
+; CHECK-NEXT:    v_writelane_b32 v40, s33, 0
 ; CHECK-NEXT:    s_mov_b32 s33, s32
 ; CHECK-NEXT:    s_add_i32 s32, s32, 0x400
+; CHECK-NEXT:    ; implicit-def: $vgpr0
 ; CHECK-NEXT:  .Ltmp0:
 ; CHECK-NEXT:    .loc 0 31 3 prologue_end ; lane-info.cpp:31:3
-; CHECK-NEXT:    v_writelane_b32 v40, s30, 0
-; CHECK-NEXT:    v_writelane_b32 v40, s31, 1
+; CHECK-NEXT:    v_writelane_b32 v0, s30, 0
+; CHECK-NEXT:    v_writelane_b32 v0, s31, 1
+; CHECK-NEXT:    buffer_store_dword v0, off, s[0:3], s33 ; 4-byte Folded Spill
 ; CHECK-NEXT:    s_getpc_b64 s[16:17]
 ; CHECK-NEXT:    s_add_u32 s16, s16, _ZL13sleep_foreverv@gotpcrel32@lo+4
 ; CHECK-NEXT:    s_addc_u32 s17, s17, _ZL13sleep_foreverv@gotpcrel32@hi+12
@@ -35,14 +37,16 @@
 ; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
 ; CHECK-NEXT:    s_swappc_b64 s[30:31], s[16:17]
 ; CHECK-NEXT:  .Ltmp1:
+; CHECK-NEXT:    buffer_load_dword v0, off, s[0:3], s33 ; 4-byte Folded Reload
 ; 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:    s_waitcnt vmcnt(0)
+; CHECK-NEXT:    v_readlane_b32 s31, v0, 1
+; CHECK-NEXT:    v_readlane_b32 s30, v0, 0
 ; CHECK-NEXT:    s_add_i32 s32, s32, 0xfffffc00
-; CHECK-NEXT:    v_readlane_b32 s33, v41, 0
+; CHECK-NEXT:    v_readlane_b32 s33, v40, 0
 ; CHECK-NEXT:    s_or_saveexec_b64 s[4:5], -1
-; CHECK-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
-; CHECK-NEXT:    buffer_load_dword v41, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
+; CHECK-NEXT:    buffer_load_dword v0, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
+; CHECK-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:8 ; 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]
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 +107,273 @@
 ; 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:    buffer_store_dword v0, off, s[92:95], 0 offset:12 ; 4-byte Folded Spill
 ; 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:    buffer_store_dword v0, off, s[92:95], 0 offset:8 ; 4-byte Folded Spill
 ; 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:    buffer_store_dword v0, off, s[92:95], 0 offset:4 ; 4-byte Folded Spill
 ; 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:    buffer_load_dword v0, off, s[92:95], 0 offset:4 ; 4-byte Folded Reload
+; GCN-NEXT:    buffer_load_dword v1, off, s[92:95], 0 offset:12 ; 4-byte Folded Reload
+; GCN-NEXT:    buffer_load_dword v2, off, s[92:95], 0 offset:8 ; 4-byte Folded Reload
+; 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 +458,17 @@
 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 s28, SCRATCH_RSRC_DWORD0
+; GCN-NEXT:    s_mov_b32 s29, SCRATCH_RSRC_DWORD1
+; GCN-NEXT:    s_mov_b32 s30, -1
+; GCN-NEXT:    s_mov_b32 s31, 0xe8f000
+; GCN-NEXT:    s_add_u32 s28, s28, s3
+; GCN-NEXT:    s_addc_u32 s29, s29, 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 +542,33 @@
 ; 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:    buffer_store_dword v0, off, s[28:31], 0 offset:8 ; 4-byte Folded Spill
 ; 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:    buffer_store_dword v0, off, s[28:31], 0 offset:4 ; 4-byte Folded Spill
 ; 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:    buffer_load_dword v0, off, s[28:31], 0 offset:8 ; 4-byte Folded Reload
+; GCN-NEXT:    buffer_load_dword v1, off, s[28:31], 0 offset:4 ; 4-byte Folded Reload
+; 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 +714,168 @@
 ; 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:    buffer_store_dword v0, off, s[52:55], 0 offset:8 ; 4-byte Folded Spill
 ; 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:    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_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:    buffer_load_dword v0, off, s[52:55], 0 offset:4 ; 4-byte Folded Reload
+; GCN-NEXT:    buffer_load_dword v1, off, s[52:55], 0 offset:8 ; 4-byte Folded Reload
+; 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 +941,144 @@
 ; 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:    buffer_store_dword v0, off, s[52:55], 0 offset:8 ; 4-byte Folded Spill
 ; 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:    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_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:    buffer_load_dword v1, off, s[52:55], 0 offset:4 ; 4-byte Folded Reload
+; GCN-NEXT:    buffer_load_dword v2, off, s[52:55], 0 offset:8 ; 4-byte Folded Reload
+; 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 +1088,24 @@
 ; 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-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 +25,171 @@
 ; 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:    buffer_store_dword v0, off, s[0:3], 0 offset:8 ; 4-byte Folded Spill
 ; 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:    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_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:    buffer_load_dword v0, off, s[0:3], 0 offset:4 ; 4-byte Folded Reload
+; GCN-NEXT:    buffer_load_dword v1, off, s[0:3], 0 offset:8 ; 4-byte Folded Reload
+; 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-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
@@ -1,4 +1,4 @@
-# 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
@@ -21,7 +21,7 @@
 
 
 # CHECK: BUFFER_STORE_DWORD_OFFSET killed $vgpr1, $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr32, 0, 0, 0, 0, implicit $exec :: (store (s32) into %stack.2, addrspace 5)
-# CHECK-NEXT: $vgpr1 = V_WRITELANE_B32 killed $sgpr10, 0, undef $vgpr1
+# CHECK-NEXT: $vgpr1 = V_WRITELANE_B32 $sgpr10, 0, undef $vgpr1
 # CHECK-NEXT: BUFFER_STORE_DWORD_OFFSET killed $vgpr1, $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr32, 8, 0, 0, 0, implicit $exec :: (store (s32) into %stack.0, addrspace 5)
 # CHECK-NEXT: $vgpr1 = BUFFER_LOAD_DWORD_OFFSET $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr32, 0, 0, 0, 0, implicit $exec :: (load (s32) from %stack.2, addrspace 5)
 
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
@@ -5,16 +5,23 @@
   ret void
 }
 
+; SGPRs are now spilled into virtual VGPRs and regalloc takes care of finding
+; physical VGPRs and this test shouldn't take the high regsiter for spill lanes.
