[AMDGPU] Move call clobbered return address registers s[30:31] to callee saved range

[RamNalamothu]

Summary

Currently the return address ABI registers s[30:31], which fall in the call
clobbered register range, are added as a live-in on the function entry to
preserve its value when we have calls so that it gets saved and restored
around the calls.

But the DWARF unwind information (CFI) needs to track where the return address
resides in a frame and the above approach makes it difficult to track the
return address when the CFI information is emitted during the frame lowering,
due to the involvment of understanding the control flow.

This patch moves the return address ABI registers s[30:31] into callee saved
registers range and stops adding live-in for return address registers, so that
the CFI machinery will know where the return address resides when CSR
save/restore happen during the frame lowering.

And doing the above poses an issue that now the return instruction uses undefined
register sgpr30_sgpr31. This is resolved by hiding the return address register
use by the return instruction through the SI_RETURN pseudo instruction, which
doesn't take any input operands, until the SI_RETURN pseudo gets lowered to the
S_SETPC_B64_return during the expandPostRAPseudo().

As an added benefit, this patch simplifies overall return instruction handling.

Note: The AMDGPU CFI changes are there only in the downstream code and another
version of this patch will be posted for review for the downstream code.

[sebastian-ne]

llvm/lib/Target/AMDGPU/SILowerSGPRSpills.cpp

94–96

Can the MVT be deduced from getRegSizeInBits(*getPhysRegClass(Reg))? I think that would look cleaner.

232

Why is the return address only marked as callee-save if this function has calls? I assume it can’t it be overwritten if there are no calls because it is marked as reserved?

Coming back here after reading the tests, I think s[30:31] should also be marked as callee-save if there are no calls. It shouldn’t cause harm if there are no calls (the register is reserved anyway) and it should fix the tests.

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

897

This test is not working with this patch. It has no calls, so s[30:31] are not marked as callee-save. But they are overwritten in inline assembly.
Before this patch, they were saved and restored. With this patch this is not the case and the return will jump to an invalid address.

llvm/test/CodeGen/AMDGPU/call-preserved-registers.ll

75–84

I think this test is supposed to clobber the return address in s[30:31].

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

419

I think these tests should continue to work without the need to introduce a call.

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

127

I think this test purposefully clobbered the return address in s[30:31].

[sebastian-ne]

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

26–49

This codes gets quite a bit worse. But I’m not sure if we can do anything about it.

[RamNalamothu]

llvm/lib/Target/AMDGPU/SILowerSGPRSpills.cpp

94–96

Can the MVT be deduced from getRegSizeInBits(*getPhysRegClass(Reg))? I think that would look cleaner.

Will check, thank you.

232

Why is the return address only marked as callee-save if this function has calls? I assume it can’t it be overwritten if there are no calls because it is marked as reserved?

Yes.

Coming back here after reading the tests, I think s[30:31] should also be marked as callee-save if there are no calls. It shouldn’t cause harm if there are no calls (the register is reserved anyway) and it should fix the tests.

AFAIK that's the eventual goal and there is a discussion to pick the ABI registers for return address from the callee saved range. And these changes should work with no/minimal changes.

However, I don't think I understand how the tests will get fixed automatically if the return address registers are marked as callee saved because even the SGPR CSR save/restore happens during SGPRSpillLowering which is where I am manually adding s[30:31] to the CSR list.

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

419

I think these tests should continue to work without the need to introduce a call.

Thanks. I will take a re-look at it. I don't recall now why I added the call.

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

127

I think this test purposefully clobbered the return address in s[30:31].

I believe the intent here is to test the behavior when non callee saved registers are clobbered inside inline asm but not particularly s[30:31] and I guess s[30:31] happened to be used since they are the boundary of non callee save range. The same goes for the other tests which were clobbering s[30:31].

As of now there is no policy on how the inline asm could use the ABI registers. I spoke to @arsenm and he is aware of this test changes.

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

26–49

This codes gets quite a bit worse. But I’m not sure if we can do anything about it.

Yeah. This is a known scenario where previously the s[30:31] were being spilled to another SGPR pair, which was used to return, but now since the return address should always be in s[30:31] it has to be spilled into either VGPR or memory.

For now, we will have to live with it.

May be we could do something during SGPRSpillLowering to use the available SGPRs, not only for this case but in general?
I didn't check if that capability is already in place and s[30:31] is a corner case which is not handled.

[scott.linder]

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

26–49

I'm not sure I understand how the old code is correct to begin with; does it rely on knowledge about the called function?

I think in a "fixed function ABI" world we can't just clobber (in this example) s[36:37], right?

[sebastian-ne]

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

26–49

As you said, it relies on knowledge about the called function.
The IPRA (Inter Procedural Register Allocation) pass looks at direct callees and removes registers from the call-clobbered list if the called function does not touch them.

[arsenm]

llvm/lib/Target/AMDGPU/SILowerSGPRSpills.cpp

232

I don't think this register should be reserved, which avoids the need to do this

llvm/lib/Target/AMDGPU/SIRegisterInfo.cpp

686–689

I don't think it's necessary to reserve this

[RamNalamothu]

Rebase and fix tests in llvm/test/CodeGen/AMDGPU/callee-frame-setup.ll.

[RamNalamothu]

llvm/lib/Target/AMDGPU/SILowerSGPRSpills.cpp

94–96

getRegSizeInBits(*getPhysRegClass(Reg)) is not useful as getPhysRegClass(Reg) gives base register class i.e. SReg_32.

/// Return the 'base' register class for this register.
/// e.g. SGPR0 => SReg_32, VGPR => VGPR_32 SGPR0_SGPR1 -> SReg_32, etc.
const TargetRegisterClass *getPhysRegClass(MCRegister Reg) const;

llvm/lib/Target/AMDGPU/SIRegisterInfo.cpp

686–689

But otherwise the register allocation thinks it is available for allocation as s[30:31] are scratch registers at the moment and its usage with return instruction is hidden with using SI_RETURN through the register allocation and we don't emit Live-ins anymore.

And we will have to reserve it here even after these registers actually get picked from CSR range because we want to emit save/restore exclusively so that the unwind info machinery knows it and can generate CFI as needed for return address. That's the reason why all of this got started right.

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

419

My bad. I understood them wrongly.

Yes, you are right. Removed the calls and fixed the tests accordingly.

[arsenm]

llvm/lib/Target/AMDGPU/SIRegisterInfo.cpp

686–689

I think something special happens with the return address we are ignoring, but I haven't followed it all the way through.

AArch64 passes LR to the TargetRegisterInfo constructor, where we're passing the dummy PC_REG. If you replace that with the real return address, does that help?

[arsenm]

llvm/lib/Target/AMDGPU/SIRegisterInfo.cpp

686–689

Even if not, AArch64 does not reserve the return address

[RamNalamothu]

llvm/lib/Target/AMDGPU/SIRegisterInfo.cpp

686–689

I looked at it already and it appears there is nothing magical about it except getting MCRegisterInfo::getRARegister() to return that register from the constructor.

And we get the same by hardcoding s[30:31] in SIRegisterInfo::getReturnAddressReg().

Though I didn't go through everything that Hexagon does, but they also reserve the return address register.

def retflag : SDNode<"HexagonISD::RET_FLAG", SDTNone,
                      [SDNPHasChain, SDNPOptInGlue, SDNPVariadic]>;
def: Pat<(retflag),   (PS_jmpret (i32 R31))>;

[sebastian-ne]

llvm/lib/Target/AMDGPU/SILowerSGPRSpills.cpp

232

Why is the return address not callee-save if there are no calls?

Does removing the && MFI.hasCalls() condition have any bad effects?

[RamNalamothu]

llvm/lib/Target/AMDGPU/SILowerSGPRSpills.cpp

232

No functional issues but its just that it will be unnecessary to save/restore when the registers are not clobbered i.e. when we have no calls.

[sebastian-ne]

llvm/lib/Target/AMDGPU/SILowerSGPRSpills.cpp

232

But if we have no calls and the register is not clobbered, marking the registers as callee-save shouldn’t save and restore them?

I think it doesn’t hurt to remove this condition. On the contrary, I expect that removing the hasCalls condition will fix the case where s30 or s31 are overwritten in inline-assembly.

[RamNalamothu]

llvm/lib/Target/AMDGPU/SILowerSGPRSpills.cpp

232

Sorry, I didn't get your point well.

Are you saying we should save/restore s[30:31] even when we don't have calls i.e. even when s[30:31] is not modified in the current function? Does that help in any way, other than the inline asm scenario?

For the inline asm scenario, as I already mentioned. I heard that we currently don't have any policy and the ABI registers cannot be clobbered like any other scratch register right?
And even if we want to handle that scenario, I guess a blanket removal of hasCalls() condition is not the best/optimal IMO because we will be emitting unnecessary save/restore if the code inside inline asm does not actually modify s[30:31].

[sebastian-ne]

llvm/lib/Target/AMDGPU/SILowerSGPRSpills.cpp

232

Ups, ignore my previous comment. I should have read the code again before making that comment.

What I thought this code would do, is add s[30:31] to the callee-save register list if this condition is satisfied. I did not see that it is already spilling it, based on the condition.

Let me try what happens if s[30:31] is (unconditionally) added to the CSR list and the code here is removed.

[sebastian-ne]

llvm/lib/Target/AMDGPU/SILowerSGPRSpills.cpp

232

I tested it and I think marking s[30:31] as callee-save register instead of explicitly saving them here is better.

1. We don’t need hasCall to check if s[30:31] may be overwritten.
2. We shouldn’t need to mark the register as reserved (didn’t test this).
3. It “just works”, even if there is inline assembly overwriting s30 or s31 (and there are no unnecessary saves/restores if it isn’t overwritten).

The last point also means you don’t need to add work-arounds to tests by introducing artificial calls or changing the registers in inline assembly, because it will keep working as is.

See this sketched commit: https://github.com/Flakebi/amdvlk-llvm-project/commit/240040601e0f8849acb85141142408b3e6a32ff9
E.g. it fixed the test here: https://github.com/Flakebi/amdvlk-llvm-project/commit/240040601e0f8849acb85141142408b3e6a32ff9#diff-25d99c15c21155a13557b29a1c0edfa792c19040f02a0d7cd676865b46275713L75-R88

[RamNalamothu]

llvm/lib/Target/AMDGPU/SILowerSGPRSpills.cpp

232

What you did there is you just took a part of this patch and doing that is incorrect. If you still want to do those experiments, please take the full patch and try making your changes on top of that.

This patch makes changes to return instruction and its use of return address during the instruction selection. I think you missed to fully understand that aspect.

Please read again my earlier response to @arsenm which I copied here.

I don't think it's necessary to reserve this

But otherwise the register allocation thinks it is available for allocation as s[30:31] are scratch registers at the moment and its usage with return instruction is hidden with using SI_RETURN through the register allocation and we don't emit Live-ins anymore.

And we will have to reserve it here even after these registers actually get picked from CSR range because we want to emit save/restore exclusively so that the unwind info machinery knows it and can generate CFI as needed for return address. That's the reason why all of this got started right.

[sebastian-ne]

llvm/lib/Target/AMDGPU/SILowerSGPRSpills.cpp

232

I used arc patch to get the diff (https://github.com/Flakebi/amdvlk-llvm-project/commit/240040601e0f8849acb85141142408b3e6a32ff9) and based my commit on top of that, so I hope it got all of your changes?

Unfortunately I’m not familiar with CFI (I guess it refers to call-frame-information and not control-flow-integrity?). Are there more upcoming patches that need the explicit spill here and can’t rely on the llvm infrastructure for CSRs?

I like that this patch simplifies the call-related code in amdgpu, but as you probably noticed, I’m not that happy that inline assembly that touches s[30:31] now ends up with non-working code.

If this is an intended change (and I don’t feel authoritative to say if this is fine or not), I think it should be mentioned in the commit message that s[30:31] is not saved anymore if used in inline assembly. And the tests that intend to test that (like @void_func_void_clobber_s30_s31) should probably just be removed.

[RamNalamothu]

llvm/lib/Target/AMDGPU/SILowerSGPRSpills.cpp

232

Ah, my bad. I didn't see the parent commit.

It is call frame information and there will be additional CFI changes but only in the downstream. They are not/can't be up-streamed yet.

This patch is written with the assumption that s[30:31] can't be moved to CSR range right now. Like I mentioned in the beginning, there is already a discussion to pick a different register pair (i.e. s[28:29]) for return address which will be part of the (new) CSR range.

Yes, the condition check in SILowerSGPRSpills can be removed if we can move s[30:31] to CSR range now. @t-tye @arsenm what do you say?

However s[30:31] spill should not come from the register allocation and must be generated only during SGPR/frame lowering for the unwind info (CFI) generation. For this, I still believe s[30:31] has to be reserved so that the register allocation won't interfere with that. Or am I missing something here @scott.linder ?

[RamNalamothu]

llvm/lib/Target/AMDGPU/SILowerSGPRSpills.cpp

232

@t-tye @arsenm @scott.linder ping.

[RamNalamothu]

Rebase onto ToT and also made the following changes to address review feedback.

• Move s[30:31] into CSR range
• Do not reserve s[30:31] during register allocation

[sebastian-ne]

Thanks for making these changes!
The patch looks good to me, I’ll leave it for others to accept.

