[AMDGPU] Preserve only the inactive lanes of scratch vgprs

[cdevadas]

Summary

In general, a callee is free to use the scratch registers without
preserving its previous state. However, the VGPR used for SGPR
spilling can potentially have its inactive lanes overwritten by
the writelane instructions. When the function returns, it can
cause unexpected behavior if the VGPR value is not preserved
appropriately.

The current scheme to preserve the inactive lanes of such
scratch VGPRs is not done rightly. It preserves all lanes
and causes the outgoing values (if any) getting overwritten
by the epilog restores. It thus corrupts the return value.

To avoid such situations, this patch ensures we preserve only
the inactive lanes.

[arsenm]

llvm/lib/Target/AMDGPU/SIFrameLowering.cpp

876

I'd rename IsCalleeSavedReg to EnableInactiveLanes

1107–1108

Needs a comment explaining we're going to get 2 exec flip and restore blocks

llvm/lib/Target/AMDGPU/SIMachineFunctionInfo.cpp

290

This needs to check the MRI callee saved regs, not TRI directly for dynamically changed CSRs

llvm/test/CodeGen/AMDGPU/GlobalISel/image-waterfall-loop-O0.ll

148

In the case of lane exit we'll end up restoring values that weren't preserved, but I guess that's OK since those lanes are exiting as dead anyway

[cdevadas]

Suggestions incorporated.

[arsenm]

llvm/lib/Target/AMDGPU/SIMachineFunctionInfo.cpp

284–287

I don't really like having to split these into two arrays like this but I don't have a better suggestion

290

Actually I've seen this mistake too many times. We should probably rename the TRI version and hide it

[cdevadas]

code rebase

llvm/lib/Target/AMDGPU/SIFrameLowering.cpp

case TargetStackID::SGPRSpill:
     return true;
   case TargetStackID::ScalableVector:
   case TargetStackID::WasmLocal:
     return false;
   }
   llvm_unreachable("Invalid TargetStackID::Value");
 }
 
-// Activate all lanes, returns saved exec.
+// Activate only the inactive lanes when \p EnableInactiveLanes is true.
+// Otherwise, activate all lanes. It returns the saved exec.
 static Register buildScratchExecCopy(LivePhysRegs &LiveRegs,
                                      MachineFunction &MF,
                                      MachineBasicBlock &MBB,
                                      MachineBasicBlock::iterator MBBI,
-                                     const DebugLoc &DL, bool IsProlog) {
+                                     const DebugLoc &DL, bool IsProlog,
+                                     bool EnableInactiveLanes) {
   Register ScratchExecCopy;
   MachineRegisterInfo &MRI = MF.getRegInfo();
   const GCNSubtarget &ST = MF.getSubtarget<GCNSubtarget>();
   const SIInstrInfo *TII = ST.getInstrInfo();
   const SIRegisterInfo &TRI = TII->getRegisterInfo();
   SIMachineFunctionInfo *FuncInfo = MF.getInfo<SIMachineFunctionInfo>();
 
   initLiveRegs(LiveRegs, TRI, FuncInfo, MF, MBB, MBBI, IsProlog);
 
   ScratchExecCopy = findScratchNonCalleeSaveRegister(
       MRI, LiveRegs, *TRI.getWaveMaskRegClass());
   if (!ScratchExecCopy)
     report_fatal_error("failed to find free scratch register");
 
   LiveRegs.addReg(ScratchExecCopy);
 
-  const unsigned OrSaveExec =
-      ST.isWave32() ? AMDGPU::S_OR_SAVEEXEC_B32 : AMDGPU::S_OR_SAVEEXEC_B64;
-  auto SaveExec = BuildMI(MBB, MBBI, DL, TII->get(OrSaveExec), ScratchExecCopy)
-    .addImm(-1);
+  const unsigned SaveExecOpc =
+      ST.isWave32() ? (EnableInactiveLanes ? AMDGPU::S_XOR_SAVEEXEC_B32
+                                           : AMDGPU::S_OR_SAVEEXEC_B32)
+                    : (EnableInactiveLanes ? AMDGPU::S_XOR_SAVEEXEC_B64
+                                           : AMDGPU::S_OR_SAVEEXEC_B64);
+  auto SaveExec =
+      BuildMI(MBB, MBBI, DL, TII->get(SaveExecOpc), ScratchExecCopy).addImm(-1);
   SaveExec->getOperand(3).setIsDead(); // Mark SCC as dead.
 
   return ScratchExecCopy;
 }
 
 void SIFrameLowering::emitPrologue(MachineFunction &MF,
                                    MachineBasicBlock &MBB) const {
   SIMachineFunctionInfo *FuncInfo = MF.getInfo<SIMachineFunctionInfo>();
void SIFrameLowering::emitPrologue(MachineFunction &MF,
   bool HasFP = false;
   bool HasBP = false;
   uint32_t NumBytes = MFI.getStackSize();
   uint32_t RoundedSize = NumBytes;
   // To avoid clobbering VGPRs in lanes that weren't active on function entry,
   // turn on all lanes before doing the spill to memory.
   Register ScratchExecCopy;
 
-  // Spill Whole-Wave Mode VGPRs.
-  for (const auto &Reg : FuncInfo->getWWMSpills()) {
+  // Spill Whole-Wave Mode VGPRs. Save only the inactive lanes of the scratch
+  // registers. However, save all lanes of callee-saved VGPRs. Due to this, we
+  // might end up flipping the EXEC bits twice.
+  SmallVector<std::pair<Register, int>, 2> WWMCalleeSavedRegs, WWMScratchRegs;
+  FuncInfo->splitWWMSpillRegisters(MF, WWMCalleeSavedRegs, WWMScratchRegs);
+  if (!WWMScratchRegs.empty())
+    ScratchExecCopy =
+        buildScratchExecCopy(LiveRegs, MF, MBB, MBBI, DL,
+                             /*IsProlog*/ true, /*EnableInactiveLanes*/ true);
+
+  auto StoreWWMRegisters =
+      [&](SmallVectorImpl<std::pair<Register, int>> &WWMRegs) {
+        for (const auto &Reg : WWMRegs) {
           Register VGPR = Reg.first;
           int FI = Reg.second;
-    if (!ScratchExecCopy)
-      ScratchExecCopy = buildScratchExecCopy(LiveRegs, MF, MBB, MBBI, DL,
-                                             /*IsProlog*/ true);
+          buildPrologSpill(ST, TRI, *FuncInfo, LiveRegs, MF, MBB, MBBI, DL,
+                           VGPR, FI);
+        }
+      };
 
-    buildPrologSpill(ST, TRI, *FuncInfo, LiveRegs, MF, MBB, MBBI, DL, VGPR, FI);
+  StoreWWMRegisters(WWMScratchRegs);
+  if (!WWMCalleeSavedRegs.empty()) {
+    if (ScratchExecCopy) {
+      unsigned MovOpc = ST.isWave32() ? AMDGPU::S_MOV_B32 : AMDGPU::S_MOV_B64;
+      MCRegister Exec = ST.isWave32() ? AMDGPU::EXEC_LO : AMDGPU::EXEC;
+      BuildMI(MBB, MBBI, DL, TII->get(MovOpc), Exec).addImm(-1);
+    } else {
+      ScratchExecCopy = buildScratchExecCopy(LiveRegs, MF, MBB, MBBI, DL,
+                                             /*IsProlog*/ true,
+                                             /*EnableInactiveLanes*/ false);
+    }
   }
 
+  StoreWWMRegisters(WWMCalleeSavedRegs);
   if (ScratchExecCopy) {
     // FIXME: Split block and make terminator.
     unsigned ExecMov = ST.isWave32() ? AMDGPU::S_MOV_B32 : AMDGPU::S_MOV_B64;
     MCRegister Exec = ST.isWave32() ? AMDGPU::EXEC_LO : AMDGPU::EXEC;
     BuildMI(MBB, MBBI, DL, TII->get(ExecMov), Exec)
         .addReg(ScratchExecCopy, RegState::Kill);
     LiveRegs.addReg(ScratchExecCopy);
   }
void SIFrameLowering::emitEpilogue(MachineFunction &MF,
   }
 
   for (const auto &Spill : FuncInfo->getPrologEpilogSGPRSpills()) {
     PrologEpilogSGPRSpillBuilder SB(Spill.first, Spill.second, MBB, MBBI, DL,
                                     TII, TRI, LiveRegs);
     SB.restore();
   }
 
+  // Restore Whole-Wave Mode VGPRs. Restore only the inactive lanes of the
+  // scratch registers. However, restore all lanes of callee-saved VGPRs. Due to
+  // this, we might end up flipping the EXEC bits twice.
   Register ScratchExecCopy;
-  for (const auto &Reg : FuncInfo->getWWMSpills()) {
+  SmallVector<std::pair<Register, int>, 2> WWMCalleeSavedRegs, WWMScratchRegs;
+  FuncInfo->splitWWMSpillRegisters(MF, WWMCalleeSavedRegs, WWMScratchRegs);
+  if (!WWMScratchRegs.empty())
+    ScratchExecCopy =
+        buildScratchExecCopy(LiveRegs, MF, MBB, MBBI, DL, /*IsProlog*/ false,
+                             /*EnableInactiveLanes*/ true);
+
+  auto RestoreWWMRegisters =
+      [&](SmallVectorImpl<std::pair<Register, int>> &WWMRegs) {
+        for (const auto &Reg : WWMRegs) {
           Register VGPR = Reg.first;
           int FI = Reg.second;
-    if (!ScratchExecCopy)
+          buildEpilogRestore(ST, TRI, *FuncInfo, LiveRegs, MF, MBB, MBBI, DL,
+                             VGPR, FI);
+        }
+      };
+  RestoreWWMRegisters(WWMScratchRegs);
+  if (!WWMCalleeSavedRegs.empty()) {
+    if (ScratchExecCopy) {
+      unsigned MovOpc = ST.isWave32() ? AMDGPU::S_MOV_B32 : AMDGPU::S_MOV_B64;
+      MCRegister Exec = ST.isWave32() ? AMDGPU::EXEC_LO : AMDGPU::EXEC;
+      BuildMI(MBB, MBBI, DL, TII->get(MovOpc), Exec).addImm(-1);
+    } else {
       ScratchExecCopy =
-          buildScratchExecCopy(LiveRegs, MF, MBB, MBBI, DL, /*IsProlog*/ false);
-
-    buildEpilogRestore(ST, TRI, *FuncInfo, LiveRegs, MF, MBB, MBBI, DL, VGPR,
-                       FI);
+          buildScratchExecCopy(LiveRegs, MF, MBB, MBBI, DL, /*IsProlog*/ false,
+                               /*EnableInactiveLanes*/ false);
+    }
   }
 
+  RestoreWWMRegisters(WWMCalleeSavedRegs);
   if (ScratchExecCopy) {
     // FIXME: Split block and make terminator.
     unsigned ExecMov = ST.isWave32() ? AMDGPU::S_MOV_B32 : AMDGPU::S_MOV_B64;
     MCRegister Exec = ST.isWave32() ? AMDGPU::EXEC_LO : AMDGPU::EXEC;
     BuildMI(MBB, MBBI, DL, TII->get(ExecMov), Exec)
         .addReg(ScratchExecCopy, RegState::Kill);
   }
 }

llvm/lib/Target/AMDGPU/SIMachineFunctionInfo.h

public:
   Register getVGPRForAGPRCopy() const {
     return VGPRForAGPRCopy;
   }
 
   void setVGPRForAGPRCopy(Register NewVGPRForAGPRCopy) {
     VGPRForAGPRCopy = NewVGPRForAGPRCopy;
   }
 
-  bool isCalleeSavedReg(const MCPhysReg *CSRegs, MCPhysReg Reg);
+  bool isCalleeSavedReg(const MCPhysReg *CSRegs, MCPhysReg Reg) const;
 
 public:
   SIMachineFunctionInfo(const MachineFunction &MF);
   SIMachineFunctionInfo(const SIMachineFunctionInfo &MFI) = default;
 
   MachineFunctionInfo *
   clone(BumpPtrAllocator &Allocator, MachineFunction &DestMF,
         const DenseMap<MachineBasicBlock *, MachineBasicBlock *> &Src2DstMBB)
getPrologEpilogSGPRSpillToVGPRLanes(int FrameIndex) const {
     return (I == PrologEpilogSGPRSpillToVGPRLanes.end())
                ? ArrayRef<SIRegisterInfo::SpilledReg>()
                : makeArrayRef(I->second);
   }
 
   void allocateWWMSpill(MachineFunction &MF, Register VGPR, uint64_t Size = 4,
                         Align Alignment = Align(4));
 
+  void splitWWMSpillRegisters(
+      MachineFunction &MF,
+      SmallVectorImpl<std::pair<Register, int>> &CalleeSavedRegs,
+      SmallVectorImpl<std::pair<Register, int>> &ScratchRegs) const;
+
   ArrayRef<MCPhysReg> getAGPRSpillVGPRs() const {
     return SpillAGPR;
   }
 
   ArrayRef<MCPhysReg> getVGPRSpillAGPRs() const {
     return SpillVGPR;
   }
 

llvm/lib/Target/AMDGPU/SIMachineFunctionInfo.cpp

void SIMachineFunctionInfo::allocateWWMSpill(MachineFunction &MF, Register VGPR,
   // Skip if it is an entry function or the register is already added.
   if (isEntryFunction() || WWMSpills.count(VGPR))
     return;
 
   WWMSpills.insert(std::make_pair(
       VGPR, MF.getFrameInfo().CreateSpillStackObject(Size, Alignment)));
 }
 
+// Separate out the callee-saved and scratch registers.
+void SIMachineFunctionInfo::splitWWMSpillRegisters(
+    MachineFunction &MF,
+    SmallVectorImpl<std::pair<Register, int>> &CalleeSavedRegs,
+    SmallVectorImpl<std::pair<Register, int>> &ScratchRegs) const {
+  const MCPhysReg *CSRegs = MF.getRegInfo().getCalleeSavedRegs();
+  for (auto &Reg : WWMSpills) {
+    if (isCalleeSavedReg(CSRegs, Reg.first))
+      CalleeSavedRegs.push_back(Reg);
+    else
+      ScratchRegs.push_back(Reg);
+  }
+}
+
 bool SIMachineFunctionInfo::isCalleeSavedReg(const MCPhysReg *CSRegs,
-                                             MCPhysReg Reg) {
+                                             MCPhysReg Reg) const {
   for (unsigned I = 0; CSRegs[I]; ++I) {
     if (CSRegs[I] == Reg)
       return true;
   }
 
   return false;
 }
 

llvm/test/CodeGen/AMDGPU/GlobalISel/image-waterfall-loop-O0.ll

 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc -global-isel -O0 -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1031 -verify-machineinstrs -o - %s | FileCheck %s
 
 ; Make sure the waterfall loop does not fail the verifier after regalloc fast
 define <4 x float> @waterfall_loop(<8 x i32> %vgpr_srd) {
 ; CHECK-LABEL: waterfall_loop:
 ; CHECK:       ; %bb.0: ; %bb
 ; 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:    s_xor_saveexec_b32 s4, -1
 ; CHECK-NEXT:    buffer_store_dword v8, off, s[0:3], s32 offset:44 ; 4-byte Folded Spill
 ; CHECK-NEXT:    s_mov_b32 exec_lo, s4
 ; CHECK-NEXT:    v_mov_b32_e32 v15, v1
 ; CHECK-NEXT:    v_mov_b32_e32 v14, v2
 ; CHECK-NEXT:    v_mov_b32_e32 v13, v3
 ; CHECK-NEXT:    v_mov_b32_e32 v12, v4
 ; CHECK-NEXT:    v_mov_b32_e32 v11, v5
 ; CHECK-NEXT:    v_mov_b32_e32 v10, v6

 ; CHECK-NEXT:    v_readlane_b32 s4, v8, 4
 ; CHECK-NEXT:    s_mov_b32 exec_lo, s4
 ; CHECK-NEXT:  ; %bb.4:
 ; CHECK-NEXT:    buffer_load_dword v0, off, s[0:3], s32 offset:40 ; 4-byte Folded Reload
 ; CHECK-NEXT:    ; implicit-def: $sgpr4
 ; CHECK-NEXT:    v_mov_b32_e32 v1, s4
 ; CHECK-NEXT:    v_mov_b32_e32 v2, s4
 ; CHECK-NEXT:    v_mov_b32_e32 v3, s4
-; CHECK-NEXT:    s_or_saveexec_b32 s4, -1
+; CHECK-NEXT:    s_xor_saveexec_b32 s4, -1
 ; CHECK-NEXT:    buffer_load_dword v8, off, s[0:3], s32 offset:44 ; 4-byte Folded Reload
 ; CHECK-NEXT:    s_mov_b32 exec_lo, s4
 ; CHECK-NEXT:    s_waitcnt vmcnt(0)
 ; CHECK-NEXT:    s_waitcnt_vscnt null, 0x0
 ; CHECK-NEXT:    s_setpc_b64 s[30:31]
 bb:
   %ret = tail call <4 x float> @llvm.amdgcn.image.sample.2d.v4f32.f32(i32 1, float 0.000000e+00, float 0.000000e+00, <8 x i32> %vgpr_srd, <4 x i32> zeroinitializer, i1 false, i32 0, i32 0)
   ret <4 x float> %ret
 }
 
 declare <4 x float> @llvm.amdgcn.image.sample.2d.v4f32.f32(i32 immarg, float, float, <8 x i32>, <4 x i32>, i1 immarg, i32 immarg, i32 immarg) #0
 
 attributes #0 = { nounwind readonly willreturn }

llvm/test/CodeGen/AMDGPU/bf16.ll

 entry:
   ret <16 x bfloat> %in
 }
 
 define void @test_call(bfloat %in, ptr addrspace(5) %out) {
 ; GCN-LABEL: test_call:
 ; GCN:       ; %bb.0: ; %entry
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GCN-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_store_dword v2, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_mov_b32 s8, s33
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
 ; GCN-NEXT:    s_waitcnt expcnt(0)
 ; GCN-NEXT:    v_writelane_b32 v2, s30, 0
 ; GCN-NEXT:    v_writelane_b32 v2, s31, 1
 ; GCN-NEXT:    s_getpc_b64 s[4:5]
 ; GCN-NEXT:    s_add_u32 s4, s4, test_arg_store@gotpcrel32@lo+4
 ; GCN-NEXT:    s_addc_u32 s5, s5, test_arg_store@gotpcrel32@hi+12
 ; GCN-NEXT:    s_load_dwordx2 s[4:5], s[4:5], 0x0
 ; GCN-NEXT:    s_waitcnt lgkmcnt(0)
 ; GCN-NEXT:    s_swappc_b64 s[30:31], s[4:5]
 ; GCN-NEXT:    v_lshrrev_b32_e32 v0, 16, v0
 ; GCN-NEXT:    buffer_store_short v0, v1, s[0:3], 0 offen
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
 ; GCN-NEXT:    v_readlane_b32 s31, v2, 1
 ; GCN-NEXT:    v_readlane_b32 s30, v2, 0
 ; GCN-NEXT:    s_addk_i32 s32, 0xfc00
 ; GCN-NEXT:    s_mov_b32 s33, s8
-; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GCN-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_load_dword v2, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0)
 ; GCN-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX7-LABEL: test_call:
 ; GFX7:       ; %bb.0: ; %entry
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX7-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX7-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX7-NEXT:    buffer_store_dword v2, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX7-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX7-NEXT:    s_mov_b32 s8, s33
 ; GFX7-NEXT:    s_mov_b32 s33, s32
 ; GFX7-NEXT:    s_addk_i32 s32, 0x400
 ; GFX7-NEXT:    s_getpc_b64 s[4:5]
 ; GFX7-NEXT:    s_add_u32 s4, s4, test_arg_store@gotpcrel32@lo+4
 ; GFX7-NEXT:    s_addc_u32 s5, s5, test_arg_store@gotpcrel32@hi+12
 ; GFX7-NEXT:    s_load_dwordx2 s[4:5], s[4:5], 0x0
 ; GFX7-NEXT:    v_writelane_b32 v2, s30, 0
 ; GFX7-NEXT:    v_writelane_b32 v2, s31, 1
 ; GFX7-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX7-NEXT:    s_swappc_b64 s[30:31], s[4:5]
 ; GFX7-NEXT:    v_lshrrev_b32_e32 v0, 16, v0
 ; GFX7-NEXT:    buffer_store_short v0, v1, s[0:3], 0 offen
 ; GFX7-NEXT:    s_waitcnt vmcnt(0)
 ; GFX7-NEXT:    v_readlane_b32 s31, v2, 1
 ; GFX7-NEXT:    v_readlane_b32 s30, v2, 0
 ; GFX7-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX7-NEXT:    s_mov_b32 s33, s8
-; GFX7-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX7-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX7-NEXT:    buffer_load_dword v2, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX7-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX7-NEXT:    s_waitcnt vmcnt(0)
 ; GFX7-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX8-LABEL: test_call:
 ; GFX8:       ; %bb.0: ; %entry
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX8-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX8-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX8-NEXT:    buffer_store_dword v2, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX8-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX8-NEXT:    s_mov_b32 s6, s33
 ; GFX8-NEXT:    s_mov_b32 s33, s32
 ; GFX8-NEXT:    s_addk_i32 s32, 0x400
 ; GFX8-NEXT:    s_getpc_b64 s[4:5]
 ; GFX8-NEXT:    s_add_u32 s4, s4, test_arg_store@gotpcrel32@lo+4
 ; GFX8-NEXT:    s_addc_u32 s5, s5, test_arg_store@gotpcrel32@hi+12
 ; GFX8-NEXT:    s_load_dwordx2 s[4:5], s[4:5], 0x0
 ; GFX8-NEXT:    v_writelane_b32 v2, s30, 0
 ; GFX8-NEXT:    v_writelane_b32 v2, s31, 1
 ; GFX8-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX8-NEXT:    s_swappc_b64 s[30:31], s[4:5]
 ; GFX8-NEXT:    v_lshrrev_b32_e32 v0, 16, v0
 ; GFX8-NEXT:    buffer_store_short v0, v1, s[0:3], 0 offen
 ; GFX8-NEXT:    s_waitcnt vmcnt(0)
 ; GFX8-NEXT:    v_readlane_b32 s31, v2, 1
 ; GFX8-NEXT:    v_readlane_b32 s30, v2, 0
 ; GFX8-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX8-NEXT:    s_mov_b32 s33, s6
-; GFX8-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX8-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX8-NEXT:    buffer_load_dword v2, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX8-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX8-NEXT:    s_waitcnt vmcnt(0)
 ; GFX8-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX9-LABEL: test_call:
 ; GFX9:       ; %bb.0: ; %entry
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX9-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX9-NEXT:    buffer_store_dword v2, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX9-NEXT:    s_mov_b32 s6, s33
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    s_getpc_b64 s[4:5]
 ; GFX9-NEXT:    s_add_u32 s4, s4, test_arg_store@gotpcrel32@lo+4
 ; GFX9-NEXT:    s_addc_u32 s5, s5, test_arg_store@gotpcrel32@hi+12
 ; GFX9-NEXT:    s_load_dwordx2 s[4:5], s[4:5], 0x0
 ; 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[4:5]
 ; GFX9-NEXT:    buffer_store_short_d16_hi v0, v1, s[0:3], 0 offen
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    v_readlane_b32 s31, v2, 1
 ; GFX9-NEXT:    v_readlane_b32 s30, v2, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    s_mov_b32 s33, s6
-; GFX9-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX9-NEXT:    buffer_load_dword v2, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: test_call:
 ; GFX10:       ; %bb.0: ; %entry
 ; 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:    s_xor_saveexec_b32 s4, -1
 ; GFX10-NEXT:    buffer_store_dword v2, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s4
 ; GFX10-NEXT:    s_mov_b32 s6, s33
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    s_getpc_b64 s[4:5]
 ; GFX10-NEXT:    s_add_u32 s4, s4, test_arg_store@gotpcrel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s5, s5, test_arg_store@gotpcrel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v2, s30, 0
 ; GFX10-NEXT:    s_load_dwordx2 s[4:5], s[4:5], 0x0
 ; GFX10-NEXT:    v_writelane_b32 v2, s31, 1
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[4:5]
 ; GFX10-NEXT:    buffer_store_short_d16_hi v0, v1, s[0:3], 0 offen
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    v_readlane_b32 s31, v2, 1
 ; GFX10-NEXT:    v_readlane_b32 s30, v2, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    s_mov_b32 s33, s6
-; GFX10-NEXT:    s_or_saveexec_b32 s4, -1
+; GFX10-NEXT:    s_xor_saveexec_b32 s4, -1
 ; GFX10-NEXT:    buffer_load_dword v2, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s4
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    s_setpc_b64 s[30:31]
 entry:
   %result = call bfloat @test_arg_store(bfloat %in)
   store volatile bfloat %result, ptr addrspace(5) %out
   ret void
 }
 
 define void @test_call_v2bf16(<2 x bfloat> %in, ptr addrspace(5) %out) {
 ; GCN-LABEL: test_call_v2bf16:
 ; GCN:       ; %bb.0: ; %entry
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GCN-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_store_dword v3, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_mov_b32 s8, s33
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
 ; GCN-NEXT:    s_waitcnt expcnt(0)
 ; GCN-NEXT:    v_writelane_b32 v3, s30, 0
 ; GCN-NEXT:    v_writelane_b32 v3, s31, 1