 ; GCN-LABEL: {{^}}spill_sgpr_with_no_lower_vgpr_available:
-; GCN:  buffer_store_dword v255, off, s[0:3], s32
+; GCN: buffer_store_dword [[LANE_VGPR:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, s{{[0-9]+}} offset:[[IDX_OFF:[0-9]+]] ; 4-byte Folded Spill
+; GCN-NOT:  buffer_store_dword v255, off, s[0:3], s32
 ; GCN:  s_mov_b32 [[TMP_SGPR:s[0-9]+]], s33
-; GCN:  v_writelane_b32 v255, s30, 0
-; GCN:  v_writelane_b32 v255, s31, 1
+; GCN:  v_writelane_b32 [[LANE_VGPR]], s30, 0
+; GCN:  v_writelane_b32 [[LANE_VGPR]], s31, 1
+; GCN-NOT:  v_writelane_b32 v255, s30, 0
+; GCN-NOT:  v_writelane_b32 v255, s31, 1
 ; GCN:  s_swappc_b64 s[30:31], s[4:5]
-; GCN:  v_readlane_b32 s31, v255, 1
-; GCN:  v_readlane_b32 s30, v255, 0
+; GCN:  v_readlane_b32 s31, [[LANE_VGPR]], 1
+; GCN:  v_readlane_b32 s30, [[LANE_VGPR]], 0
+; GCN-NOT:  v_readlane_b32 s31, v255, 1
+; GCN-NOT:  v_readlane_b32 s30, v255, 0
 ; GCN:  s_mov_b32 s33, [[TMP_SGPR]]
-; GCN: ; NumVgprs: 256
+; GCN: ; NumVgprs: 255
 
 define void @spill_sgpr_with_no_lower_vgpr_available() #0 {
   %alloca = alloca i32, align 4, addrspace(5)
@@ -52,13 +59,18 @@
 }
 
 ; GCN-LABEL: {{^}}spill_to_lowest_available_vgpr:
-; GCN:  buffer_store_dword v254, off, s[0:3], s32
+; GCN: buffer_store_dword [[LANE_VGPR:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, s{{[0-9]+}} offset:[[IDX_OFF:[0-9]+]] ; 4-byte Folded Spill
+; GCN-NOT:  buffer_store_dword v254, off, s[0:3], s32
 ; GCN:  s_mov_b32 [[TMP_SGPR:s[0-9]+]], s33
-; GCN:  v_writelane_b32 v254, s30, 0
-; GCN:  v_writelane_b32 v254, s31, 1
+; GCN:  v_writelane_b32 [[LANE_VGPR]], s30, 0
+; GCN:  v_writelane_b32 [[LANE_VGPR]], s31, 1
+; GCN-NOT:  v_writelane_b32 v254, s30, 0
+; GCN-NOT:  v_writelane_b32 v254, s31, 1
 ; GCN:  s_swappc_b64 s[30:31], s[4:5]
-; GCN:  v_readlane_b32 s31, v254, 1
-; GCN:  v_readlane_b32 s30, v254, 0
+; GCN:  v_readlane_b32 s31, [[LANE_VGPR]], 1
+; GCN:  v_readlane_b32 s30, [[LANE_VGPR]], 0
+; GCN-NOT:  v_readlane_b32 s31, v254, 1
+; GCN-NOT:  v_readlane_b32 s30, v254, 0
 ; GCN:  s_mov_b32 s33, [[TMP_SGPR]]
 
 define void @spill_to_lowest_available_vgpr() #0 {
@@ -97,10 +109,14 @@
 }
 
 ; GCN-LABEL: {{^}}spill_sgpr_with_sgpr_uses:
+; GCN: buffer_store_dword [[LANE_VGPR:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, s{{[0-9]+}} offset:[[IDX_OFF:[0-9]+]] ; 4-byte Folded Spill
 ; GCN-NOT:  buffer_store_dword v255, off, s[0:3], s32
+; GCN-NOT:  buffer_store_dword v254, off, s[0:3], s32
 ; GCN: ; def s4
-; GCN: v_writelane_b32 v254, s4, 0
-; GCN: v_readlane_b32 s4, v254, 0
+; GCN:  v_writelane_b32 [[LANE_VGPR]], s4, 0
+; GCN:  v_readlane_b32 s4, [[LANE_VGPR]], 0
+; GCN-NOT: v_writelane_b32 v254, s4, 0
+; GCN-NOT: v_readlane_b32 s4, v254, 0
 ; GCN: ; use s4
 
 define void @spill_sgpr_with_sgpr_uses() #0 {
@@ -192,22 +208,22 @@
 
 ; GCN-LABEL: {{^}}spill_sgpr_no_free_vgpr:
 ; GCN: v_writelane_b32 [[A:v[0-9]+]], s34, 0
-; GCN: buffer_store_dword [[A]], off, s[0:3], s32
-; GCN: v_writelane_b32 [[B:v[0-9]+]], s35, 0
-; GCN: buffer_store_dword [[B]], off, s[0:3], s32
-; GCN: v_writelane_b32 [[C:v[0-9]+]], s36, 0
-; GCN: buffer_store_dword [[C]], off, s[0:3], s32
-; GCN: v_writelane_b32 [[D:v[0-9]+]], s37, 0
-; GCN: buffer_store_dword [[D]], off, s[0:3], s32
+; GCN: v_writelane_b32 [[A]], s35, 1
+; GCN: v_writelane_b32 [[A]], s36, 2
+; GCN: v_writelane_b32 [[A]], s37, 3
+; GCN: buffer_store_dword [[A]], off, s{{\[[0-9]+:[0-9]+\]}}, s{{[0-9]+}} offset:[[IDX_OFF:[0-9]+]] ; 4-byte Folded Spill
+; GCN-NOT: v_writelane_b32 v{{[0-9]+}}, s35, 0
+; GCN-NOT: v_writelane_b32 v{{[0-9]+}}, s36, 0
+; GCN-NOT: v_writelane_b32 v{{[0-9]+}}, s37, 0
 ; GCN: #ASMEND
-; GCN: buffer_load_dword [[E:v[0-9]+]]
-; GCN: v_readlane_b32 s37, [[E]], 0
-; GCN: buffer_load_dword [[F:v[0-9]+]]
-; GCN: v_readlane_b32 s36, [[F]], 0
-; GCN: buffer_load_dword [[G:v[0-9]+]]
-; GCN: v_readlane_b32 s35, [[G]], 0
-; GCN: buffer_load_dword [[H:v[0-9]+]]
-; GCN: v_readlane_b32 s34, [[H]], 0
+; GCN-NOT: v_readlane_b32 s37, v{{[0-9]+}}, 0
+; GCN-NOT: v_readlane_b32 s36, v{{[0-9]+}}, 0
+; GCN-NOT: v_readlane_b32 s35, v{{[0-9]+}}, 0
+; GCN: buffer_load_dword [[B:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, s{{[0-9]+}}  offset:[[IDX_OFF]] ; 4-byte Folded Reload
+; GCN: v_readlane_b32 s37, [[B]], 3
+; GCN: v_readlane_b32 s36, [[B]], 2
+; GCN: v_readlane_b32 s35, [[B]], 1
+; GCN: v_readlane_b32 s34, [[B]], 0
 
 define void @spill_sgpr_no_free_vgpr(<4 x i32> addrspace(1)* %out, <4 x i32> addrspace(1)* %in) #0 {
   %a = load <4 x i32>, <4 x i32> addrspace(1)* %in
@@ -282,15 +298,15 @@
 }
 
 ; GCN-LABEL: {{^}}spill_sgpr_no_free_vgpr_ipra:
-; GCN: v_writelane_b32 v0, s30, 0
-; GCN: buffer_store_dword v0, off
-; GCN: v_writelane_b32 v0, s31, 0
-; GCN: buffer_store_dword v0, off
+; GCN: v_writelane_b32 [[A:v[0-9]+]], s30, 0
+; GCN: v_writelane_b32 [[A]], s31, 1
+; GCN: buffer_store_dword [[A]], off, s{{\[[0-9]+:[0-9]+\]}}, s{{[0-9]+}} offset:[[IDX_OFF:[0-9]+]] ; 4-byte Folded Spill
+; GCN-NOT: v_writelane_b32 v{{[0-9]+}}, s31, 0
 ; GCN: swappc
-; GCN: buffer_load_dword v0, off
-; GCN: v_readlane_b32 s31, v0, 0
-; GCN: buffer_load_dword v0, off
-; GCN: v_readlane_b32 s30, v0, 0
+; GCN-NOT: v_readlane_b32 s31, v{{[0-9]+}}, 0
+; GCN: buffer_load_dword [[B:v[0-9]+]], off, s{{\[[0-9]+:[0-9]+\]}}, s{{[0-9]+}}  offset:[[IDX_OFF]] ; 4-byte Folded Reload
+; GCN: v_readlane_b32 s31, [[B]], 1
+; GCN: v_readlane_b32 s30, [[B]], 0
 define void @spill_sgpr_no_free_vgpr_ipra() #0 {
   call void @child_function_ipra()
   ret void
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
@@ -7,15 +7,13 @@
 
 ; 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, s100
 ; 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
@@ -210,15 +210,15 @@
 ; 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
@@ -2,19 +2,19 @@
 
 ; GCN-LABEL: {{^}}spill_csr_s5_copy:
 ; GCN: s_or_saveexec_b64
-; GCN-NEXT: buffer_store_dword v40, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
-; GCN-NEXT: buffer_store_dword v41, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
+; GCN-NEXT: buffer_store_dword v0, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
+; GCN-NEXT: buffer_store_dword v40, off, s[0:3], s32 offset:12 ; 4-byte Folded Spill
 ; GCN-NEXT: s_mov_b64 exec
-; GCN: v_writelane_b32 v41, s33, 0
+; GCN: v_writelane_b32 v40, s33, 0
 ; 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 s33, v41, 0
+; GCN: v_readlane_b32 s33, v40, 0
 ; GCN: s_or_saveexec_b64
-; GCN-NEXT: buffer_load_dword v40, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
-; GCN-NEXT: buffer_load_dword v41, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
+; GCN-NEXT: buffer_load_dword v0, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
+; GCN-NEXT: buffer_load_dword v40, off, s[0:3], s32 offset:12 ; 4-byte Folded Reload
 ; GCN: s_mov_b64 exec
 ; GCN: s_setpc_b64
 define void @spill_csr_s5_copy() #0 {
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, 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_OR_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, 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, 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_OR_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, 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, 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, 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, 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, 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, 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, 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, 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, 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, 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, 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, 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, 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-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
@@ -10,10 +10,12 @@
   bb.0:
     liveins: $sgpr50
     ; CHECK-LABEL: name: spill_csr_sgpr_argument
-    ; CHECK: liveins: $sgpr50, $vgpr0
-    ; CHECK: $vgpr0 = V_WRITELANE_B32 $sgpr50, 0, $vgpr0
-    ; CHECK: S_NOP 0, implicit $sgpr50
-    ; CHECK: $sgpr50 = S_MOV_B32 0
+    ; CHECK: liveins: $sgpr50
+    ; CHECK-NEXT: {{  $}}
+    ; 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
     $sgpr50 = S_MOV_B32 0
 
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
@@ -1,7 +1,8 @@
 ; 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:
@@ -18,31 +19,19 @@
 ; 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_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:    ;;#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_waitcnt vmcnt(0)
 ; GFX10-NEXT:    v_readlane_b32 s0, v0, 0
@@ -50,10 +39,6 @@
 ; 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
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-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
@@ -13,6 +13,7 @@
 ; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -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:    ; implicit-def: $vgpr0
 ; GCN-NEXT:    v_writelane_b32 v0, s35, 0