[arsenm]

llvm/lib/Target/AMDGPU/SIInstrInfo.cpp

2048

If we're going to just hardcode this to s[30:31], probably should do away with getReturnAddressReg

2051–2056

There's a copyImplicitOperands helper for this

[RamNalamothu]

Rebase onto ToT and use copyImplicitOps.

[RamNalamothu]

I think I have addressed all the outstanding feedback. Is this good to go now?

llvm/lib/Target/AMDGPU/SIInstrInfo.cpp

2048

I believe using getReturnAddressReg() doesn't hurt and any future updates can be made easily without changing all the places.

[arsenm]

llvm/lib/Target/AMDGPU/SIInstrInfo.cpp

2052

Should break or return true, I don't like switch fallthrough here

[RamNalamothu]

llvm/lib/Target/AMDGPU/SIInstrInfo.cpp

2052

Ah, thanks for catching it. Fixed in the committed changes.

[ronlieb]

hi Ram, landed patch breaks https://lab.llvm.org/buildbot/#/builders/193/builds/3514

[RamNalamothu]

Fixed an issue exposed by OpenMP LIT test failures.

determineCalleeSavesSGPR() now explicitly adds S30 and S31 to CSR list
when we have calls and this is needed for cases like D117243.

[RamNalamothu]

Fixed the blender application failure seen in the downstream testing.

Added the test llvm/test/CodeGen/AMDGPU/ipra-return-address-save-restore.ll to catch that scenario.

[arsenm]

llvm/lib/Target/AMDGPU/SIFrameLowering.cpp

1351 ↗

(On Diff #413656)

Referring to phabricator reviews in a comment is weird

1359 ↗

(On Diff #413656)

Can just directly check isPhysRegModified(RetAddrReg)

[RamNalamothu]

Address review feedback.

[RamNalamothu]

llvm/lib/Target/AMDGPU/SIFrameLowering.cpp

1351 ↗

(On Diff #413656)

I have seen this in many places of clang/llvm and in fact in AMDGPU's source/tests as well. But if you say this is not a recommended practice, I will remove those references while landing the patch.

1359 ↗

(On Diff #413656)

Fixed. Thanks.

[arsenm]

I'm not sure I understand why a separate isPhysRegUsed check is needed since ultimately that's what IPRA is doing, but I don't see how it could hurt

[RamNalamothu]

In D114652#3367310, @arsenm wrote:

I'm not sure I understand why a separate isPhysRegUsed check is needed since ultimately that's what IPRA is doing, but I don't see how it could hurt

The separate isPhysRegModified check is needed because we will have the SI_RETURN pseudo, which doesn't have return address operand/use, when IPRA's register usage collection runs and that hides the return address use.
Since the caller function save/restores return address registers, as the call instruction clobbers them, and return address registers use in the callee function is hidden, the IPRA/register allocator concludes that return address registers can be clobbered without save/restoring them in the callee function if the callee function doesn't have any function calls.

Hope that helps.

llvm/lib/Target/AMDGPU/AMDGPUCallLowering.cpp

return determineAndHandleAssignments(RetHandler, Assigner, SplitRetInfos, B,
                                        CC, F.isVarArg());
 }
 
 bool AMDGPUCallLowering::lowerReturn(MachineIRBuilder &B, const Value *Val,
                                      ArrayRef<Register> VRegs,
                                      FunctionLoweringInfo &FLI) const {
 
   MachineFunction &MF = B.getMF();
-  MachineRegisterInfo &MRI = MF.getRegInfo();
   SIMachineFunctionInfo *MFI = MF.getInfo<SIMachineFunctionInfo>();
   MFI->setIfReturnsVoid(!Val);
 
   assert(!Val == VRegs.empty() && "Return value without a vreg");
 
   CallingConv::ID CC = B.getMF().getFunction().getCallingConv();
   const bool IsShader = AMDGPU::isShader(CC);
   const bool IsWaveEnd =
       (IsShader && MFI->returnsVoid()) || AMDGPU::isKernel(CC);
   if (IsWaveEnd) {
     B.buildInstr(AMDGPU::S_ENDPGM)
       .addImm(0);
     return true;
   }
 
-  auto const &ST = MF.getSubtarget<GCNSubtarget>();
-
-  unsigned ReturnOpc = 0;
-  if (IsShader)
-    ReturnOpc = AMDGPU::SI_RETURN_TO_EPILOG;
-  else if (CC == CallingConv::AMDGPU_Gfx)
-    ReturnOpc = AMDGPU::S_SETPC_B64_return_gfx;
-  else
-    ReturnOpc = AMDGPU::S_SETPC_B64_return;
-
+  unsigned ReturnOpc =
+      IsShader ? AMDGPU::SI_RETURN_TO_EPILOG : AMDGPU::SI_RETURN;
   auto Ret = B.buildInstrNoInsert(ReturnOpc);
-  Register ReturnAddrVReg;
-  if (ReturnOpc == AMDGPU::S_SETPC_B64_return) {
-    ReturnAddrVReg = MRI.createVirtualRegister(&AMDGPU::CCR_SGPR_64RegClass);
-    Ret.addUse(ReturnAddrVReg);
-  } else if (ReturnOpc == AMDGPU::S_SETPC_B64_return_gfx) {
-    ReturnAddrVReg =
-        MRI.createVirtualRegister(&AMDGPU::Gfx_CCR_SGPR_64RegClass);
-    Ret.addUse(ReturnAddrVReg);
-  }
 
   if (!FLI.CanLowerReturn)
     insertSRetStores(B, Val->getType(), VRegs, FLI.DemoteRegister);
   else if (!lowerReturnVal(B, Val, VRegs, Ret))
     return false;
 
-  if (ReturnOpc == AMDGPU::S_SETPC_B64_return ||
-      ReturnOpc == AMDGPU::S_SETPC_B64_return_gfx) {
-    const SIRegisterInfo *TRI = ST.getRegisterInfo();
-    Register LiveInReturn = MF.addLiveIn(TRI->getReturnAddressReg(MF),
-                                         &AMDGPU::SGPR_64RegClass);
-    B.buildCopy(ReturnAddrVReg, LiveInReturn);
-  }
-
   // TODO: Handle CalleeSavedRegsViaCopy.
 
   B.insertInstr(Ret);
   return true;
 }
 
 void AMDGPUCallLowering::lowerParameterPtr(Register DstReg, MachineIRBuilder &B,
                                            uint64_t Offset) const {
bool AMDGPUCallLowering::lowerFormalArguments(
   const SIRegisterInfo *TRI = Subtarget.getRegisterInfo();
   const DataLayout &DL = F.getParent()->getDataLayout();
 
   Info->allocateModuleLDSGlobal(F.getParent());
 
   SmallVector<CCValAssign, 16> ArgLocs;
   CCState CCInfo(CC, F.isVarArg(), MF, ArgLocs, F.getContext());
 
-  if (!IsEntryFunc) {
-    Register ReturnAddrReg = TRI->getReturnAddressReg(MF);
-    Register LiveInReturn = MF.addLiveIn(ReturnAddrReg,
-                                         &AMDGPU::SGPR_64RegClass);
-    MBB.addLiveIn(ReturnAddrReg);
-    B.buildCopy(LiveInReturn, ReturnAddrReg);
-  }
-
   if (Info->hasImplicitBufferPtr()) {
     Register ImplicitBufferPtrReg = Info->addImplicitBufferPtr(*TRI);
     MF.addLiveIn(ImplicitBufferPtrReg, &AMDGPU::SGPR_64RegClass);
     CCInfo.AllocateReg(ImplicitBufferPtrReg);
   }
 
   SmallVector<ArgInfo, 32> SplitArgs;
   unsigned Idx = 0;

llvm/lib/Target/AMDGPU/AMDGPUISelLowering.h

enum NodeType : unsigned {
   ENDPGM,
 
   // Return to a shader part's epilog code.
   RETURN_TO_EPILOG,
 
   // Return with values from a non-entry function.
   RET_FLAG,
 
-  // Return with values from a non-entry function (AMDGPU_Gfx CC).
-  RET_GFX_FLAG,
-
   DWORDADDR,
   FRACT,
 
   /// CLAMP value between 0.0 and 1.0. NaN clamped to 0, following clamp output
   /// modifier behavior with dx10_enable.
   CLAMP,
 
   // This is SETCC with the full mask result which is used for a compare with a

llvm/lib/Target/AMDGPU/AMDGPUISelLowering.cpp

const char* AMDGPUTargetLowering::getTargetNodeName(unsigned Opcode) const {
   // AMDGPU DAG nodes
   NODE_NAME_CASE(IF)
   NODE_NAME_CASE(ELSE)
   NODE_NAME_CASE(LOOP)
   NODE_NAME_CASE(CALL)
   NODE_NAME_CASE(TC_RETURN)
   NODE_NAME_CASE(TRAP)
   NODE_NAME_CASE(RET_FLAG)
-  NODE_NAME_CASE(RET_GFX_FLAG)
   NODE_NAME_CASE(RETURN_TO_EPILOG)
   NODE_NAME_CASE(ENDPGM)
   NODE_NAME_CASE(DWORDADDR)
   NODE_NAME_CASE(FRACT)
   NODE_NAME_CASE(SETCC)
   NODE_NAME_CASE(SETREG)
   NODE_NAME_CASE(DENORM_MODE)
   NODE_NAME_CASE(FMA_W_CHAIN)

llvm/lib/Target/AMDGPU/AMDGPUInstrInfo.td

 // Call/Return DAG Nodes
 //===----------------------------------------------------------------------===//
 def AMDGPUendpgm : SDNode<"AMDGPUISD::ENDPGM", SDTNone,
     [SDNPHasChain, SDNPOptInGlue]>;
 
 def AMDGPUreturn_to_epilog : SDNode<"AMDGPUISD::RETURN_TO_EPILOG", SDTNone,
     [SDNPHasChain, SDNPOptInGlue, SDNPVariadic]>;
 
-def AMDGPUret_flag : SDNode<"AMDGPUISD::RET_FLAG", SDTypeProfile<0, 1, [SDTCisPtrTy<0>]>,
-  [SDNPHasChain, SDNPOptInGlue, SDNPVariadic]
->;
-
-def AMDGPUret_gfx_flag : SDNode<"AMDGPUISD::RET_GFX_FLAG", SDTypeProfile<0, 1, [SDTCisPtrTy<0>]>,
+def AMDGPUret_flag : SDNode<"AMDGPUISD::RET_FLAG", SDTNone,
   [SDNPHasChain, SDNPOptInGlue, SDNPVariadic]
 >;
 
 
 //===----------------------------------------------------------------------===//
 // Intrinsic/Custom node compatibility PatFrags
 //===----------------------------------------------------------------------===//
 

llvm/lib/Target/AMDGPU/AMDGPUMCInstLower.cpp

 
 void AMDGPUMCInstLower::lower(const MachineInstr *MI, MCInst &OutMI) const {
   unsigned Opcode = MI->getOpcode();
   const auto *TII = static_cast<const SIInstrInfo*>(ST.getInstrInfo());
 
   // FIXME: Should be able to handle this with emitPseudoExpansionLowering. We
   // need to select it to the subtarget specific version, and there's no way to
   // do that with a single pseudo source operation.
-  if (Opcode == AMDGPU::S_SETPC_B64_return ||
-      Opcode == AMDGPU::S_SETPC_B64_return_gfx)
+  if (Opcode == AMDGPU::S_SETPC_B64_return)
     Opcode = AMDGPU::S_SETPC_B64;
   else if (Opcode == AMDGPU::SI_CALL) {
     // SI_CALL is just S_SWAPPC_B64 with an additional operand to track the
     // called function (which we need to remove here).
     OutMI.setOpcode(TII->pseudoToMCOpcode(AMDGPU::S_SWAPPC_B64));
     MCOperand Dest, Src;
     lowerOperand(MI->getOperand(0), Dest);
     lowerOperand(MI->getOperand(1), Src);

llvm/lib/Target/AMDGPU/SIISelLowering.cpp

SITargetLowering::LowerReturn(SDValue Chain, CallingConv::ID CallConv,
 
   // Analyze outgoing return values.
   CCInfo.AnalyzeReturn(Outs, CCAssignFnForReturn(CallConv, isVarArg));
 
   SDValue Flag;
   SmallVector<SDValue, 48> RetOps;
   RetOps.push_back(Chain); // Operand #0 = Chain (updated below)
 
-  // Add return address for callable functions.
-  if (!Info->isEntryFunction()) {
-    const SIRegisterInfo *TRI = getSubtarget()->getRegisterInfo();
-    SDValue ReturnAddrReg = CreateLiveInRegister(
-      DAG, &AMDGPU::SReg_64RegClass, TRI->getReturnAddressReg(MF), MVT::i64);
-
-    SDValue ReturnAddrVirtualReg =
-        DAG.getRegister(MF.getRegInfo().createVirtualRegister(
-                            CallConv != CallingConv::AMDGPU_Gfx
-                                ? &AMDGPU::CCR_SGPR_64RegClass
-                                : &AMDGPU::Gfx_CCR_SGPR_64RegClass),
-                        MVT::i64);
-    Chain =
-        DAG.getCopyToReg(Chain, DL, ReturnAddrVirtualReg, ReturnAddrReg, Flag);
-    Flag = Chain.getValue(1);
-    RetOps.push_back(ReturnAddrVirtualReg);
-  }
-
   // Copy the result values into the output registers.
   for (unsigned I = 0, RealRVLocIdx = 0, E = RVLocs.size(); I != E;
        ++I, ++RealRVLocIdx) {
     CCValAssign &VA = RVLocs[I];
     assert(VA.isRegLoc() && "Can only return in registers!");
     // TODO: Partially return in registers if return values don't fit.
     SDValue Arg = OutVals[RealRVLocIdx];
 