 ; GCN-NEXT:    buffer_store_short v1, v4, s[0:3], 0 offen
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
 ; GCN-NEXT:    buffer_store_short v0, v2, s[0:3], 0 offen
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
 ; GCN-NEXT:    v_readlane_b32 s31, v3, 1
 ; GCN-NEXT:    v_readlane_b32 s30, v3, 0
 ; GCN-NEXT:    s_addk_i32 s32, 0xfc00
 ; GCN-NEXT:    s_mov_b32 s33, s8
-; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GCN-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_load_dword v3, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0)
 ; GCN-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX7-LABEL: test_call_v2bf16:
 ; GFX7:       ; %bb.0: ; %entry
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX7-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX7-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX7-NEXT:    buffer_store_dword v3, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX7-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX7-NEXT:    s_mov_b32 s8, s33
 ; GFX7-NEXT:    s_mov_b32 s33, s32
 ; GFX7-NEXT:    s_addk_i32 s32, 0x400
 ; GFX7-NEXT:    s_getpc_b64 s[4:5]
 ; GFX7-NEXT:    s_add_u32 s4, s4, test_arg_store_v2bf16@gotpcrel32@lo+4
 ; GFX7-NEXT:    s_addc_u32 s5, s5, test_arg_store_v2bf16@gotpcrel32@hi+12
 ; GFX7-NEXT:    s_load_dwordx2 s[4:5], s[4:5], 0x0
 ; GFX7-NEXT:    v_writelane_b32 v3, s30, 0
 ; GFX7-NEXT:    v_writelane_b32 v3, s31, 1
 ; GFX7-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX7-NEXT:    s_swappc_b64 s[30:31], s[4:5]
 ; GFX7-NEXT:    v_lshrrev_b32_e32 v1, 16, v1
 ; GFX7-NEXT:    v_add_i32_e32 v4, vcc, 2, v2
 ; GFX7-NEXT:    v_lshrrev_b32_e32 v0, 16, v0
 ; GFX7-NEXT:    buffer_store_short v1, v4, s[0:3], 0 offen
 ; GFX7-NEXT:    s_waitcnt vmcnt(0)
 ; GFX7-NEXT:    buffer_store_short v0, v2, s[0:3], 0 offen
 ; GFX7-NEXT:    s_waitcnt vmcnt(0)
 ; GFX7-NEXT:    v_readlane_b32 s31, v3, 1
 ; GFX7-NEXT:    v_readlane_b32 s30, v3, 0
 ; GFX7-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX7-NEXT:    s_mov_b32 s33, s8
-; GFX7-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX7-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX7-NEXT:    buffer_load_dword v3, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX7-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX7-NEXT:    s_waitcnt vmcnt(0)
 ; GFX7-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX8-LABEL: test_call_v2bf16:
 ; GFX8:       ; %bb.0: ; %entry
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX8-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX8-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX8-NEXT:    buffer_store_dword v2, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX8-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX8-NEXT:    s_mov_b32 s6, s33
 ; GFX8-NEXT:    s_mov_b32 s33, s32
 ; GFX8-NEXT:    s_addk_i32 s32, 0x400
 ; GFX8-NEXT:    s_getpc_b64 s[4:5]
 ; GFX8-NEXT:    s_add_u32 s4, s4, test_arg_store_v2bf16@gotpcrel32@lo+4
 ; GFX8-NEXT:    s_addc_u32 s5, s5, test_arg_store_v2bf16@gotpcrel32@hi+12
 ; GFX8-NEXT:    s_load_dwordx2 s[4:5], s[4:5], 0x0
 ; GFX8-NEXT:    v_writelane_b32 v2, s30, 0
 ; GFX8-NEXT:    v_writelane_b32 v2, s31, 1
 ; GFX8-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX8-NEXT:    s_swappc_b64 s[30:31], s[4:5]
 ; GFX8-NEXT:    buffer_store_dword v0, v1, s[0:3], 0 offen
 ; GFX8-NEXT:    s_waitcnt vmcnt(0)
 ; GFX8-NEXT:    v_readlane_b32 s31, v2, 1
 ; GFX8-NEXT:    v_readlane_b32 s30, v2, 0
 ; GFX8-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX8-NEXT:    s_mov_b32 s33, s6
-; GFX8-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX8-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX8-NEXT:    buffer_load_dword v2, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX8-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX8-NEXT:    s_waitcnt vmcnt(0)
 ; GFX8-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX9-LABEL: test_call_v2bf16:
 ; GFX9:       ; %bb.0: ; %entry
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX9-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX9-NEXT:    buffer_store_dword v2, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX9-NEXT:    s_mov_b32 s6, s33
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    s_getpc_b64 s[4:5]
 ; GFX9-NEXT:    s_add_u32 s4, s4, test_arg_store_v2bf16@gotpcrel32@lo+4
 ; GFX9-NEXT:    s_addc_u32 s5, s5, test_arg_store_v2bf16@gotpcrel32@hi+12
 ; GFX9-NEXT:    s_load_dwordx2 s[4:5], s[4:5], 0x0
 ; 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[4:5]
 ; GFX9-NEXT:    buffer_store_dword v0, v1, s[0:3], 0 offen
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    v_readlane_b32 s31, v2, 1
 ; GFX9-NEXT:    v_readlane_b32 s30, v2, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    s_mov_b32 s33, s6
-; GFX9-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX9-NEXT:    buffer_load_dword v2, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: test_call_v2bf16:
 ; GFX10:       ; %bb.0: ; %entry
 ; 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:    s_xor_saveexec_b32 s4, -1
 ; GFX10-NEXT:    buffer_store_dword v2, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s4
 ; GFX10-NEXT:    s_mov_b32 s6, s33
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    s_getpc_b64 s[4:5]
 ; GFX10-NEXT:    s_add_u32 s4, s4, test_arg_store_v2bf16@gotpcrel32@lo+4
 ; GFX10-NEXT:    s_addc_u32 s5, s5, test_arg_store_v2bf16@gotpcrel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v2, s30, 0
 ; GFX10-NEXT:    s_load_dwordx2 s[4:5], s[4:5], 0x0
 ; GFX10-NEXT:    v_writelane_b32 v2, s31, 1
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX10-NEXT:    s_swappc_b64 s[30:31], s[4:5]
 ; GFX10-NEXT:    buffer_store_dword v0, v1, s[0:3], 0 offen
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    v_readlane_b32 s31, v2, 1
 ; GFX10-NEXT:    v_readlane_b32 s30, v2, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    s_mov_b32 s33, s6
-; GFX10-NEXT:    s_or_saveexec_b32 s4, -1
+; GFX10-NEXT:    s_xor_saveexec_b32 s4, -1
 ; GFX10-NEXT:    buffer_load_dword v2, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s4
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    s_setpc_b64 s[30:31]
 entry:
   %result = call <2 x bfloat> @test_arg_store_v2bf16(<2 x bfloat> %in)
   store volatile <2 x bfloat> %result, ptr addrspace(5) %out
   ret void
 }
 
 define void @test_call_v3bf16(<3 x bfloat> %in, ptr addrspace(5) %out) {
 ; GCN-LABEL: test_call_v3bf16:
 ; GCN:       ; %bb.0: ; %entry
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GCN-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_store_dword v4, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_mov_b32 s8, s33
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
 ; GCN-NEXT:    s_waitcnt expcnt(0)
 ; GCN-NEXT:    v_writelane_b32 v4, s30, 0
 ; GCN-NEXT:    v_writelane_b32 v4, s31, 1

 ; GCN-NEXT:    buffer_store_short v2, v5, s[0:3], 0 offen
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
 ; GCN-NEXT:    buffer_store_dword v0, v3, s[0:3], 0 offen
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
 ; GCN-NEXT:    v_readlane_b32 s31, v4, 1
 ; GCN-NEXT:    v_readlane_b32 s30, v4, 0
 ; GCN-NEXT:    s_addk_i32 s32, 0xfc00
 ; GCN-NEXT:    s_mov_b32 s33, s8
-; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GCN-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_load_dword v4, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0)
 ; GCN-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX7-LABEL: test_call_v3bf16:
 ; GFX7:       ; %bb.0: ; %entry
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX7-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX7-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX7-NEXT:    buffer_store_dword v4, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX7-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX7-NEXT:    s_mov_b32 s8, s33
 ; GFX7-NEXT:    s_mov_b32 s33, s32
 ; GFX7-NEXT:    s_addk_i32 s32, 0x400
 ; GFX7-NEXT:    s_getpc_b64 s[4:5]
 ; GFX7-NEXT:    s_add_u32 s4, s4, test_arg_store_v2bf16@gotpcrel32@lo+4
 ; GFX7-NEXT:    s_addc_u32 s5, s5, test_arg_store_v2bf16@gotpcrel32@hi+12

 ; GFX7-NEXT:    buffer_store_short v1, v2, s[0:3], 0 offen
 ; GFX7-NEXT:    s_waitcnt vmcnt(0)
 ; GFX7-NEXT:    buffer_store_dword v0, v3, s[0:3], 0 offen
 ; GFX7-NEXT:    s_waitcnt vmcnt(0)
 ; GFX7-NEXT:    v_readlane_b32 s31, v4, 1
 ; GFX7-NEXT:    v_readlane_b32 s30, v4, 0
 ; GFX7-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX7-NEXT:    s_mov_b32 s33, s8
-; GFX7-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX7-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX7-NEXT:    buffer_load_dword v4, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX7-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX7-NEXT:    s_waitcnt vmcnt(0)
 ; GFX7-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX8-LABEL: test_call_v3bf16:
 ; GFX8:       ; %bb.0: ; %entry
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX8-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX8-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX8-NEXT:    buffer_store_dword v3, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX8-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX8-NEXT:    s_mov_b32 s6, s33
 ; GFX8-NEXT:    s_mov_b32 s33, s32
 ; GFX8-NEXT:    s_addk_i32 s32, 0x400
 ; GFX8-NEXT:    s_getpc_b64 s[4:5]
 ; GFX8-NEXT:    s_add_u32 s4, s4, test_arg_store_v2bf16@gotpcrel32@lo+4
 ; GFX8-NEXT:    s_addc_u32 s5, s5, test_arg_store_v2bf16@gotpcrel32@hi+12
 ; GFX8-NEXT:    s_load_dwordx2 s[4:5], s[4:5], 0x0
 ; GFX8-NEXT:    v_writelane_b32 v3, s30, 0
 ; GFX8-NEXT:    v_and_b32_e32 v1, 0xffff, v1
 ; GFX8-NEXT:    v_writelane_b32 v3, s31, 1
 ; GFX8-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX8-NEXT:    s_swappc_b64 s[30:31], s[4:5]
 ; GFX8-NEXT:    v_add_u32_e32 v4, vcc, 4, v2
 ; GFX8-NEXT:    buffer_store_short v1, v4, s[0:3], 0 offen
 ; GFX8-NEXT:    s_waitcnt vmcnt(0)
 ; GFX8-NEXT:    buffer_store_dword v0, v2, s[0:3], 0 offen
 ; GFX8-NEXT:    s_waitcnt vmcnt(0)
 ; GFX8-NEXT:    v_readlane_b32 s31, v3, 1
 ; GFX8-NEXT:    v_readlane_b32 s30, v3, 0
 ; GFX8-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX8-NEXT:    s_mov_b32 s33, s6
-; GFX8-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX8-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX8-NEXT:    buffer_load_dword v3, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX8-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX8-NEXT:    s_waitcnt vmcnt(0)
 ; GFX8-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX9-LABEL: test_call_v3bf16:
 ; GFX9:       ; %bb.0: ; %entry
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX9-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX9-NEXT:    buffer_store_dword v3, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX9-NEXT:    s_mov_b32 s6, s33
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_and_b32_e32 v4, 0xffff0000, v0
 ; GFX9-NEXT:    s_mov_b32 s4, 0xffff
 ; GFX9-NEXT:    v_and_or_b32 v0, v0, s4, v4

 ; GFX9-NEXT:    buffer_store_short v1, v2, s[0:3], 0 offen offset:4
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    buffer_store_dword v0, v2, s[0:3], 0 offen
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    v_readlane_b32 s31, v3, 1
 ; GFX9-NEXT:    v_readlane_b32 s30, v3, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    s_mov_b32 s33, s6
-; GFX9-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX9-NEXT:    buffer_load_dword v3, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: test_call_v3bf16:
 ; GFX10:       ; %bb.0: ; %entry
 ; 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:    s_xor_saveexec_b32 s4, -1
 ; GFX10-NEXT:    buffer_store_dword v3, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s4
 ; GFX10-NEXT:    s_mov_b32 s6, s33
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    s_getpc_b64 s[4:5]
 ; GFX10-NEXT:    s_add_u32 s4, s4, test_arg_store_v2bf16@gotpcrel32@lo+4

 ; GFX10-NEXT:    buffer_store_short v1, v2, s[0:3], 0 offen offset:4
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    buffer_store_dword v0, v2, s[0:3], 0 offen
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    v_readlane_b32 s31, v3, 1
 ; GFX10-NEXT:    v_readlane_b32 s30, v3, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    s_mov_b32 s33, s6
-; GFX10-NEXT:    s_or_saveexec_b32 s4, -1
+; GFX10-NEXT:    s_xor_saveexec_b32 s4, -1
 ; GFX10-NEXT:    buffer_load_dword v3, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s4
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    s_setpc_b64 s[30:31]
 entry:
   %result = call <3 x bfloat> @test_arg_store_v2bf16(<3 x bfloat> %in)
   store volatile <3 x bfloat> %result, ptr addrspace(5) %out
   ret void
 }
 
 define void @test_call_v4bf16(<4 x bfloat> %in, ptr addrspace(5) %out) {
 ; GCN-LABEL: test_call_v4bf16:
 ; GCN:       ; %bb.0: ; %entry
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GCN-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_store_dword v5, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_mov_b32 s8, s33
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
 ; GCN-NEXT:    s_waitcnt expcnt(0)
 ; GCN-NEXT:    v_writelane_b32 v5, s30, 0
 ; GCN-NEXT:    v_writelane_b32 v5, s31, 1

 ; GCN-NEXT:    buffer_store_short v1, v8, s[0:3], 0 offen
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
 ; GCN-NEXT:    buffer_store_short v0, v4, s[0:3], 0 offen
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
 ; GCN-NEXT:    v_readlane_b32 s31, v5, 1
 ; GCN-NEXT:    v_readlane_b32 s30, v5, 0
 ; GCN-NEXT:    s_addk_i32 s32, 0xfc00
 ; GCN-NEXT:    s_mov_b32 s33, s8
-; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GCN-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_load_dword v5, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0)
 ; GCN-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX7-LABEL: test_call_v4bf16:
 ; GFX7:       ; %bb.0: ; %entry
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX7-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX7-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX7-NEXT:    buffer_store_dword v5, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX7-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX7-NEXT:    s_mov_b32 s8, s33
 ; GFX7-NEXT:    s_mov_b32 s33, s32
 ; GFX7-NEXT:    s_addk_i32 s32, 0x400
 ; GFX7-NEXT:    s_getpc_b64 s[4:5]
 ; GFX7-NEXT:    s_add_u32 s4, s4, test_arg_store_v2bf16@gotpcrel32@lo+4
 ; GFX7-NEXT:    s_addc_u32 s5, s5, test_arg_store_v2bf16@gotpcrel32@hi+12

 ; GFX7-NEXT:    buffer_store_short v1, v2, s[0:3], 0 offen
 ; GFX7-NEXT:    s_waitcnt vmcnt(0)
 ; GFX7-NEXT:    buffer_store_short v0, v4, s[0:3], 0 offen
 ; GFX7-NEXT:    s_waitcnt vmcnt(0)
 ; GFX7-NEXT:    v_readlane_b32 s31, v5, 1
 ; GFX7-NEXT:    v_readlane_b32 s30, v5, 0
 ; GFX7-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX7-NEXT:    s_mov_b32 s33, s8
-; GFX7-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX7-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX7-NEXT:    buffer_load_dword v5, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX7-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX7-NEXT:    s_waitcnt vmcnt(0)
 ; GFX7-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX8-LABEL: test_call_v4bf16:
 ; GFX8:       ; %bb.0: ; %entry
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX8-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX8-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX8-NEXT:    buffer_store_dword v3, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX8-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX8-NEXT:    s_mov_b32 s6, s33
 ; GFX8-NEXT:    s_mov_b32 s33, s32
 ; GFX8-NEXT:    s_addk_i32 s32, 0x400
 ; GFX8-NEXT:    s_getpc_b64 s[4:5]
 ; GFX8-NEXT:    s_add_u32 s4, s4, test_arg_store_v2bf16@gotpcrel32@lo+4
 ; GFX8-NEXT:    s_addc_u32 s5, s5, test_arg_store_v2bf16@gotpcrel32@hi+12

 ; GFX8-NEXT:    s_waitcnt vmcnt(0)
 ; GFX8-NEXT:    v_add_u32_e32 v0, vcc, 2, v2
 ; GFX8-NEXT:    buffer_store_short v4, v0, s[0:3], 0 offen
 ; GFX8-NEXT:    s_waitcnt vmcnt(0)
 ; GFX8-NEXT:    v_readlane_b32 s31, v3, 1
 ; GFX8-NEXT:    v_readlane_b32 s30, v3, 0
 ; GFX8-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX8-NEXT:    s_mov_b32 s33, s6
-; GFX8-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX8-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX8-NEXT:    buffer_load_dword v3, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX8-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX8-NEXT:    s_waitcnt vmcnt(0)
 ; GFX8-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX9-LABEL: test_call_v4bf16:
 ; GFX9:       ; %bb.0: ; %entry
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX9-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX9-NEXT:    buffer_store_dword v3, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX9-NEXT:    s_mov_b32 s6, s33
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    s_getpc_b64 s[4:5]
 ; GFX9-NEXT:    s_add_u32 s4, s4, test_arg_store_v2bf16@gotpcrel32@lo+4
 ; GFX9-NEXT:    s_addc_u32 s5, s5, test_arg_store_v2bf16@gotpcrel32@hi+12

 ; GFX9-NEXT:    buffer_store_short_d16_hi v0, v2, s[0:3], 0 offen offset:2
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    buffer_store_short v0, v2, s[0:3], 0 offen
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    v_readlane_b32 s31, v3, 1
 ; GFX9-NEXT:    v_readlane_b32 s30, v3, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    s_mov_b32 s33, s6
-; GFX9-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX9-NEXT:    buffer_load_dword v3, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: test_call_v4bf16:
 ; GFX10:       ; %bb.0: ; %entry
 ; 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:    s_xor_saveexec_b32 s4, -1
 ; GFX10-NEXT:    buffer_store_dword v3, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s4
 ; GFX10-NEXT:    s_mov_b32 s6, s33
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    s_getpc_b64 s[4:5]
 ; GFX10-NEXT:    s_add_u32 s4, s4, test_arg_store_v2bf16@gotpcrel32@lo+4

 ; GFX10-NEXT:    buffer_store_short_d16_hi v0, v2, s[0:3], 0 offen offset:2
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    buffer_store_short v0, v2, s[0:3], 0 offen
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    v_readlane_b32 s31, v3, 1
 ; GFX10-NEXT:    v_readlane_b32 s30, v3, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    s_mov_b32 s33, s6
-; GFX10-NEXT:    s_or_saveexec_b32 s4, -1
+; GFX10-NEXT:    s_xor_saveexec_b32 s4, -1
 ; GFX10-NEXT:    buffer_load_dword v3, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s4
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    s_setpc_b64 s[30:31]
 entry:
   %result = call <4 x bfloat> @test_arg_store_v2bf16(<4 x bfloat> %in)
   store volatile <4 x bfloat> %result, ptr addrspace(5) %out
   ret void
 }
 
 define void @test_call_v8bf16(<8 x bfloat> %in, ptr addrspace(5) %out) {
 ; GCN-LABEL: test_call_v8bf16:
 ; GCN:       ; %bb.0: ; %entry
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GCN-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_store_dword v9, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_mov_b32 s8, s33
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
 ; GCN-NEXT:    s_waitcnt expcnt(0)
 ; GCN-NEXT:    v_writelane_b32 v9, s30, 0
 ; GCN-NEXT:    v_writelane_b32 v9, s31, 1

 ; GCN-NEXT:    buffer_store_short v1, v16, s[0:3], 0 offen
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
 ; GCN-NEXT:    buffer_store_short v0, v8, s[0:3], 0 offen
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
 ; GCN-NEXT:    v_readlane_b32 s31, v9, 1
 ; GCN-NEXT:    v_readlane_b32 s30, v9, 0
 ; GCN-NEXT:    s_addk_i32 s32, 0xfc00
 ; GCN-NEXT:    s_mov_b32 s33, s8
-; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GCN-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_load_dword v9, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0)
 ; GCN-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX7-LABEL: test_call_v8bf16:
 ; GFX7:       ; %bb.0: ; %entry
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX7-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX7-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX7-NEXT:    buffer_store_dword v9, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX7-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX7-NEXT:    s_mov_b32 s8, s33
 ; GFX7-NEXT:    s_mov_b32 s33, s32
 ; GFX7-NEXT:    s_addk_i32 s32, 0x400
 ; GFX7-NEXT:    s_getpc_b64 s[4:5]
 ; GFX7-NEXT:    s_add_u32 s4, s4, test_arg_store_v2bf16@gotpcrel32@lo+4
 ; GFX7-NEXT:    s_addc_u32 s5, s5, test_arg_store_v2bf16@gotpcrel32@hi+12

 ; GFX7-NEXT:    buffer_store_short v1, v2, s[0:3], 0 offen
 ; GFX7-NEXT:    s_waitcnt vmcnt(0)
 ; GFX7-NEXT:    buffer_store_short v0, v8, s[0:3], 0 offen
 ; GFX7-NEXT:    s_waitcnt vmcnt(0)
 ; GFX7-NEXT:    v_readlane_b32 s31, v9, 1
 ; GFX7-NEXT:    v_readlane_b32 s30, v9, 0
 ; GFX7-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX7-NEXT:    s_mov_b32 s33, s8
-; GFX7-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX7-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX7-NEXT:    buffer_load_dword v9, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX7-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX7-NEXT:    s_waitcnt vmcnt(0)
 ; GFX7-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX8-LABEL: test_call_v8bf16:
 ; GFX8:       ; %bb.0: ; %entry
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX8-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX8-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX8-NEXT:    buffer_store_dword v5, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX8-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX8-NEXT:    s_mov_b32 s6, s33
 ; GFX8-NEXT:    s_mov_b32 s33, s32
 ; GFX8-NEXT:    s_addk_i32 s32, 0x400
 ; GFX8-NEXT:    s_getpc_b64 s[4:5]
 ; GFX8-NEXT:    s_add_u32 s4, s4, test_arg_store_v2bf16@gotpcrel32@lo+4
 ; GFX8-NEXT:    s_addc_u32 s5, s5, test_arg_store_v2bf16@gotpcrel32@hi+12

 ; GFX8-NEXT:    s_waitcnt vmcnt(0)
 ; GFX8-NEXT:    v_add_u32_e32 v0, vcc, 2, v4
 ; GFX8-NEXT:    buffer_store_short v6, v0, s[0:3], 0 offen
 ; GFX8-NEXT:    s_waitcnt vmcnt(0)
 ; GFX8-NEXT:    v_readlane_b32 s31, v5, 1
 ; GFX8-NEXT:    v_readlane_b32 s30, v5, 0
 ; GFX8-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX8-NEXT:    s_mov_b32 s33, s6
-; GFX8-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX8-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX8-NEXT:    buffer_load_dword v5, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX8-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX8-NEXT:    s_waitcnt vmcnt(0)
 ; GFX8-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX9-LABEL: test_call_v8bf16:
 ; GFX9:       ; %bb.0: ; %entry
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX9-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX9-NEXT:    buffer_store_dword v5, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX9-NEXT:    s_mov_b32 s6, s33
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    s_getpc_b64 s[4:5]
 ; GFX9-NEXT:    s_add_u32 s4, s4, test_arg_store_v2bf16@gotpcrel32@lo+4
 ; GFX9-NEXT:    s_addc_u32 s5, s5, test_arg_store_v2bf16@gotpcrel32@hi+12

 ; GFX9-NEXT:    buffer_store_short_d16_hi v0, v4, s[0:3], 0 offen offset:2
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    buffer_store_short v0, v4, s[0:3], 0 offen
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    v_readlane_b32 s31, v5, 1
 ; GFX9-NEXT:    v_readlane_b32 s30, v5, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    s_mov_b32 s33, s6
-; GFX9-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX9-NEXT:    buffer_load_dword v5, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: test_call_v8bf16:
 ; GFX10:       ; %bb.0: ; %entry
 ; 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:    s_xor_saveexec_b32 s4, -1
 ; GFX10-NEXT:    buffer_store_dword v5, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s4
 ; GFX10-NEXT:    s_mov_b32 s6, s33
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    s_getpc_b64 s[4:5]
 ; GFX10-NEXT:    s_add_u32 s4, s4, test_arg_store_v2bf16@gotpcrel32@lo+4

 ; GFX10-NEXT:    buffer_store_short_d16_hi v0, v4, s[0:3], 0 offen offset:2
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    buffer_store_short v0, v4, s[0:3], 0 offen
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    v_readlane_b32 s31, v5, 1
 ; GFX10-NEXT:    v_readlane_b32 s30, v5, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    s_mov_b32 s33, s6
-; GFX10-NEXT:    s_or_saveexec_b32 s4, -1
+; GFX10-NEXT:    s_xor_saveexec_b32 s4, -1
 ; GFX10-NEXT:    buffer_load_dword v5, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s4
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    s_setpc_b64 s[30:31]
 entry:
   %result = call <8 x bfloat> @test_arg_store_v2bf16(<8 x bfloat> %in)
   store volatile <8 x bfloat> %result, ptr addrspace(5) %out
   ret void
 }
 
 define void @test_call_v16bf16(<16 x bfloat> %in, ptr addrspace(5) %out) {
 ; GCN-LABEL: test_call_v16bf16:
 ; GCN:       ; %bb.0: ; %entry
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GCN-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_store_dword v17, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_mov_b32 s8, s33
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
 ; GCN-NEXT:    s_waitcnt expcnt(0)
 ; GCN-NEXT:    v_writelane_b32 v17, s30, 0
 ; GCN-NEXT:    v_writelane_b32 v17, s31, 1