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ;;#ASMEND
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
@@ -20,8 +20,9 @@
 ; GCN-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GCN-NEXT:    buffer_store_dword v1, off, s[0:3], s32 ; 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_mov_b32 s36, s33
+; GCN-NEXT:    v_writelane_b32 v1, s4, 0
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
 ; GCN-NEXT:    v_writelane_b32 v1, s30, 1
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
@@ -13,6 +13,7 @@
 ; GCN-NEXT:    v_writelane_b32 v41, s33, 0
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; 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
@@ -185,35 +186,36 @@
 ; GCN:       ; %bb.0: ; %bb
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GCN-NEXT:    s_or_saveexec_b64 s[16:17], -1
-; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s32 offset:20 ; 4-byte Folded Spill
-; GCN-NEXT:    buffer_store_dword v46, off, s[0:3], s32 offset:24 ; 4-byte Folded Spill
+; GCN-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:20 ; 4-byte Folded Spill
+; GCN-NEXT:    buffer_store_dword v45, off, s[0:3], s32 offset:24 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[16:17]
-; GCN-NEXT:    v_writelane_b32 v46, s33, 0
+; GCN-NEXT:    v_writelane_b32 v45, s33, 0
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; 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, s48, 15
-; GCN-NEXT:    v_writelane_b32 v40, s49, 16
-; 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, s48, 15
+; GCN-NEXT:    v_writelane_b32 v0, s49, 16
+; GCN-NEXT:    v_mov_b32_e32 v40, v31
 ; GCN-NEXT:    s_mov_b32 s44, s14
 ; GCN-NEXT:    s_mov_b32 s45, s13
 ; GCN-NEXT:    s_mov_b32 s46, s12
@@ -224,24 +226,24 @@
 ; 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[34:35], 0, v44
+; GCN-NEXT:    v_cmp_eq_f32_e64 s[34:35], 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
@@ -250,8 +252,8 @@
 ; 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
@@ -282,7 +284,7 @@
 ; GCN-NEXT:    s_mov_b32 s12, s46
 ; GCN-NEXT:    s_mov_b32 s13, s45
 ; GCN-NEXT:    s_mov_b32 s14, 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
@@ -297,10 +299,10 @@
 ; 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
@@ -11,7 +11,7 @@
 ; GFX9:       ; %bb.0: ; %main_body
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    s_or_saveexec_b64 s[4:5], -1
-; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 offset:16 ; 4-byte Folded Spill
+; GFX9-NEXT:    buffer_store_dword v44, off, s[0:3], s32 offset:16 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v45, off, s[0:3], s32 offset:20 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX9-NEXT:    s_mov_b32 s4, 0
@@ -29,10 +29,10 @@
 ; GFX9-NEXT:    s_mov_b32 s9, s4
 ; GFX9-NEXT:    s_mov_b32 s10, s4
 ; GFX9-NEXT:    s_mov_b32 s11, s4
-; 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
@@ -41,30 +41,31 @@
 ; 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:    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:    s_addk_i32 s32, 0xf800
 ; GFX9-NEXT:    v_readlane_b32 s33, v45, 0
 ; GFX9-NEXT:    s_or_saveexec_b64 s[4:5], -1
-; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:16 ; 4-byte Folded Reload
+; GFX9-NEXT:    buffer_load_dword v44, off, s[0:3], s32 offset:16 ; 4-byte Folded Reload
 ; GFX9-NEXT:    buffer_load_dword v45, off, s[0:3], s32 offset:20 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
@@ -75,7 +76,7 @@
 ; GFX10-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    s_or_saveexec_b32 s4, -1
-; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s32 offset:16 ; 4-byte Folded Spill
+; GFX10-NEXT:    buffer_store_dword v44, off, s[0:3], s32 offset:16 ; 4-byte Folded Spill
 ; GFX10-NEXT:    buffer_store_dword v45, off, s[0:3], s32 offset:20 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s4
@@ -94,10 +95,10 @@
 ; GFX10-NEXT:    s_mov_b32 s9, s4
 ; GFX10-NEXT:    s_mov_b32 s10, s4
 ; GFX10-NEXT:    s_mov_b32 s11, s4
-; 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
@@ -106,33 +107,34 @@
 ; 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:    s_waitcnt_depctr 0xffe3
 ; 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:    s_addk_i32 s32, 0xfc00
 ; GFX10-NEXT:    v_readlane_b32 s33, v45, 0
 ; GFX10-NEXT:    s_or_saveexec_b32 s4, -1
 ; GFX10-NEXT:    s_clause 0x1
-; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:16
+; GFX10-NEXT:    buffer_load_dword v44, off, s[0:3], s32 offset:16
 ; GFX10-NEXT:    buffer_load_dword v45, off, s[0:3], s32 offset:20
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s4