SITargetLowering::LowerReturn(SDValue Chain, CallingConv::ID CallConv,
   }
 
   // Update chain and glue.
   RetOps[0] = Chain;
   if (Flag.getNode())
     RetOps.push_back(Flag);
 
   unsigned Opc = AMDGPUISD::ENDPGM;
-  if (!IsWaveEnd) {
-    if (IsShader)
-      Opc = AMDGPUISD::RETURN_TO_EPILOG;
-    else if (CallConv == CallingConv::AMDGPU_Gfx)
-      Opc = AMDGPUISD::RET_GFX_FLAG;
-    else
-      Opc = AMDGPUISD::RET_FLAG;
-  }
-
+  if (!IsWaveEnd)
+    Opc = IsShader ? AMDGPUISD::RETURN_TO_EPILOG : AMDGPUISD::RET_FLAG;
   return DAG.getNode(Opc, DL, MVT::Other, RetOps);
 }
 
 SDValue SITargetLowering::LowerCallResult(
     SDValue Chain, SDValue InFlag, CallingConv::ID CallConv, bool IsVarArg,
     const SmallVectorImpl<ISD::InputArg> &Ins, const SDLoc &DL,
     SelectionDAG &DAG, SmallVectorImpl<SDValue> &InVals, bool IsThisReturn,
     SDValue ThisVal) const {
SDValue SITargetLowering::LowerCall(CallLoweringInfo &CLI,
   // and flag operands which copy the outgoing args into the appropriate regs.
   SDValue InFlag;
   for (auto &RegToPass : RegsToPass) {
     Chain = DAG.getCopyToReg(Chain, DL, RegToPass.first,
                              RegToPass.second, InFlag);
     InFlag = Chain.getValue(1);
   }
 
 
-  SDValue PhysReturnAddrReg;
-  if (IsTailCall) {
-    // Since the return is being combined with the call, we need to pass on the
-    // return address.
-
-    const SIRegisterInfo *TRI = getSubtarget()->getRegisterInfo();
-    SDValue ReturnAddrReg = CreateLiveInRegister(
-      DAG, &AMDGPU::SReg_64RegClass, TRI->getReturnAddressReg(MF), MVT::i64);
-
-    PhysReturnAddrReg = DAG.getRegister(TRI->getReturnAddressReg(MF),
-                                        MVT::i64);
-    Chain = DAG.getCopyToReg(Chain, DL, PhysReturnAddrReg, ReturnAddrReg, InFlag);
-    InFlag = Chain.getValue(1);
-  }
-
   // We don't usually want to end the call-sequence here because we would tidy
   // the frame up *after* the call, however in the ABI-changing tail-call case
   // we've carefully laid out the parameters so that when sp is reset they'll be
   // in the correct location.
   if (IsTailCall && !IsSibCall) {
     Chain = DAG.getCALLSEQ_END(Chain,
                                DAG.getTargetConstant(NumBytes, DL, MVT::i32),
                                DAG.getTargetConstant(0, DL, MVT::i32),
if (GlobalAddressSDNode *GSD = dyn_cast<GlobalAddressSDNode>(Callee)) {
     Ops.push_back(DAG.getTargetConstant(0, DL, MVT::i64));
   }
 
   if (IsTailCall) {
     // Each tail call may have to adjust the stack by a different amount, so
     // this information must travel along with the operation for eventual
     // consumption by emitEpilogue.
     Ops.push_back(DAG.getTargetConstant(FPDiff, DL, MVT::i32));
-
-    Ops.push_back(PhysReturnAddrReg);
   }
 
   // Add argument registers to the end of the list so that they are known live
   // into the call.
   for (auto &RegToPass : RegsToPass) {
     Ops.push_back(DAG.getRegister(RegToPass.first,
                                   RegToPass.second.getValueType()));
   }

llvm/lib/Target/AMDGPU/SIInsertWaitcnts.cpp

if (MI.getOpcode() == AMDGPU::BUFFER_WBINVL1 ||
       MI.getOpcode() == AMDGPU::BUFFER_GL1_INV) {
     Wait.VmCnt = 0;
   }
 
   // All waits must be resolved at call return.
   // NOTE: this could be improved with knowledge of all call sites or
   //   with knowledge of the called routines.
   if (MI.getOpcode() == AMDGPU::SI_RETURN_TO_EPILOG ||
+      MI.getOpcode() == AMDGPU::SI_RETURN ||
       MI.getOpcode() == AMDGPU::S_SETPC_B64_return ||
-      MI.getOpcode() == AMDGPU::S_SETPC_B64_return_gfx ||
       (MI.isReturn() && MI.isCall() && !callWaitsOnFunctionEntry(MI))) {
     Wait = Wait.combined(AMDGPU::Waitcnt::allZero(ST->hasVscnt()));
   }
   // Resolve vm waits before gs-done.
   else if ((MI.getOpcode() == AMDGPU::S_SENDMSG ||
             MI.getOpcode() == AMDGPU::S_SENDMSGHALT) &&
            ((MI.getOperand(0).getImm() & AMDGPU::SendMsg::ID_MASK_) ==
             AMDGPU::SendMsg::ID_GS_DONE)) {

llvm/lib/Target/AMDGPU/SIInstrInfo.cpp

bool SIInstrInfo::expandPostRAPseudo(MachineInstr &MI) const {
   }
   case AMDGPU::EXIT_STRICT_WWM:
   case AMDGPU::EXIT_STRICT_WQM: {
     // This only gets its own opcode so that SIPreAllocateWWMRegs can tell when
     // WWM/STICT_WQM is exited.
     MI.setDesc(get(ST.isWave32() ? AMDGPU::S_MOV_B32 : AMDGPU::S_MOV_B64));
     break;
   }
+  case AMDGPU::SI_RETURN: {
+    const MachineFunction *MF = MBB.getParent();
+    const GCNSubtarget &ST = MF->getSubtarget<GCNSubtarget>();
+    const SIRegisterInfo *TRI = ST.getRegisterInfo();
+    // Hiding the return address use with SI_RETURN may lead to extra kills in
+    // the function and missing live-ins. We are fine in practice because callee
+    // saved register handling ensures the register value is restored before
+    // RET, but we need the undef flag here to appease the MachineVerifier
+    // liveness checks.
+    MachineInstrBuilder MIB =
+        BuildMI(MBB, MI, DL, get(AMDGPU::S_SETPC_B64_return))
+            .addReg(TRI->getReturnAddressReg(*MF), RegState::Undef);
+
+    const MCInstrDesc &Desc = MI.getDesc();
+    for (unsigned i = Desc.getNumOperands(), e = MI.getNumOperands(); i != e;
+         ++i) {
+      const MachineOperand &MO = MI.getOperand(i);
+      assert(MO.isReg() && MO.getReg());
+      MIB.add(MO);
+    }
+    MI.eraseFromParent();
+  }
   }
   return true;
 }
 
 std::pair<MachineInstr*, MachineInstr*>
 SIInstrInfo::expandMovDPP64(MachineInstr &MI) const {
   assert (MI.getOpcode() == AMDGPU::V_MOV_B64_DPP_PSEUDO);
 

llvm/lib/Target/AMDGPU/SIInstructions.td

def SI_RETURN_TO_EPILOG : SPseudoInstSI <
   let isReturn = 1;
   let hasNoSchedulingInfo = 1;
   let DisableWQM = 1;
   let FixedSize = 1;
 }
 
 // Return for returning function calls.
 def SI_RETURN : SPseudoInstSI <
-  (outs), (ins), [],
+  (outs), (ins), [(AMDGPUret_flag)],
   "; return"> {
   let isTerminator = 1;
   let isBarrier = 1;
   let isReturn = 1;
   let SchedRW = [WriteBranch];
 }
 
 // Return for returning function calls without output register.

llvm/lib/Target/AMDGPU/SILowerSGPRSpills.cpp

 /// Insert restore code for the callee-saved registers used in the function.
 static void insertCSRSaves(MachineBasicBlock &SaveBlock,
                            ArrayRef<CalleeSavedInfo> CSI,
                            LiveIntervals *LIS) {
   MachineFunction &MF = *SaveBlock.getParent();
   const TargetInstrInfo &TII = *MF.getSubtarget().getInstrInfo();
   const TargetFrameLowering *TFI = MF.getSubtarget().getFrameLowering();
   const TargetRegisterInfo *TRI = MF.getSubtarget().getRegisterInfo();
+  const GCNSubtarget &ST = MF.getSubtarget<GCNSubtarget>();
+  const SIRegisterInfo *RI = ST.getRegisterInfo();
 
   MachineBasicBlock::iterator I = SaveBlock.begin();
   if (!TFI->spillCalleeSavedRegisters(SaveBlock, I, CSI, TRI)) {
     const MachineRegisterInfo &MRI = MF.getRegInfo();
 
     for (const CalleeSavedInfo &CS : CSI) {
       // Insert the spill to the stack frame.
       MCRegister Reg = CS.getReg();
 
       MachineInstrSpan MIS(I, &SaveBlock);
-      const TargetRegisterClass *RC =
-        TRI->getMinimalPhysRegClass(Reg, MVT::i32);
+      const TargetRegisterClass *RC = TRI->getMinimalPhysRegClass(
+          Reg, Reg == RI->getReturnAddressReg(MF) ? MVT::i64 : MVT::i32);
 
       // If this value was already livein, we probably have a direct use of the
       // incoming register value, so don't kill at the spill point. This happens
       // since we pass some special inputs (workgroup IDs) in the callee saved
       // range.
       const bool IsLiveIn = MRI.isLiveIn(Reg);
       TII.storeRegToStackSlot(SaveBlock, I, Reg, !IsLiveIn, CS.getFrameIdx(),
                               RC, TRI);
 

 /// Insert restore code for the callee-saved registers used in the function.
 static void insertCSRRestores(MachineBasicBlock &RestoreBlock,
                               MutableArrayRef<CalleeSavedInfo> CSI,
                               LiveIntervals *LIS) {
   MachineFunction &MF = *RestoreBlock.getParent();
   const TargetInstrInfo &TII = *MF.getSubtarget().getInstrInfo();
   const TargetFrameLowering *TFI = MF.getSubtarget().getFrameLowering();
   const TargetRegisterInfo *TRI = MF.getSubtarget().getRegisterInfo();
-
+  const GCNSubtarget &ST = MF.getSubtarget<GCNSubtarget>();
+  const SIRegisterInfo *RI = ST.getRegisterInfo();
   // Restore all registers immediately before the return and any
   // terminators that precede it.
   MachineBasicBlock::iterator I = RestoreBlock.getFirstTerminator();
 
   // FIXME: Just emit the readlane/writelane directly
   if (!TFI->restoreCalleeSavedRegisters(RestoreBlock, I, CSI, TRI)) {
     for (const CalleeSavedInfo &CI : reverse(CSI)) {
       unsigned Reg = CI.getReg();
-      const TargetRegisterClass *RC =
-        TRI->getMinimalPhysRegClass(Reg, MVT::i32);
+      const TargetRegisterClass *RC = TRI->getMinimalPhysRegClass(
+          Reg, Reg == RI->getReturnAddressReg(MF) ? MVT::i64 : MVT::i32);
 
       TII.loadRegFromStackSlot(RestoreBlock, I, Reg, CI.getFrameIdx(), RC, TRI);
       assert(I != RestoreBlock.begin() &&
              "loadRegFromStackSlot didn't insert any code!");
       // Insert in reverse order.  loadRegFromStackSlot can insert
       // multiple instructions.
 
       if (LIS) {
for (unsigned I = 0; CSRegs[I]; ++I) {
           TRI->getMinimalPhysRegClass(Reg, MVT::i32);
         int JunkFI = MFI.CreateStackObject(TRI->getSpillSize(*RC),
                                            TRI->getSpillAlign(*RC), true);
 
         CSI.push_back(CalleeSavedInfo(Reg, JunkFI));
       }
     }
 
+    // Add return address register, which is call clobbered, to the CSR list if
+    // we have calls.
+    const SIMachineFunctionInfo *FuncInfo = MF.getInfo<SIMachineFunctionInfo>();
+    Register Reg = TRI->getReturnAddressReg(MF);
+    if (!FuncInfo->isEntryFunction() && MFI.hasCalls()) {
+      const TargetRegisterClass *RC =
+          TRI->getMinimalPhysRegClass(Reg, MVT::i64);
+      int JunkFI = MFI.CreateStackObject(TRI->getSpillSize(*RC),
+                                         TRI->getSpillAlign(*RC), true);
+      CSI.push_back(CalleeSavedInfo(Reg, JunkFI));
+    }
+
     if (!CSI.empty()) {
       for (MachineBasicBlock *SaveBlock : SaveBlocks)
         insertCSRSaves(*SaveBlock, CSI, LIS);
 