 ; GCN-NEXT:    buffer_store_short v1, v14, s[0:3], 0 offen
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
 ; GCN-NEXT:    buffer_store_short v0, v16, s[0:3], 0 offen
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
 ; GCN-NEXT:    v_readlane_b32 s31, v17, 1
 ; GCN-NEXT:    v_readlane_b32 s30, v17, 0
 ; GCN-NEXT:    s_addk_i32 s32, 0xfc00
 ; GCN-NEXT:    s_mov_b32 s33, s8
-; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GCN-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_load_dword v17, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0)
 ; GCN-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX7-LABEL: test_call_v16bf16:
 ; GFX7:       ; %bb.0: ; %entry
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX7-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX7-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX7-NEXT:    buffer_store_dword v17, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX7-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX7-NEXT:    s_mov_b32 s8, s33
 ; GFX7-NEXT:    s_mov_b32 s33, s32
 ; GFX7-NEXT:    s_addk_i32 s32, 0x400
 ; GFX7-NEXT:    s_getpc_b64 s[4:5]
 ; GFX7-NEXT:    s_add_u32 s4, s4, test_arg_store_v2bf16@gotpcrel32@lo+4
 ; GFX7-NEXT:    s_addc_u32 s5, s5, test_arg_store_v2bf16@gotpcrel32@hi+12

 ; GFX7-NEXT:    buffer_store_short v1, v2, s[0:3], 0 offen
 ; GFX7-NEXT:    s_waitcnt vmcnt(0)
 ; GFX7-NEXT:    buffer_store_short v0, v16, s[0:3], 0 offen
 ; GFX7-NEXT:    s_waitcnt vmcnt(0)
 ; GFX7-NEXT:    v_readlane_b32 s31, v17, 1
 ; GFX7-NEXT:    v_readlane_b32 s30, v17, 0
 ; GFX7-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX7-NEXT:    s_mov_b32 s33, s8
-; GFX7-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX7-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX7-NEXT:    buffer_load_dword v17, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX7-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX7-NEXT:    s_waitcnt vmcnt(0)
 ; GFX7-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX8-LABEL: test_call_v16bf16:
 ; GFX8:       ; %bb.0: ; %entry
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX8-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX8-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX8-NEXT:    buffer_store_dword v9, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX8-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX8-NEXT:    s_mov_b32 s6, s33
 ; GFX8-NEXT:    s_mov_b32 s33, s32
 ; GFX8-NEXT:    s_addk_i32 s32, 0x400
 ; GFX8-NEXT:    s_getpc_b64 s[4:5]
 ; GFX8-NEXT:    s_add_u32 s4, s4, test_arg_store_v2bf16@gotpcrel32@lo+4
 ; GFX8-NEXT:    s_addc_u32 s5, s5, test_arg_store_v2bf16@gotpcrel32@hi+12

 ; GFX8-NEXT:    s_waitcnt vmcnt(0)
 ; GFX8-NEXT:    v_add_u32_e32 v0, vcc, 2, v8
 ; GFX8-NEXT:    buffer_store_short v10, v0, s[0:3], 0 offen
 ; GFX8-NEXT:    s_waitcnt vmcnt(0)
 ; GFX8-NEXT:    v_readlane_b32 s31, v9, 1
 ; GFX8-NEXT:    v_readlane_b32 s30, v9, 0
 ; GFX8-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX8-NEXT:    s_mov_b32 s33, s6
-; GFX8-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX8-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX8-NEXT:    buffer_load_dword v9, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX8-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX8-NEXT:    s_waitcnt vmcnt(0)
 ; GFX8-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX9-LABEL: test_call_v16bf16:
 ; GFX9:       ; %bb.0: ; %entry
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX9-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX9-NEXT:    buffer_store_dword v9, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX9-NEXT:    s_mov_b32 s6, s33
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    s_getpc_b64 s[4:5]
 ; GFX9-NEXT:    s_add_u32 s4, s4, test_arg_store_v2bf16@gotpcrel32@lo+4
 ; GFX9-NEXT:    s_addc_u32 s5, s5, test_arg_store_v2bf16@gotpcrel32@hi+12

 ; GFX9-NEXT:    buffer_store_short_d16_hi v0, v8, s[0:3], 0 offen offset:2
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    buffer_store_short v0, v8, s[0:3], 0 offen
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    v_readlane_b32 s31, v9, 1
 ; GFX9-NEXT:    v_readlane_b32 s30, v9, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    s_mov_b32 s33, s6
-; GFX9-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX9-NEXT:    buffer_load_dword v9, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: test_call_v16bf16:
 ; GFX10:       ; %bb.0: ; %entry
 ; 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:    s_xor_saveexec_b32 s4, -1
 ; GFX10-NEXT:    buffer_store_dword v9, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s4
 ; GFX10-NEXT:    s_mov_b32 s6, s33
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    s_getpc_b64 s[4:5]
 ; GFX10-NEXT:    s_add_u32 s4, s4, test_arg_store_v2bf16@gotpcrel32@lo+4

 ; GFX10-NEXT:    buffer_store_short_d16_hi v0, v8, s[0:3], 0 offen offset:2
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    buffer_store_short v0, v8, s[0:3], 0 offen
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    v_readlane_b32 s31, v9, 1
 ; GFX10-NEXT:    v_readlane_b32 s30, v9, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    s_mov_b32 s33, s6
-; GFX10-NEXT:    s_or_saveexec_b32 s4, -1
+; GFX10-NEXT:    s_xor_saveexec_b32 s4, -1
 ; GFX10-NEXT:    buffer_load_dword v9, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s4
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    s_setpc_b64 s[30:31]
 entry:
   %result = call <16 x bfloat> @test_arg_store_v2bf16(<16 x bfloat> %in)
   store volatile <16 x bfloat> %result, ptr addrspace(5) %out

llvm/test/CodeGen/AMDGPU/branch-relax-spill.ll

bb3:
   tail call void asm sideeffect "; reg use $0", "{vcc_hi}"(i32 %vcc_hi) #0
   ret void
 }
 
 define void @spill_func(i32 addrspace(1)* %arg) #0 {
 ; CHECK-LABEL: spill_func:
 ; 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:    s_xor_saveexec_b64 s[4:5], -1
 ; CHECK-NEXT:    buffer_store_dword v0, off, s[0:3], s32 ; 4-byte Folded Spill
 ; CHECK-NEXT:    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:    s_waitcnt expcnt(1)
 ; CHECK-NEXT:    v_writelane_b32 v0, s30, 0
 ; CHECK-NEXT:    v_writelane_b32 v0, s31, 1
 ; CHECK-NEXT:    v_writelane_b32 v0, s33, 2
 ; CHECK-NEXT:    v_writelane_b32 v0, s34, 3

 ; CHECK-NEXT:    v_readlane_b32 s38, v0, 7
 ; CHECK-NEXT:    v_readlane_b32 s37, v0, 6
 ; CHECK-NEXT:    v_readlane_b32 s36, v0, 5
 ; CHECK-NEXT:    v_readlane_b32 s35, v0, 4
 ; CHECK-NEXT:    v_readlane_b32 s34, v0, 3
 ; CHECK-NEXT:    v_readlane_b32 s33, v0, 2
 ; CHECK-NEXT:    v_readlane_b32 s31, v0, 1
 ; CHECK-NEXT:    v_readlane_b32 s30, v0, 0
-; CHECK-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; CHECK-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; CHECK-NEXT:    buffer_load_dword v0, off, s[0:3], s32 ; 4-byte Folded Reload
 ; CHECK-NEXT:    buffer_load_dword v1, 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]
 entry:
   %cnd = tail call i32 @llvm.amdgcn.workgroup.id.x() #0
   %sgpr0 = tail call i32 asm sideeffect "s_mov_b32 s0, 0", "={s0}"() #0

llvm/test/CodeGen/AMDGPU/callee-frame-setup.ll

define void @callee_func_sgpr_spill_no_calls(i32 %in) #0 {
   ret void
 }
 
 ; Has no spilled CSR VGPRs used for SGPR spilling, so no need to
 ; enable all lanes and restore.
 
 ; GCN-LABEL: {{^}}spill_only_csr_sgpr:
 ; GCN: s_waitcnt
-; GCN-NEXT: s_or_saveexec_b64
+; GCN-NEXT: s_xor_saveexec_b64
 ; 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: v_writelane_b32 v0, s42, 0
 ; GCN-NEXT: ;;#ASMSTART
 ; GCN-NEXT: ; clobber s42
 ; GCN-NEXT: ;;#ASMEND
 ; GCN-NEXT: v_readlane_b32 s42, v0, 0
-; GCN-NEXT: s_or_saveexec_b64
+; GCN-NEXT: s_xor_saveexec_b64
 ; MUBUF-NEXT: buffer_load_dword v0, off, s[0:3], s32 ; 4-byte Folded Reload
 ; FLATSCR-NEXT: scratch_load_dword v0, off, s32 ; 4-byte Folded Reload
 ; GCN-NEXT: s_mov_b64 exec,
 ; GCN-NEXT: s_waitcnt vmcnt(0)
 ; GCN-NEXT: s_setpc_b64
 define void @spill_only_csr_sgpr() {
   call void asm sideeffect "; clobber s42", "~{s42}"()
   ret void
define void @callee_with_stack_no_fp_elim_csr_vgpr() #1 {
   store volatile i32 0, ptr addrspace(5) %alloca
   call void asm sideeffect "; clobber v41", "~{v41}"()
   ret void
 }
 
 ; Use a copy to a free SGPR instead of introducing a second CSR VGPR.
 ; GCN-LABEL: {{^}}last_lane_vgpr_for_fp_csr:
 ; GCN: s_waitcnt
-; GCN-NEXT: s_or_saveexec_b64 [[COPY_EXEC0:s\[[0-9]+:[0-9]+\]]], -1{{$}}
+; GCN-NEXT: s_xor_saveexec_b64 [[COPY_EXEC0:s\[[0-9]+:[0-9]+\]]], -1{{$}}
 ; MUBUF-NEXT:   buffer_store_dword [[CSR_VGPR:v[0-9]+]], off, s[0:3], 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: s_mov_b32 s33, s32
 ; GCN: v_writelane_b32 v0
 ; 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
 ; GCN: v_writelane_b32 v0
 ; MUBUF:   buffer_store_dword v{{[0-9]+}}, off, s[0:3], s33 offset:4
 ; FLATSCR: scratch_store_dword off, v{{[0-9]+}}, s33 offset:4
 ; GCN: ;;#ASMSTART
 ; GCN: v_writelane_b32 v0
 
 ; MUBUF:        s_addk_i32 s32, 0x400
 ; MUBUF:        s_addk_i32 s32, 0xfc00
 ; FLATSCR:      s_add_i32 s32, s32, 16
 ; FLATSCR:      s_add_i32 s32, s32, -16
 ; GCN-NEXT: s_mov_b32 s33, [[TMP_SGPR]]
-; GCN-NEXT: s_or_saveexec_b64 [[COPY_EXEC1:s\[[0-9]+:[0-9]+\]]], -1{{$}}
+; GCN-NEXT: s_xor_saveexec_b64 [[COPY_EXEC1:s\[[0-9]+:[0-9]+\]]], -1{{$}}
 ; MUBUF-NEXT:   buffer_load_dword [[CSR_VGPR]], off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
 ; FLATSCR-NEXT: scratch_load_dword [[CSR_VGPR]], off, s32 offset:8 ; 4-byte Folded Reload
 ; GCN-NEXT: s_mov_b64 exec, [[COPY_EXEC1]]
 ; GCN-NEXT: s_waitcnt vmcnt(0)
 ; GCN-NEXT: s_setpc_b64
 define void @last_lane_vgpr_for_fp_csr() #1 {
   %alloca = alloca i32, addrspace(5)
   store volatile i32 0, ptr addrspace(5) %alloca
   call void asm sideeffect "; clobber v41", "~{v41}"()
   call void asm sideeffect "",
     "~{s40},~{s41},~{s42},~{s43},~{s44},~{s45},~{s46},~{s47},~{s48},~{s49}
     ,~{s50},~{s51},~{s52},~{s53},~{s54},~{s55},~{s56},~{s57},~{s58},~{s59}
     ,~{s60},~{s61},~{s62},~{s63},~{s64},~{s65},~{s66},~{s67},~{s68},~{s69}
     ,~{s70},~{s71},~{s72},~{s73},~{s74},~{s75},~{s76},~{s77},~{s78},~{s79}
     ,~{s80},~{s81},~{s82},~{s83},~{s84},~{s85},~{s86},~{s87},~{s88},~{s89}
     ,~{s90},~{s91},~{s92},~{s93},~{s94},~{s95},~{s96},~{s97},~{s98},~{s99}
     ,~{s100},~{s101},~{s102}"() #1
 
   ret void
 }
 
 ; Use a copy to a free SGPR instead of introducing a second CSR VGPR.
 ; GCN-LABEL: {{^}}no_new_vgpr_for_fp_csr:
 ; GCN: s_waitcnt
-; GCN-NEXT: s_or_saveexec_b64 [[COPY_EXEC0:s\[[0-9]+:[0-9]+\]]], -1{{$}}
+; GCN-NEXT: s_xor_saveexec_b64 [[COPY_EXEC0:s\[[0-9]+:[0-9]+\]]], -1{{$}}
 ; MUBUF-NEXT:   buffer_store_dword [[CSR_VGPR:v[0-9]+]], off, s[0:3], 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,
 ; FLATSCR: s_mov_b32 [[FP_COPY:s[0-9]+]], s33
 ; 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
 ; GCN: v_writelane_b32 v0,
 ; MUBUF:   buffer_store_dword
 ; FLATSCR: scratch_store_dword
 ; GCN: ;;#ASMSTART
 ; GCN: v_writelane_b32 v0,
 ; MUBUF:   buffer_load_dword v41, off, s[0:3], s33 ; 4-byte Folded Reload
 ; FLATSCR: scratch_load_dword v41, off, s33 ; 4-byte Folded Reload
 ; MUBUF:        s_addk_i32 s32, 0x400
 ; FLATSCR:      s_add_i32 s32, s32, 16
 ; GCN-COUNT-64: v_readlane_b32 s{{[0-9]+}}, v0
 ; MUBUF-NEXT:   s_addk_i32 s32, 0xfc00
 ; FLATSCR-NEXT: s_add_i32 s32, s32, -16
 ; GCN-NEXT: s_mov_b32 s33, [[FP_COPY]]
-; GCN-NEXT: s_or_saveexec_b64 [[COPY_EXEC1:s\[[0-9]+:[0-9]+\]]], -1{{$}}
+; GCN-NEXT: s_xor_saveexec_b64 [[COPY_EXEC1:s\[[0-9]+:[0-9]+\]]], -1{{$}}
 ; MUBUF-NEXT:   buffer_load_dword [[CSR_VGPR]], off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
 ; FLATSCR-NEXT: scratch_load_dword [[CSR_VGPR]], off, s32 offset:8 ; 4-byte Folded Reload
 ; GCN-NEXT: s_mov_b64 exec, [[COPY_EXEC1]]
 ; GCN-NEXT: s_waitcnt vmcnt(0)
 ; GCN-NEXT: s_setpc_b64
 define void @no_new_vgpr_for_fp_csr() #1 {
   %alloca = alloca i32, addrspace(5)
   store volatile i32 0, ptr addrspace(5) %alloca

 define void @realign_stack_no_fp_elim() #1 {
   %alloca = alloca i32, align 8192, addrspace(5)
   store volatile i32 0, ptr addrspace(5) %alloca
   ret void
 }
 
 ; GCN-LABEL: {{^}}no_unused_non_csr_sgpr_for_fp:
 ; GCN: s_waitcnt
-; GCN-NEXT: s_or_saveexec_b64 [[COPY_EXEC0:s\[[0-9]+:[0-9]+\]]], -1{{$}}
+; GCN-NEXT: s_xor_saveexec_b64 [[COPY_EXEC0:s\[[0-9]+:[0-9]+\]]], -1{{$}}
 ; MUBUF-NEXT:   buffer_store_dword [[CSR_VGPR:v[0-9]+]], off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; FLATSCR-NEXT: scratch_store_dword off, [[CSR_VGPR:v[0-9]+]], s32 offset:4 ; 4-byte Folded Spill
 ; GCN-NEXT: s_mov_b64 exec, [[COPY_EXEC0]]
 ; GCN-NEXT: s_mov_b32 vcc_lo, s33
 ; GCN-NEXT: s_mov_b32 s33, s32
 ; GCN: v_writelane_b32 [[CSR_VGPR]], s30, 0
 ; GCN: v_mov_b32_e32 [[ZERO:v[0-9]+]], 0
 ; MUBUF:        s_addk_i32 s32, 0x300
 ; FLATSCR:      s_add_i32 s32, s32, 12
 ; GCN: v_writelane_b32 [[CSR_VGPR]], s31, 1
 ; MUBUF:   buffer_store_dword [[ZERO]], off, s[0:3], s33{{$}}
 ; FLATSCR: scratch_store_dword off, [[ZERO]], s33{{$}}
 ; GCN-NEXT:     s_waitcnt vmcnt(0)
 ; GCN: ;;#ASMSTART
 ; GCN:     v_readlane_b32 s31, [[CSR_VGPR]], 1
 ; GCN:     v_readlane_b32 s30, [[CSR_VGPR]], 0
 ; MUBUF:   s_addk_i32 s32, 0xfd00
 ; FLATSCR: s_add_i32 s32, s32, -12
 ; GCN-NEXT: s_mov_b32 s33, vcc_lo
-; GCN-NEXT: s_or_saveexec_b64 [[COPY_EXEC1:s\[[0-9]+:[0-9]+\]]], -1{{$}}
+; GCN-NEXT: s_xor_saveexec_b64 [[COPY_EXEC1:s\[[0-9]+:[0-9]+\]]], -1{{$}}
 ; MUBUF-NEXT:   buffer_load_dword [[CSR_VGPR]], off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
 ; FLATSCR-NEXT: scratch_load_dword [[CSR_VGPR]], off, s32 offset:4 ; 4-byte Folded Reload
 ; GCN-NEXT: s_mov_b64 exec, [[COPY_EXEC1]]
 ; GCN-NEXT: s_waitcnt vmcnt(0)
 ; GCN-NEXT: s_setpc_b64 s[30:31]
 define void @no_unused_non_csr_sgpr_for_fp() #1 {
   %alloca = alloca i32, addrspace(5)
   store volatile i32 0, ptr addrspace(5) %alloca
call void asm sideeffect "; clobber all VGPRs",
     ,~{v30},~{v31},~{v32},~{v33},~{v34},~{v35},~{v36},~{v37},~{v38},~{v39}"()
   ret void
 }
 
 ; If we have a reserved VGPR that can be used for SGPR spills, we may still
 ; need to spill the FP to memory if there are no free lanes in the reserved
 ; VGPR.
 ; GCN-LABEL: {{^}}callee_need_to_spill_fp_to_memory_full_reserved_vgpr:
-; MUBUF:   s_or_saveexec_b64 [[COPY_EXEC1:s\[[0-9]+:[0-9]+\]]], -1{{$}}
+; MUBUF:   s_xor_saveexec_b64 [[COPY_EXEC1:s\[[0-9]+:[0-9]+\]]], -1{{$}}
 ; MUBUF:   s_mov_b64 exec, [[COPY_EXEC1]]
 ; MUBUF:   v_mov_b32_e32 [[TMP_VGPR1:v[0-9]+]], s33
 ; MUBUF:   buffer_store_dword [[TMP_VGPR1]], off, s[0:3], s32 offset:[[OFF:[0-9]+]]
 ; GCN-NOT: v_writelane_b32 v40, s33
 ; MUBUF:   s_mov_b32 s33, s32
 ; FLATSCR: s_mov_b32 s33, s0
 ; GCN-NOT: v_readlane_b32 s33, v40
 ; MUBUF:   buffer_load_dword [[TMP_VGPR2:v[0-9]+]], off, s[0:3], s32 offset:[[OFF]]
 ; MUBUF:   v_readfirstlane_b32 s33, [[TMP_VGPR2]]
-; MUBUF:   s_or_saveexec_b64 [[COPY_EXEC2:s\[[0-9]+:[0-9]+\]]], -1{{$}}
+; MUBUF:   s_xor_saveexec_b64 [[COPY_EXEC2:s\[[0-9]+:[0-9]+\]]], -1{{$}}
 ; MUBUF:   s_mov_b64 exec, [[COPY_EXEC2]]
 ; GCN:     s_setpc_b64
 define void @callee_need_to_spill_fp_to_memory_full_reserved_vgpr() #3 {
   call void asm sideeffect "; clobber nonpreserved SGPRs and 64 CSRs",
     "~{s4},~{s5},~{s6},~{s7},~{s8},~{s9}
     ,~{s10},~{s11},~{s12},~{s13},~{s14},~{s15},~{s16},~{s17},~{s18},~{s19}
     ,~{s20},~{s21},~{s22},~{s23},~{s24},~{s25},~{s26},~{s27},~{s28},~{s29}
     ,~{s40},~{s41},~{s42},~{s43},~{s44},~{s45},~{s46},~{s47},~{s48},~{s49}
call void asm sideeffect "; clobber all VGPRs except CSR v40",
     ,~{v20},~{v21},~{v22},~{v23},~{v24},~{v25},~{v26},~{v27},~{v28},~{v29}
     ,~{v30},~{v31},~{v32},~{v33},~{v34},~{v35},~{v36},~{v37},~{v38},~{v39}"()
   ret void
 }
 
 ; If the size of the offset exceeds the MUBUF offset field we need another
 ; scratch VGPR to hold the offset.
 ; GCN-LABEL: {{^}}spill_fp_to_memory_scratch_reg_needed_mubuf_offset
-; MUBUF: s_or_saveexec_b64 s[4:5], -1
+; MUBUF: s_xor_saveexec_b64 s[4:5], -1
 ; MUBUF-NEXT: s_add_i32 [[SCRATCH_SGPR:s[0-9]+]], s32, 0x40100
 ; 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: 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

llvm/test/CodeGen/AMDGPU/csr-sgpr-spill-live-ins.mir

machineFunctionInfo:
   scratchRSrcReg: $sgpr0_sgpr1_sgpr2_sgpr3
   stackPtrOffsetReg: $sgpr32
 body: |
   ; CHECK-LABEL: name: def_csr_sgpr
   ; CHECK: bb.0:
   ; CHECK-NEXT:   successors: %bb.1(0x80000000)
   ; CHECK-NEXT:   liveins: $sgpr42, $sgpr43, $sgpr46, $sgpr47, $vgpr0
   ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   $sgpr4_sgpr5 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+  ; CHECK-NEXT:   $sgpr4_sgpr5 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
   ; CHECK-NEXT:   BUFFER_STORE_DWORD_OFFSET $vgpr0, $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr32, 0, 0, 0, implicit $exec :: (store (s32) into %stack.4, addrspace 5)
   ; CHECK-NEXT:   $exec = S_MOV_B64 killed $sgpr4_sgpr5
   ; CHECK-NEXT:   $vgpr0 = V_WRITELANE_B32 $sgpr42, 0, $vgpr0
   ; CHECK-NEXT:   $vgpr0 = V_WRITELANE_B32 $sgpr43, 1, $vgpr0
   ; CHECK-NEXT:   $vgpr0 = V_WRITELANE_B32 $sgpr46, 2, $vgpr0
   ; CHECK-NEXT:   $vgpr0 = V_WRITELANE_B32 $sgpr47, 3, $vgpr0
   ; CHECK-NEXT:   S_NOP 0
   ; CHECK-NEXT: {{  $}}

llvm/test/CodeGen/AMDGPU/gfx-callable-preserved-registers.ll

; GFX11-NEXT:    s_setpc_b64 s[30:31]
   call amdgpu_gfx void @external_void_func_void()
   ret void
 }
 