@@ -167,32 +169,33 @@
 ; GFX9:       ; %bb.0: ; %main_body
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    s_or_saveexec_b64 s[4:5], -1
-; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 offset:20 ; 4-byte Folded Spill
+; GFX9-NEXT:    buffer_store_dword v45, off, s[0:3], s32 offset:20 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v46, off, s[0:3], s32 offset:24 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[4:5]
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
-; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    v_writelane_b32 v40, s36, 2
-; GFX9-NEXT:    v_writelane_b32 v40, s37, 3
-; GFX9-NEXT:    v_writelane_b32 v40, s38, 4
-; GFX9-NEXT:    v_writelane_b32 v40, s39, 5
-; GFX9-NEXT:    v_writelane_b32 v40, s40, 6
-; GFX9-NEXT:    v_writelane_b32 v40, s41, 7
+; GFX9-NEXT:    ; implicit-def: $vgpr45
 ; GFX9-NEXT:    v_writelane_b32 v46, s33, 0
+; GFX9-NEXT:    v_writelane_b32 v45, s30, 0
+; GFX9-NEXT:    v_writelane_b32 v45, s31, 1
+; GFX9-NEXT:    v_writelane_b32 v45, s36, 2
+; GFX9-NEXT:    v_writelane_b32 v45, s37, 3
+; GFX9-NEXT:    v_writelane_b32 v45, s38, 4
+; GFX9-NEXT:    v_writelane_b32 v45, s39, 5
+; GFX9-NEXT:    v_writelane_b32 v45, s40, 6
+; GFX9-NEXT:    v_writelane_b32 v45, s41, 7
 ; GFX9-NEXT:    s_mov_b32 s33, s32
-; GFX9-NEXT:    v_writelane_b32 v40, s42, 8
+; GFX9-NEXT:    v_writelane_b32 v45, s42, 8
 ; GFX9-NEXT:    s_mov_b32 s36, 0
-; 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_writelane_b32 v40, s43, 9
-; 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:    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_writelane_b32 v45, s43, 9
+; 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:    s_mov_b32 s37, s36
 ; GFX9-NEXT:    s_mov_b32 s38, s36
 ; GFX9-NEXT:    s_mov_b32 s39, s36
@@ -200,7 +203,7 @@
 ; GFX9-NEXT:    s_mov_b32 s41, s36
 ; GFX9-NEXT:    s_mov_b32 s42, s36
 ; GFX9-NEXT:    s_mov_b32 s43, s36
-; GFX9-NEXT:    image_gather4_c_b_cl v[0:3], v[41:45], s[36:43], s[4:7] dmask:0x1
+; GFX9-NEXT:    image_gather4_c_b_cl v[0:3], v[40:44], s[36:43], s[4:7] dmask:0x1
 ; GFX9-NEXT:    s_addk_i32 s32, 0x800
 ; GFX9-NEXT:    s_getpc_b64 s[4:5]
 ; GFX9-NEXT:    s_add_u32 s4, s4, extern_func@gotpcrel32@lo+4
@@ -210,27 +213,27 @@
 ; 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[36:43], s[4:7] dmask:0x1
+; GFX9-NEXT:    image_gather4_c_b_cl v[0:3], v[40:44], s[36:43], 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 s43, v40, 9
-; GFX9-NEXT:    v_readlane_b32 s42, v40, 8
-; GFX9-NEXT:    v_readlane_b32 s41, v40, 7
-; GFX9-NEXT:    v_readlane_b32 s40, v40, 6
-; GFX9-NEXT:    v_readlane_b32 s39, v40, 5
-; GFX9-NEXT:    v_readlane_b32 s38, v40, 4
-; GFX9-NEXT:    v_readlane_b32 s37, v40, 3
-; GFX9-NEXT:    v_readlane_b32 s36, v40, 2
-; 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 s43, v45, 9
+; GFX9-NEXT:    v_readlane_b32 s42, v45, 8
+; GFX9-NEXT:    v_readlane_b32 s41, v45, 7
+; GFX9-NEXT:    v_readlane_b32 s40, v45, 6
+; GFX9-NEXT:    v_readlane_b32 s39, v45, 5
+; GFX9-NEXT:    v_readlane_b32 s38, v45, 4
+; GFX9-NEXT:    v_readlane_b32 s37, v45, 3
+; GFX9-NEXT:    v_readlane_b32 s36, v45, 2
+; GFX9-NEXT:    v_readlane_b32 s31, v45, 1
+; GFX9-NEXT:    v_readlane_b32 s30, v45, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xf800
 ; GFX9-NEXT:    v_readlane_b32 s33, v46, 0
 ; GFX9-NEXT:    s_or_saveexec_b64 s[4:5], -1
-; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:20 ; 4-byte Folded Reload
+; GFX9-NEXT:    buffer_load_dword v45, off, s[0:3], s32 offset:20 ; 4-byte Folded Reload
 ; GFX9-NEXT:    buffer_load_dword v46, off, s[0:3], s32 offset:24 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
@@ -241,40 +244,41 @@
 ; GFX10-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    s_or_saveexec_b32 s4, -1
-; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s32 offset:20 ; 4-byte Folded Spill
+; GFX10-NEXT:    buffer_store_dword v45, off, s[0:3], s32 offset:20 ; 4-byte Folded Spill
 ; GFX10-NEXT:    buffer_store_dword v46, off, s[0:3], s32 offset:24 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s4
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    ; implicit-def: $vgpr45
 ; GFX10-NEXT:    v_writelane_b32 v46, s33, 0
+; GFX10-NEXT:    v_writelane_b32 v45, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
-; 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:    v_writelane_b32 v40, s31, 1
+; 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:    s_addk_i32 s32, 0x400
-; 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, s36, 2