       // Add live ins to save blocks.
       assert(SaveBlocks.size() == 1 && "shrink wrapping not fully implemented");
       updateLiveness(MF, CSI);
 

llvm/lib/Target/AMDGPU/SIRegisterInfo.cpp

for (MCPhysReg Reg : MFI->getAGPRSpillVGPRs())
     reserveRegisterTuples(Reserved, Reg);
 
   for (MCPhysReg Reg : MFI->getVGPRSpillAGPRs())
     reserveRegisterTuples(Reserved, Reg);
 
   for (auto SSpill : MFI->getSGPRSpillVGPRs())
     reserveRegisterTuples(Reserved, SSpill.VGPR);
 
+  // The return address register is call clobbered and the CFI needs to track
+  // its location. Hence it is handled specially.
+  if (!MFI->isEntryFunction())
+    reserveRegisterTuples(Reserved, getReturnAddressReg(MF));
+
   return Reserved;
 }
 
 bool SIRegisterInfo::shouldRealignStack(const MachineFunction &MF) const {
   const SIMachineFunctionInfo *Info = MF.getInfo<SIMachineFunctionInfo>();
   // On entry, the base address is 0, so it can't possibly need any more
   // alignment.
 

llvm/lib/Target/AMDGPU/SIRegisterInfo.td

 }
 
 def SGPR_64 : RegisterClass<"AMDGPU", [v2i32, i64, v2f32, f64, v4i16, v4f16], 32,
                             (add SGPR_64Regs)> {
   let CopyCost = 1;
   let AllocationPriority = 11;
 }
 
-// CCR (call clobbered registers) SGPR 64-bit registers
-def CCR_SGPR_64 : RegisterClass<"AMDGPU", SGPR_64.RegTypes, 32,
-                                (add (trunc SGPR_64, 16))> {
-  let CopyCost = SGPR_64.CopyCost;
-  let AllocationPriority = SGPR_64.AllocationPriority;
-}
-
-// Call clobbered 64-bit SGPRs for AMDGPU_Gfx CC
-def Gfx_CCR_SGPR_64 : RegisterClass<"AMDGPU", SGPR_64.RegTypes, 32,
-                                (add (trunc (shl SGPR_64, 15), 1), // s[30:31]
-                                     (trunc (shl SGPR_64, 18), 14))> { // s[36:37]-s[s62:63]
-  let CopyCost = SGPR_64.CopyCost;
-  let AllocationPriority = SGPR_64.AllocationPriority;
-}
-
 def TTMP_64 : RegisterClass<"AMDGPU", [v2i32, i64, f64, v4i16, v4f16], 32,
                             (add TTMP_64Regs)> {
   let isAllocatable = 0;
 }
 
 def SReg_64_XEXEC : RegisterClass<"AMDGPU", [v2i32, i64, v2f32, f64, i1, v4i16, v4f16], 32,
   (add SGPR_64, VCC, FLAT_SCR, XNACK_MASK, TTMP_64, TBA, TMA)> {
   let CopyCost = 1;

llvm/lib/Target/AMDGPU/SOPInstructions.td

 
 // no input, 64-bit output.
 class SOP1_64_0 <string opName, list<dag> pattern=[]> : SOP1_Pseudo <
   opName, (outs SReg_64:$sdst), (ins), "$sdst", pattern> {
   let has_src0 = 0;
 }
 
 // 64-bit input, no output
-class SOP1_1 <string opName, RegisterClass rc = SReg_64, list<dag> pattern=[]> : SOP1_Pseudo <
-  opName, (outs), (ins rc:$src0), "$src0", pattern> {
+class SOP1_1 <string opName, list<dag> pattern=[]> : SOP1_Pseudo <
+  opName, (outs), (ins SReg_64:$src0), "$src0", pattern> {
   let has_sdst = 0;
 }
 
 
 let isMoveImm = 1 in {
   let isReMaterializable = 1, isAsCheapAsAMove = 1 in {
     def S_MOV_B32 : SOP1_32 <"s_mov_b32">;
     def S_MOV_B64 : SOP1_64 <"s_mov_b64">;

 let isTerminator = 1, isBarrier = 1, SchedRW = [WriteBranch] in {
 
 let isBranch = 1, isIndirectBranch = 1 in {
 def S_SETPC_B64 : SOP1_1  <"s_setpc_b64">;
 } // End isBranch = 1, isIndirectBranch = 1
 
 let isReturn = 1 in {
 // Define variant marked as return rather than branch.
-def S_SETPC_B64_return : SOP1_1<"", CCR_SGPR_64, [(AMDGPUret_flag i64:$src0)]>;
-def S_SETPC_B64_return_gfx : SOP1_1<"", Gfx_CCR_SGPR_64, [(AMDGPUret_gfx_flag i64:$src0)]>;
+def S_SETPC_B64_return : SOP1_1<"">;
 }
 } // End isTerminator = 1, isBarrier = 1
 
 let isCall = 1 in {
 def S_SWAPPC_B64 : SOP1_64 <"s_swappc_b64"
 >;
 }
 

llvm/test/CodeGen/AMDGPU/GlobalISel/combine-fma-add-mul-post-legalize.mir

 # NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
 # RUN: llc -global-isel -march=amdgcn -mcpu=gfx900 -run-pass=amdgpu-postlegalizer-combiner %s -o - | FileCheck -check-prefix=GFX9 %s
 # RUN: llc -global-isel -march=amdgcn -mcpu=gfx900 -run-pass=amdgpu-postlegalizer-combiner -fp-contract=fast %s -o - | FileCheck -check-prefix=GFX9-CONTRACT %s
 # RUN: llc -global-isel -march=amdgcn -mcpu=gfx900 -run-pass=amdgpu-postlegalizer-combiner --denormal-fp-math=preserve-sign %s -o - | FileCheck -check-prefix=GFX9-DENORM %s
 # RUN: llc -global-isel -march=amdgcn -mcpu=gfx900 -run-pass=amdgpu-postlegalizer-combiner -enable-unsafe-fp-math %s -o - | FileCheck -check-prefix=GFX9-UNSAFE %s
 # RUN: llc -global-isel -march=amdgcn -mcpu=gfx1010 -run-pass=amdgpu-postlegalizer-combiner %s -o - | FileCheck -check-prefix=GFX10 %s
 # RUN: llc -global-isel -march=amdgcn -mcpu=gfx1010 -run-pass=amdgpu-postlegalizer-combiner -fp-contract=fast %s -o - | FileCheck -check-prefix=GFX10-CONTRACT %s
 # RUN: llc -global-isel -march=amdgcn -mcpu=gfx1010 -run-pass=amdgpu-postlegalizer-combiner --denormal-fp-math=preserve-sign %s -o - | FileCheck -check-prefix=GFX10-DENORM %s
 # RUN: llc -global-isel -march=amdgcn -mcpu=gfx1010 -run-pass=amdgpu-postlegalizer-combiner -enable-unsafe-fp-math %s -o - | FileCheck -check-prefix=GFX10-UNSAFE %s
 
 ---
 name:            test_f32_add_mul
 body:             |
   bb.1.entry:
-    liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
+    liveins: $vgpr0, $vgpr1, $vgpr2
 
     ; GFX9-LABEL: name: test_f32_add_mul
     ; GFX9: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX9: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9: [[FMUL:%[0-9]+]]:_(s32) = G_FMUL [[COPY]], [[COPY1]]
     ; GFX9: [[FADD:%[0-9]+]]:_(s32) = G_FADD [[FMUL]], [[COPY2]]
     ; GFX9: $vgpr0 = COPY [[FADD]](s32)
-    ; GFX9: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX9-CONTRACT-LABEL: name: test_f32_add_mul
     ; GFX9-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX9-CONTRACT: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-CONTRACT: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[COPY]], [[COPY1]], [[COPY2]]
     ; GFX9-CONTRACT: $vgpr0 = COPY [[FMA]](s32)
-    ; GFX9-CONTRACT: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9-CONTRACT: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX9-DENORM-LABEL: name: test_f32_add_mul
     ; GFX9-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX9-DENORM: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-DENORM: [[FMUL:%[0-9]+]]:_(s32) = G_FMUL [[COPY]], [[COPY1]]
     ; GFX9-DENORM: [[FADD:%[0-9]+]]:_(s32) = G_FADD [[FMUL]], [[COPY2]]
     ; GFX9-DENORM: $vgpr0 = COPY [[FADD]](s32)
-    ; GFX9-DENORM: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9-DENORM: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX9-UNSAFE-LABEL: name: test_f32_add_mul
     ; GFX9-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX9-UNSAFE: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-UNSAFE: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[COPY]], [[COPY1]], [[COPY2]]
     ; GFX9-UNSAFE: $vgpr0 = COPY [[FMA]](s32)
-    ; GFX9-UNSAFE: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9-UNSAFE: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-LABEL: name: test_f32_add_mul
     ; GFX10: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX10: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10: [[FMUL:%[0-9]+]]:_(s32) = G_FMUL [[COPY]], [[COPY1]]
     ; GFX10: [[FADD:%[0-9]+]]:_(s32) = G_FADD [[FMUL]], [[COPY2]]
     ; GFX10: $vgpr0 = COPY [[FADD]](s32)
-    ; GFX10: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-CONTRACT-LABEL: name: test_f32_add_mul
     ; GFX10-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX10-CONTRACT: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-CONTRACT: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[COPY]], [[COPY1]], [[COPY2]]
     ; GFX10-CONTRACT: $vgpr0 = COPY [[FMA]](s32)
-    ; GFX10-CONTRACT: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10-CONTRACT: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-DENORM-LABEL: name: test_f32_add_mul
     ; GFX10-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX10-DENORM: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-DENORM: [[FMUL:%[0-9]+]]:_(s32) = G_FMUL [[COPY]], [[COPY1]]
     ; GFX10-DENORM: [[FADD:%[0-9]+]]:_(s32) = G_FADD [[FMUL]], [[COPY2]]
     ; GFX10-DENORM: $vgpr0 = COPY [[FADD]](s32)
-    ; GFX10-DENORM: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10-DENORM: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-UNSAFE-LABEL: name: test_f32_add_mul
     ; GFX10-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX10-UNSAFE: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-UNSAFE: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[COPY]], [[COPY1]], [[COPY2]]
     ; GFX10-UNSAFE: $vgpr0 = COPY [[FMA]](s32)
-    ; GFX10-UNSAFE: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10-UNSAFE: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     %0:_(s32) = COPY $vgpr0
     %1:_(s32) = COPY $vgpr1
     %2:_(s32) = COPY $vgpr2
-    %3:sgpr_64 = COPY $sgpr30_sgpr31
     %4:_(s32) = G_FMUL %0, %1
     %5:_(s32) = G_FADD %4, %2
     $vgpr0 = COPY %5(s32)
-    %6:ccr_sgpr_64 = COPY %3
-    S_SETPC_B64_return %6, implicit $vgpr0
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
 ...
 
 ---
 name:            test_f32_add_mul_rhs
 body:             |
   bb.1.entry:
-    liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
+    liveins: $vgpr0, $vgpr1, $vgpr2
 
     ; GFX9-LABEL: name: test_f32_add_mul_rhs
     ; GFX9: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX9: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9: [[FMUL:%[0-9]+]]:_(s32) = G_FMUL [[COPY]], [[COPY1]]
     ; GFX9: [[FADD:%[0-9]+]]:_(s32) = G_FADD [[COPY2]], [[FMUL]]
     ; GFX9: $vgpr0 = COPY [[FADD]](s32)
-    ; GFX9: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX9-CONTRACT-LABEL: name: test_f32_add_mul_rhs
     ; GFX9-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX9-CONTRACT: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-CONTRACT: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[COPY]], [[COPY1]], [[COPY2]]
     ; GFX9-CONTRACT: $vgpr0 = COPY [[FMA]](s32)
-    ; GFX9-CONTRACT: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9-CONTRACT: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX9-DENORM-LABEL: name: test_f32_add_mul_rhs
     ; GFX9-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX9-DENORM: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-DENORM: [[FMUL:%[0-9]+]]:_(s32) = G_FMUL [[COPY]], [[COPY1]]
     ; GFX9-DENORM: [[FADD:%[0-9]+]]:_(s32) = G_FADD [[COPY2]], [[FMUL]]
     ; GFX9-DENORM: $vgpr0 = COPY [[FADD]](s32)
-    ; GFX9-DENORM: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9-DENORM: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX9-UNSAFE-LABEL: name: test_f32_add_mul_rhs
     ; GFX9-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX9-UNSAFE: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-UNSAFE: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[COPY]], [[COPY1]], [[COPY2]]
     ; GFX9-UNSAFE: $vgpr0 = COPY [[FMA]](s32)
-    ; GFX9-UNSAFE: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9-UNSAFE: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-LABEL: name: test_f32_add_mul_rhs
     ; GFX10: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX10: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10: [[FMUL:%[0-9]+]]:_(s32) = G_FMUL [[COPY]], [[COPY1]]
     ; GFX10: [[FADD:%[0-9]+]]:_(s32) = G_FADD [[COPY2]], [[FMUL]]
     ; GFX10: $vgpr0 = COPY [[FADD]](s32)
-    ; GFX10: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-CONTRACT-LABEL: name: test_f32_add_mul_rhs
     ; GFX10-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX10-CONTRACT: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-CONTRACT: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[COPY]], [[COPY1]], [[COPY2]]
     ; GFX10-CONTRACT: $vgpr0 = COPY [[FMA]](s32)
-    ; GFX10-CONTRACT: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10-CONTRACT: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-DENORM-LABEL: name: test_f32_add_mul_rhs
     ; GFX10-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX10-DENORM: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-DENORM: [[FMUL:%[0-9]+]]:_(s32) = G_FMUL [[COPY]], [[COPY1]]
     ; GFX10-DENORM: [[FADD:%[0-9]+]]:_(s32) = G_FADD [[COPY2]], [[FMUL]]
     ; GFX10-DENORM: $vgpr0 = COPY [[FADD]](s32)
-    ; GFX10-DENORM: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10-DENORM: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-UNSAFE-LABEL: name: test_f32_add_mul_rhs
     ; GFX10-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX10-UNSAFE: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-UNSAFE: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[COPY]], [[COPY1]], [[COPY2]]
     ; GFX10-UNSAFE: $vgpr0 = COPY [[FMA]](s32)
-    ; GFX10-UNSAFE: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10-UNSAFE: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     %0:_(s32) = COPY $vgpr0
     %1:_(s32) = COPY $vgpr1
     %2:_(s32) = COPY $vgpr2
-    %3:sgpr_64 = COPY $sgpr30_sgpr31
     %4:_(s32) = G_FMUL %0, %1
     %5:_(s32) = G_FADD %2, %4
     $vgpr0 = COPY %5(s32)
-    %6:ccr_sgpr_64 = COPY %3
-    S_SETPC_B64_return %6, implicit $vgpr0
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
 ...
 