 define amdgpu_gfx void @void_func_void_clobber_s28_s29() #1 {
 ; GFX9-LABEL: void_func_void_clobber_s28_s29:
 ; 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:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v0, s28, 0
 ; GFX9-NEXT:    v_writelane_b32 v0, s29, 1
 ; GFX9-NEXT:    v_writelane_b32 v0, s30, 2
 ; GFX9-NEXT:    v_writelane_b32 v0, s31, 3
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; clobber
 ; GFX9-NEXT:    ;;#ASMEND
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; clobber
 ; GFX9-NEXT:    ;;#ASMEND
 ; GFX9-NEXT:    v_readlane_b32 s31, v0, 3
 ; GFX9-NEXT:    v_readlane_b32 s30, v0, 2
 ; GFX9-NEXT:    v_readlane_b32 s29, v0, 1
 ; GFX9-NEXT:    v_readlane_b32 s28, v0, 0
-; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v0, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: void_func_void_clobber_s28_s29:
 ; GFX10:       ; %bb.0:
 ; 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:    s_xor_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_store_dword v0, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; GFX10-NEXT:    v_writelane_b32 v0, s28, 0
 ; GFX10-NEXT:    v_writelane_b32 v0, s29, 1
 ; GFX10-NEXT:    v_writelane_b32 v0, s30, 2
 ; GFX10-NEXT:    v_writelane_b32 v0, s31, 3
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; clobber
 ; GFX10-NEXT:    ;;#ASMEND
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; clobber
 ; GFX10-NEXT:    ;;#ASMEND
 ; GFX10-NEXT:    v_readlane_b32 s31, v0, 3
 ; GFX10-NEXT:    v_readlane_b32 s30, v0, 2
 ; GFX10-NEXT:    v_readlane_b32 s29, v0, 1
 ; GFX10-NEXT:    v_readlane_b32 s28, v0, 0
-; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
+; GFX10-NEXT:    s_xor_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v0, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX11-LABEL: void_func_void_clobber_s28_s29:
 ; GFX11:       ; %bb.0:
 ; GFX11-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX11-NEXT:    s_waitcnt_vscnt null, 0x0
-; GFX11-NEXT:    s_or_saveexec_b32 s0, -1
+; GFX11-NEXT:    s_xor_saveexec_b32 s0, -1
 ; GFX11-NEXT:    scratch_store_b32 off, v0, s32 ; 4-byte Folded Spill
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s0
 ; GFX11-NEXT:    v_writelane_b32 v0, s28, 0
 ; GFX11-NEXT:    v_writelane_b32 v0, s29, 1
 ; GFX11-NEXT:    v_writelane_b32 v0, s30, 2
 ; GFX11-NEXT:    v_writelane_b32 v0, s31, 3
 ; GFX11-NEXT:    ;;#ASMSTART
 ; GFX11-NEXT:    ; clobber
 ; GFX11-NEXT:    ;;#ASMEND
 ; GFX11-NEXT:    ;;#ASMSTART
 ; GFX11-NEXT:    ; clobber
 ; GFX11-NEXT:    ;;#ASMEND
 ; GFX11-NEXT:    s_delay_alu instid0(VALU_DEP_1)
 ; GFX11-NEXT:    v_readlane_b32 s31, v0, 3
 ; GFX11-NEXT:    v_readlane_b32 s30, v0, 2
 ; GFX11-NEXT:    v_readlane_b32 s29, v0, 1
 ; GFX11-NEXT:    v_readlane_b32 s28, v0, 0
-; GFX11-NEXT:    s_or_saveexec_b32 s0, -1
+; GFX11-NEXT:    s_xor_saveexec_b32 s0, -1
 ; GFX11-NEXT:    scratch_load_b32 v0, off, s32 ; 4-byte Folded Reload
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s0
 ; GFX11-NEXT:    s_waitcnt vmcnt(0)
 ; GFX11-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX11-NEXT:    s_setpc_b64 s[30:31]
   call void asm sideeffect "; clobber", "~{s[30:31]}"() #0
 ; GCN:          v_writelane_b32 v0, s28, 0
 ; GCN:          v_writelane_b32 v0, s29, 1
; GFX11-NEXT:    s_setpc_b64 s[30:31]
   call void asm sideeffect "; use $0", "{v40}"(i32 %v40)
   ret void
 }
 
 define hidden void @void_func_void_clobber_s33() #1 {
 ; GFX9-LABEL: void_func_void_clobber_s33:
 ; 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:    s_xor_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:    v_writelane_b32 v0, s33, 0
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; clobber
 ; GFX9-NEXT:    ;;#ASMEND
 ; GFX9-NEXT:    v_readlane_b32 s33, v0, 0
-; GFX9-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX9-NEXT:    buffer_load_dword v0, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: void_func_void_clobber_s33:
 ; GFX10:       ; %bb.0:
 ; 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:    s_xor_saveexec_b32 s4, -1
 ; GFX10-NEXT:    buffer_store_dword v0, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s4
 ; GFX10-NEXT:    v_writelane_b32 v0, s33, 0
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; clobber
 ; GFX10-NEXT:    ;;#ASMEND
 ; GFX10-NEXT:    v_readlane_b32 s33, v0, 0
-; GFX10-NEXT:    s_or_saveexec_b32 s4, -1
+; GFX10-NEXT:    s_xor_saveexec_b32 s4, -1
 ; GFX10-NEXT:    buffer_load_dword v0, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s4
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX11-LABEL: void_func_void_clobber_s33:
 ; GFX11:       ; %bb.0:
 ; GFX11-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX11-NEXT:    s_waitcnt_vscnt null, 0x0
-; GFX11-NEXT:    s_or_saveexec_b32 s0, -1
+; GFX11-NEXT:    s_xor_saveexec_b32 s0, -1
 ; GFX11-NEXT:    scratch_store_b32 off, v0, s32 ; 4-byte Folded Spill
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s0
 ; GFX11-NEXT:    v_writelane_b32 v0, s33, 0
 ; GFX11-NEXT:    ;;#ASMSTART
 ; GFX11-NEXT:    ; clobber
 ; GFX11-NEXT:    ;;#ASMEND
 ; GFX11-NEXT:    s_delay_alu instid0(VALU_DEP_1)
 ; GFX11-NEXT:    v_readlane_b32 s33, v0, 0
-; GFX11-NEXT:    s_or_saveexec_b32 s0, -1
+; GFX11-NEXT:    s_xor_saveexec_b32 s0, -1
 ; GFX11-NEXT:    scratch_load_b32 v0, off, s32 ; 4-byte Folded Reload
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s0
 ; GFX11-NEXT:    s_waitcnt vmcnt(0)
 ; GFX11-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX11-NEXT:    s_setpc_b64 s[30:31]
   call void asm sideeffect "; clobber", "~{s33}"() #0
   ret void
 }
 
 define hidden void @void_func_void_clobber_s34() #1 {
 ; GFX9-LABEL: void_func_void_clobber_s34:
 ; 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:    s_xor_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:    v_writelane_b32 v0, s34, 0
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; clobber
 ; GFX9-NEXT:    ;;#ASMEND
 ; GFX9-NEXT:    v_readlane_b32 s34, v0, 0
-; GFX9-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GFX9-NEXT:    buffer_load_dword v0, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: void_func_void_clobber_s34:
 ; GFX10:       ; %bb.0:
 ; 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:    s_xor_saveexec_b32 s4, -1
 ; GFX10-NEXT:    buffer_store_dword v0, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s4
 ; GFX10-NEXT:    v_writelane_b32 v0, s34, 0
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; clobber
 ; GFX10-NEXT:    ;;#ASMEND
 ; GFX10-NEXT:    v_readlane_b32 s34, v0, 0
-; GFX10-NEXT:    s_or_saveexec_b32 s4, -1
+; GFX10-NEXT:    s_xor_saveexec_b32 s4, -1
 ; GFX10-NEXT:    buffer_load_dword v0, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s4
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX11-LABEL: void_func_void_clobber_s34:
 ; GFX11:       ; %bb.0:
 ; GFX11-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX11-NEXT:    s_waitcnt_vscnt null, 0x0
-; GFX11-NEXT:    s_or_saveexec_b32 s0, -1
+; GFX11-NEXT:    s_xor_saveexec_b32 s0, -1
 ; GFX11-NEXT:    scratch_store_b32 off, v0, s32 ; 4-byte Folded Spill
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s0
 ; GFX11-NEXT:    v_writelane_b32 v0, s34, 0
 ; GFX11-NEXT:    ;;#ASMSTART
 ; GFX11-NEXT:    ; clobber
 ; GFX11-NEXT:    ;;#ASMEND
 ; GFX11-NEXT:    s_delay_alu instid0(VALU_DEP_1)
 ; GFX11-NEXT:    v_readlane_b32 s34, v0, 0
-; GFX11-NEXT:    s_or_saveexec_b32 s0, -1
+; GFX11-NEXT:    s_xor_saveexec_b32 s0, -1
 ; GFX11-NEXT:    scratch_load_b32 v0, off, s32 ; 4-byte Folded Reload
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s0
 ; GFX11-NEXT:    s_waitcnt vmcnt(0)
 ; GFX11-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX11-NEXT:    s_setpc_b64 s[30:31]
   call void asm sideeffect "; clobber", "~{s34}"() #0
   ret void
 }

llvm/test/CodeGen/AMDGPU/gfx-callable-return-types.ll

 entry:
   ret i1 1
 }
 
 define amdgpu_gfx void @call_i1() #0 {
 ; GFX9-LABEL: call_i1:
 ; GFX9:       ; %bb.0: ; %entry
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v1, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_mov_b32 s36, s33
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    s_getpc_b64 s[34:35]
 ; 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:    v_writelane_b32 v1, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v1, s31, 1
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s31, v1, 1
 ; GFX9-NEXT:    v_readlane_b32 s30, v1, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    s_mov_b32 s33, s36
-; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v1, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: call_i1:
 ; GFX10:       ; %bb.0: ; %entry
 ; 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:    s_xor_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_store_dword v1, 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:    s_mov_b32 s36, s33
 ; 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, 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:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; 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]
 ; GFX10-NEXT:    v_readlane_b32 s31, v1, 1
 ; GFX10-NEXT:    v_readlane_b32 s30, v1, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    s_mov_b32 s33, s36
-; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
+; GFX10-NEXT:    s_xor_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v1, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX11-LABEL: call_i1:
 ; GFX11:       ; %bb.0: ; %entry
 ; GFX11-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX11-NEXT:    s_waitcnt_vscnt null, 0x0
-; GFX11-NEXT:    s_or_saveexec_b32 s0, -1
+; GFX11-NEXT:    s_xor_saveexec_b32 s0, -1
 ; GFX11-NEXT:    scratch_store_b32 off, v1, s32 ; 4-byte Folded Spill
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s0
 ; GFX11-NEXT:    s_mov_b32 s2, s33
 ; GFX11-NEXT:    s_mov_b32 s33, s32
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, return_i1@gotpcrel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, return_i1@gotpcrel32@hi+12
 ; GFX11-NEXT:    v_writelane_b32 v1, s30, 0
 ; GFX11-NEXT:    s_load_b64 s[0:1], s[0:1], 0x0
 ; GFX11-NEXT:    v_writelane_b32 v1, s31, 1
 ; GFX11-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX11-NEXT:    s_delay_alu instid0(VALU_DEP_1)
 ; GFX11-NEXT:    v_readlane_b32 s31, v1, 1
 ; GFX11-NEXT:    v_readlane_b32 s30, v1, 0
 ; GFX11-NEXT:    s_add_i32 s32, s32, -16
 ; GFX11-NEXT:    s_mov_b32 s33, s2
-; GFX11-NEXT:    s_or_saveexec_b32 s0, -1
+; GFX11-NEXT:    s_xor_saveexec_b32 s0, -1
 ; GFX11-NEXT:    scratch_load_b32 v1, off, s32 ; 4-byte Folded Reload
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s0
 ; GFX11-NEXT:    s_waitcnt vmcnt(0)
 ; GFX11-NEXT:    s_setpc_b64 s[30:31]
 entry:
   call amdgpu_gfx i1 @return_i1()
   ret void
 }

 entry:
   ret i16 10
 }
 
 define amdgpu_gfx void @call_i16() #0 {
 ; GFX9-LABEL: call_i16:
 ; GFX9:       ; %bb.0: ; %entry
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v1, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_mov_b32 s36, s33
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    s_getpc_b64 s[34:35]
 ; 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:    v_writelane_b32 v1, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v1, s31, 1
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s31, v1, 1
 ; GFX9-NEXT:    v_readlane_b32 s30, v1, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    s_mov_b32 s33, s36
-; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v1, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: call_i16:
 ; GFX10:       ; %bb.0: ; %entry
 ; 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:    s_xor_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_store_dword v1, 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:    s_mov_b32 s36, s33
 ; 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, 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:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; 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]
 ; GFX10-NEXT:    v_readlane_b32 s31, v1, 1
 ; GFX10-NEXT:    v_readlane_b32 s30, v1, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    s_mov_b32 s33, s36
-; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
+; GFX10-NEXT:    s_xor_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v1, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX11-LABEL: call_i16:
 ; GFX11:       ; %bb.0: ; %entry
 ; GFX11-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX11-NEXT:    s_waitcnt_vscnt null, 0x0
-; GFX11-NEXT:    s_or_saveexec_b32 s0, -1
+; GFX11-NEXT:    s_xor_saveexec_b32 s0, -1
 ; GFX11-NEXT:    scratch_store_b32 off, v1, s32 ; 4-byte Folded Spill
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s0
 ; GFX11-NEXT:    s_mov_b32 s2, s33
 ; GFX11-NEXT:    s_mov_b32 s33, s32
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, return_i16@gotpcrel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, return_i16@gotpcrel32@hi+12
 ; GFX11-NEXT:    v_writelane_b32 v1, s30, 0
 ; GFX11-NEXT:    s_load_b64 s[0:1], s[0:1], 0x0
 ; GFX11-NEXT:    v_writelane_b32 v1, s31, 1
 ; GFX11-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX11-NEXT:    s_delay_alu instid0(VALU_DEP_1)
 ; GFX11-NEXT:    v_readlane_b32 s31, v1, 1
 ; GFX11-NEXT:    v_readlane_b32 s30, v1, 0
 ; GFX11-NEXT:    s_add_i32 s32, s32, -16
 ; GFX11-NEXT:    s_mov_b32 s33, s2
-; GFX11-NEXT:    s_or_saveexec_b32 s0, -1
+; GFX11-NEXT:    s_xor_saveexec_b32 s0, -1
 ; GFX11-NEXT:    scratch_load_b32 v1, off, s32 ; 4-byte Folded Reload
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s0
 ; GFX11-NEXT:    s_waitcnt vmcnt(0)
 ; GFX11-NEXT:    s_setpc_b64 s[30:31]
 entry:
   call amdgpu_gfx i16 @return_i16()
   ret void
 }

 entry:
   ret <2 x i16> <i16 1, i16 2>
 }
 
 define amdgpu_gfx void @call_2xi16() #0 {
 ; GFX9-LABEL: call_2xi16:
 ; GFX9:       ; %bb.0: ; %entry
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v1, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_mov_b32 s36, s33
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    s_getpc_b64 s[34:35]
 ; 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:    v_writelane_b32 v1, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v1, s31, 1
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s31, v1, 1
 ; GFX9-NEXT:    v_readlane_b32 s30, v1, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    s_mov_b32 s33, s36
-; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v1, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: call_2xi16:
 ; GFX10:       ; %bb.0: ; %entry
 ; 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:    s_xor_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_store_dword v1, 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:    s_mov_b32 s36, s33
 ; 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, 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:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; 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]
 ; GFX10-NEXT:    v_readlane_b32 s31, v1, 1
 ; GFX10-NEXT:    v_readlane_b32 s30, v1, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    s_mov_b32 s33, s36
-; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
+; GFX10-NEXT:    s_xor_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v1, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX11-LABEL: call_2xi16:
 ; GFX11:       ; %bb.0: ; %entry
 ; GFX11-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX11-NEXT:    s_waitcnt_vscnt null, 0x0
-; GFX11-NEXT:    s_or_saveexec_b32 s0, -1
+; GFX11-NEXT:    s_xor_saveexec_b32 s0, -1
 ; GFX11-NEXT:    scratch_store_b32 off, v1, s32 ; 4-byte Folded Spill
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s0
 ; GFX11-NEXT:    s_mov_b32 s2, s33
 ; GFX11-NEXT:    s_mov_b32 s33, s32
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, return_2xi16@gotpcrel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, return_2xi16@gotpcrel32@hi+12
 ; GFX11-NEXT:    v_writelane_b32 v1, s30, 0
 ; GFX11-NEXT:    s_load_b64 s[0:1], s[0:1], 0x0
 ; GFX11-NEXT:    v_writelane_b32 v1, s31, 1
 ; GFX11-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX11-NEXT:    s_delay_alu instid0(VALU_DEP_1)
 ; GFX11-NEXT:    v_readlane_b32 s31, v1, 1
 ; GFX11-NEXT:    v_readlane_b32 s30, v1, 0
 ; GFX11-NEXT:    s_add_i32 s32, s32, -16
 ; GFX11-NEXT:    s_mov_b32 s33, s2
-; GFX11-NEXT:    s_or_saveexec_b32 s0, -1
+; GFX11-NEXT:    s_xor_saveexec_b32 s0, -1
 ; GFX11-NEXT:    scratch_load_b32 v1, off, s32 ; 4-byte Folded Reload
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s0
 ; GFX11-NEXT:    s_waitcnt vmcnt(0)
 ; GFX11-NEXT:    s_setpc_b64 s[30:31]
 entry:
   call amdgpu_gfx <2 x i16> @return_2xi16()
   ret void
 }

 entry:
   ret <3 x i16> <i16 1, i16 2, i16 3>
 }
 
 define amdgpu_gfx void @call_3xi16() #0 {
 ; GFX9-LABEL: call_3xi16:
 ; GFX9:       ; %bb.0: ; %entry
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v2, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_mov_b32 s36, s33
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    s_getpc_b64 s[34:35]
 ; 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:    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]
 ; GFX9-NEXT:    v_readlane_b32 s31, v2, 1
 ; GFX9-NEXT:    v_readlane_b32 s30, v2, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    s_mov_b32 s33, s36
-; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v2, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: call_3xi16:
 ; GFX10:       ; %bb.0: ; %entry
 ; 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:    s_xor_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_store_dword v2, 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:    s_mov_b32 s36, s33
 ; 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, 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:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; 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]
 ; GFX10-NEXT:    v_readlane_b32 s31, v2, 1
 ; GFX10-NEXT:    v_readlane_b32 s30, v2, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    s_mov_b32 s33, s36
-; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
+; GFX10-NEXT:    s_xor_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v2, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX11-LABEL: call_3xi16:
 ; GFX11:       ; %bb.0: ; %entry
 ; GFX11-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX11-NEXT:    s_waitcnt_vscnt null, 0x0
-; GFX11-NEXT:    s_or_saveexec_b32 s0, -1
+; GFX11-NEXT:    s_xor_saveexec_b32 s0, -1
 ; GFX11-NEXT:    scratch_store_b32 off, v2, s32 ; 4-byte Folded Spill
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s0
 ; GFX11-NEXT:    s_mov_b32 s2, s33
 ; GFX11-NEXT:    s_mov_b32 s33, s32
 ; GFX11-NEXT:    s_add_i32 s32, s32, 16
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, return_3xi16@gotpcrel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, return_3xi16@gotpcrel32@hi+12
 ; GFX11-NEXT:    v_writelane_b32 v2, s30, 0
 ; GFX11-NEXT:    s_load_b64 s[0:1], s[0:1], 0x0
 ; GFX11-NEXT:    v_writelane_b32 v2, s31, 1
 ; GFX11-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX11-NEXT:    s_delay_alu instid0(VALU_DEP_1)
 ; GFX11-NEXT:    v_readlane_b32 s31, v2, 1
 ; GFX11-NEXT:    v_readlane_b32 s30, v2, 0
 ; GFX11-NEXT:    s_add_i32 s32, s32, -16
 ; GFX11-NEXT:    s_mov_b32 s33, s2
-; GFX11-NEXT:    s_or_saveexec_b32 s0, -1
+; GFX11-NEXT:    s_xor_saveexec_b32 s0, -1
 ; GFX11-NEXT:    scratch_load_b32 v2, off, s32 ; 4-byte Folded Reload
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s0
 ; GFX11-NEXT:    s_waitcnt vmcnt(0)
 ; GFX11-NEXT:    s_setpc_b64 s[30:31]
 entry:
   call amdgpu_gfx <3 x i16> @return_3xi16()
   ret void
 }

 entry:
   ret <512 x i32> zeroinitializer
 }
 
 define amdgpu_gfx void @call_512xi32() #0 {
 ; GFX9-LABEL: call_512xi32:
 ; GFX9:       ; %bb.0: ; %entry
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v2, off, s[0:3], s32 offset:2048 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_mov_b32 s36, s33
 ; GFX9-NEXT:    s_add_i32 s33, s32, 0x1ffc0
 ; GFX9-NEXT:    s_and_b32 s33, s33, 0xfffe0000
 ; GFX9-NEXT:    s_add_i32 s32, s32, 0x60000
 ; GFX9-NEXT:    s_getpc_b64 s[34:35]
 ; 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:    v_lshrrev_b32_e64 v0, 6, s33
 ; 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]
 ; GFX9-NEXT:    v_readlane_b32 s31, v2, 1
 ; GFX9-NEXT:    v_readlane_b32 s30, v2, 0
 ; GFX9-NEXT:    s_add_i32 s32, s32, 0xfffa0000
 ; GFX9-NEXT:    s_mov_b32 s33, s36
-; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v2, off, s[0:3], s32 offset:2048 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: call_512xi32:
 ; GFX10:       ; %bb.0: ; %entry
 ; 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:    s_xor_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_store_dword v2, off, s[0:3], s32 offset:2048 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; GFX10-NEXT:    s_mov_b32 s36, s33
 ; GFX10-NEXT:    s_add_i32 s33, s32, 0xffe0
 ; GFX10-NEXT:    s_add_i32 s32, s32, 0x30000
 ; GFX10-NEXT:    s_and_b32 s33, s33, 0xffff0000
 ; 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:    v_lshrrev_b32_e64 v0, 5, s33
 ; 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]
 ; GFX10-NEXT:    v_readlane_b32 s31, v2, 1
 ; GFX10-NEXT:    v_readlane_b32 s30, v2, 0
 ; GFX10-NEXT:    s_add_i32 s32, s32, 0xfffd0000
 ; GFX10-NEXT:    s_mov_b32 s33, s36
-; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
+; GFX10-NEXT:    s_xor_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v2, off, s[0:3], s32 offset:2048 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX11-LABEL: call_512xi32:
 ; GFX11:       ; %bb.0: ; %entry
 ; GFX11-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX11-NEXT:    s_waitcnt_vscnt null, 0x0
-; GFX11-NEXT:    s_or_saveexec_b32 s0, -1
+; GFX11-NEXT:    s_xor_saveexec_b32 s0, -1
 ; GFX11-NEXT:    scratch_store_b32 off, v5, s32 offset:2048 ; 4-byte Folded Spill
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s0
 ; GFX11-NEXT:    s_mov_b32 s34, s33
 ; GFX11-NEXT:    s_add_i32 s33, s32, 0x7ff
 ; GFX11-NEXT:    s_addk_i32 s32, 0x1800
 ; GFX11-NEXT:    s_and_b32 s33, s33, 0xfffff800
 ; GFX11-NEXT:    s_getpc_b64 s[0:1]
 ; GFX11-NEXT:    s_add_u32 s0, s0, return_512xi32@gotpcrel32@lo+4
 ; GFX11-NEXT:    s_addc_u32 s1, s1, return_512xi32@gotpcrel32@hi+12
 ; GFX11-NEXT:    v_writelane_b32 v5, s30, 0
 ; GFX11-NEXT:    s_load_b64 s[0:1], s[0:1], 0x0
 ; GFX11-NEXT:    v_mov_b32_e32 v0, s33
 ; GFX11-NEXT:    v_writelane_b32 v5, s31, 1
 ; GFX11-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX11-NEXT:    s_swappc_b64 s[30:31], s[0:1]
 ; GFX11-NEXT:    s_delay_alu instid0(VALU_DEP_1)
 ; GFX11-NEXT:    v_readlane_b32 s31, v5, 1
 ; GFX11-NEXT:    v_readlane_b32 s30, v5, 0
 ; GFX11-NEXT:    s_addk_i32 s32, 0xe800
 ; GFX11-NEXT:    s_mov_b32 s33, s34
-; GFX11-NEXT:    s_or_saveexec_b32 s0, -1
+; GFX11-NEXT:    s_xor_saveexec_b32 s0, -1
 ; GFX11-NEXT:    scratch_load_b32 v5, off, s32 offset:2048 ; 4-byte Folded Reload
 ; GFX11-NEXT:    s_mov_b32 exec_lo, s0
 ; GFX11-NEXT:    s_waitcnt vmcnt(0)
 ; GFX11-NEXT:    s_setpc_b64 s[30:31]
 entry:
   call amdgpu_gfx <512 x i32> @return_512xi32()
   ret void
 }
 
 attributes #0 = { nounwind }

llvm/test/CodeGen/AMDGPU/need-fp-from-vgpr-spills.ll

 
 ; Has no stack objects, but introduces them due to the CSR spill. We
 ; see the FP modified in the callee with IPRA. We should not have
 ; redundant spills of s33 or assert.
 define internal fastcc void @csr_vgpr_spill_fp_callee() #0 {
 ; CHECK-LABEL: csr_vgpr_spill_fp_callee:
 ; 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:    s_xor_saveexec_b64 s[4:5], -1
 ; CHECK-NEXT:    buffer_store_dword v1, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; CHECK-NEXT:    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:    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:    s_mov_b64 s[10:11], s[2:3]
 ; CHECK-NEXT:    s_mov_b64 s[8:9], s[0:1]
 ; 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:    ;;#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:    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:    s_xor_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:    s_mov_b64 exec, s[4:5]
 ; CHECK-NEXT:    s_waitcnt vmcnt(0)
 ; CHECK-NEXT:    s_setpc_b64 s[30:31]
 bb:
   call fastcc void @callee_has_fp()
   call void asm sideeffect "; clobber csr v40", "~{v40}"()
   ret void
bb:
   ret void
 }
 