+; GFX10-NEXT:    v_writelane_b32 v45, s31, 1
+; GFX10-NEXT:    v_mov_b32_e32 v40, v16
+; 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_writelane_b32 v45, s36, 2
 ; GFX10-NEXT:    s_mov_b32 s36, 0
-; GFX10-NEXT:    v_mov_b32_e32 v44, v13
-; GFX10-NEXT:    v_mov_b32_e32 v45, v12
-; GFX10-NEXT:    v_writelane_b32 v40, s37, 3
+; GFX10-NEXT:    v_mov_b32_e32 v44, v12
+; GFX10-NEXT:    v_writelane_b32 v45, s37, 3
 ; GFX10-NEXT:    s_mov_b32 s37, s36
-; GFX10-NEXT:    v_writelane_b32 v40, s38, 4
+; GFX10-NEXT:    v_writelane_b32 v45, s38, 4
 ; GFX10-NEXT:    s_mov_b32 s38, s36
-; GFX10-NEXT:    v_writelane_b32 v40, s39, 5
+; GFX10-NEXT:    v_writelane_b32 v45, s39, 5
 ; GFX10-NEXT:    s_mov_b32 s39, s36
-; GFX10-NEXT:    v_writelane_b32 v40, s40, 6
+; GFX10-NEXT:    v_writelane_b32 v45, s40, 6
 ; GFX10-NEXT:    s_mov_b32 s40, s36
-; GFX10-NEXT:    v_writelane_b32 v40, s41, 7
+; GFX10-NEXT:    v_writelane_b32 v45, s41, 7
 ; GFX10-NEXT:    s_mov_b32 s41, s36
-; GFX10-NEXT:    v_writelane_b32 v40, s42, 8
+; GFX10-NEXT:    v_writelane_b32 v45, s42, 8
 ; GFX10-NEXT:    s_mov_b32 s42, s36
-; GFX10-NEXT:    v_writelane_b32 v40, s43, 9
+; GFX10-NEXT:    v_writelane_b32 v45, s43, 9
 ; GFX10-NEXT:    s_mov_b32 s43, s36
 ; GFX10-NEXT:    image_gather4_c_b_cl v[0:3], v[12:16], s[36:43], s[4:7] dmask:0x1 dim:SQ_RSRC_IMG_2D
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
@@ -286,28 +290,28 @@
 ; 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[36:43], 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[36:43], 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 s43, v40, 9
-; GFX10-NEXT:    v_readlane_b32 s42, v40, 8
-; GFX10-NEXT:    v_readlane_b32 s41, v40, 7
-; GFX10-NEXT:    v_readlane_b32 s40, v40, 6
-; GFX10-NEXT:    v_readlane_b32 s39, v40, 5
-; GFX10-NEXT:    v_readlane_b32 s38, v40, 4
-; GFX10-NEXT:    v_readlane_b32 s37, v40, 3
-; GFX10-NEXT:    v_readlane_b32 s36, v40, 2
-; 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 s43, v45, 9
+; GFX10-NEXT:    v_readlane_b32 s42, v45, 8
+; GFX10-NEXT:    v_readlane_b32 s41, v45, 7
+; GFX10-NEXT:    v_readlane_b32 s40, v45, 6
+; GFX10-NEXT:    v_readlane_b32 s39, v45, 5
+; GFX10-NEXT:    v_readlane_b32 s38, v45, 4
+; GFX10-NEXT:    v_readlane_b32 s37, v45, 3
+; GFX10-NEXT:    v_readlane_b32 s36, v45, 2
+; GFX10-NEXT:    v_readlane_b32 s31, v45, 1
+; GFX10-NEXT:    v_readlane_b32 s30, v45, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX10-NEXT:    v_readlane_b32 s33, v46, 0
 ; GFX10-NEXT:    s_or_saveexec_b32 s4, -1
 ; GFX10-NEXT:    s_clause 0x1
-; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:20
+; GFX10-NEXT:    buffer_load_dword v45, off, s[0:3], s32 offset:20
 ; GFX10-NEXT:    buffer_load_dword v46, off, s[0:3], s32 offset:24
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s4
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
@@ -133,9 +133,10 @@
 ; GFX9-O0:       ; %bb.0: ; %entry
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-O0-NEXT:    s_or_saveexec_b64 s[34:35], -1
-; GFX9-O0-NEXT:    buffer_store_dword v5, 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 v3, off, s[0:3], s32 offset:20 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    buffer_store_dword v0, 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:    s_mov_b32 s36, s4
 ; GFX9-O0-NEXT:    ; kill: def $sgpr36 killed $sgpr36 def $sgpr36_sgpr37_sgpr38_sgpr39
@@ -144,10 +145,12 @@
 ; 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 v5, s40, 0
-; GFX9-O0-NEXT:    v_writelane_b32 v5, s41, 1
-; GFX9-O0-NEXT:    v_writelane_b32 v5, s42, 2
-; GFX9-O0-NEXT:    v_writelane_b32 v5, s43, 3
+; GFX9-O0-NEXT:    ; implicit-def: $vgpr3
+; 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:    buffer_store_dword v3, off, s[0:3], s32 offset:16 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    s_mov_b32 s34, 0
 ; GFX9-O0-NEXT:    buffer_load_dwordx2 v[3:4], off, s[36:39], s34
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
@@ -170,9 +173,12 @@
 ; 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_load_dword v0, off, s[0:3], s32 offset:16 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_mov_b64 s[34:35], exec
-; GFX9-O0-NEXT:    v_writelane_b32 v5, s34, 4
-; GFX9-O0-NEXT:    v_writelane_b32 v5, s35, 5
+; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
+; GFX9-O0-NEXT:    v_writelane_b32 v0, s34, 4
+; GFX9-O0-NEXT:    v_writelane_b32 v0, s35, 5
+; GFX9-O0-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:16 ; 4-byte Folded Spill
 ; 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
@@ -196,13 +202,15 @@