 ---
 name:            test_half_add_mul
 body:             |
   bb.1.entry:
-    liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
+    liveins: $vgpr0, $vgpr1, $vgpr2
 
     ; GFX9-LABEL: name: test_half_add_mul
     ; GFX9: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX9: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX9: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX9: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9: [[FMUL:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC]], [[TRUNC1]]
     ; GFX9: [[FADD:%[0-9]+]]:_(s16) = G_FADD [[FMUL]], [[TRUNC2]]
     ; GFX9: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD]](s16)
     ; GFX9: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX9: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX9-CONTRACT-LABEL: name: test_half_add_mul
     ; GFX9-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-CONTRACT: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX9-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-CONTRACT: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX9-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-CONTRACT: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX9-CONTRACT: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-CONTRACT: [[FMA:%[0-9]+]]:_(s16) = G_FMA [[TRUNC]], [[TRUNC1]], [[TRUNC2]]
     ; GFX9-CONTRACT: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FMA]](s16)
     ; GFX9-CONTRACT: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX9-CONTRACT: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9-CONTRACT: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX9-DENORM-LABEL: name: test_half_add_mul
     ; GFX9-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-DENORM: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX9-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-DENORM: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX9-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-DENORM: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX9-DENORM: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-DENORM: [[FMUL:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC]], [[TRUNC1]]
     ; GFX9-DENORM: [[FADD:%[0-9]+]]:_(s16) = G_FADD [[FMUL]], [[TRUNC2]]
     ; GFX9-DENORM: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD]](s16)
     ; GFX9-DENORM: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX9-DENORM: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9-DENORM: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX9-UNSAFE-LABEL: name: test_half_add_mul
     ; GFX9-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-UNSAFE: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX9-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-UNSAFE: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX9-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-UNSAFE: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX9-UNSAFE: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-UNSAFE: [[FMA:%[0-9]+]]:_(s16) = G_FMA [[TRUNC]], [[TRUNC1]], [[TRUNC2]]
     ; GFX9-UNSAFE: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FMA]](s16)
     ; GFX9-UNSAFE: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX9-UNSAFE: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9-UNSAFE: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-LABEL: name: test_half_add_mul
     ; GFX10: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX10: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX10: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX10: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10: [[FMUL:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC]], [[TRUNC1]]
     ; GFX10: [[FADD:%[0-9]+]]:_(s16) = G_FADD [[FMUL]], [[TRUNC2]]
     ; GFX10: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD]](s16)
     ; GFX10: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX10: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-CONTRACT-LABEL: name: test_half_add_mul
     ; GFX10-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-CONTRACT: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX10-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-CONTRACT: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX10-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-CONTRACT: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX10-CONTRACT: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-CONTRACT: [[FMA:%[0-9]+]]:_(s16) = G_FMA [[TRUNC]], [[TRUNC1]], [[TRUNC2]]
     ; GFX10-CONTRACT: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FMA]](s16)
     ; GFX10-CONTRACT: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX10-CONTRACT: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10-CONTRACT: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-DENORM-LABEL: name: test_half_add_mul
     ; GFX10-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-DENORM: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX10-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-DENORM: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX10-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-DENORM: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX10-DENORM: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-DENORM: [[FMUL:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC]], [[TRUNC1]]
     ; GFX10-DENORM: [[FADD:%[0-9]+]]:_(s16) = G_FADD [[FMUL]], [[TRUNC2]]
     ; GFX10-DENORM: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD]](s16)
     ; GFX10-DENORM: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX10-DENORM: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10-DENORM: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-UNSAFE-LABEL: name: test_half_add_mul
     ; GFX10-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-UNSAFE: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX10-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-UNSAFE: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX10-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-UNSAFE: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX10-UNSAFE: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-UNSAFE: [[FMA:%[0-9]+]]:_(s16) = G_FMA [[TRUNC]], [[TRUNC1]], [[TRUNC2]]
     ; GFX10-UNSAFE: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FMA]](s16)
     ; GFX10-UNSAFE: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX10-UNSAFE: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10-UNSAFE: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     %4:_(s32) = COPY $vgpr0
     %0:_(s16) = G_TRUNC %4(s32)
     %5:_(s32) = COPY $vgpr1
     %1:_(s16) = G_TRUNC %5(s32)
     %6:_(s32) = COPY $vgpr2
     %2:_(s16) = G_TRUNC %6(s32)
-    %3:sgpr_64 = COPY $sgpr30_sgpr31
     %7:_(s16) = G_FMUL %0, %1
     %8:_(s16) = G_FADD %7, %2
     %10:_(s32) = G_ANYEXT %8(s16)
     $vgpr0 = COPY %10(s32)
-    %9:ccr_sgpr_64 = COPY %3
-    S_SETPC_B64_return %9, implicit $vgpr0
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
 ...
 
 ---
 name:            test_half_add_mul_rhs
 body:             |
   bb.1.entry:
-    liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
+    liveins: $vgpr0, $vgpr1, $vgpr2
 
     ; GFX9-LABEL: name: test_half_add_mul_rhs
     ; GFX9: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX9: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX9: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX9: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9: [[FMUL:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC]], [[TRUNC1]]
     ; GFX9: [[FADD:%[0-9]+]]:_(s16) = G_FADD [[TRUNC2]], [[FMUL]]
     ; GFX9: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD]](s16)
     ; GFX9: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX9: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX9-CONTRACT-LABEL: name: test_half_add_mul_rhs
     ; GFX9-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-CONTRACT: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX9-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-CONTRACT: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX9-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-CONTRACT: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX9-CONTRACT: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-CONTRACT: [[FMA:%[0-9]+]]:_(s16) = G_FMA [[TRUNC]], [[TRUNC1]], [[TRUNC2]]
     ; GFX9-CONTRACT: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FMA]](s16)
     ; GFX9-CONTRACT: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX9-CONTRACT: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9-CONTRACT: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX9-DENORM-LABEL: name: test_half_add_mul_rhs
     ; GFX9-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-DENORM: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX9-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-DENORM: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX9-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-DENORM: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX9-DENORM: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-DENORM: [[FMUL:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC]], [[TRUNC1]]
     ; GFX9-DENORM: [[FADD:%[0-9]+]]:_(s16) = G_FADD [[TRUNC2]], [[FMUL]]
     ; GFX9-DENORM: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD]](s16)
     ; GFX9-DENORM: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX9-DENORM: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9-DENORM: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX9-UNSAFE-LABEL: name: test_half_add_mul_rhs
     ; GFX9-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-UNSAFE: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX9-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-UNSAFE: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX9-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-UNSAFE: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX9-UNSAFE: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-UNSAFE: [[FMA:%[0-9]+]]:_(s16) = G_FMA [[TRUNC]], [[TRUNC1]], [[TRUNC2]]
     ; GFX9-UNSAFE: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FMA]](s16)
     ; GFX9-UNSAFE: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX9-UNSAFE: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9-UNSAFE: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-LABEL: name: test_half_add_mul_rhs
     ; GFX10: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX10: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX10: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX10: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10: [[FMUL:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC]], [[TRUNC1]]
     ; GFX10: [[FADD:%[0-9]+]]:_(s16) = G_FADD [[TRUNC2]], [[FMUL]]
     ; GFX10: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD]](s16)
     ; GFX10: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX10: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-CONTRACT-LABEL: name: test_half_add_mul_rhs
     ; GFX10-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-CONTRACT: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX10-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-CONTRACT: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX10-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-CONTRACT: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX10-CONTRACT: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-CONTRACT: [[FMA:%[0-9]+]]:_(s16) = G_FMA [[TRUNC]], [[TRUNC1]], [[TRUNC2]]
     ; GFX10-CONTRACT: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FMA]](s16)
     ; GFX10-CONTRACT: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX10-CONTRACT: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10-CONTRACT: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-DENORM-LABEL: name: test_half_add_mul_rhs
     ; GFX10-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-DENORM: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX10-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-DENORM: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX10-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-DENORM: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX10-DENORM: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-DENORM: [[FMUL:%[0-9]+]]:_(s16) = G_FMUL [[TRUNC]], [[TRUNC1]]
     ; GFX10-DENORM: [[FADD:%[0-9]+]]:_(s16) = G_FADD [[TRUNC2]], [[FMUL]]
     ; GFX10-DENORM: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD]](s16)
     ; GFX10-DENORM: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX10-DENORM: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10-DENORM: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-UNSAFE-LABEL: name: test_half_add_mul_rhs
     ; GFX10-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-UNSAFE: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX10-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-UNSAFE: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX10-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-UNSAFE: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX10-UNSAFE: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-UNSAFE: [[FMA:%[0-9]+]]:_(s16) = G_FMA [[TRUNC]], [[TRUNC1]], [[TRUNC2]]
     ; GFX10-UNSAFE: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FMA]](s16)
     ; GFX10-UNSAFE: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX10-UNSAFE: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10-UNSAFE: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     %4:_(s32) = COPY $vgpr0
     %0:_(s16) = G_TRUNC %4(s32)
     %5:_(s32) = COPY $vgpr1
     %1:_(s16) = G_TRUNC %5(s32)
     %6:_(s32) = COPY $vgpr2
     %2:_(s16) = G_TRUNC %6(s32)
-    %3:sgpr_64 = COPY $sgpr30_sgpr31
     %7:_(s16) = G_FMUL %0, %1
     %8:_(s16) = G_FADD %2, %7
     %10:_(s32) = G_ANYEXT %8(s16)
     $vgpr0 = COPY %10(s32)
-    %9:ccr_sgpr_64 = COPY %3
-    S_SETPC_B64_return %9, implicit $vgpr0
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
 ...
 
 ---
 name:            test_double_add_mul
 body:             |
   bb.1.entry:
-    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $sgpr30_sgpr31
+    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
 