 ; Same, except with a tail call.
 define internal fastcc void @csr_vgpr_spill_fp_tailcall_callee() #0 {
 ; CHECK-LABEL: csr_vgpr_spill_fp_tailcall_callee:
 ; 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:    s_xor_saveexec_b64 s[4:5], -1
 ; CHECK-NEXT:    buffer_store_dword v1, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; CHECK-NEXT:    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:    ;;#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:    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:    s_xor_saveexec_b64 s[6:7], -1
 ; CHECK-NEXT:    buffer_load_dword v1, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
 ; CHECK-NEXT:    s_mov_b64 exec, s[6:7]
 ; CHECK-NEXT:    s_setpc_b64 s[4:5]
 bb:
   call void asm sideeffect "; clobber csr v40", "~{v40}"()
   tail call fastcc void @callee_has_fp()
   ret void
 }

 entry:
   ret i32 0
 }
 
 define hidden i32 @caller_save_vgpr_spill_fp_tail_call() #0 {
 ; CHECK-LABEL: caller_save_vgpr_spill_fp_tail_call:
 ; CHECK:       ; %bb.0: ; %entry
 ; CHECK-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; CHECK-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; CHECK-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; CHECK-NEXT:    buffer_store_dword v1, off, s[0:3], s32 ; 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:    v_writelane_b32 v1, s30, 0
 ; CHECK-NEXT:    v_writelane_b32 v1, s31, 1
 ; CHECK-NEXT:    s_getpc_b64 s[4:5]
 ; CHECK-NEXT:    s_add_u32 s4, s4, tail_call@rel32@lo+4
 ; CHECK-NEXT:    s_addc_u32 s5, s5, tail_call@rel32@hi+12
 ; CHECK-NEXT:    s_mov_b64 s[10:11], s[2:3]
 ; CHECK-NEXT:    s_mov_b64 s[8:9], s[0:1]
 ; 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, v1, 1
 ; CHECK-NEXT:    v_readlane_b32 s30, v1, 0
 ; 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:    s_xor_saveexec_b64 s[4:5], -1
 ; CHECK-NEXT:    buffer_load_dword v1, off, s[0:3], s32 ; 4-byte Folded Reload
 ; CHECK-NEXT:    s_mov_b64 exec, s[4:5]
 ; CHECK-NEXT:    s_waitcnt vmcnt(0)
 ; CHECK-NEXT:    s_setpc_b64 s[30:31]
 entry:
   %call = call i32 @tail_call()
   ret i32 %call
 }
 
 define hidden i32 @caller_save_vgpr_spill_fp() #0 {
 ; CHECK-LABEL: caller_save_vgpr_spill_fp:
 ; 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:    s_xor_saveexec_b64 s[4:5], -1
 ; CHECK-NEXT:    buffer_store_dword v2, off, s[0:3], s32 ; 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:    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
 ; CHECK-NEXT:    s_mov_b64 s[10:11], s[2:3]
 ; CHECK-NEXT:    s_mov_b64 s[8:9], s[0:1]
 ; CHECK-NEXT:    s_mov_b64 s[0:1], s[8:9]
 ; CHECK-NEXT:    s_mov_b64 s[2:3], s[10:11]
 ; CHECK-NEXT:    s_swappc_b64 s[30:31], s[4:5]
 ; CHECK-NEXT:    v_readlane_b32 s31, v2, 1
 ; CHECK-NEXT:    v_readlane_b32 s30, v2, 0
 ; CHECK-NEXT:    s_add_i32 s32, s32, 0xfffffc00
 ; CHECK-NEXT:    s_mov_b32 s33, s7
-; CHECK-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; CHECK-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; CHECK-NEXT:    buffer_load_dword v2, off, s[0:3], s32 ; 4-byte Folded Reload
 ; CHECK-NEXT:    s_mov_b64 exec, s[4:5]
 ; CHECK-NEXT:    s_waitcnt vmcnt(0)
 ; CHECK-NEXT:    s_setpc_b64 s[30:31]
 entry:
   %call = call i32 @caller_save_vgpr_spill_fp_tail_call()
   ret i32 %call
 }

llvm/test/CodeGen/AMDGPU/pei-scavenge-sgpr-carry-out.mir

 
 body:             |
   bb.0:
     liveins: $vgpr1
 
     ; CHECK-LABEL: name: scavenge_sgpr_pei_no_sgprs
     ; CHECK: liveins: $vgpr1, $vgpr2
     ; CHECK-NEXT: {{  $}}
-    ; CHECK-NEXT: $sgpr4_sgpr5 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+    ; CHECK-NEXT: $sgpr4_sgpr5 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
     ; CHECK-NEXT: $sgpr6 = S_ADD_I32 $sgpr32, 1048832, implicit-def dead $scc
     ; CHECK-NEXT: BUFFER_STORE_DWORD_OFFSET $vgpr2, $sgpr0_sgpr1_sgpr2_sgpr3, killed $sgpr6, 0, 0, 0, implicit $exec :: (store (s32) into %stack.3, addrspace 5)
     ; CHECK-NEXT: $exec = S_MOV_B64 killed $sgpr4_sgpr5
     ; CHECK-NEXT: $vgpr2 = V_WRITELANE_B32 $sgpr33, 0, undef $vgpr2
     ; CHECK-NEXT: $sgpr33 = frame-setup S_ADD_I32 $sgpr32, 524224, implicit-def $scc
     ; CHECK-NEXT: $sgpr33 = frame-setup S_AND_B32 killed $sgpr33, 4294443008, implicit-def dead $scc
     ; CHECK-NEXT: $sgpr32 = frame-setup S_ADD_I32 $sgpr32, 2097152, implicit-def dead $scc
     ; CHECK-NEXT: S_NOP 0, implicit-def $sgpr4, implicit-def $sgpr5, implicit-def $sgpr6, implicit-def $sgpr7, implicit-def $sgpr8, implicit-def $sgpr9, implicit-def $sgpr10, implicit-def $sgpr11, implicit-def $sgpr12, implicit-def $sgpr13, implicit-def $sgpr14, implicit-def $sgpr15, implicit-def $sgpr16, implicit-def $sgpr17, implicit-def $sgpr18, implicit-def $sgpr19, implicit-def $sgpr20, implicit-def $sgpr21, implicit-def $sgpr22, implicit-def $sgpr23, implicit-def $sgpr24, implicit-def $sgpr25, implicit-def $sgpr26, implicit-def $sgpr27, implicit-def $sgpr28, implicit-def $sgpr29, implicit-def $sgpr30, implicit-def $sgpr31, implicit-def $vcc
     ; CHECK-NEXT: $sgpr33 = S_LSHR_B32 $sgpr33, 6, implicit-def $scc
     ; CHECK-NEXT: $sgpr33 = S_ADD_I32 killed $sgpr33, 8192, implicit-def $scc
     ; CHECK-NEXT: $vgpr0 = COPY killed $sgpr33
     ; CHECK-NEXT: $sgpr33 = S_ADD_I32 killed $sgpr33, -8192, implicit-def $scc
     ; CHECK-NEXT: $sgpr33 = S_LSHL_B32 $sgpr33, 6, implicit-def $scc
     ; CHECK-NEXT: $sgpr33 = S_LSHR_B32 $sgpr33, 6, implicit-def $scc
     ; CHECK-NEXT: $sgpr33 = S_ADD_I32 killed $sgpr33, 16384, implicit-def $scc
     ; CHECK-NEXT: $vgpr3 = COPY killed $sgpr33
     ; CHECK-NEXT: $sgpr33 = S_ADD_I32 killed $sgpr33, -16384, implicit-def $scc
     ; CHECK-NEXT: $sgpr33 = S_LSHL_B32 $sgpr33, 6, implicit-def $scc
     ; CHECK-NEXT: $vgpr0 = V_OR_B32_e32 killed $vgpr3, $vgpr1, implicit $exec, implicit $sgpr4, implicit $sgpr5, implicit $sgpr6, implicit $sgpr7, implicit $sgpr8, implicit $sgpr9, implicit $sgpr10, implicit $sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $sgpr15, implicit $sgpr16, implicit $sgpr17, implicit $sgpr18, implicit $sgpr19, implicit $sgpr20, implicit $sgpr21, implicit $sgpr22, implicit $sgpr23, implicit $sgpr24, implicit $sgpr25, implicit $sgpr26, implicit $sgpr27, implicit $sgpr28, implicit $sgpr29, implicit $sgpr30, implicit $sgpr31
     ; CHECK-NEXT: $sgpr32 = frame-destroy S_ADD_I32 $sgpr32, -2097152, implicit-def dead $scc
     ; CHECK-NEXT: $sgpr33 = V_READLANE_B32 $vgpr2, 0
-    ; CHECK-NEXT: $sgpr4_sgpr5 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+    ; CHECK-NEXT: $sgpr4_sgpr5 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
     ; CHECK-NEXT: $sgpr6 = S_ADD_I32 $sgpr32, 1048832, implicit-def dead $scc
     ; CHECK-NEXT: $vgpr2 = BUFFER_LOAD_DWORD_OFFSET $sgpr0_sgpr1_sgpr2_sgpr3, killed $sgpr6, 0, 0, 0, implicit $exec :: (load (s32) from %stack.3, addrspace 5)
     ; CHECK-NEXT: $exec = S_MOV_B64 killed $sgpr4_sgpr5
     ; CHECK-NEXT: S_ENDPGM 0, implicit $vcc
     S_NOP 0, implicit-def $sgpr4, implicit-def $sgpr5, implicit-def $sgpr6, implicit-def $sgpr7, implicit-def $sgpr8, implicit-def $sgpr9, implicit-def $sgpr10, implicit-def $sgpr11, implicit-def $sgpr12, implicit-def $sgpr13, implicit-def $sgpr14, implicit-def $sgpr15, implicit-def $sgpr16, implicit-def $sgpr17, implicit-def $sgpr18, implicit-def $sgpr19, implicit-def $sgpr20, implicit-def $sgpr21, implicit-def $sgpr22, implicit-def $sgpr23, implicit-def $sgpr24, implicit-def $sgpr25, implicit-def $sgpr26, implicit-def $sgpr27, implicit-def $sgpr28, implicit-def $sgpr29, implicit-def $sgpr30, implicit-def $sgpr31, implicit-def $vcc
     $vgpr0 = V_MOV_B32_e32 %stack.0, implicit $exec
     $vgpr0 = V_OR_B32_e32 %stack.1, $vgpr1, implicit $exec, implicit $sgpr4, implicit $sgpr5, implicit $sgpr6, implicit $sgpr7, implicit $sgpr8, implicit $sgpr9, implicit $sgpr10, implicit $sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $sgpr15, implicit $sgpr16, implicit $sgpr17, implicit $sgpr18, implicit $sgpr19, implicit $sgpr20, implicit $sgpr21, implicit $sgpr22, implicit $sgpr23, implicit $sgpr24, implicit $sgpr25, implicit $sgpr26, implicit $sgpr27, implicit $sgpr28, implicit $sgpr29, implicit $sgpr30, implicit $sgpr31
     S_ENDPGM 0, implicit $vcc

llvm/test/CodeGen/AMDGPU/pei-scavenge-sgpr-gfx9.mir

 
 body:             |
   bb.0:
     liveins: $vgpr1
 
     ; MUBUF-LABEL: name: scavenge_sgpr_pei_no_sgprs
     ; MUBUF: liveins: $vgpr1, $vgpr2
     ; MUBUF-NEXT: {{  $}}
-    ; MUBUF-NEXT: $sgpr4_sgpr5 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+    ; MUBUF-NEXT: $sgpr4_sgpr5 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
     ; MUBUF-NEXT: $sgpr6 = S_ADD_I32 $sgpr32, 1048832, implicit-def dead $scc
     ; MUBUF-NEXT: BUFFER_STORE_DWORD_OFFSET $vgpr2, $sgpr0_sgpr1_sgpr2_sgpr3, killed $sgpr6, 0, 0, 0, implicit $exec :: (store (s32) into %stack.3, addrspace 5)
     ; MUBUF-NEXT: $exec = S_MOV_B64 killed $sgpr4_sgpr5
     ; MUBUF-NEXT: $vgpr2 = V_WRITELANE_B32 $sgpr33, 0, undef $vgpr2
     ; MUBUF-NEXT: $sgpr33 = frame-setup S_ADD_I32 $sgpr32, 524224, implicit-def $scc
     ; MUBUF-NEXT: $sgpr33 = frame-setup S_AND_B32 killed $sgpr33, 4294443008, implicit-def dead $scc
     ; MUBUF-NEXT: $sgpr32 = frame-setup S_ADD_I32 $sgpr32, 2097152, implicit-def dead $scc
     ; MUBUF-NEXT: S_NOP 0, implicit-def $sgpr4, implicit-def $sgpr5, implicit-def $sgpr6, implicit-def $sgpr7, implicit-def $sgpr8, implicit-def $sgpr9, implicit-def $sgpr10, implicit-def $sgpr11, implicit-def $sgpr12, implicit-def $sgpr13, implicit-def $sgpr14, implicit-def $sgpr15, implicit-def $sgpr16, implicit-def $sgpr17, implicit-def $sgpr18, implicit-def $sgpr19, implicit-def $sgpr20, implicit-def $sgpr21, implicit-def $sgpr22, implicit-def $sgpr23, implicit-def $sgpr24, implicit-def $sgpr25, implicit-def $sgpr26, implicit-def $sgpr27, implicit-def $sgpr28, implicit-def $sgpr29, implicit-def $sgpr30, implicit-def $sgpr31, implicit-def $vcc
     ; MUBUF-NEXT: $vgpr0 = V_LSHRREV_B32_e64 6, $sgpr33, implicit $exec
     ; MUBUF-NEXT: $vgpr0 = V_ADD_U32_e32 8192, killed $vgpr0, implicit $exec
     ; MUBUF-NEXT: $vgpr3 = V_LSHRREV_B32_e64 6, $sgpr33, implicit $exec
     ; MUBUF-NEXT: $vgpr3 = V_ADD_U32_e32 16384, killed $vgpr3, implicit $exec
     ; MUBUF-NEXT: $vgpr0 = V_OR_B32_e32 killed $vgpr3, $vgpr1, implicit $exec, implicit $sgpr4, implicit $sgpr5, implicit $sgpr6, implicit $sgpr7, implicit $sgpr8, implicit $sgpr9, implicit $sgpr10, implicit $sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $sgpr15, implicit $sgpr16, implicit $sgpr17, implicit $sgpr18, implicit $sgpr19, implicit $sgpr20, implicit $sgpr21, implicit $sgpr22, implicit $sgpr23, implicit $sgpr24, implicit $sgpr25, implicit $sgpr26, implicit $sgpr27, implicit $sgpr28, implicit $sgpr29, implicit $sgpr30, implicit $sgpr31
     ; MUBUF-NEXT: $sgpr32 = frame-destroy S_ADD_I32 $sgpr32, -2097152, implicit-def dead $scc
     ; MUBUF-NEXT: $sgpr33 = V_READLANE_B32 $vgpr2, 0
-    ; MUBUF-NEXT: $sgpr4_sgpr5 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+    ; MUBUF-NEXT: $sgpr4_sgpr5 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
     ; MUBUF-NEXT: $sgpr6 = S_ADD_I32 $sgpr32, 1048832, implicit-def dead $scc
     ; MUBUF-NEXT: $vgpr2 = BUFFER_LOAD_DWORD_OFFSET $sgpr0_sgpr1_sgpr2_sgpr3, killed $sgpr6, 0, 0, 0, implicit $exec :: (load (s32) from %stack.3, addrspace 5)
     ; MUBUF-NEXT: $exec = S_MOV_B64 killed $sgpr4_sgpr5
     ; MUBUF-NEXT: S_ENDPGM 0, implicit $vcc
     ; FLATSCR-LABEL: name: scavenge_sgpr_pei_no_sgprs
     ; FLATSCR: liveins: $vgpr1, $vgpr2
     ; FLATSCR-NEXT: {{  $}}
-    ; FLATSCR-NEXT: $sgpr4_sgpr5 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+    ; FLATSCR-NEXT: $sgpr4_sgpr5 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
     ; FLATSCR-NEXT: $sgpr6 = S_ADD_I32 $sgpr32, 16388, implicit-def dead $scc
     ; FLATSCR-NEXT: SCRATCH_STORE_DWORD_SADDR $vgpr2, killed $sgpr6, 0, 0, implicit $exec, implicit $flat_scr :: (store (s32) into %stack.3, addrspace 5)
     ; FLATSCR-NEXT: $exec = S_MOV_B64 killed $sgpr4_sgpr5
     ; FLATSCR-NEXT: $vgpr2 = V_WRITELANE_B32 $sgpr33, 0, undef $vgpr2
     ; FLATSCR-NEXT: $sgpr33 = frame-setup S_ADD_I32 $sgpr32, 8191, implicit-def $scc
     ; FLATSCR-NEXT: $sgpr33 = frame-setup S_AND_B32 killed $sgpr33, 4294959104, implicit-def dead $scc
     ; FLATSCR-NEXT: $sgpr32 = frame-setup S_ADD_I32 $sgpr32, 32768, implicit-def dead $scc
     ; FLATSCR-NEXT: S_NOP 0, implicit-def $sgpr4, implicit-def $sgpr5, implicit-def $sgpr6, implicit-def $sgpr7, implicit-def $sgpr8, implicit-def $sgpr9, implicit-def $sgpr10, implicit-def $sgpr11, implicit-def $sgpr12, implicit-def $sgpr13, implicit-def $sgpr14, implicit-def $sgpr15, implicit-def $sgpr16, implicit-def $sgpr17, implicit-def $sgpr18, implicit-def $sgpr19, implicit-def $sgpr20, implicit-def $sgpr21, implicit-def $sgpr22, implicit-def $sgpr23, implicit-def $sgpr24, implicit-def $sgpr25, implicit-def $sgpr26, implicit-def $sgpr27, implicit-def $sgpr28, implicit-def $sgpr29, implicit-def $sgpr30, implicit-def $sgpr31, implicit-def $vcc
     ; FLATSCR-NEXT: $sgpr33 = S_ADD_I32 $sgpr33, 8192, implicit-def $scc
     ; FLATSCR-NEXT: $vgpr0 = V_MOV_B32_e32 $sgpr33, implicit $exec
     ; FLATSCR-NEXT: $sgpr33 = S_ADD_I32 $sgpr33, -8192, implicit-def $scc
     ; FLATSCR-NEXT: $sgpr33 = S_ADD_I32 $sgpr33, 16384, implicit-def $scc
     ; FLATSCR-NEXT: $vgpr0 = V_OR_B32_e32 $sgpr33, $vgpr1, implicit $exec, implicit $sgpr4, implicit $sgpr5, implicit $sgpr6, implicit $sgpr7, implicit $sgpr8, implicit $sgpr9, implicit $sgpr10, implicit $sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $sgpr15, implicit $sgpr16, implicit $sgpr17, implicit $sgpr18, implicit $sgpr19, implicit $sgpr20, implicit $sgpr21, implicit $sgpr22, implicit $sgpr23, implicit $sgpr24, implicit $sgpr25, implicit $sgpr26, implicit $sgpr27, implicit $sgpr28, implicit $sgpr29, implicit $sgpr30, implicit $sgpr31
     ; FLATSCR-NEXT: $sgpr33 = S_ADD_I32 $sgpr33, -16384, implicit-def $scc
     ; FLATSCR-NEXT: $sgpr32 = frame-destroy S_ADD_I32 $sgpr32, -32768, implicit-def dead $scc
     ; FLATSCR-NEXT: $sgpr33 = V_READLANE_B32 $vgpr2, 0
-    ; FLATSCR-NEXT: $sgpr4_sgpr5 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+    ; FLATSCR-NEXT: $sgpr4_sgpr5 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
     ; FLATSCR-NEXT: $sgpr6 = S_ADD_I32 $sgpr32, 16388, implicit-def dead $scc
     ; FLATSCR-NEXT: $vgpr2 = SCRATCH_LOAD_DWORD_SADDR killed $sgpr6, 0, 0, implicit $exec, implicit $flat_scr :: (load (s32) from %stack.3, addrspace 5)
     ; FLATSCR-NEXT: $exec = S_MOV_B64 killed $sgpr4_sgpr5
     ; FLATSCR-NEXT: S_ENDPGM 0, implicit $vcc
     S_NOP 0, implicit-def $sgpr4, implicit-def $sgpr5, implicit-def $sgpr6, implicit-def $sgpr7, implicit-def $sgpr8, implicit-def $sgpr9, implicit-def $sgpr10, implicit-def $sgpr11, implicit-def $sgpr12, implicit-def $sgpr13, implicit-def $sgpr14, implicit-def $sgpr15, implicit-def $sgpr16, implicit-def $sgpr17, implicit-def $sgpr18, implicit-def $sgpr19, implicit-def $sgpr20, implicit-def $sgpr21, implicit-def $sgpr22, implicit-def $sgpr23, implicit-def $sgpr24, implicit-def $sgpr25, implicit-def $sgpr26, implicit-def $sgpr27, implicit-def $sgpr28, implicit-def $sgpr29, implicit-def $sgpr30, implicit-def $sgpr31, implicit-def $vcc
     $vgpr0 = V_MOV_B32_e32 %stack.0, implicit $exec
     $vgpr0 = V_OR_B32_e32 %stack.1, $vgpr1, implicit $exec, implicit $sgpr4, implicit $sgpr5, implicit $sgpr6, implicit $sgpr7, implicit $sgpr8, implicit $sgpr9, implicit $sgpr10, implicit $sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $sgpr15, implicit $sgpr16, implicit $sgpr17, implicit $sgpr18, implicit $sgpr19, implicit $sgpr20, implicit $sgpr21, implicit $sgpr22, implicit $sgpr23, implicit $sgpr24, implicit $sgpr25, implicit $sgpr26, implicit $sgpr27, implicit $sgpr28, implicit $sgpr29, implicit $sgpr30, implicit $sgpr31
     S_ENDPGM 0, implicit $vcc
 ...

llvm/test/CodeGen/AMDGPU/pei-scavenge-sgpr.mir

 
 body:             |
   bb.0:
     liveins: $vgpr1
 
     ; CHECK-LABEL: name: scavenge_sgpr_pei
     ; CHECK: liveins: $vgpr1, $vgpr2
     ; CHECK-NEXT: {{  $}}
-    ; CHECK-NEXT: $sgpr4_sgpr5 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+    ; CHECK-NEXT: $sgpr4_sgpr5 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
     ; CHECK-NEXT: $sgpr6 = S_ADD_I32 $sgpr32, 262400, implicit-def dead $scc
     ; CHECK-NEXT: BUFFER_STORE_DWORD_OFFSET $vgpr2, $sgpr0_sgpr1_sgpr2_sgpr3, killed $sgpr6, 0, 0, 0, implicit $exec :: (store (s32) into %stack.2, addrspace 5)
     ; CHECK-NEXT: $exec = S_MOV_B64 killed $sgpr4_sgpr5
     ; CHECK-NEXT: $vgpr2 = V_WRITELANE_B32 $sgpr33, 0, undef $vgpr2
     ; CHECK-NEXT: $sgpr33 = frame-setup S_ADD_I32 $sgpr32, 262080, implicit-def $scc
     ; CHECK-NEXT: $sgpr33 = frame-setup S_AND_B32 killed $sgpr33, 4294705152, implicit-def dead $scc
     ; CHECK-NEXT: $sgpr32 = frame-setup S_ADD_I32 $sgpr32, 786432, implicit-def dead $scc
     ; CHECK-NEXT: S_NOP 0, implicit-def $sgpr4, implicit-def $sgpr5, implicit-def $sgpr6, implicit-def $sgpr7, implicit-def $sgpr8, implicit-def $sgpr9, implicit-def $sgpr10, implicit-def $sgpr11, implicit-def $sgpr12, implicit-def $sgpr13, implicit-def $sgpr14, implicit-def $sgpr15, implicit-def $sgpr16, implicit-def $sgpr17, implicit-def $sgpr18, implicit-def $sgpr19, implicit-def $sgpr20, implicit-def $sgpr21, implicit-def $sgpr22, implicit-def $sgpr23, implicit-def $sgpr24, implicit-def $sgpr25, implicit-def $sgpr26, implicit-def $sgpr27, implicit-def $sgpr28, implicit-def $sgpr29, implicit-def $sgpr30, implicit-def $sgpr31, implicit-def $vcc
     ; CHECK-NEXT: $sgpr33 = S_LSHR_B32 $sgpr33, 6, implicit-def $scc
     ; CHECK-NEXT: $sgpr33 = S_ADD_I32 killed $sgpr33, 4096, implicit-def $scc
     ; CHECK-NEXT: $vgpr3 = COPY killed $sgpr33
     ; CHECK-NEXT: $sgpr33 = S_ADD_I32 killed $sgpr33, -4096, implicit-def $scc
     ; CHECK-NEXT: $sgpr33 = S_LSHL_B32 $sgpr33, 6, implicit-def $scc
     ; CHECK-NEXT: $vgpr0 = V_OR_B32_e32 killed $vgpr3, $vgpr1, implicit $exec, implicit $sgpr4, implicit $sgpr5, implicit $sgpr6, implicit $sgpr7, implicit $sgpr8, implicit $sgpr9, implicit $sgpr10, implicit $sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $sgpr15, implicit $sgpr16, implicit $sgpr17, implicit $sgpr18, implicit $sgpr19, implicit $sgpr20, implicit $sgpr21, implicit $sgpr22, implicit $sgpr23, implicit $sgpr24, implicit $sgpr25, implicit $sgpr26, implicit $sgpr27, implicit $sgpr28, implicit $sgpr29, implicit $sgpr30, implicit $sgpr31
     ; CHECK-NEXT: $sgpr32 = frame-destroy S_ADD_I32 $sgpr32, -786432, implicit-def dead $scc
     ; CHECK-NEXT: $sgpr33 = V_READLANE_B32 $vgpr2, 0
-    ; CHECK-NEXT: $sgpr4_sgpr5 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+    ; CHECK-NEXT: $sgpr4_sgpr5 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
     ; CHECK-NEXT: $sgpr6 = S_ADD_I32 $sgpr32, 262400, implicit-def dead $scc
     ; CHECK-NEXT: $vgpr2 = BUFFER_LOAD_DWORD_OFFSET $sgpr0_sgpr1_sgpr2_sgpr3, killed $sgpr6, 0, 0, 0, implicit $exec :: (load (s32) from %stack.2, addrspace 5)
     ; CHECK-NEXT: $exec = S_MOV_B64 killed $sgpr4_sgpr5
     ; CHECK-NEXT: S_ENDPGM 0, implicit $vcc
     S_NOP 0, implicit-def $sgpr4, implicit-def $sgpr5, implicit-def $sgpr6, implicit-def $sgpr7, implicit-def $sgpr8, implicit-def $sgpr9, implicit-def $sgpr10, implicit-def $sgpr11, implicit-def $sgpr12, implicit-def $sgpr13, implicit-def $sgpr14, implicit-def $sgpr15, implicit-def $sgpr16, implicit-def $sgpr17, implicit-def $sgpr18, implicit-def $sgpr19, implicit-def $sgpr20, implicit-def $sgpr21, implicit-def $sgpr22, implicit-def $sgpr23, implicit-def $sgpr24, implicit-def $sgpr25, implicit-def $sgpr26, implicit-def $sgpr27, implicit-def $sgpr28, implicit-def $sgpr29, implicit-def $sgpr30, implicit-def $sgpr31, implicit-def $vcc
     $vgpr0 = V_OR_B32_e32 %stack.0, $vgpr1, implicit $exec, implicit $sgpr4, implicit $sgpr5, implicit $sgpr6, implicit $sgpr7, implicit $sgpr8, implicit $sgpr9, implicit $sgpr10, implicit $sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $sgpr15, implicit $sgpr16, implicit $sgpr17, implicit $sgpr18, implicit $sgpr19, implicit $sgpr20, implicit $sgpr21, implicit $sgpr22, implicit $sgpr23, implicit $sgpr24, implicit $sgpr25, implicit $sgpr26, implicit $sgpr27, implicit $sgpr28, implicit $sgpr29, implicit $sgpr30, implicit $sgpr31
     S_ENDPGM 0, implicit $vcc
 ...