 ; 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:  .LBB1_2: ; %merge
-; GFX9-O0-NEXT:    v_readlane_b32 s34, v5, 4
-; GFX9-O0-NEXT:    v_readlane_b32 s35, v5, 5
+; GFX9-O0-NEXT:    buffer_load_dword v0, off, s[0:3], s32 offset:16 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
+; GFX9-O0-NEXT:    v_readlane_b32 s34, v0, 4
+; GFX9-O0-NEXT:    v_readlane_b32 s35, v0, 5
 ; GFX9-O0-NEXT:    s_or_b64 exec, exec, s[34:35]
-; GFX9-O0-NEXT:    v_readlane_b32 s36, v5, 0
-; GFX9-O0-NEXT:    v_readlane_b32 s37, v5, 1
-; GFX9-O0-NEXT:    v_readlane_b32 s38, v5, 2
-; GFX9-O0-NEXT:    v_readlane_b32 s39, v5, 3
+; GFX9-O0-NEXT:    v_readlane_b32 s36, v0, 0
+; GFX9-O0-NEXT:    v_readlane_b32 s37, v0, 1
+; GFX9-O0-NEXT:    v_readlane_b32 s38, v0, 2
+; GFX9-O0-NEXT:    v_readlane_b32 s39, v0, 3
 ; GFX9-O0-NEXT:    buffer_load_dword v0, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    buffer_load_dword v3, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
@@ -215,9 +223,10 @@
 ; 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_or_saveexec_b64 s[34:35], -1
-; GFX9-O0-NEXT:    buffer_load_dword v5, 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 v3, off, s[0:3], s32 offset:20 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v0, 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]
@@ -332,15 +341,17 @@
 ; GFX9-O0:       ; %bb.0:
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-O0-NEXT:    s_or_saveexec_b64 s[34:35], -1
-; GFX9-O0-NEXT:    buffer_store_dword v3, off, s[0:3], s32 ; 4-byte Folded Spill
-; GFX9-O0-NEXT:    buffer_store_dword v2, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
-; GFX9-O0-NEXT:    buffer_store_dword v1, off, s[0:3], s32 offset:8 ; 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:    buffer_store_dword v2, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    buffer_store_dword v1, off, s[0:3], s32 offset:12 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-O0-NEXT:    s_mov_b32 s35, s33
 ; GFX9-O0-NEXT:    s_mov_b32 s33, s32
-; 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:    buffer_store_dword v0, off, s[0:3], s33 ; 4-byte Folded Spill
 ; 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
@@ -367,14 +378,16 @@
 ; 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:    s_add_i32 s32, s32, 0xfffffc00
+; GFX9-O0-NEXT:    buffer_load_dword v0, off, s[0:3], s33 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
+; GFX9-O0-NEXT:    v_readlane_b32 s31, v0, 1
+; GFX9-O0-NEXT:    v_readlane_b32 s30, v0, 0
+; GFX9-O0-NEXT:    s_add_i32 s32, s32, 0xfffff800
 ; GFX9-O0-NEXT:    s_mov_b32 s33, s35
 ; GFX9-O0-NEXT:    s_or_saveexec_b64 s[34:35], -1
-; GFX9-O0-NEXT:    buffer_load_dword v3, off, s[0:3], s32 ; 4-byte Folded Reload
-; GFX9-O0-NEXT:    buffer_load_dword v2, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
-; GFX9-O0-NEXT:    buffer_load_dword v1, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v0, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v2, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v1, off, s[0:3], s32 offset:12 ; 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]
@@ -387,8 +400,9 @@
 ; GFX9-O3-NEXT:    buffer_store_dword v2, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    buffer_store_dword v1, off, s[0:3], s32 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_mov_b32 s38, s33
+; GFX9-O3-NEXT:    v_writelane_b32 v3, s30, 0
 ; GFX9-O3-NEXT:    s_mov_b32 s33, s32
 ; GFX9-O3-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-O3-NEXT:    v_writelane_b32 v3, s31, 1
@@ -511,36 +525,37 @@
 ; GFX9-O0:       ; %bb.0:
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-O0-NEXT:    s_or_saveexec_b64 s[34:35], -1
-; GFX9-O0-NEXT:    buffer_store_dword v10, off, s[0:3], s32 ; 4-byte Folded Spill
-; GFX9-O0-NEXT:    buffer_store_dword v8, off, s[0:3], s32 offset:4 ; 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:    buffer_store_dword v8, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
-; GFX9-O0-NEXT:    buffer_store_dword v9, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
-; GFX9-O0-NEXT:    buffer_store_dword v2, off, s[0:3], s32 offset:12 ; 4-byte Folded Spill
-; GFX9-O0-NEXT:    buffer_store_dword v3, off, s[0:3], s32 offset:16 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    buffer_store_dword v9, off, s[0:3], s32 offset:12 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    buffer_store_dword v2, off, s[0:3], s32 offset:16 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    buffer_store_dword v3, off, s[0:3], s32 offset:20 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