     ; GFX9-LABEL: name: test_double_add_mul
     ; GFX9: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX9: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX9: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX9: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9: [[FMUL:%[0-9]+]]:_(s64) = G_FMUL [[MV]], [[MV1]]
     ; GFX9: [[FADD:%[0-9]+]]:_(s64) = G_FADD [[FMUL]], [[MV2]]
     ; GFX9: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](s64)
     ; GFX9: $vgpr0 = COPY [[UV]](s32)
     ; GFX9: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX9: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX9-CONTRACT-LABEL: name: test_double_add_mul
     ; GFX9-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-CONTRACT: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX9-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-CONTRACT: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-CONTRACT: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX9-CONTRACT: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-CONTRACT: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-CONTRACT: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX9-CONTRACT: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-CONTRACT: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[MV]], [[MV1]], [[MV2]]
     ; GFX9-CONTRACT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](s64)
     ; GFX9-CONTRACT: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-CONTRACT: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX9-CONTRACT: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9-CONTRACT: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX9-DENORM-LABEL: name: test_double_add_mul
     ; GFX9-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-DENORM: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX9-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-DENORM: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-DENORM: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX9-DENORM: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-DENORM: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-DENORM: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX9-DENORM: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-DENORM: [[FMUL:%[0-9]+]]:_(s64) = G_FMUL [[MV]], [[MV1]]
     ; GFX9-DENORM: [[FADD:%[0-9]+]]:_(s64) = G_FADD [[FMUL]], [[MV2]]
     ; GFX9-DENORM: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](s64)
     ; GFX9-DENORM: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-DENORM: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX9-DENORM: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9-DENORM: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX9-UNSAFE-LABEL: name: test_double_add_mul
     ; GFX9-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-UNSAFE: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX9-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-UNSAFE: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-UNSAFE: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX9-UNSAFE: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-UNSAFE: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-UNSAFE: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX9-UNSAFE: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-UNSAFE: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[MV]], [[MV1]], [[MV2]]
     ; GFX9-UNSAFE: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](s64)
     ; GFX9-UNSAFE: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-UNSAFE: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX9-UNSAFE: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9-UNSAFE: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-LABEL: name: test_double_add_mul
     ; GFX10: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX10: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX10: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX10: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10: [[FMUL:%[0-9]+]]:_(s64) = G_FMUL [[MV]], [[MV1]]
     ; GFX10: [[FADD:%[0-9]+]]:_(s64) = G_FADD [[FMUL]], [[MV2]]
     ; GFX10: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](s64)
     ; GFX10: $vgpr0 = COPY [[UV]](s32)
     ; GFX10: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX10: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-CONTRACT-LABEL: name: test_double_add_mul
     ; GFX10-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-CONTRACT: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX10-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-CONTRACT: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-CONTRACT: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX10-CONTRACT: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-CONTRACT: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-CONTRACT: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX10-CONTRACT: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-CONTRACT: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[MV]], [[MV1]], [[MV2]]
     ; GFX10-CONTRACT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](s64)
     ; GFX10-CONTRACT: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-CONTRACT: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX10-CONTRACT: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10-CONTRACT: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-DENORM-LABEL: name: test_double_add_mul
     ; GFX10-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-DENORM: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX10-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-DENORM: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-DENORM: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX10-DENORM: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-DENORM: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-DENORM: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX10-DENORM: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-DENORM: [[FMUL:%[0-9]+]]:_(s64) = G_FMUL [[MV]], [[MV1]]
     ; GFX10-DENORM: [[FADD:%[0-9]+]]:_(s64) = G_FADD [[FMUL]], [[MV2]]
     ; GFX10-DENORM: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](s64)
     ; GFX10-DENORM: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-DENORM: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX10-DENORM: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10-DENORM: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-UNSAFE-LABEL: name: test_double_add_mul
     ; GFX10-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-UNSAFE: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX10-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-UNSAFE: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-UNSAFE: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX10-UNSAFE: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-UNSAFE: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-UNSAFE: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX10-UNSAFE: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-UNSAFE: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[MV]], [[MV1]], [[MV2]]
     ; GFX10-UNSAFE: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](s64)
     ; GFX10-UNSAFE: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-UNSAFE: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX10-UNSAFE: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10-UNSAFE: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     %4:_(s32) = COPY $vgpr0
     %5:_(s32) = COPY $vgpr1
     %0:_(s64) = G_MERGE_VALUES %4(s32), %5(s32)
     %6:_(s32) = COPY $vgpr2
     %7:_(s32) = COPY $vgpr3
     %1:_(s64) = G_MERGE_VALUES %6(s32), %7(s32)
     %8:_(s32) = COPY $vgpr4
     %9:_(s32) = COPY $vgpr5
     %2:_(s64) = G_MERGE_VALUES %8(s32), %9(s32)
-    %3:sgpr_64 = COPY $sgpr30_sgpr31
     %10:_(s64) = G_FMUL %0, %1
     %11:_(s64) = G_FADD %10, %2
     %13:_(s32), %14:_(s32) = G_UNMERGE_VALUES %11(s64)
     $vgpr0 = COPY %13(s32)
     $vgpr1 = COPY %14(s32)
-    %12:ccr_sgpr_64 = COPY %3
-    S_SETPC_B64_return %12, implicit $vgpr0, implicit $vgpr1
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
 ...
 
 ---
 name:            test_double_add_mul_rhs
 body:             |
   bb.1.entry:
-    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $sgpr30_sgpr31
+    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
 
     ; GFX9-LABEL: name: test_double_add_mul_rhs
     ; GFX9: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX9: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX9: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX9: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9: [[FMUL:%[0-9]+]]:_(s64) = G_FMUL [[MV]], [[MV1]]
     ; GFX9: [[FADD:%[0-9]+]]:_(s64) = G_FADD [[MV2]], [[FMUL]]
     ; GFX9: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](s64)
     ; GFX9: $vgpr0 = COPY [[UV]](s32)
     ; GFX9: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX9: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX9-CONTRACT-LABEL: name: test_double_add_mul_rhs
     ; GFX9-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-CONTRACT: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX9-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-CONTRACT: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-CONTRACT: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX9-CONTRACT: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-CONTRACT: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-CONTRACT: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX9-CONTRACT: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-CONTRACT: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[MV]], [[MV1]], [[MV2]]
     ; GFX9-CONTRACT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](s64)
     ; GFX9-CONTRACT: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-CONTRACT: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX9-CONTRACT: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9-CONTRACT: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX9-DENORM-LABEL: name: test_double_add_mul_rhs
     ; GFX9-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-DENORM: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX9-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-DENORM: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-DENORM: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX9-DENORM: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-DENORM: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-DENORM: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX9-DENORM: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-DENORM: [[FMUL:%[0-9]+]]:_(s64) = G_FMUL [[MV]], [[MV1]]
     ; GFX9-DENORM: [[FADD:%[0-9]+]]:_(s64) = G_FADD [[MV2]], [[FMUL]]
     ; GFX9-DENORM: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](s64)
     ; GFX9-DENORM: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-DENORM: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX9-DENORM: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9-DENORM: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX9-UNSAFE-LABEL: name: test_double_add_mul_rhs
     ; GFX9-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-UNSAFE: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX9-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-UNSAFE: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-UNSAFE: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX9-UNSAFE: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-UNSAFE: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-UNSAFE: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX9-UNSAFE: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-UNSAFE: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[MV]], [[MV1]], [[MV2]]
     ; GFX9-UNSAFE: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](s64)
     ; GFX9-UNSAFE: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-UNSAFE: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX9-UNSAFE: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9-UNSAFE: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-LABEL: name: test_double_add_mul_rhs
     ; GFX10: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX10: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX10: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX10: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10: [[FMUL:%[0-9]+]]:_(s64) = G_FMUL [[MV]], [[MV1]]
     ; GFX10: [[FADD:%[0-9]+]]:_(s64) = G_FADD [[MV2]], [[FMUL]]
     ; GFX10: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](s64)
     ; GFX10: $vgpr0 = COPY [[UV]](s32)
     ; GFX10: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX10: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-CONTRACT-LABEL: name: test_double_add_mul_rhs
     ; GFX10-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-CONTRACT: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX10-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-CONTRACT: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-CONTRACT: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX10-CONTRACT: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-CONTRACT: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-CONTRACT: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX10-CONTRACT: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-CONTRACT: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[MV]], [[MV1]], [[MV2]]
     ; GFX10-CONTRACT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](s64)
     ; GFX10-CONTRACT: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-CONTRACT: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX10-CONTRACT: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10-CONTRACT: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-DENORM-LABEL: name: test_double_add_mul_rhs
     ; GFX10-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-DENORM: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX10-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-DENORM: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-DENORM: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX10-DENORM: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-DENORM: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-DENORM: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX10-DENORM: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-DENORM: [[FMUL:%[0-9]+]]:_(s64) = G_FMUL [[MV]], [[MV1]]
     ; GFX10-DENORM: [[FADD:%[0-9]+]]:_(s64) = G_FADD [[MV2]], [[FMUL]]
     ; GFX10-DENORM: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](s64)
     ; GFX10-DENORM: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-DENORM: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX10-DENORM: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10-DENORM: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-UNSAFE-LABEL: name: test_double_add_mul_rhs
     ; GFX10-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-UNSAFE: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX10-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-UNSAFE: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-UNSAFE: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX10-UNSAFE: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-UNSAFE: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-UNSAFE: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX10-UNSAFE: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-UNSAFE: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[MV]], [[MV1]], [[MV2]]
     ; GFX10-UNSAFE: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](s64)
     ; GFX10-UNSAFE: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-UNSAFE: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX10-UNSAFE: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10-UNSAFE: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     %4:_(s32) = COPY $vgpr0
     %5:_(s32) = COPY $vgpr1
     %0:_(s64) = G_MERGE_VALUES %4(s32), %5(s32)
     %6:_(s32) = COPY $vgpr2
     %7:_(s32) = COPY $vgpr3
     %1:_(s64) = G_MERGE_VALUES %6(s32), %7(s32)
     %8:_(s32) = COPY $vgpr4
     %9:_(s32) = COPY $vgpr5
     %2:_(s64) = G_MERGE_VALUES %8(s32), %9(s32)
-    %3:sgpr_64 = COPY $sgpr30_sgpr31
     %10:_(s64) = G_FMUL %0, %1
     %11:_(s64) = G_FADD %2, %10
     %13:_(s32), %14:_(s32) = G_UNMERGE_VALUES %11(s64)
     $vgpr0 = COPY %13(s32)
     $vgpr1 = COPY %14(s32)
-    %12:ccr_sgpr_64 = COPY %3
-    S_SETPC_B64_return %12, implicit $vgpr0, implicit $vgpr1
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
 ...
 
 ---
 name:            test_4xfloat_add_mul
 body:             |
   bb.1.entry:
-    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7, $vgpr8, $vgpr9, $vgpr10, $vgpr11, $sgpr30_sgpr31
+    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7, $vgpr8, $vgpr9, $vgpr10
 