llvm/test/CodeGen/AMDGPU/pei-scavenge-vgpr-spill.mir

 
 body:             |
   bb.0:
     liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15, $vgpr16_vgpr17_vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31, $vgpr32_vgpr33_vgpr34_vgpr35_vgpr36_vgpr37_vgpr38_vgpr39_vgpr40_vgpr41_vgpr42_vgpr43_vgpr44_vgpr45_vgpr46_vgpr47, $vgpr48_vgpr49_vgpr50_vgpr51_vgpr52_vgpr53_vgpr54_vgpr55_vgpr56_vgpr57_vgpr58_vgpr59_vgpr60_vgpr61_vgpr62_vgpr63, $vgpr64_vgpr65_vgpr66_vgpr67_vgpr68_vgpr69_vgpr70_vgpr71_vgpr72_vgpr73_vgpr74_vgpr75_vgpr76_vgpr77_vgpr78_vgpr79, $vgpr80_vgpr81_vgpr82_vgpr83_vgpr84_vgpr85_vgpr86_vgpr87_vgpr88_vgpr89_vgpr90_vgpr91_vgpr92_vgpr93_vgpr94_vgpr95, $vgpr96_vgpr97_vgpr98_vgpr99_vgpr100_vgpr101_vgpr102_vgpr103_vgpr104_vgpr105_vgpr106_vgpr107_vgpr108_vgpr109_vgpr110_vgpr111, $vgpr112_vgpr113_vgpr114_vgpr115_vgpr116_vgpr117_vgpr118_vgpr119_vgpr120_vgpr121_vgpr122_vgpr123_vgpr124_vgpr125_vgpr126_vgpr127, $vgpr128_vgpr129_vgpr130_vgpr131_vgpr132_vgpr133_vgpr134_vgpr135_vgpr136_vgpr137_vgpr138_vgpr139_vgpr140_vgpr141_vgpr142_vgpr143, $vgpr144_vgpr145_vgpr146_vgpr147_vgpr148_vgpr149_vgpr150_vgpr151_vgpr152_vgpr153_vgpr154_vgpr155_vgpr156_vgpr157_vgpr158_vgpr159, $vgpr160_vgpr161_vgpr162_vgpr163_vgpr164_vgpr165_vgpr166_vgpr167_vgpr168_vgpr169_vgpr170_vgpr171_vgpr172_vgpr173_vgpr174_vgpr175, $vgpr176_vgpr177_vgpr178_vgpr179_vgpr180_vgpr181_vgpr182_vgpr183_vgpr184_vgpr185_vgpr186_vgpr187_vgpr188_vgpr189_vgpr190_vgpr191, $vgpr192_vgpr193_vgpr194_vgpr195_vgpr196_vgpr197_vgpr198_vgpr199_vgpr200_vgpr201_vgpr202_vgpr203_vgpr204_vgpr205_vgpr206_vgpr207, $vgpr208_vgpr209_vgpr210_vgpr211_vgpr212_vgpr213_vgpr214_vgpr215_vgpr216_vgpr217_vgpr218_vgpr219_vgpr220_vgpr221_vgpr222_vgpr223, $vgpr224_vgpr225_vgpr226_vgpr227_vgpr228_vgpr229_vgpr230_vgpr231_vgpr232_vgpr233_vgpr234_vgpr235_vgpr236_vgpr237_vgpr238_vgpr239, $vgpr240_vgpr241_vgpr242_vgpr243_vgpr244_vgpr245_vgpr246_vgpr247, $vgpr248_vgpr249_vgpr250_vgpr251, $vgpr252_vgpr253_vgpr254_vgpr255
 
     ; GFX8-LABEL: name: pei_scavenge_vgpr_spill
     ; GFX8: liveins: $vgpr2, $vgpr248_vgpr249_vgpr250_vgpr251, $vgpr252_vgpr253_vgpr254_vgpr255, $vgpr240_vgpr241_vgpr242_vgpr243_vgpr244_vgpr245_vgpr246_vgpr247, $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15, $vgpr16_vgpr17_vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31, $vgpr32_vgpr33_vgpr34_vgpr35_vgpr36_vgpr37_vgpr38_vgpr39_vgpr40_vgpr41_vgpr42_vgpr43_vgpr44_vgpr45_vgpr46_vgpr47, $vgpr48_vgpr49_vgpr50_vgpr51_vgpr52_vgpr53_vgpr54_vgpr55_vgpr56_vgpr57_vgpr58_vgpr59_vgpr60_vgpr61_vgpr62_vgpr63, $vgpr64_vgpr65_vgpr66_vgpr67_vgpr68_vgpr69_vgpr70_vgpr71_vgpr72_vgpr73_vgpr74_vgpr75_vgpr76_vgpr77_vgpr78_vgpr79, $vgpr80_vgpr81_vgpr82_vgpr83_vgpr84_vgpr85_vgpr86_vgpr87_vgpr88_vgpr89_vgpr90_vgpr91_vgpr92_vgpr93_vgpr94_vgpr95, $vgpr96_vgpr97_vgpr98_vgpr99_vgpr100_vgpr101_vgpr102_vgpr103_vgpr104_vgpr105_vgpr106_vgpr107_vgpr108_vgpr109_vgpr110_vgpr111, $vgpr112_vgpr113_vgpr114_vgpr115_vgpr116_vgpr117_vgpr118_vgpr119_vgpr120_vgpr121_vgpr122_vgpr123_vgpr124_vgpr125_vgpr126_vgpr127, $vgpr128_vgpr129_vgpr130_vgpr131_vgpr132_vgpr133_vgpr134_vgpr135_vgpr136_vgpr137_vgpr138_vgpr139_vgpr140_vgpr141_vgpr142_vgpr143, $vgpr144_vgpr145_vgpr146_vgpr147_vgpr148_vgpr149_vgpr150_vgpr151_vgpr152_vgpr153_vgpr154_vgpr155_vgpr156_vgpr157_vgpr158_vgpr159, $vgpr160_vgpr161_vgpr162_vgpr163_vgpr164_vgpr165_vgpr166_vgpr167_vgpr168_vgpr169_vgpr170_vgpr171_vgpr172_vgpr173_vgpr174_vgpr175, $vgpr176_vgpr177_vgpr178_vgpr179_vgpr180_vgpr181_vgpr182_vgpr183_vgpr184_vgpr185_vgpr186_vgpr187_vgpr188_vgpr189_vgpr190_vgpr191, $vgpr192_vgpr193_vgpr194_vgpr195_vgpr196_vgpr197_vgpr198_vgpr199_vgpr200_vgpr201_vgpr202_vgpr203_vgpr204_vgpr205_vgpr206_vgpr207, $vgpr208_vgpr209_vgpr210_vgpr211_vgpr212_vgpr213_vgpr214_vgpr215_vgpr216_vgpr217_vgpr218_vgpr219_vgpr220_vgpr221_vgpr222_vgpr223, $vgpr224_vgpr225_vgpr226_vgpr227_vgpr228_vgpr229_vgpr230_vgpr231_vgpr232_vgpr233_vgpr234_vgpr235_vgpr236_vgpr237_vgpr238_vgpr239
     ; GFX8-NEXT: {{  $}}
-    ; GFX8-NEXT: $sgpr4_sgpr5 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+    ; GFX8-NEXT: $sgpr4_sgpr5 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
     ; GFX8-NEXT: $sgpr6 = S_ADD_I32 $sgpr32, 1048832, implicit-def dead $scc
     ; GFX8-NEXT: BUFFER_STORE_DWORD_OFFSET $vgpr2, $sgpr0_sgpr1_sgpr2_sgpr3, killed $sgpr6, 0, 0, 0, implicit $exec :: (store (s32) into %stack.3, addrspace 5)
     ; GFX8-NEXT: $exec = S_MOV_B64 killed $sgpr4_sgpr5
     ; GFX8-NEXT: $vgpr2 = V_WRITELANE_B32 $sgpr33, 0, undef $vgpr2
     ; GFX8-NEXT: $sgpr33 = frame-setup S_ADD_I32 $sgpr32, 524224, implicit-def $scc
     ; GFX8-NEXT: $sgpr33 = frame-setup S_AND_B32 killed $sgpr33, 4294443008, implicit-def dead $scc
     ; GFX8-NEXT: $sgpr32 = frame-setup S_ADD_I32 $sgpr32, 2097152, implicit-def dead $scc
     ; GFX8-NEXT: $vgpr0 = V_LSHRREV_B32_e64 6, $sgpr33, implicit $exec
     ; GFX8-NEXT: $vcc_lo = S_MOV_B32 8192
     ; GFX8-NEXT: $vgpr0, dead $vcc = V_ADD_CO_U32_e64 killed $vcc_lo, killed $vgpr0, 0, implicit $exec
     ; GFX8-NEXT: $vgpr3 = V_LSHRREV_B32_e64 6, $sgpr33, implicit $exec
     ; GFX8-NEXT: $vcc_lo = S_MOV_B32 16384
     ; GFX8-NEXT: $vgpr3, dead $vcc = V_ADD_CO_U32_e64 killed $vcc_lo, killed $vgpr3, 0, implicit $exec
     ; GFX8-NEXT: $vgpr0 = V_OR_B32_e32 killed $vgpr3, $vgpr1, implicit $exec
     ; GFX8-NEXT: $sgpr32 = frame-destroy S_ADD_I32 $sgpr32, -2097152, implicit-def dead $scc
     ; GFX8-NEXT: $sgpr33 = V_READLANE_B32 $vgpr2, 0
-    ; GFX8-NEXT: $sgpr4_sgpr5 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+    ; GFX8-NEXT: $sgpr4_sgpr5 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
     ; GFX8-NEXT: $sgpr6 = S_ADD_I32 $sgpr32, 1048832, implicit-def dead $scc
     ; GFX8-NEXT: $vgpr2 = BUFFER_LOAD_DWORD_OFFSET $sgpr0_sgpr1_sgpr2_sgpr3, killed $sgpr6, 0, 0, 0, implicit $exec :: (load (s32) from %stack.3, addrspace 5)
     ; GFX8-NEXT: $exec = S_MOV_B64 killed $sgpr4_sgpr5
     ; GFX8-NEXT: S_ENDPGM 0, amdgpu_allvgprs
     ; GFX9-LABEL: name: pei_scavenge_vgpr_spill
     ; GFX9: liveins: $vgpr2, $vgpr248_vgpr249_vgpr250_vgpr251, $vgpr252_vgpr253_vgpr254_vgpr255, $vgpr240_vgpr241_vgpr242_vgpr243_vgpr244_vgpr245_vgpr246_vgpr247, $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15, $vgpr16_vgpr17_vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31, $vgpr32_vgpr33_vgpr34_vgpr35_vgpr36_vgpr37_vgpr38_vgpr39_vgpr40_vgpr41_vgpr42_vgpr43_vgpr44_vgpr45_vgpr46_vgpr47, $vgpr48_vgpr49_vgpr50_vgpr51_vgpr52_vgpr53_vgpr54_vgpr55_vgpr56_vgpr57_vgpr58_vgpr59_vgpr60_vgpr61_vgpr62_vgpr63, $vgpr64_vgpr65_vgpr66_vgpr67_vgpr68_vgpr69_vgpr70_vgpr71_vgpr72_vgpr73_vgpr74_vgpr75_vgpr76_vgpr77_vgpr78_vgpr79, $vgpr80_vgpr81_vgpr82_vgpr83_vgpr84_vgpr85_vgpr86_vgpr87_vgpr88_vgpr89_vgpr90_vgpr91_vgpr92_vgpr93_vgpr94_vgpr95, $vgpr96_vgpr97_vgpr98_vgpr99_vgpr100_vgpr101_vgpr102_vgpr103_vgpr104_vgpr105_vgpr106_vgpr107_vgpr108_vgpr109_vgpr110_vgpr111, $vgpr112_vgpr113_vgpr114_vgpr115_vgpr116_vgpr117_vgpr118_vgpr119_vgpr120_vgpr121_vgpr122_vgpr123_vgpr124_vgpr125_vgpr126_vgpr127, $vgpr128_vgpr129_vgpr130_vgpr131_vgpr132_vgpr133_vgpr134_vgpr135_vgpr136_vgpr137_vgpr138_vgpr139_vgpr140_vgpr141_vgpr142_vgpr143, $vgpr144_vgpr145_vgpr146_vgpr147_vgpr148_vgpr149_vgpr150_vgpr151_vgpr152_vgpr153_vgpr154_vgpr155_vgpr156_vgpr157_vgpr158_vgpr159, $vgpr160_vgpr161_vgpr162_vgpr163_vgpr164_vgpr165_vgpr166_vgpr167_vgpr168_vgpr169_vgpr170_vgpr171_vgpr172_vgpr173_vgpr174_vgpr175, $vgpr176_vgpr177_vgpr178_vgpr179_vgpr180_vgpr181_vgpr182_vgpr183_vgpr184_vgpr185_vgpr186_vgpr187_vgpr188_vgpr189_vgpr190_vgpr191, $vgpr192_vgpr193_vgpr194_vgpr195_vgpr196_vgpr197_vgpr198_vgpr199_vgpr200_vgpr201_vgpr202_vgpr203_vgpr204_vgpr205_vgpr206_vgpr207, $vgpr208_vgpr209_vgpr210_vgpr211_vgpr212_vgpr213_vgpr214_vgpr215_vgpr216_vgpr217_vgpr218_vgpr219_vgpr220_vgpr221_vgpr222_vgpr223, $vgpr224_vgpr225_vgpr226_vgpr227_vgpr228_vgpr229_vgpr230_vgpr231_vgpr232_vgpr233_vgpr234_vgpr235_vgpr236_vgpr237_vgpr238_vgpr239
     ; GFX9-NEXT: {{  $}}
-    ; GFX9-NEXT: $sgpr4_sgpr5 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+    ; GFX9-NEXT: $sgpr4_sgpr5 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
     ; GFX9-NEXT: $sgpr6 = S_ADD_I32 $sgpr32, 1048832, implicit-def dead $scc
     ; GFX9-NEXT: BUFFER_STORE_DWORD_OFFSET $vgpr2, $sgpr0_sgpr1_sgpr2_sgpr3, killed $sgpr6, 0, 0, 0, implicit $exec :: (store (s32) into %stack.3, addrspace 5)
     ; GFX9-NEXT: $exec = S_MOV_B64 killed $sgpr4_sgpr5
     ; GFX9-NEXT: $vgpr2 = V_WRITELANE_B32 $sgpr33, 0, undef $vgpr2
     ; GFX9-NEXT: $sgpr33 = frame-setup S_ADD_I32 $sgpr32, 524224, implicit-def $scc
     ; GFX9-NEXT: $sgpr33 = frame-setup S_AND_B32 killed $sgpr33, 4294443008, implicit-def dead $scc
     ; GFX9-NEXT: $sgpr32 = frame-setup S_ADD_I32 $sgpr32, 2097152, implicit-def dead $scc
     ; GFX9-NEXT: $vgpr0 = V_LSHRREV_B32_e64 6, $sgpr33, implicit $exec
     ; GFX9-NEXT: $vgpr0 = V_ADD_U32_e32 8192, killed $vgpr0, implicit $exec
     ; GFX9-NEXT: $vgpr3 = V_LSHRREV_B32_e64 6, $sgpr33, implicit $exec
     ; GFX9-NEXT: $vgpr3 = V_ADD_U32_e32 16384, killed $vgpr3, implicit $exec
     ; GFX9-NEXT: $vgpr0 = V_OR_B32_e32 killed $vgpr3, $vgpr1, implicit $exec
     ; GFX9-NEXT: $sgpr32 = frame-destroy S_ADD_I32 $sgpr32, -2097152, implicit-def dead $scc
     ; GFX9-NEXT: $sgpr33 = V_READLANE_B32 $vgpr2, 0
-    ; GFX9-NEXT: $sgpr4_sgpr5 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+    ; GFX9-NEXT: $sgpr4_sgpr5 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
     ; GFX9-NEXT: $sgpr6 = S_ADD_I32 $sgpr32, 1048832, implicit-def dead $scc
     ; GFX9-NEXT: $vgpr2 = BUFFER_LOAD_DWORD_OFFSET $sgpr0_sgpr1_sgpr2_sgpr3, killed $sgpr6, 0, 0, 0, implicit $exec :: (load (s32) from %stack.3, addrspace 5)
     ; GFX9-NEXT: $exec = S_MOV_B64 killed $sgpr4_sgpr5
     ; GFX9-NEXT: S_ENDPGM 0, amdgpu_allvgprs
     ; GFX9-FLATSCR-LABEL: name: pei_scavenge_vgpr_spill
     ; GFX9-FLATSCR: liveins: $vgpr2, $vgpr248_vgpr249_vgpr250_vgpr251, $vgpr252_vgpr253_vgpr254_vgpr255, $vgpr240_vgpr241_vgpr242_vgpr243_vgpr244_vgpr245_vgpr246_vgpr247, $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15, $vgpr16_vgpr17_vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31, $vgpr32_vgpr33_vgpr34_vgpr35_vgpr36_vgpr37_vgpr38_vgpr39_vgpr40_vgpr41_vgpr42_vgpr43_vgpr44_vgpr45_vgpr46_vgpr47, $vgpr48_vgpr49_vgpr50_vgpr51_vgpr52_vgpr53_vgpr54_vgpr55_vgpr56_vgpr57_vgpr58_vgpr59_vgpr60_vgpr61_vgpr62_vgpr63, $vgpr64_vgpr65_vgpr66_vgpr67_vgpr68_vgpr69_vgpr70_vgpr71_vgpr72_vgpr73_vgpr74_vgpr75_vgpr76_vgpr77_vgpr78_vgpr79, $vgpr80_vgpr81_vgpr82_vgpr83_vgpr84_vgpr85_vgpr86_vgpr87_vgpr88_vgpr89_vgpr90_vgpr91_vgpr92_vgpr93_vgpr94_vgpr95, $vgpr96_vgpr97_vgpr98_vgpr99_vgpr100_vgpr101_vgpr102_vgpr103_vgpr104_vgpr105_vgpr106_vgpr107_vgpr108_vgpr109_vgpr110_vgpr111, $vgpr112_vgpr113_vgpr114_vgpr115_vgpr116_vgpr117_vgpr118_vgpr119_vgpr120_vgpr121_vgpr122_vgpr123_vgpr124_vgpr125_vgpr126_vgpr127, $vgpr128_vgpr129_vgpr130_vgpr131_vgpr132_vgpr133_vgpr134_vgpr135_vgpr136_vgpr137_vgpr138_vgpr139_vgpr140_vgpr141_vgpr142_vgpr143, $vgpr144_vgpr145_vgpr146_vgpr147_vgpr148_vgpr149_vgpr150_vgpr151_vgpr152_vgpr153_vgpr154_vgpr155_vgpr156_vgpr157_vgpr158_vgpr159, $vgpr160_vgpr161_vgpr162_vgpr163_vgpr164_vgpr165_vgpr166_vgpr167_vgpr168_vgpr169_vgpr170_vgpr171_vgpr172_vgpr173_vgpr174_vgpr175, $vgpr176_vgpr177_vgpr178_vgpr179_vgpr180_vgpr181_vgpr182_vgpr183_vgpr184_vgpr185_vgpr186_vgpr187_vgpr188_vgpr189_vgpr190_vgpr191, $vgpr192_vgpr193_vgpr194_vgpr195_vgpr196_vgpr197_vgpr198_vgpr199_vgpr200_vgpr201_vgpr202_vgpr203_vgpr204_vgpr205_vgpr206_vgpr207, $vgpr208_vgpr209_vgpr210_vgpr211_vgpr212_vgpr213_vgpr214_vgpr215_vgpr216_vgpr217_vgpr218_vgpr219_vgpr220_vgpr221_vgpr222_vgpr223, $vgpr224_vgpr225_vgpr226_vgpr227_vgpr228_vgpr229_vgpr230_vgpr231_vgpr232_vgpr233_vgpr234_vgpr235_vgpr236_vgpr237_vgpr238_vgpr239
     ; GFX9-FLATSCR-NEXT: {{  $}}
-    ; GFX9-FLATSCR-NEXT: $sgpr4_sgpr5 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+    ; GFX9-FLATSCR-NEXT: $sgpr4_sgpr5 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
     ; GFX9-FLATSCR-NEXT: $sgpr6 = S_ADD_I32 $sgpr32, 16388, implicit-def dead $scc
     ; GFX9-FLATSCR-NEXT: SCRATCH_STORE_DWORD_SADDR $vgpr2, killed $sgpr6, 0, 0, implicit $exec, implicit $flat_scr :: (store (s32) into %stack.3, addrspace 5)
     ; GFX9-FLATSCR-NEXT: $exec = S_MOV_B64 killed $sgpr4_sgpr5
     ; GFX9-FLATSCR-NEXT: $vgpr2 = V_WRITELANE_B32 $sgpr33, 0, undef $vgpr2
     ; GFX9-FLATSCR-NEXT: $sgpr33 = frame-setup S_ADD_I32 $sgpr32, 8191, implicit-def $scc
     ; GFX9-FLATSCR-NEXT: $sgpr33 = frame-setup S_AND_B32 killed $sgpr33, 4294959104, implicit-def dead $scc
     ; GFX9-FLATSCR-NEXT: $sgpr32 = frame-setup S_ADD_I32 $sgpr32, 32768, implicit-def dead $scc
     ; GFX9-FLATSCR-NEXT: $vcc_lo = S_ADD_I32 $sgpr33, 8192, implicit-def $scc
     ; GFX9-FLATSCR-NEXT: $vgpr0 = V_MOV_B32_e32 killed $vcc_lo, implicit $exec
     ; GFX9-FLATSCR-NEXT: $vcc_hi = S_ADD_I32 $sgpr33, 16384, implicit-def $scc
     ; GFX9-FLATSCR-NEXT: $vgpr0 = V_OR_B32_e32 killed $vcc_hi, $vgpr1, implicit $exec
     ; GFX9-FLATSCR-NEXT: $sgpr32 = frame-destroy S_ADD_I32 $sgpr32, -32768, implicit-def dead $scc
     ; GFX9-FLATSCR-NEXT: $sgpr33 = V_READLANE_B32 $vgpr2, 0
-    ; GFX9-FLATSCR-NEXT: $sgpr4_sgpr5 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+    ; GFX9-FLATSCR-NEXT: $sgpr4_sgpr5 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
     ; GFX9-FLATSCR-NEXT: $sgpr6 = S_ADD_I32 $sgpr32, 16388, implicit-def dead $scc
     ; GFX9-FLATSCR-NEXT: $vgpr2 = SCRATCH_LOAD_DWORD_SADDR killed $sgpr6, 0, 0, implicit $exec, implicit $flat_scr :: (load (s32) from %stack.3, addrspace 5)
     ; GFX9-FLATSCR-NEXT: $exec = S_MOV_B64 killed $sgpr4_sgpr5
     ; GFX9-FLATSCR-NEXT: S_ENDPGM 0, amdgpu_allvgprs
     $vgpr0 = V_MOV_B32_e32 %stack.0, implicit $exec
     $vgpr0 = V_OR_B32_e32 %stack.1, $vgpr1, implicit $exec
     S_ENDPGM 0, amdgpu_allvgprs
 ...