-; GFX9-O0-NEXT:    buffer_store_dword v4, off, s[0:3], s32 offset:20 ; 4-byte Folded Spill
-; GFX9-O0-NEXT:    buffer_store_dword v3, off, s[0:3], s32 offset:24 ; 4-byte Folded Spill
-; GFX9-O0-NEXT:    buffer_store_dword v2, off, s[0:3], s32 offset:28 ; 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 v3, 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_waitcnt vmcnt(0)
-; GFX9-O0-NEXT:    buffer_store_dword v3, off, s[0:3], s32 offset:32 ; 4-byte Folded Spill
-; GFX9-O0-NEXT:    buffer_store_dword v4, off, s[0:3], s32 offset:36 ; 4-byte Folded Spill
-; GFX9-O0-NEXT:    buffer_store_dword v5, off, s[0:3], s32 offset:40 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    buffer_store_dword v3, off, s[0:3], s32 offset:36 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    buffer_store_dword v4, off, s[0:3], s32 offset:40 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    buffer_store_dword v5, off, s[0:3], s32 offset:44 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-O0-NEXT:    s_mov_b32 s42, s33
 ; GFX9-O0-NEXT:    s_mov_b32 s33, s32
-; 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
@@ -552,8 +567,9 @@
 ; 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:    buffer_store_dword v0, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v2, v8
 ; GFX9-O0-NEXT:    s_mov_b32 s34, 32
 ; GFX9-O0-NEXT:    ; implicit-def: $sgpr36_sgpr37
@@ -570,12 +586,14 @@
 ; 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:    buffer_load_dword v6, off, s[0:3], s33 ; 4-byte Folded Reload
+; 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:    v_mov_b32_e32 v3, v1
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v4, v8
@@ -587,31 +605,33 @@
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v1, 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:    s_add_i32 s32, s32, 0xfffff400
+; GFX9-O0-NEXT:    buffer_load_dword v0, off, s[0:3], s33 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
+; GFX9-O0-NEXT:    v_readlane_b32 s31, v0, 1
+; GFX9-O0-NEXT:    v_readlane_b32 s30, v0, 0
+; GFX9-O0-NEXT:    s_add_i32 s32, s32, 0xfffff000
 ; GFX9-O0-NEXT:    s_mov_b32 s33, s42
 ; GFX9-O0-NEXT:    s_or_saveexec_b64 s[34:35], -1
-; GFX9-O0-NEXT:    buffer_load_dword v10, off, s[0:3], s32 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v0, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
-; GFX9-O0-NEXT:    buffer_load_dword v8, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v8, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
-; GFX9-O0-NEXT:    buffer_load_dword v9, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v9, off, s[0:3], s32 offset:12 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
-; GFX9-O0-NEXT:    buffer_load_dword v2, off, s[0:3], s32 offset:12 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v2, off, s[0:3], s32 offset:16 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
-; GFX9-O0-NEXT:    buffer_load_dword v3, off, s[0:3], s32 offset:16 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v3, off, s[0:3], s32 offset:20 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
-; GFX9-O0-NEXT:    buffer_load_dword v4, 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:    s_nop 0
-; GFX9-O0-NEXT:    buffer_load_dword v3, off, s[0:3], s32 offset:24 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v3, off, s[0:3], s32 offset:28 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
-; GFX9-O0-NEXT:    buffer_load_dword v2, 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_nop 0
-; GFX9-O0-NEXT:    buffer_load_dword v3, off, s[0:3], s32 offset:32 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v3, off, s[0:3], s32 offset:36 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
-; GFX9-O0-NEXT:    buffer_load_dword v4, off, s[0:3], s32 offset:36 ; 4-byte Folded Reload
-; GFX9-O0-NEXT:    buffer_load_dword v5, off, s[0:3], s32 offset:40 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v4, off, s[0:3], s32 offset:40 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    buffer_load_dword v5, off, s[0:3], s32 offset:44 ; 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]
@@ -630,8 +650,9 @@
 ; GFX9-O3-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-O3-NEXT:    buffer_store_dword v3, off, s[0:3], s32 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_mov_b32 s40, s33
+; GFX9-O3-NEXT:    v_writelane_b32 v8, s30, 0
 ; GFX9-O3-NEXT:    s_mov_b32 s33, s32
 ; GFX9-O3-NEXT:    s_addk_i32 s32, 0x800
 ; GFX9-O3-NEXT:    v_writelane_b32 v8, s31, 1