     ; GFX9-LABEL: name: test_4xfloat_add_mul
     ; GFX9: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9: [[BUILD_VECTOR:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
     ; GFX9: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX9: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX9: [[BUILD_VECTOR1:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY4]](s32), [[COPY5]](s32), [[COPY6]](s32), [[COPY7]](s32)
     ; GFX9: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX9: [[COPY9:%[0-9]+]]:_(s32) = COPY $vgpr9
     ; GFX9: [[COPY10:%[0-9]+]]:_(s32) = COPY $vgpr10
     ; GFX9: [[COPY11:%[0-9]+]]:_(s32) = COPY $vgpr11
     ; GFX9: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY8]](s32), [[COPY9]](s32), [[COPY10]](s32), [[COPY11]](s32)
-    ; GFX9: [[COPY12:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9: [[FMUL:%[0-9]+]]:_(<4 x s32>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX9: [[FADD:%[0-9]+]]:_(<4 x s32>) = G_FADD [[FMUL]], [[BUILD_VECTOR2]]
     ; GFX9: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<4 x s32>)
     ; GFX9: $vgpr0 = COPY [[UV]](s32)
     ; GFX9: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9: $vgpr2 = COPY [[UV2]](s32)
     ; GFX9: $vgpr3 = COPY [[UV3]](s32)
-    ; GFX9: [[COPY13:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY12]]
-    ; GFX9: S_SETPC_B64_return [[COPY13]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+    ; GFX9: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
     ; GFX9-CONTRACT-LABEL: name: test_4xfloat_add_mul
     ; GFX9-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-CONTRACT: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-CONTRACT: [[BUILD_VECTOR:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
     ; GFX9-CONTRACT: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-CONTRACT: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-CONTRACT: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX9-CONTRACT: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX9-CONTRACT: [[BUILD_VECTOR1:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY4]](s32), [[COPY5]](s32), [[COPY6]](s32), [[COPY7]](s32)
     ; GFX9-CONTRACT: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX9-CONTRACT: [[COPY9:%[0-9]+]]:_(s32) = COPY $vgpr9
     ; GFX9-CONTRACT: [[COPY10:%[0-9]+]]:_(s32) = COPY $vgpr10
     ; GFX9-CONTRACT: [[COPY11:%[0-9]+]]:_(s32) = COPY $vgpr11
     ; GFX9-CONTRACT: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY8]](s32), [[COPY9]](s32), [[COPY10]](s32), [[COPY11]](s32)
-    ; GFX9-CONTRACT: [[COPY12:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-CONTRACT: [[FMUL:%[0-9]+]]:_(<4 x s32>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX9-CONTRACT: [[FADD:%[0-9]+]]:_(<4 x s32>) = G_FADD [[FMUL]], [[BUILD_VECTOR2]]
     ; GFX9-CONTRACT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<4 x s32>)
     ; GFX9-CONTRACT: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-CONTRACT: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9-CONTRACT: $vgpr2 = COPY [[UV2]](s32)
     ; GFX9-CONTRACT: $vgpr3 = COPY [[UV3]](s32)
-    ; GFX9-CONTRACT: [[COPY13:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY12]]
-    ; GFX9-CONTRACT: S_SETPC_B64_return [[COPY13]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+    ; GFX9-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
     ; GFX9-DENORM-LABEL: name: test_4xfloat_add_mul
     ; GFX9-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-DENORM: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-DENORM: [[BUILD_VECTOR:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
     ; GFX9-DENORM: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-DENORM: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-DENORM: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX9-DENORM: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX9-DENORM: [[BUILD_VECTOR1:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY4]](s32), [[COPY5]](s32), [[COPY6]](s32), [[COPY7]](s32)
     ; GFX9-DENORM: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX9-DENORM: [[COPY9:%[0-9]+]]:_(s32) = COPY $vgpr9
     ; GFX9-DENORM: [[COPY10:%[0-9]+]]:_(s32) = COPY $vgpr10
     ; GFX9-DENORM: [[COPY11:%[0-9]+]]:_(s32) = COPY $vgpr11
     ; GFX9-DENORM: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY8]](s32), [[COPY9]](s32), [[COPY10]](s32), [[COPY11]](s32)
-    ; GFX9-DENORM: [[COPY12:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-DENORM: [[FMUL:%[0-9]+]]:_(<4 x s32>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX9-DENORM: [[FADD:%[0-9]+]]:_(<4 x s32>) = G_FADD [[FMUL]], [[BUILD_VECTOR2]]
     ; GFX9-DENORM: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<4 x s32>)
     ; GFX9-DENORM: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-DENORM: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9-DENORM: $vgpr2 = COPY [[UV2]](s32)
     ; GFX9-DENORM: $vgpr3 = COPY [[UV3]](s32)
-    ; GFX9-DENORM: [[COPY13:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY12]]
-    ; GFX9-DENORM: S_SETPC_B64_return [[COPY13]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+    ; GFX9-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
     ; GFX9-UNSAFE-LABEL: name: test_4xfloat_add_mul
     ; GFX9-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-UNSAFE: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-UNSAFE: [[BUILD_VECTOR:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
     ; GFX9-UNSAFE: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-UNSAFE: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-UNSAFE: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX9-UNSAFE: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX9-UNSAFE: [[BUILD_VECTOR1:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY4]](s32), [[COPY5]](s32), [[COPY6]](s32), [[COPY7]](s32)
     ; GFX9-UNSAFE: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX9-UNSAFE: [[COPY9:%[0-9]+]]:_(s32) = COPY $vgpr9
     ; GFX9-UNSAFE: [[COPY10:%[0-9]+]]:_(s32) = COPY $vgpr10
     ; GFX9-UNSAFE: [[COPY11:%[0-9]+]]:_(s32) = COPY $vgpr11
     ; GFX9-UNSAFE: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY8]](s32), [[COPY9]](s32), [[COPY10]](s32), [[COPY11]](s32)
-    ; GFX9-UNSAFE: [[COPY12:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-UNSAFE: [[FMUL:%[0-9]+]]:_(<4 x s32>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX9-UNSAFE: [[FADD:%[0-9]+]]:_(<4 x s32>) = G_FADD [[FMUL]], [[BUILD_VECTOR2]]
     ; GFX9-UNSAFE: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<4 x s32>)
     ; GFX9-UNSAFE: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-UNSAFE: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9-UNSAFE: $vgpr2 = COPY [[UV2]](s32)
     ; GFX9-UNSAFE: $vgpr3 = COPY [[UV3]](s32)
-    ; GFX9-UNSAFE: [[COPY13:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY12]]
-    ; GFX9-UNSAFE: S_SETPC_B64_return [[COPY13]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+    ; GFX9-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
     ; GFX10-LABEL: name: test_4xfloat_add_mul
     ; GFX10: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10: [[BUILD_VECTOR:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
     ; GFX10: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX10: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX10: [[BUILD_VECTOR1:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY4]](s32), [[COPY5]](s32), [[COPY6]](s32), [[COPY7]](s32)
     ; GFX10: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX10: [[COPY9:%[0-9]+]]:_(s32) = COPY $vgpr9
     ; GFX10: [[COPY10:%[0-9]+]]:_(s32) = COPY $vgpr10
     ; GFX10: [[COPY11:%[0-9]+]]:_(s32) = COPY $vgpr11
     ; GFX10: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY8]](s32), [[COPY9]](s32), [[COPY10]](s32), [[COPY11]](s32)
-    ; GFX10: [[COPY12:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10: [[FMUL:%[0-9]+]]:_(<4 x s32>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX10: [[FADD:%[0-9]+]]:_(<4 x s32>) = G_FADD [[FMUL]], [[BUILD_VECTOR2]]
     ; GFX10: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<4 x s32>)
     ; GFX10: $vgpr0 = COPY [[UV]](s32)
     ; GFX10: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10: $vgpr2 = COPY [[UV2]](s32)
     ; GFX10: $vgpr3 = COPY [[UV3]](s32)
-    ; GFX10: [[COPY13:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY12]]
-    ; GFX10: S_SETPC_B64_return [[COPY13]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+    ; GFX10: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
     ; GFX10-CONTRACT-LABEL: name: test_4xfloat_add_mul
     ; GFX10-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-CONTRACT: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-CONTRACT: [[BUILD_VECTOR:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
     ; GFX10-CONTRACT: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-CONTRACT: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-CONTRACT: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX10-CONTRACT: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX10-CONTRACT: [[BUILD_VECTOR1:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY4]](s32), [[COPY5]](s32), [[COPY6]](s32), [[COPY7]](s32)
     ; GFX10-CONTRACT: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX10-CONTRACT: [[COPY9:%[0-9]+]]:_(s32) = COPY $vgpr9
     ; GFX10-CONTRACT: [[COPY10:%[0-9]+]]:_(s32) = COPY $vgpr10
     ; GFX10-CONTRACT: [[COPY11:%[0-9]+]]:_(s32) = COPY $vgpr11
     ; GFX10-CONTRACT: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY8]](s32), [[COPY9]](s32), [[COPY10]](s32), [[COPY11]](s32)
-    ; GFX10-CONTRACT: [[COPY12:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-CONTRACT: [[FMUL:%[0-9]+]]:_(<4 x s32>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX10-CONTRACT: [[FADD:%[0-9]+]]:_(<4 x s32>) = G_FADD [[FMUL]], [[BUILD_VECTOR2]]
     ; GFX10-CONTRACT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<4 x s32>)
     ; GFX10-CONTRACT: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-CONTRACT: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10-CONTRACT: $vgpr2 = COPY [[UV2]](s32)
     ; GFX10-CONTRACT: $vgpr3 = COPY [[UV3]](s32)
-    ; GFX10-CONTRACT: [[COPY13:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY12]]
-    ; GFX10-CONTRACT: S_SETPC_B64_return [[COPY13]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+    ; GFX10-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
     ; GFX10-DENORM-LABEL: name: test_4xfloat_add_mul
     ; GFX10-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-DENORM: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-DENORM: [[BUILD_VECTOR:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
     ; GFX10-DENORM: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-DENORM: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-DENORM: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX10-DENORM: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX10-DENORM: [[BUILD_VECTOR1:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY4]](s32), [[COPY5]](s32), [[COPY6]](s32), [[COPY7]](s32)
     ; GFX10-DENORM: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX10-DENORM: [[COPY9:%[0-9]+]]:_(s32) = COPY $vgpr9
     ; GFX10-DENORM: [[COPY10:%[0-9]+]]:_(s32) = COPY $vgpr10
     ; GFX10-DENORM: [[COPY11:%[0-9]+]]:_(s32) = COPY $vgpr11
     ; GFX10-DENORM: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY8]](s32), [[COPY9]](s32), [[COPY10]](s32), [[COPY11]](s32)
-    ; GFX10-DENORM: [[COPY12:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-DENORM: [[FMUL:%[0-9]+]]:_(<4 x s32>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX10-DENORM: [[FADD:%[0-9]+]]:_(<4 x s32>) = G_FADD [[FMUL]], [[BUILD_VECTOR2]]
     ; GFX10-DENORM: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<4 x s32>)
     ; GFX10-DENORM: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-DENORM: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10-DENORM: $vgpr2 = COPY [[UV2]](s32)
     ; GFX10-DENORM: $vgpr3 = COPY [[UV3]](s32)
-    ; GFX10-DENORM: [[COPY13:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY12]]
-    ; GFX10-DENORM: S_SETPC_B64_return [[COPY13]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+    ; GFX10-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
     ; GFX10-UNSAFE-LABEL: name: test_4xfloat_add_mul
     ; GFX10-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-UNSAFE: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-UNSAFE: [[BUILD_VECTOR:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
     ; GFX10-UNSAFE: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-UNSAFE: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-UNSAFE: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX10-UNSAFE: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX10-UNSAFE: [[BUILD_VECTOR1:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY4]](s32), [[COPY5]](s32), [[COPY6]](s32), [[COPY7]](s32)
     ; GFX10-UNSAFE: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX10-UNSAFE: [[COPY9:%[0-9]+]]:_(s32) = COPY $vgpr9
     ; GFX10-UNSAFE: [[COPY10:%[0-9]+]]:_(s32) = COPY $vgpr10
     ; GFX10-UNSAFE: [[COPY11:%[0-9]+]]:_(s32) = COPY $vgpr11
     ; GFX10-UNSAFE: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY8]](s32), [[COPY9]](s32), [[COPY10]](s32), [[COPY11]](s32)
-    ; GFX10-UNSAFE: [[COPY12:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-UNSAFE: [[FMUL:%[0-9]+]]:_(<4 x s32>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX10-UNSAFE: [[FADD:%[0-9]+]]:_(<4 x s32>) = G_FADD [[FMUL]], [[BUILD_VECTOR2]]
     ; GFX10-UNSAFE: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<4 x s32>)
     ; GFX10-UNSAFE: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-UNSAFE: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10-UNSAFE: $vgpr2 = COPY [[UV2]](s32)
     ; GFX10-UNSAFE: $vgpr3 = COPY [[UV3]](s32)
-    ; GFX10-UNSAFE: [[COPY13:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY12]]
-    ; GFX10-UNSAFE: S_SETPC_B64_return [[COPY13]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+    ; GFX10-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
     %4:_(s32) = COPY $vgpr0
     %5:_(s32) = COPY $vgpr1
     %6:_(s32) = COPY $vgpr2
     %7:_(s32) = COPY $vgpr3
     %0:_(<4 x s32>) = G_BUILD_VECTOR %4(s32), %5(s32), %6(s32), %7(s32)
     %8:_(s32) = COPY $vgpr4
     %9:_(s32) = COPY $vgpr5
     %10:_(s32) = COPY $vgpr6
     %11:_(s32) = COPY $vgpr7
     %1:_(<4 x s32>) = G_BUILD_VECTOR %8(s32), %9(s32), %10(s32), %11(s32)
     %12:_(s32) = COPY $vgpr8
     %13:_(s32) = COPY $vgpr9
     %14:_(s32) = COPY $vgpr10
     %15:_(s32) = COPY $vgpr11
     %2:_(<4 x s32>) = G_BUILD_VECTOR %12(s32), %13(s32), %14(s32), %15(s32)
-    %3:sgpr_64 = COPY $sgpr30_sgpr31
     %16:_(<4 x s32>) = G_FMUL %0, %1
     %17:_(<4 x s32>) = G_FADD %16, %2
     %19:_(s32), %20:_(s32), %21:_(s32), %22:_(s32) = G_UNMERGE_VALUES %17(<4 x s32>)
     $vgpr0 = COPY %19(s32)
     $vgpr1 = COPY %20(s32)
     $vgpr2 = COPY %21(s32)
     $vgpr3 = COPY %22(s32)
-    %18:ccr_sgpr_64 = COPY %3
-    S_SETPC_B64_return %18, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
 ...
 
 ---
 name:            test_3xfloat_add_mul_rhs
 body:             |
   bb.1.entry:
-    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7, $vgpr8, $sgpr30_sgpr31
+    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7, $vgpr8
 