llvm/test/CodeGen/AMDGPU/preserve-only-inactive-lane.mir

+# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
+# RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx906 -verify-machineinstrs -run-pass=prologepilog -o - %s | FileCheck -check-prefix=GCN %s
+
+# The scratch register vgpr0 is used here to spill sgpr35 register.
+# The function also returns a value that goes via. vgpr0.
+# This test is to ensure the vgpr0 store/restore inserted at the prolog/epilog
+# preserves only its inactive lanes to ensure the outgoing value will not be overwritten.
+
+---
+name: preserve_scratch_vgpr_inactive_lanes
+tracksRegLiveness: true
+machineFunctionInfo:
+  scratchRSrcReg:  $sgpr0_sgpr1_sgpr2_sgpr3
+  frameOffsetReg:  $sgpr33
+  stackPtrOffsetReg:  $sgpr32
+
+body:             |
+  bb.0:
+    liveins: $sgpr35, $vgpr0, $sgpr30_sgpr31
+    ; GCN-LABEL: name: preserve_scratch_vgpr_inactive_lanes
+    ; GCN: liveins: $sgpr35, $vgpr0, $sgpr30_sgpr31
+    ; GCN-NEXT: {{  $}}
+    ; GCN-NEXT: $sgpr4_sgpr5 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+    ; GCN-NEXT: BUFFER_STORE_DWORD_OFFSET $vgpr0, $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr32, 0, 0, 0, implicit $exec :: (store (s32) into %stack.0, addrspace 5)
+    ; GCN-NEXT: $exec = S_MOV_B64 killed $sgpr4_sgpr5
+    ; GCN-NEXT: renamable $vgpr0 = V_WRITELANE_B32 $sgpr35, 0, killed $vgpr0
+    ; GCN-NEXT: $sgpr35 = S_MOV_B32 5
+    ; GCN-NEXT: $sgpr35 = V_READLANE_B32 $vgpr0, 0
+    ; GCN-NEXT: renamable $vgpr0 = V_MOV_B32_e32 10, implicit $exec
+    ; GCN-NEXT: $sgpr4_sgpr5 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+    ; GCN-NEXT: $vgpr0 = BUFFER_LOAD_DWORD_OFFSET $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr32, 0, 0, 0, implicit $exec :: (load (s32) from %stack.0, addrspace 5)
+    ; GCN-NEXT: $exec = S_MOV_B64 killed $sgpr4_sgpr5
+    ; GCN-NEXT: S_SETPC_B64_return killed renamable $sgpr30_sgpr31, implicit $vgpr0
+    renamable $vgpr0 = V_WRITELANE_B32 $sgpr35, 0, killed $vgpr0
+    $sgpr35 = S_MOV_B32 5
+    $sgpr35 = V_READLANE_B32 $vgpr0, 0
+    renamable $vgpr0 = V_MOV_B32_e32 10, implicit $exec
+    S_SETPC_B64_return killed renamable $sgpr30_sgpr31, implicit $vgpr0
+...

llvm/test/CodeGen/AMDGPU/sgpr-spills-split-regalloc.ll

; GCN-NEXT:    s_setpc_b64 s[4:5]
   musttail call void @child_function()
   ret void
 }
 
 define void @spill_sgpr_no_free_vgpr(<4 x i32> addrspace(1)* %out, <4 x i32> addrspace(1)* %in) #0 {
 ; GCN-LABEL: spill_sgpr_no_free_vgpr:
 ; 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:    s_xor_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_store_dword v4, off, s[0:3], s32 offset:464 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s32 offset:444 ; 4-byte Folded Spill
 ; GCN-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:440 ; 4-byte Folded Spill
 ; GCN-NEXT:    buffer_store_dword v42, off, s[0:3], s32 offset:436 ; 4-byte Folded Spill
 ; GCN-NEXT:    buffer_store_dword v43, off, s[0:3], s32 offset:432 ; 4-byte Folded Spill
 ; GCN-NEXT:    buffer_store_dword v44, off, s[0:3], s32 offset:428 ; 4-byte Folded Spill
 ; GCN-NEXT:    buffer_store_dword v45, off, s[0:3], s32 offset:424 ; 4-byte Folded Spill

 ; GCN-NEXT:    buffer_load_dword v47, off, s[0:3], s32 offset:416 ; 4-byte Folded Reload
 ; GCN-NEXT:    buffer_load_dword v46, off, s[0:3], s32 offset:420 ; 4-byte Folded Reload
 ; GCN-NEXT:    buffer_load_dword v45, off, s[0:3], s32 offset:424 ; 4-byte Folded Reload
 ; GCN-NEXT:    buffer_load_dword v44, off, s[0:3], s32 offset:428 ; 4-byte Folded Reload
 ; GCN-NEXT:    buffer_load_dword v43, off, s[0:3], s32 offset:432 ; 4-byte Folded Reload
 ; GCN-NEXT:    buffer_load_dword v42, off, s[0:3], s32 offset:436 ; 4-byte Folded Reload
 ; GCN-NEXT:    buffer_load_dword v41, off, s[0:3], s32 offset:440 ; 4-byte Folded Reload
 ; GCN-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:444 ; 4-byte Folded Reload
-; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GCN-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_load_dword v4, off, s[0:3], s32 offset:464 ; 4-byte Folded Reload
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
 ; GCN-NEXT:    s_setpc_b64 s[30:31]
   %a = load <4 x i32>, <4 x i32> addrspace(1)* %in
   call void asm sideeffect "",
   "~{v6},~{v7},~{v8},~{v9}
   ,~{v10},~{v11},~{v12},~{v13},~{v14},~{v15},~{v16},~{v17},~{v18},~{v19}

llvm/test/CodeGen/AMDGPU/spill-reg-tuple-super-reg-use.mir

 
 body:             |
   bb.0:
     liveins: $sgpr0, $sgpr1, $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $sgpr7, $vgpr0, $vgpr1, $vgpr2, $vgpr3
 
     ; GCN-LABEL: name: spill_sgpr128_use_subreg
     ; GCN: liveins: $sgpr0, $sgpr1, $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $sgpr7, $vgpr0, $vgpr1, $vgpr2, $vgpr3
     ; GCN-NEXT: {{  $}}
-    ; GCN-NEXT: $sgpr8_sgpr9 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+    ; GCN-NEXT: $sgpr8_sgpr9 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
     ; GCN-NEXT: BUFFER_STORE_DWORD_OFFSET $vgpr0, $sgpr100_sgpr101_sgpr102_sgpr103, $sgpr32, 0, 0, 0, implicit $exec :: (store (s32) into %stack.1, addrspace 5)
     ; GCN-NEXT: $exec = S_MOV_B64 killed $sgpr8_sgpr9
     ; GCN-NEXT: renamable $sgpr1 = COPY $sgpr2
     ; GCN-NEXT: $vgpr0 = V_WRITELANE_B32 $sgpr0, 0, $vgpr0, implicit-def $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr0_sgpr1_sgpr2_sgpr3
     ; GCN-NEXT: $vgpr0 = V_WRITELANE_B32 $sgpr1, 1, $vgpr0, implicit $sgpr0_sgpr1_sgpr2_sgpr3
     ; GCN-NEXT: $vgpr0 = V_WRITELANE_B32 $sgpr2, 2, $vgpr0, implicit $sgpr0_sgpr1_sgpr2_sgpr3
     ; GCN-NEXT: $vgpr0 = V_WRITELANE_B32 $sgpr3, 3, $vgpr0, implicit $sgpr0_sgpr1_sgpr2_sgpr3
     ; GCN-NEXT: renamable $sgpr8 = COPY killed renamable $sgpr1
-    ; GCN-NEXT: $sgpr0_sgpr1 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+    ; GCN-NEXT: $sgpr0_sgpr1 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
     ; GCN-NEXT: $vgpr0 = BUFFER_LOAD_DWORD_OFFSET $sgpr100_sgpr101_sgpr102_sgpr103, $sgpr32, 0, 0, 0, implicit $exec :: (load (s32) from %stack.1, addrspace 5)
     ; GCN-NEXT: $exec = S_MOV_B64 killed $sgpr0_sgpr1
     ; GCN-NEXT: S_ENDPGM 0, implicit $sgpr8
     renamable $sgpr1 = COPY $sgpr2
     SI_SPILL_S128_SAVE renamable $sgpr0_sgpr1_sgpr2_sgpr3, %stack.0, implicit $exec, implicit $sgpr32 :: (store (s128) into %stack.0, align 4, addrspace 5)
     renamable $sgpr8 = COPY killed renamable $sgpr1
     S_ENDPGM 0, implicit $sgpr8
 ...

 
 body:             |
   bb.0:
     liveins: $sgpr0, $sgpr1, $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $sgpr7, $vgpr0, $vgpr1, $vgpr2, $vgpr3
 
     ; GCN-LABEL: name: spill_sgpr128_use_kill
     ; GCN: liveins: $sgpr0, $sgpr1, $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $sgpr7, $vgpr0, $vgpr1, $vgpr2, $vgpr3
     ; GCN-NEXT: {{  $}}
-    ; GCN-NEXT: $sgpr8_sgpr9 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+    ; GCN-NEXT: $sgpr8_sgpr9 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
     ; GCN-NEXT: BUFFER_STORE_DWORD_OFFSET $vgpr0, $sgpr100_sgpr101_sgpr102_sgpr103, $sgpr32, 0, 0, 0, implicit $exec :: (store (s32) into %stack.1, addrspace 5)
     ; GCN-NEXT: $exec = S_MOV_B64 killed $sgpr8_sgpr9
     ; GCN-NEXT: renamable $sgpr1 = COPY $sgpr2
     ; GCN-NEXT: $vgpr0 = V_WRITELANE_B32 $sgpr0, 0, $vgpr0, implicit-def $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr0_sgpr1_sgpr2_sgpr3
     ; GCN-NEXT: $vgpr0 = V_WRITELANE_B32 $sgpr1, 1, $vgpr0, implicit $sgpr0_sgpr1_sgpr2_sgpr3
     ; GCN-NEXT: $vgpr0 = V_WRITELANE_B32 $sgpr2, 2, $vgpr0, implicit $sgpr0_sgpr1_sgpr2_sgpr3
     ; GCN-NEXT: $vgpr0 = V_WRITELANE_B32 killed $sgpr3, 3, $vgpr0, implicit killed $sgpr0_sgpr1_sgpr2_sgpr3
-    ; GCN-NEXT: $sgpr0_sgpr1 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+    ; GCN-NEXT: $sgpr0_sgpr1 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
     ; GCN-NEXT: $vgpr0 = BUFFER_LOAD_DWORD_OFFSET $sgpr100_sgpr101_sgpr102_sgpr103, $sgpr32, 0, 0, 0, implicit $exec :: (load (s32) from %stack.1, addrspace 5)
     ; GCN-NEXT: $exec = S_MOV_B64 killed $sgpr0_sgpr1
     ; GCN-NEXT: S_ENDPGM 0
     renamable $sgpr1 = COPY $sgpr2
     SI_SPILL_S128_SAVE renamable killed $sgpr0_sgpr1_sgpr2_sgpr3, %stack.0, implicit $exec, implicit $sgpr32 :: (store (s128) into %stack.0, align 4, addrspace 5)
     S_ENDPGM 0
 ...
 

llvm/test/CodeGen/AMDGPU/spill-writelane-vgprs.ll

 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx908 -verify-machineinstrs -o - %s | FileCheck -check-prefix=GCN %s
 ; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx90a -verify-machineinstrs -o - %s | FileCheck -check-prefix=GCN %s
 
 ; Callee must preserve the VGPR modified by writelane even if it is marked Caller-saved.
 
 declare i32 @llvm.amdgcn.writelane(i32, i32, i32)
 
 define void @sgpr_spill_writelane() {
 ; GCN-LABEL: sgpr_spill_writelane:
 ; 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:    s_xor_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:    v_writelane_b32 v0, s35, 0
 ; GCN-NEXT:    ;;#ASMSTART
 ; GCN-NEXT:    ;;#ASMEND
 ; GCN-NEXT:    v_readlane_b32 s35, v0, 0
-; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GCN-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_load_dword v0, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
 ; GCN-NEXT:    s_setpc_b64 s[30:31]
   call void asm sideeffect "", "~{s35}"()
   ret void
 }
 
 ; FIXME: The writelane intrinsic doesn't really overwrite any inactive lanes
 ; and hence there is no need to preserve the VGPR it modifies.
 define void @device_writelane_intrinsic(i32 addrspace(1)* %out, i32 %src) {
 ; GCN-LABEL: device_writelane_intrinsic:
 ; 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:    s_xor_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_store_dword v3, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    v_mov_b32_e32 v3, 15
 ; GCN-NEXT:    v_readfirstlane_b32 s4, v2
 ; GCN-NEXT:    v_writelane_b32 v3, s4, 23
 ; GCN-NEXT:    global_store_dword v[0:1], v3, off
-; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
+; GCN-NEXT:    s_xor_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_load_dword v3, off, s[0:3], s32 ; 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]
   %writelane = call i32 @llvm.amdgcn.writelane(i32 %src, i32 23, i32 15)
   store i32 %writelane, i32 addrspace(1)* %out, align 4
   ret void
 }

llvm/test/CodeGen/AMDGPU/stack-realign.ll

call void asm sideeffect "",
     ,~{vcc_hi}"() #0
   ret void
 }
 
 define void @no_free_regs_spill_bp_to_memory(<32 x i32> %a, i32 %b) #5 {
 ; If there are no free SGPRs or VGPRs available we must spill the BP to memory.
 
 ; GCN-LABEL: no_free_regs_spill_bp_to_mem
-; GCN: s_or_saveexec_b64 s[4:5], -1
+; GCN: s_xor_saveexec_b64 s[4:5], -1
+; GCN: buffer_store_dword v39, off, s[0:3], s32
 ; GCN: v_mov_b32_e32 v0, s34
 ; GCN: buffer_store_dword v0, off, s[0:3], s32
 ; GCN: v_mov_b32_e32 v0, s33
 ; GCN-DAG: buffer_store_dword v0, off, s[0:3], s32
   %local_val = alloca i32, align 128, addrspace(5)
   store volatile i32 %b, ptr addrspace(5) %local_val, align 128
 
   call void asm sideeffect "; clobber nonpreserved SGPRs and 64 CSRs",
; GCN-DAG: buffer_store_dword v0, off, s[0:3], s32
   ret void
 }
 
 define void @spill_bp_to_memory_scratch_reg_needed_mubuf_offset(<32 x i32> %a, i32 %b, ptr addrspace(5) byval([4096 x i8]) align 4 %arg) #5 {
 ; If the size of the offset exceeds the MUBUF offset field we need another
 ; scratch VGPR to hold the offset.
 
 ; GCN-LABEL: spill_bp_to_memory_scratch_reg_needed_mubuf_offset
-; GCN: s_or_saveexec_b64 s[4:5], -1
+; GCN: s_xor_saveexec_b64 s[4:5], -1
 ; GCN-NEXT: s_add_i32 s6, s32, 0x42100
 ; GCN-NEXT: buffer_store_dword v39, off, s[0:3], s6 ; 4-byte Folded Spill
 ; GCN-NEXT: s_mov_b64 exec, s[4:5]
 ; GCN-NEXT: v_mov_b32_e32 v0, s34
 ; GCN-NOT: v_mov_b32_e32 v0, 0x1088
 ; GCN-NEXT: s_add_i32 s6, s32, 0x42300
 ; GCN-NEXT: buffer_store_dword v0, off, s[0:3], s6 ; 4-byte Folded Spill
 ; GCN-NEXT: v_mov_b32_e32 v0, s33

llvm/test/CodeGen/AMDGPU/tail-call-amdgpu-gfx.ll

; GCN-NEXT:    s_setpc_b64 s[30:31]
   %add = fadd float %v.arg0, %s.arg1
   ret float %add
 }
 
 define amdgpu_gfx float @caller(float %arg0) {
 ; GCN-LABEL: caller:
 ; GCN:       ; %bb.0:
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GCN-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GCN-NEXT:    s_xor_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:    s_mov_b32 s36, s33
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
 ; GCN-NEXT:    v_writelane_b32 v1, s30, 1
 ; GCN-NEXT:    v_add_f32_e32 v0, 1.0, v0
 ; GCN-NEXT:    s_mov_b32 s4, 2.0
 ; GCN-NEXT:    v_writelane_b32 v1, s31, 2
 ; GCN-NEXT:    s_getpc_b64 s[34:35]
 ; GCN-NEXT:    s_add_u32 s34, s34, callee@rel32@lo+4
 ; GCN-NEXT:    s_addc_u32 s35, s35, callee@rel32@hi+12
 ; GCN-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GCN-NEXT:    v_readlane_b32 s31, v1, 2
 ; GCN-NEXT:    v_readlane_b32 s30, v1, 1
 ; GCN-NEXT:    v_readlane_b32 s4, v1, 0
 ; GCN-NEXT:    s_addk_i32 s32, 0xfc00
 ; GCN-NEXT:    s_mov_b32 s33, s36
-; GCN-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GCN-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GCN-NEXT:    buffer_load_dword v1, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GCN-NEXT:    s_mov_b64 exec, s[34:35]
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
 ; GCN-NEXT:    s_setpc_b64 s[30:31]
   %add = fadd float %arg0, 1.0
   %call = tail call amdgpu_gfx float @callee(float %add, float inreg 2.0)
   ret float %call
 }

llvm/test/CodeGen/AMDGPU/track-spilled-vgpr-liveness.mir

machineFunctionInfo:
   stackPtrOffsetReg: '$sgpr32'
 body:             |
   bb.0:
     liveins: $sgpr42, $vgpr0, $sgpr30_sgpr31
 
     ; GCN-LABEL: name: vgpr_use_after_prolog_spill
     ; GCN: liveins: $sgpr42, $vgpr0, $sgpr30_sgpr31
     ; GCN-NEXT: {{  $}}
-    ; GCN-NEXT: $sgpr4_sgpr5 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+    ; GCN-NEXT: $sgpr4_sgpr5 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
     ; GCN-NEXT: BUFFER_STORE_DWORD_OFFSET $vgpr0, $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr32, 0, 0, 0, implicit $exec :: (store (s32) into %stack.0, addrspace 5)
     ; GCN-NEXT: $exec = S_MOV_B64 killed $sgpr4_sgpr5
     ; GCN-NEXT: $vgpr0 = V_ADD_U32_e32 8192, killed $vgpr0, implicit $exec
     ; GCN-NEXT: $vgpr0 = V_WRITELANE_B32 killed $sgpr42, 0, $vgpr0
     ; GCN-NEXT: S_NOP 0, implicit-def $vgpr0
-    ; GCN-NEXT: $sgpr4_sgpr5 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+    ; GCN-NEXT: $sgpr4_sgpr5 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
     ; GCN-NEXT: $vgpr0 = BUFFER_LOAD_DWORD_OFFSET $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr32, 0, 0, 0, implicit $exec :: (load (s32) from %stack.0, addrspace 5)
     ; GCN-NEXT: $exec = S_MOV_B64 killed $sgpr4_sgpr5
     ; GCN-NEXT: S_SETPC_B64_return $sgpr30_sgpr31
     $vgpr0 = V_ADD_U32_e32 8192, killed $vgpr0, implicit $exec
     $vgpr0 = V_WRITELANE_B32 killed $sgpr42, 0, $vgpr0
     S_NOP 0, implicit-def $vgpr0
     S_SETPC_B64_return $sgpr30_sgpr31
 ...
 
 ---
 name:            livein_vgpr_def_after_prolog_spill
 tracksRegLiveness: true
 machineFunctionInfo:
   scratchRSrcReg:  '$sgpr0_sgpr1_sgpr2_sgpr3'
   frameOffsetReg:  '$sgpr33'
   stackPtrOffsetReg: '$sgpr32'
 body:             |
   bb.0:
     liveins: $sgpr42, $vgpr0, $vgpr1, $sgpr30_sgpr31
 
     ; GCN-LABEL: name: livein_vgpr_def_after_prolog_spill
     ; GCN: liveins: $sgpr42, $vgpr0, $vgpr1, $sgpr30_sgpr31
     ; GCN-NEXT: {{  $}}
-    ; GCN-NEXT: $sgpr4_sgpr5 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+    ; GCN-NEXT: $sgpr4_sgpr5 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
     ; GCN-NEXT: BUFFER_STORE_DWORD_OFFSET $vgpr0, $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr32, 0, 0, 0, implicit $exec :: (store (s32) into %stack.0, addrspace 5)
     ; GCN-NEXT: $exec = S_MOV_B64 killed $sgpr4_sgpr5
     ; GCN-NEXT: $vgpr0 = V_ADD_U32_e32 8192, killed $vgpr1, implicit $exec
     ; GCN-NEXT: $vgpr0 = V_WRITELANE_B32 killed $sgpr42, 0, $vgpr0
     ; GCN-NEXT: S_NOP 0, implicit-def $vgpr0
-    ; GCN-NEXT: $sgpr4_sgpr5 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+    ; GCN-NEXT: $sgpr4_sgpr5 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
     ; GCN-NEXT: $vgpr0 = BUFFER_LOAD_DWORD_OFFSET $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr32, 0, 0, 0, implicit $exec :: (load (s32) from %stack.0, addrspace 5)
     ; GCN-NEXT: $exec = S_MOV_B64 killed $sgpr4_sgpr5
     ; GCN-NEXT: S_SETPC_B64_return $sgpr30_sgpr31
     $vgpr0 = V_ADD_U32_e32 8192, killed $vgpr1, implicit $exec
     $vgpr0 = V_WRITELANE_B32 killed $sgpr42, 0, $vgpr0
     S_NOP 0, implicit-def $vgpr0
     S_SETPC_B64_return $sgpr30_sgpr31
 ...
 
 ---
 name:            livein_vgpr_not_referred_in_entry_bb
 tracksRegLiveness: true
 machineFunctionInfo:
   scratchRSrcReg:  '$sgpr0_sgpr1_sgpr2_sgpr3'
   frameOffsetReg:  '$sgpr33'
   stackPtrOffsetReg: '$sgpr32'
 body:             |
   ; GCN-LABEL: name: livein_vgpr_not_referred_in_entry_bb
   ; GCN: bb.0:
   ; GCN-NEXT:   successors: %bb.1(0x80000000)
   ; GCN-NEXT:   liveins: $sgpr42, $vgpr0, $sgpr30_sgpr31
   ; GCN-NEXT: {{  $}}
-  ; GCN-NEXT:   $sgpr4_sgpr5 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+  ; GCN-NEXT:   $sgpr4_sgpr5 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
   ; GCN-NEXT:   BUFFER_STORE_DWORD_OFFSET $vgpr0, $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr32, 0, 0, 0, implicit $exec :: (store (s32) into %stack.0, addrspace 5)
   ; GCN-NEXT:   $exec = S_MOV_B64 killed $sgpr4_sgpr5
   ; GCN-NEXT:   S_NOP 0
   ; GCN-NEXT:   S_BRANCH %bb.1
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT: bb.1:
   ; GCN-NEXT:   liveins: $sgpr42, $vgpr0, $sgpr30_sgpr31
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT:   $vgpr0 = V_WRITELANE_B32 killed $sgpr42, 0, $vgpr0
   ; GCN-NEXT:   S_NOP 0, implicit-def $vgpr0
-  ; GCN-NEXT:   $sgpr4_sgpr5 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+  ; GCN-NEXT:   $sgpr4_sgpr5 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
   ; GCN-NEXT:   $vgpr0 = BUFFER_LOAD_DWORD_OFFSET $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr32, 0, 0, 0, implicit $exec :: (load (s32) from %stack.0, addrspace 5)
   ; GCN-NEXT:   $exec = S_MOV_B64 killed $sgpr4_sgpr5
   ; GCN-NEXT:   S_SETPC_B64_return $sgpr30_sgpr31
   bb.0:
     liveins: $sgpr42, $vgpr0, $sgpr30_sgpr31
     S_NOP 0
     S_BRANCH %bb.1
   bb.1:
     liveins: $sgpr42, $vgpr0, $sgpr30_sgpr31
     $vgpr0 = V_WRITELANE_B32 killed $sgpr42, 0, $vgpr0
     S_NOP 0, implicit-def $vgpr0
     S_SETPC_B64_return $sgpr30_sgpr31
 ...

llvm/test/CodeGen/AMDGPU/use_restore_frame_reg.mir

machineFunctionInfo:
   stackPtrOffsetReg:  $sgpr32
 
 body:             |
   ; MUBUF-LABEL: name: use_restore_frame_reg
   ; MUBUF: bb.0:
   ; MUBUF-NEXT:   successors: %bb.2(0x40000000), %bb.1(0x40000000)
   ; MUBUF-NEXT:   liveins: $vgpr1, $vgpr2
   ; MUBUF-NEXT: {{  $}}
-  ; MUBUF-NEXT:   $sgpr4_sgpr5 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+  ; MUBUF-NEXT:   $sgpr4_sgpr5 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
   ; MUBUF-NEXT:   $sgpr6 = S_ADD_I32 $sgpr32, 9961728, implicit-def dead $scc
   ; MUBUF-NEXT:   BUFFER_STORE_DWORD_OFFSET $vgpr2, $sgpr0_sgpr1_sgpr2_sgpr3, killed $sgpr6, 0, 0, 0, implicit $exec :: (store (s32) into %stack.20, addrspace 5)
   ; MUBUF-NEXT:   $exec = S_MOV_B64 killed $sgpr4_sgpr5
   ; MUBUF-NEXT:   $vgpr2 = V_WRITELANE_B32 $sgpr33, 0, undef $vgpr2
   ; MUBUF-NEXT:   $sgpr33 = frame-setup S_ADD_I32 $sgpr32, 524224, implicit-def $scc
   ; MUBUF-NEXT:   $sgpr33 = frame-setup S_AND_B32 killed $sgpr33, 4294443008, implicit-def dead $scc
   ; MUBUF-NEXT:   $sgpr32 = frame-setup S_ADD_I32 $sgpr32, 11010048, implicit-def dead $scc
   ; MUBUF-NEXT:   S_CMP_EQ_U32 0, 0, implicit-def $scc
body:             |
   ; MUBUF-NEXT: {{  $}}
   ; MUBUF-NEXT:   S_NOP 0
   ; MUBUF-NEXT: {{  $}}
   ; MUBUF-NEXT: bb.2:
   ; MUBUF-NEXT:   liveins: $vgpr2
   ; MUBUF-NEXT: {{  $}}
   ; MUBUF-NEXT:   $sgpr32 = frame-destroy S_ADD_I32 $sgpr32, -11010048, implicit-def dead $scc
   ; MUBUF-NEXT:   $sgpr33 = V_READLANE_B32 $vgpr2, 0
-  ; MUBUF-NEXT:   $sgpr4_sgpr5 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+  ; MUBUF-NEXT:   $sgpr4_sgpr5 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
   ; MUBUF-NEXT:   $sgpr6 = S_ADD_I32 $sgpr32, 9961728, implicit-def dead $scc
   ; MUBUF-NEXT:   $vgpr2 = BUFFER_LOAD_DWORD_OFFSET $sgpr0_sgpr1_sgpr2_sgpr3, killed $sgpr6, 0, 0, 0, implicit $exec :: (load (s32) from %stack.20, addrspace 5)
   ; MUBUF-NEXT:   $exec = S_MOV_B64 killed $sgpr4_sgpr5
   ; MUBUF-NEXT:   S_ENDPGM 0
   ; FLATSCR-LABEL: name: use_restore_frame_reg
   ; FLATSCR: bb.0:
   ; FLATSCR-NEXT:   successors: %bb.2(0x40000000), %bb.1(0x40000000)
   ; FLATSCR-NEXT:   liveins: $vgpr1, $vgpr2
   ; FLATSCR-NEXT: {{  $}}
-  ; FLATSCR-NEXT:   $sgpr4_sgpr5 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+  ; FLATSCR-NEXT:   $sgpr4_sgpr5 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
   ; FLATSCR-NEXT:   $sgpr6 = S_ADD_I32 $sgpr32, 155652, implicit-def dead $scc
   ; FLATSCR-NEXT:   SCRATCH_STORE_DWORD_SADDR $vgpr2, killed $sgpr6, 0, 0, implicit $exec, implicit $flat_scr :: (store (s32) into %stack.20, addrspace 5)
   ; FLATSCR-NEXT:   $exec = S_MOV_B64 killed $sgpr4_sgpr5
   ; FLATSCR-NEXT:   $vgpr2 = V_WRITELANE_B32 $sgpr33, 0, undef $vgpr2
   ; FLATSCR-NEXT:   $sgpr33 = frame-setup S_ADD_I32 $sgpr32, 8191, implicit-def $scc
   ; FLATSCR-NEXT:   $sgpr33 = frame-setup S_AND_B32 killed $sgpr33, 4294959104, implicit-def dead $scc
   ; FLATSCR-NEXT:   $sgpr32 = frame-setup S_ADD_I32 $sgpr32, 172032, implicit-def dead $scc
   ; FLATSCR-NEXT:   S_CMP_EQ_U32 0, 0, implicit-def $scc
body:             |
   ; FLATSCR-NEXT: {{  $}}
   ; FLATSCR-NEXT:   S_NOP 0
   ; FLATSCR-NEXT: {{  $}}
   ; FLATSCR-NEXT: bb.2:
   ; FLATSCR-NEXT:   liveins: $vgpr2
   ; FLATSCR-NEXT: {{  $}}
   ; FLATSCR-NEXT:   $sgpr32 = frame-destroy S_ADD_I32 $sgpr32, -172032, implicit-def dead $scc
   ; FLATSCR-NEXT:   $sgpr33 = V_READLANE_B32 $vgpr2, 0
-  ; FLATSCR-NEXT:   $sgpr4_sgpr5 = S_OR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
+  ; FLATSCR-NEXT:   $sgpr4_sgpr5 = S_XOR_SAVEEXEC_B64 -1, implicit-def $exec, implicit-def dead $scc, implicit $exec
   ; FLATSCR-NEXT:   $sgpr6 = S_ADD_I32 $sgpr32, 155652, implicit-def dead $scc
   ; FLATSCR-NEXT:   $vgpr2 = SCRATCH_LOAD_DWORD_SADDR killed $sgpr6, 0, 0, implicit $exec, implicit $flat_scr :: (load (s32) from %stack.20, addrspace 5)
   ; FLATSCR-NEXT:   $exec = S_MOV_B64 killed $sgpr4_sgpr5
   ; FLATSCR-NEXT:   S_ENDPGM 0
   bb.0:
     liveins: $vgpr1
 
     S_CMP_EQ_U32 0, 0, implicit-def $scc

llvm/test/CodeGen/AMDGPU/wwm-reserved-spill.ll

 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc -O0 -mtriple=amdgcn- -mcpu=gfx900 -amdgpu-dpp-combine=false -verify-machineinstrs < %s | FileCheck -check-prefix=GFX9-O0 %s
 ; RUN: llc -mtriple=amdgcn- -mcpu=gfx900 -amdgpu-dpp-combine=false -verify-machineinstrs < %s | FileCheck -check-prefix=GFX9-O3 %s
 
 define amdgpu_gfx void @strict_wwm_no_cfg(<4 x i32> inreg %tmp14) {
 ; GFX9-O0-LABEL: strict_wwm_no_cfg:
 ; 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:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-O0-NEXT:    buffer_store_dword v0, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    buffer_store_dword v1, 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:    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
 ; GFX9-O0-NEXT:    s_mov_b32 s37, s5
 ; GFX9-O0-NEXT:    s_mov_b32 s38, s6

 ; GFX9-O0-NEXT:    v_mov_b32_e32 v4, v0
 ; GFX9-O0-NEXT:    v_cmp_eq_u32_e64 s[40:41], v3, v4
 ; GFX9-O0-NEXT:    v_cndmask_b32_e64 v3, 0, 1, s[40:41]
 ; GFX9-O0-NEXT:    s_mov_b32 s35, 1
 ; GFX9-O0-NEXT:    v_lshlrev_b32_e64 v3, s35, v3
 ; GFX9-O0-NEXT:    s_mov_b32 s35, 2
 ; GFX9-O0-NEXT:    v_and_b32_e64 v3, v3, s35
 ; GFX9-O0-NEXT:    buffer_store_dword v3, off, s[36:39], s34 offset:4
-; GFX9-O0-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-O0-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-O0-NEXT:    buffer_load_dword v0, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    buffer_load_dword v1, 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:    s_mov_b64 exec, s[34:35]
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-O0-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX9-O3-LABEL: strict_wwm_no_cfg:
 ; GFX9-O3:       ; %bb.0:
 ; GFX9-O3-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX9-O3-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-O3-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-O3-NEXT:    buffer_store_dword v0, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    buffer_store_dword v1, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    buffer_store_dword v2, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    buffer_store_dword v3, off, s[0:3], s32 offset:12 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-O3-NEXT:    buffer_load_dwordx2 v[4:5], off, s[4:7], 0
 ; GFX9-O3-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v0, 0

 ; GFX9-O3-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v4, v1
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v5, v0
 ; GFX9-O3-NEXT:    v_cmp_eq_u32_e32 vcc, v4, v5
 ; GFX9-O3-NEXT:    v_cndmask_b32_e64 v4, 0, 1, vcc
 ; GFX9-O3-NEXT:    v_lshlrev_b32_e32 v4, 1, v4
 ; GFX9-O3-NEXT:    v_and_b32_e32 v4, 2, v4
 ; GFX9-O3-NEXT:    buffer_store_dword v4, off, s[4:7], 0 offset:4
-; GFX9-O3-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-O3-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-O3-NEXT:    buffer_load_dword v0, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-O3-NEXT:    buffer_load_dword v1, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
 ; GFX9-O3-NEXT:    buffer_load_dword v2, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
 ; GFX9-O3-NEXT:    buffer_load_dword v3, off, s[0:3], s32 offset:12 ; 4-byte Folded Reload
 ; GFX9-O3-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-O3-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-O3-NEXT:    s_setpc_b64 s[30:31]
   %tmp100 = call <2 x float> @llvm.amdgcn.raw.buffer.load.v2f32(<4 x i32> %tmp14, i32 0, i32 0, i32 0)
; GFX9-O3-NEXT:    s_setpc_b64 s[30:31]
   call void @llvm.amdgcn.raw.buffer.store.f32(float %tmp145, <4 x i32> %tmp14, i32 4, i32 0, i32 0)
   ret void
 }
 
 define amdgpu_gfx void @strict_wwm_cfg(<4 x i32> inreg %tmp14, i32 %arg) {
 ; GFX9-O0-LABEL: strict_wwm_cfg:
 ; 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:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-O0-NEXT:    buffer_store_dword v3, off, s[0:3], s32 offset:16 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    buffer_store_dword v1, off, s[0:3], s32 offset:20 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    buffer_store_dword v2, off, s[0:3], s32 offset:24 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    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
 ; GFX9-O0-NEXT:    s_mov_b32 s37, s5
 ; GFX9-O0-NEXT:    s_mov_b32 s38, s6

 ; GFX9-O0-NEXT:    v_cmp_eq_u32_e64 s[34:35], v0, v4
 ; GFX9-O0-NEXT:    v_cndmask_b32_e64 v0, 0, 1, s[34:35]
 ; GFX9-O0-NEXT:    s_mov_b32 s34, 1
 ; GFX9-O0-NEXT:    v_lshlrev_b32_e64 v0, s34, v0
 ; GFX9-O0-NEXT:    s_mov_b32 s34, 2
 ; GFX9-O0-NEXT:    v_and_b32_e64 v0, v0, s34
 ; 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:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-O0-NEXT:    buffer_load_dword v3, off, s[0:3], s32 offset:16 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    buffer_load_dword v1, off, s[0:3], s32 offset:20 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    buffer_load_dword v2, off, s[0:3], s32 offset:24 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-O0-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX9-O3-LABEL: strict_wwm_cfg:
 ; GFX9-O3:       ; %bb.0: ; %entry
 ; GFX9-O3-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX9-O3-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-O3-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-O3-NEXT:    buffer_store_dword v1, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    buffer_store_dword v2, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-O3-NEXT:    buffer_load_dwordx2 v[3:4], off, s[4:7], 0
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v5, 0
 ; GFX9-O3-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX9-O3-NEXT:    s_mov_b64 exec, s[34:35]

 ; GFX9-O3-NEXT:    v_mov_b32_e32 v5, v1
 ; GFX9-O3-NEXT:  .LBB1_2: ; %merge
 ; GFX9-O3-NEXT:    s_or_b64 exec, exec, s[34:35]
 ; GFX9-O3-NEXT:    v_cmp_eq_u32_e32 vcc, v3, v5
 ; GFX9-O3-NEXT:    v_cndmask_b32_e64 v0, 0, 1, vcc
 ; GFX9-O3-NEXT:    v_lshlrev_b32_e32 v0, 1, v0
 ; GFX9-O3-NEXT:    v_and_b32_e32 v0, 2, v0
 ; GFX9-O3-NEXT:    buffer_store_dword v0, off, s[4:7], 0 offset:4
-; GFX9-O3-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-O3-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-O3-NEXT:    buffer_load_dword v1, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-O3-NEXT:    buffer_load_dword v2, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
 ; GFX9-O3-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-O3-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-O3-NEXT:    s_setpc_b64 s[30:31]
 entry:
   %tmp100 = call <2 x float> @llvm.amdgcn.raw.buffer.load.v2f32(<4 x i32> %tmp14, i32 0, i32 0, i32 0)
   %tmp101 = bitcast <2 x float> %tmp100 to <2 x i32>
; GFX9-O3-NEXT:    s_setpc_b64 s[30:31]
   %sub = sub i32 %mul, %add
   ret i32 %sub
 }
 
 define amdgpu_gfx void @strict_wwm_call(<4 x i32> inreg %tmp14, i32 inreg %arg) {
 ; GFX9-O0-LABEL: strict_wwm_call:
 ; 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:    s_xor_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:    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_add_u32_e64 v1, v1, v2
 ; GFX9-O0-NEXT:    s_mov_b64 exec, s[40:41]
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v0, v1
 ; GFX9-O0-NEXT:    buffer_store_dword v0, off, s[36:39], s34 offset:4
 ; GFX9-O0-NEXT:    v_readlane_b32 s31, v3, 1
 ; GFX9-O0-NEXT:    v_readlane_b32 s30, v3, 0
 ; GFX9-O0-NEXT:    s_add_i32 s32, s32, 0xfffffc00
 ; GFX9-O0-NEXT:    s_mov_b32 s33, s35
-; GFX9-O0-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-O0-NEXT:    s_xor_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:    s_mov_b64 exec, s[34:35]
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-O0-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX9-O3-LABEL: strict_wwm_call:
 ; GFX9-O3:       ; %bb.0:
 ; GFX9-O3-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX9-O3-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-O3-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-O3-NEXT:    buffer_store_dword v3, off, s[0:3], s32 ; 4-byte Folded Spill
 ; 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:    s_mov_b32 s38, s33
 ; GFX9-O3-NEXT:    s_mov_b32 s33, s32
 ; GFX9-O3-NEXT:    s_addk_i32 s32, 0x400

 ; GFX9-O3-NEXT:    v_add_u32_e32 v1, v1, v2
 ; GFX9-O3-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v0, v1
 ; GFX9-O3-NEXT:    buffer_store_dword v0, off, s[4:7], 0 offset:4
 ; GFX9-O3-NEXT:    v_readlane_b32 s31, v3, 1
 ; GFX9-O3-NEXT:    v_readlane_b32 s30, v3, 0
 ; GFX9-O3-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-O3-NEXT:    s_mov_b32 s33, s38
-; GFX9-O3-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-O3-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-O3-NEXT:    buffer_load_dword v3, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-O3-NEXT:    buffer_load_dword v2, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
 ; GFX9-O3-NEXT:    buffer_load_dword v1, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
 ; GFX9-O3-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-O3-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-O3-NEXT:    s_setpc_b64 s[30:31]
   %tmp107 = tail call i32 @llvm.amdgcn.set.inactive.i32(i32 %arg, i32 0)
   %tmp134 = call amdgpu_gfx i32 @strict_wwm_called(i32 %tmp107)
; GFX9-O3-NEXT:    s_setpc_b64 s[30:31]
   %sub = sub i64 %mul, %add
   ret i64 %sub
 }
 
 define amdgpu_gfx void @strict_wwm_call_i64(<4 x i32> inreg %tmp14, i64 inreg %arg) {
 ; GFX9-O0-LABEL: strict_wwm_call_i64:
 ; GFX9-O0:       ; %bb.0:
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX9-O0-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-O0-NEXT:    s_xor_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:    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:    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:    v_mov_b32_e32 v0, v2
 ; 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:    s_mov_b32 s33, s42
-; GFX9-O0-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-O0-NEXT:    s_xor_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:    s_nop 0
 ; GFX9-O0-NEXT:    buffer_load_dword v8, off, s[0:3], s32 offset:4 ; 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:    s_nop 0
 ; GFX9-O0-NEXT:    buffer_load_dword v2, off, s[0:3], s32 offset:12 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0

 ; GFX9-O0-NEXT:    buffer_load_dword v5, off, s[0:3], s32 offset:40 ; 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]
 ;
 ; GFX9-O3-LABEL: strict_wwm_call_i64:
 ; GFX9-O3:       ; %bb.0:
 ; GFX9-O3-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX9-O3-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-O3-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-O3-NEXT:    buffer_store_dword v8, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    buffer_store_dword v6, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-O3-NEXT:    buffer_store_dword v7, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    buffer_store_dword v2, off, s[0:3], s32 offset:12 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    buffer_store_dword v3, off, s[0:3], s32 offset:16 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    buffer_store_dword v2, off, s[0:3], s32 offset:20 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    s_waitcnt vmcnt(0)

 ; GFX9-O3-NEXT:    s_mov_b64 exec, s[38:39]
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v0, v2
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v1, v3
 ; GFX9-O3-NEXT:    buffer_store_dwordx2 v[0:1], off, s[4:7], 0 offset:4
 ; GFX9-O3-NEXT:    v_readlane_b32 s31, v8, 1
 ; GFX9-O3-NEXT:    v_readlane_b32 s30, v8, 0
 ; GFX9-O3-NEXT:    s_addk_i32 s32, 0xf800
 ; GFX9-O3-NEXT:    s_mov_b32 s33, s40
-; GFX9-O3-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-O3-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-O3-NEXT:    buffer_load_dword v8, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-O3-NEXT:    s_nop 0
 ; GFX9-O3-NEXT:    buffer_load_dword v6, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
 ; GFX9-O3-NEXT:    s_nop 0
 ; GFX9-O3-NEXT:    buffer_load_dword v7, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
 ; GFX9-O3-NEXT:    s_nop 0
 ; GFX9-O3-NEXT:    buffer_load_dword v2, off, s[0:3], s32 offset:12 ; 4-byte Folded Reload
 ; GFX9-O3-NEXT:    buffer_load_dword v3, off, s[0:3], s32 offset:16 ; 4-byte Folded Reload
; GFX9-O3-NEXT:    s_setpc_b64 s[30:31]
   call void @llvm.amdgcn.raw.buffer.store.v2i32(<2 x i32> %tmp138, <4 x i32> %tmp14, i32 4, i32 0, i32 0)
   ret void
 }
 
 define amdgpu_gfx void @strict_wwm_amdgpu_cs_main(<4 x i32> inreg %desc, i32 %index) {
 ; GFX9-O0-LABEL: strict_wwm_amdgpu_cs_main:
 ; 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:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-O0-NEXT:    buffer_store_dword v1, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-O0-NEXT:    buffer_store_dword v2, off, s[0:3], s32 offset:4 ; 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
 ; GFX9-O0-NEXT:    s_mov_b32 s37, s5
 ; GFX9-O0-NEXT:    s_mov_b32 s38, s6

 ; GFX9-O0-NEXT:    ; kill: def $vgpr11 killed $vgpr11 killed $exec
 ; GFX9-O0-NEXT:    ; kill: def $vgpr9 killed $vgpr9 killed $exec
 ; GFX9-O0-NEXT:    ; kill: def $vgpr5 killed $vgpr5 def $vgpr5_vgpr6_vgpr7_vgpr8 killed $exec
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v6, v11
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v7, v10
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v8, v9
 ; GFX9-O0-NEXT:    buffer_store_dwordx4 v[5:8], v0, s[36:39], s34 offen
 ; GFX9-O0-NEXT:    buffer_store_dwordx2 v[3:4], v0, s[36:39], s34 offen offset:16
-; GFX9-O0-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-O0-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-O0-NEXT:    buffer_load_dword v1, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
 ; GFX9-O0-NEXT:    buffer_load_dword v2, off, s[0:3], s32 offset:4 ; 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]
 ;
 ; GFX9-O3-LABEL: strict_wwm_amdgpu_cs_main:
 ; GFX9-O3:       ; %bb.0:
 ; GFX9-O3-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX9-O3-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-O3-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-O3-NEXT:    buffer_store_dword v1, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-O3-NEXT:    buffer_store_dword v2, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    buffer_store_dword v3, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-O3-NEXT:    buffer_store_dword v4, off, s[0:3], s32 offset:12 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    buffer_store_dword v5, off, s[0:3], s32 offset:16 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    s_waitcnt vmcnt(0)

 ; GFX9-O3-NEXT:    v_mov_b32_e32 v7, v1
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v9, v3
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v8, v2
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v10, v4
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v11, v5
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v12, v6
 ; GFX9-O3-NEXT:    buffer_store_dwordx4 v[7:10], v0, s[4:7], 0 offen
 ; GFX9-O3-NEXT:    buffer_store_dwordx2 v[11:12], v0, s[4:7], 0 offen offset:16
-; GFX9-O3-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-O3-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-O3-NEXT:    buffer_load_dword v1, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-O3-NEXT:    s_nop 0
 ; GFX9-O3-NEXT:    buffer_load_dword v2, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
 ; GFX9-O3-NEXT:    s_nop 0
 ; GFX9-O3-NEXT:    buffer_load_dword v3, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
 ; GFX9-O3-NEXT:    s_nop 0
 ; GFX9-O3-NEXT:    buffer_load_dword v4, off, s[0:3], s32 offset:12 ; 4-byte Folded Reload
 ; GFX9-O3-NEXT:    s_nop 0
; GFX9-O3-NEXT:    s_setpc_b64 s[30:31]
   ret void
 }
 
 ; FIXME: This spills v40 and v41 twice, once in whole-wave-mode and once for the active lanes
 define amdgpu_gfx <32 x i32> @strict_wwm_callee_saves(<32 x i32> inreg %keep, ptr addrspace(5) %ptr, i64 %a, i64 %b, i64 %c, i64 %d, i64 %e) #0 {
 ; GFX9-O0-LABEL: strict_wwm_callee_saves:
 ; 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 v42, off, s[0:3], s32 offset:220 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-O0-NEXT:    buffer_store_dword v32, off, s[0:3], s32 offset:224 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-O0-NEXT:    buffer_store_dword v33, off, s[0:3], s32 offset:228 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    buffer_store_dword v34, off, s[0:3], s32 offset:232 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-O0-NEXT:    buffer_store_dword v35, off, s[0:3], s32 offset:236 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    buffer_store_dword v36, off, s[0:3], s32 offset:240 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-O0-NEXT:    buffer_store_dword v37, off, s[0:3], s32 offset:244 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    buffer_store_dword v38, off, s[0:3], s32 offset:248 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-O0-NEXT:    buffer_store_dword v39, off, s[0:3], s32 offset:252 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    buffer_store_dword v40, off, s[0:3], s32 offset:256 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-O0-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:260 ; 4-byte Folded Spill
+; GFX9-O0-NEXT:    s_mov_b64 exec, -1
+; GFX9-O0-NEXT:    buffer_store_dword v42, off, s[0:3], s32 offset:220 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-O0-NEXT:    buffer_store_dword v40, off, s[0:3], s32 offset:48 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    buffer_store_dword v41, off, s[0:3], s32 offset:44 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    buffer_store_dword v43, off, s[0:3], s32 offset:40 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    buffer_store_dword v44, off, s[0:3], s32 offset:36 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    buffer_store_dword v45, off, s[0:3], s32 offset:32 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    buffer_store_dword v46, off, s[0:3], s32 offset:28 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    buffer_store_dword v47, off, s[0:3], s32 offset:24 ; 4-byte Folded Spill

 ; GFX9-O0-NEXT:    v_readlane_b32 s64, v42, 0
 ; GFX9-O0-NEXT:    buffer_load_dword v47, off, s[0:3], s32 offset:24 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    buffer_load_dword v46, off, s[0:3], s32 offset:28 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    buffer_load_dword v45, off, s[0:3], s32 offset:32 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    buffer_load_dword v44, off, s[0:3], s32 offset:36 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    buffer_load_dword v43, off, s[0:3], s32 offset:40 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    buffer_load_dword v41, off, s[0:3], s32 offset:44 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:48 ; 4-byte Folded Reload
-; GFX9-O0-NEXT:    s_or_saveexec_b64 s[34:35], -1
-; GFX9-O0-NEXT:    buffer_load_dword v42, off, s[0:3], s32 offset:220 ; 4-byte Folded Reload
-; GFX9-O0-NEXT:    s_nop 0
+; GFX9-O0-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-O0-NEXT:    buffer_load_dword v32, off, s[0:3], s32 offset:224 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
 ; GFX9-O0-NEXT:    buffer_load_dword v33, off, s[0:3], s32 offset:228 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
 ; GFX9-O0-NEXT:    buffer_load_dword v34, off, s[0:3], s32 offset:232 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
 ; GFX9-O0-NEXT:    buffer_load_dword v35, off, s[0:3], s32 offset:236 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
 ; GFX9-O0-NEXT:    buffer_load_dword v36, off, s[0:3], s32 offset:240 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
 ; GFX9-O0-NEXT:    buffer_load_dword v37, off, s[0:3], s32 offset:244 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
 ; GFX9-O0-NEXT:    buffer_load_dword v38, off, s[0:3], s32 offset:248 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
 ; GFX9-O0-NEXT:    buffer_load_dword v39, off, s[0:3], s32 offset:252 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
 ; GFX9-O0-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:256 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
 ; GFX9-O0-NEXT:    buffer_load_dword v41, off, s[0:3], s32 offset:260 ; 4-byte Folded Reload
+; GFX9-O0-NEXT:    s_mov_b64 exec, -1
+; GFX9-O0-NEXT:    buffer_load_dword v42, off, s[0:3], s32 offset:220 ; 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]
 ;
 ; GFX9-O3-LABEL: strict_wwm_callee_saves:
 ; GFX9-O3:       ; %bb.0:
 ; GFX9-O3-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX9-O3-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-O3-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-O3-NEXT:    buffer_store_dword v32, off, s[0:3], s32 offset:32 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-O3-NEXT:    buffer_store_dword v33, off, s[0:3], s32 offset:36 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    buffer_store_dword v34, off, s[0:3], s32 offset:40 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-O3-NEXT:    buffer_store_dword v35, off, s[0:3], s32 offset:44 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    buffer_store_dword v36, off, s[0:3], s32 offset:48 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    s_waitcnt vmcnt(0)

 ; GFX9-O3-NEXT:    v_mov_b32_e32 v18, s22
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v19, s23
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v20, s24
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v21, s25
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v22, s26
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v23, s27
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v24, s28
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v25, s29
-; GFX9-O3-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-O3-NEXT:    s_xor_saveexec_b64 s[34:35], -1
 ; GFX9-O3-NEXT:    buffer_load_dword v32, off, s[0:3], s32 offset:32 ; 4-byte Folded Reload
 ; GFX9-O3-NEXT:    s_nop 0
 ; GFX9-O3-NEXT:    buffer_load_dword v33, off, s[0:3], s32 offset:36 ; 4-byte Folded Reload
 ; GFX9-O3-NEXT:    s_nop 0
 ; GFX9-O3-NEXT:    buffer_load_dword v34, off, s[0:3], s32 offset:40 ; 4-byte Folded Reload
 ; GFX9-O3-NEXT:    s_nop 0
 ; GFX9-O3-NEXT:    buffer_load_dword v35, off, s[0:3], s32 offset:44 ; 4-byte Folded Reload
 ; GFX9-O3-NEXT:    s_nop 0