     ; GFX9-LABEL: name: test_3xfloat_add_mul_rhs
     ; GFX9: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9: [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32)
     ; GFX9: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9: [[BUILD_VECTOR1:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY3]](s32), [[COPY4]](s32), [[COPY5]](s32)
     ; GFX9: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX9: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX9: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX9: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY6]](s32), [[COPY7]](s32), [[COPY8]](s32)
-    ; GFX9: [[COPY9:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9: [[FMUL:%[0-9]+]]:_(<3 x s32>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX9: [[FADD:%[0-9]+]]:_(<3 x s32>) = G_FADD [[BUILD_VECTOR2]], [[FMUL]]
     ; GFX9: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<3 x s32>)
     ; GFX9: $vgpr0 = COPY [[UV]](s32)
     ; GFX9: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9: $vgpr2 = COPY [[UV2]](s32)
-    ; GFX9: [[COPY10:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY9]]
-    ; GFX9: S_SETPC_B64_return [[COPY10]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
+    ; GFX9: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
     ; GFX9-CONTRACT-LABEL: name: test_3xfloat_add_mul_rhs
     ; GFX9-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-CONTRACT: [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32)
     ; GFX9-CONTRACT: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-CONTRACT: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-CONTRACT: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-CONTRACT: [[BUILD_VECTOR1:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY3]](s32), [[COPY4]](s32), [[COPY5]](s32)
     ; GFX9-CONTRACT: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX9-CONTRACT: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX9-CONTRACT: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX9-CONTRACT: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY6]](s32), [[COPY7]](s32), [[COPY8]](s32)
-    ; GFX9-CONTRACT: [[COPY9:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-CONTRACT: [[FMUL:%[0-9]+]]:_(<3 x s32>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX9-CONTRACT: [[FADD:%[0-9]+]]:_(<3 x s32>) = G_FADD [[BUILD_VECTOR2]], [[FMUL]]
     ; GFX9-CONTRACT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<3 x s32>)
     ; GFX9-CONTRACT: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-CONTRACT: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9-CONTRACT: $vgpr2 = COPY [[UV2]](s32)
-    ; GFX9-CONTRACT: [[COPY10:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY9]]
-    ; GFX9-CONTRACT: S_SETPC_B64_return [[COPY10]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
+    ; GFX9-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
     ; GFX9-DENORM-LABEL: name: test_3xfloat_add_mul_rhs
     ; GFX9-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-DENORM: [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32)
     ; GFX9-DENORM: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-DENORM: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-DENORM: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-DENORM: [[BUILD_VECTOR1:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY3]](s32), [[COPY4]](s32), [[COPY5]](s32)
     ; GFX9-DENORM: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX9-DENORM: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX9-DENORM: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX9-DENORM: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY6]](s32), [[COPY7]](s32), [[COPY8]](s32)
-    ; GFX9-DENORM: [[COPY9:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-DENORM: [[FMUL:%[0-9]+]]:_(<3 x s32>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX9-DENORM: [[FADD:%[0-9]+]]:_(<3 x s32>) = G_FADD [[BUILD_VECTOR2]], [[FMUL]]
     ; GFX9-DENORM: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<3 x s32>)
     ; GFX9-DENORM: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-DENORM: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9-DENORM: $vgpr2 = COPY [[UV2]](s32)
-    ; GFX9-DENORM: [[COPY10:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY9]]
-    ; GFX9-DENORM: S_SETPC_B64_return [[COPY10]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
+    ; GFX9-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
     ; GFX9-UNSAFE-LABEL: name: test_3xfloat_add_mul_rhs
     ; GFX9-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-UNSAFE: [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32)
     ; GFX9-UNSAFE: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-UNSAFE: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-UNSAFE: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-UNSAFE: [[BUILD_VECTOR1:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY3]](s32), [[COPY4]](s32), [[COPY5]](s32)
     ; GFX9-UNSAFE: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX9-UNSAFE: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX9-UNSAFE: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX9-UNSAFE: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY6]](s32), [[COPY7]](s32), [[COPY8]](s32)
-    ; GFX9-UNSAFE: [[COPY9:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-UNSAFE: [[FMUL:%[0-9]+]]:_(<3 x s32>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX9-UNSAFE: [[FADD:%[0-9]+]]:_(<3 x s32>) = G_FADD [[BUILD_VECTOR2]], [[FMUL]]
     ; GFX9-UNSAFE: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<3 x s32>)
     ; GFX9-UNSAFE: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-UNSAFE: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9-UNSAFE: $vgpr2 = COPY [[UV2]](s32)
-    ; GFX9-UNSAFE: [[COPY10:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY9]]
-    ; GFX9-UNSAFE: S_SETPC_B64_return [[COPY10]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
+    ; GFX9-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
     ; GFX10-LABEL: name: test_3xfloat_add_mul_rhs
     ; GFX10: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10: [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32)
     ; GFX10: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10: [[BUILD_VECTOR1:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY3]](s32), [[COPY4]](s32), [[COPY5]](s32)
     ; GFX10: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX10: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX10: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX10: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY6]](s32), [[COPY7]](s32), [[COPY8]](s32)
-    ; GFX10: [[COPY9:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10: [[FMUL:%[0-9]+]]:_(<3 x s32>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX10: [[FADD:%[0-9]+]]:_(<3 x s32>) = G_FADD [[BUILD_VECTOR2]], [[FMUL]]
     ; GFX10: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<3 x s32>)
     ; GFX10: $vgpr0 = COPY [[UV]](s32)
     ; GFX10: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10: $vgpr2 = COPY [[UV2]](s32)
-    ; GFX10: [[COPY10:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY9]]
-    ; GFX10: S_SETPC_B64_return [[COPY10]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
+    ; GFX10: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
     ; GFX10-CONTRACT-LABEL: name: test_3xfloat_add_mul_rhs
     ; GFX10-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-CONTRACT: [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32)
     ; GFX10-CONTRACT: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-CONTRACT: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-CONTRACT: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-CONTRACT: [[BUILD_VECTOR1:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY3]](s32), [[COPY4]](s32), [[COPY5]](s32)
     ; GFX10-CONTRACT: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX10-CONTRACT: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX10-CONTRACT: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX10-CONTRACT: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY6]](s32), [[COPY7]](s32), [[COPY8]](s32)
-    ; GFX10-CONTRACT: [[COPY9:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-CONTRACT: [[FMUL:%[0-9]+]]:_(<3 x s32>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX10-CONTRACT: [[FADD:%[0-9]+]]:_(<3 x s32>) = G_FADD [[BUILD_VECTOR2]], [[FMUL]]
     ; GFX10-CONTRACT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<3 x s32>)
     ; GFX10-CONTRACT: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-CONTRACT: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10-CONTRACT: $vgpr2 = COPY [[UV2]](s32)
-    ; GFX10-CONTRACT: [[COPY10:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY9]]
-    ; GFX10-CONTRACT: S_SETPC_B64_return [[COPY10]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
+    ; GFX10-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
     ; GFX10-DENORM-LABEL: name: test_3xfloat_add_mul_rhs
     ; GFX10-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-DENORM: [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32)
     ; GFX10-DENORM: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-DENORM: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-DENORM: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-DENORM: [[BUILD_VECTOR1:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY3]](s32), [[COPY4]](s32), [[COPY5]](s32)
     ; GFX10-DENORM: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX10-DENORM: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX10-DENORM: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX10-DENORM: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY6]](s32), [[COPY7]](s32), [[COPY8]](s32)
-    ; GFX10-DENORM: [[COPY9:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-DENORM: [[FMUL:%[0-9]+]]:_(<3 x s32>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX10-DENORM: [[FADD:%[0-9]+]]:_(<3 x s32>) = G_FADD [[BUILD_VECTOR2]], [[FMUL]]
     ; GFX10-DENORM: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<3 x s32>)
     ; GFX10-DENORM: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-DENORM: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10-DENORM: $vgpr2 = COPY [[UV2]](s32)
-    ; GFX10-DENORM: [[COPY10:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY9]]
-    ; GFX10-DENORM: S_SETPC_B64_return [[COPY10]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
+    ; GFX10-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
     ; GFX10-UNSAFE-LABEL: name: test_3xfloat_add_mul_rhs
     ; GFX10-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-UNSAFE: [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32)
     ; GFX10-UNSAFE: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-UNSAFE: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-UNSAFE: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-UNSAFE: [[BUILD_VECTOR1:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY3]](s32), [[COPY4]](s32), [[COPY5]](s32)
     ; GFX10-UNSAFE: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX10-UNSAFE: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX10-UNSAFE: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX10-UNSAFE: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY6]](s32), [[COPY7]](s32), [[COPY8]](s32)
-    ; GFX10-UNSAFE: [[COPY9:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-UNSAFE: [[FMUL:%[0-9]+]]:_(<3 x s32>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX10-UNSAFE: [[FADD:%[0-9]+]]:_(<3 x s32>) = G_FADD [[BUILD_VECTOR2]], [[FMUL]]
     ; GFX10-UNSAFE: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<3 x s32>)
     ; GFX10-UNSAFE: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-UNSAFE: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10-UNSAFE: $vgpr2 = COPY [[UV2]](s32)
-    ; GFX10-UNSAFE: [[COPY10:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY9]]
-    ; GFX10-UNSAFE: S_SETPC_B64_return [[COPY10]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
+    ; GFX10-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
     %4:_(s32) = COPY $vgpr0
     %5:_(s32) = COPY $vgpr1
     %6:_(s32) = COPY $vgpr2
     %0:_(<3 x s32>) = G_BUILD_VECTOR %4(s32), %5(s32), %6(s32)
     %7:_(s32) = COPY $vgpr3
     %8:_(s32) = COPY $vgpr4
     %9:_(s32) = COPY $vgpr5
     %1:_(<3 x s32>) = G_BUILD_VECTOR %7(s32), %8(s32), %9(s32)
     %10:_(s32) = COPY $vgpr6
     %11:_(s32) = COPY $vgpr7
     %12:_(s32) = COPY $vgpr8
     %2:_(<3 x s32>) = G_BUILD_VECTOR %10(s32), %11(s32), %12(s32)
-    %3:sgpr_64 = COPY $sgpr30_sgpr31
     %13:_(<3 x s32>) = G_FMUL %0, %1
     %14:_(<3 x s32>) = G_FADD %2, %13
     %16:_(s32), %17:_(s32), %18:_(s32) = G_UNMERGE_VALUES %14(<3 x s32>)
     $vgpr0 = COPY %16(s32)
     $vgpr1 = COPY %17(s32)
     $vgpr2 = COPY %18(s32)
-    %15:ccr_sgpr_64 = COPY %3
-    S_SETPC_B64_return %15, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
 ...
 
 ---
 name:            test_4xhalf_add_mul
 body:             |
   bb.1.entry:
-    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $sgpr30_sgpr31
+    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
 
     ; GFX9-LABEL: name: test_4xhalf_add_mul
     ; GFX9: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX9: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX9: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>)
     ; GFX9: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX9: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX9: [[CONCAT_VECTORS1:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>)
     ; GFX9: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX9: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX9: [[CONCAT_VECTORS2:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>)
-    ; GFX9: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9: [[FMUL:%[0-9]+]]:_(<4 x s16>) = G_FMUL [[CONCAT_VECTORS]], [[CONCAT_VECTORS1]]
     ; GFX9: [[FADD:%[0-9]+]]:_(<4 x s16>) = G_FADD [[FMUL]], [[CONCAT_VECTORS2]]
     ; GFX9: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[FADD]](<4 x s16>)
     ; GFX9: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX9: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX9: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX9-CONTRACT-LABEL: name: test_4xhalf_add_mul
     ; GFX9-CONTRACT: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX9-CONTRACT: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX9-CONTRACT: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>)
     ; GFX9-CONTRACT: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX9-CONTRACT: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX9-CONTRACT: [[CONCAT_VECTORS1:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>)
     ; GFX9-CONTRACT: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX9-CONTRACT: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX9-CONTRACT: [[CONCAT_VECTORS2:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>)
-    ; GFX9-CONTRACT: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-CONTRACT: [[FMUL:%[0-9]+]]:_(<4 x s16>) = G_FMUL [[CONCAT_VECTORS]], [[CONCAT_VECTORS1]]
     ; GFX9-CONTRACT: [[FADD:%[0-9]+]]:_(<4 x s16>) = G_FADD [[FMUL]], [[CONCAT_VECTORS2]]
     ; GFX9-CONTRACT: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[FADD]](<4 x s16>)
     ; GFX9-CONTRACT: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX9-CONTRACT: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX9-CONTRACT: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9-CONTRACT: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX9-DENORM-LABEL: name: test_4xhalf_add_mul
     ; GFX9-DENORM: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX9-DENORM: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX9-DENORM: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>)
     ; GFX9-DENORM: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX9-DENORM: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX9-DENORM: [[CONCAT_VECTORS1:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>)
     ; GFX9-DENORM: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX9-DENORM: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX9-DENORM: [[CONCAT_VECTORS2:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>)
-    ; GFX9-DENORM: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-DENORM: [[FMUL:%[0-9]+]]:_(<4 x s16>) = G_FMUL [[CONCAT_VECTORS]], [[CONCAT_VECTORS1]]
     ; GFX9-DENORM: [[FADD:%[0-9]+]]:_(<4 x s16>) = G_FADD [[FMUL]], [[CONCAT_VECTORS2]]
     ; GFX9-DENORM: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[FADD]](<4 x s16>)
     ; GFX9-DENORM: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX9-DENORM: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX9-DENORM: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9-DENORM: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX9-UNSAFE-LABEL: name: test_4xhalf_add_mul
     ; GFX9-UNSAFE: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX9-UNSAFE: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX9-UNSAFE: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>)
     ; GFX9-UNSAFE: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX9-UNSAFE: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX9-UNSAFE: [[CONCAT_VECTORS1:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>)
     ; GFX9-UNSAFE: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX9-UNSAFE: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX9-UNSAFE: [[CONCAT_VECTORS2:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>)
-    ; GFX9-UNSAFE: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-UNSAFE: [[FMUL:%[0-9]+]]:_(<4 x s16>) = G_FMUL [[CONCAT_VECTORS]], [[CONCAT_VECTORS1]]
     ; GFX9-UNSAFE: [[FADD:%[0-9]+]]:_(<4 x s16>) = G_FADD [[FMUL]], [[CONCAT_VECTORS2]]
     ; GFX9-UNSAFE: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[FADD]](<4 x s16>)
     ; GFX9-UNSAFE: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX9-UNSAFE: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX9-UNSAFE: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9-UNSAFE: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-LABEL: name: test_4xhalf_add_mul
     ; GFX10: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX10: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX10: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>)
     ; GFX10: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX10: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX10: [[CONCAT_VECTORS1:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>)
     ; GFX10: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX10: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX10: [[CONCAT_VECTORS2:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>)
-    ; GFX10: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10: [[FMUL:%[0-9]+]]:_(<4 x s16>) = G_FMUL [[CONCAT_VECTORS]], [[CONCAT_VECTORS1]]
     ; GFX10: [[FADD:%[0-9]+]]:_(<4 x s16>) = G_FADD [[FMUL]], [[CONCAT_VECTORS2]]
     ; GFX10: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[FADD]](<4 x s16>)
     ; GFX10: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX10: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX10: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-CONTRACT-LABEL: name: test_4xhalf_add_mul
     ; GFX10-CONTRACT: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX10-CONTRACT: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX10-CONTRACT: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>)
     ; GFX10-CONTRACT: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX10-CONTRACT: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX10-CONTRACT: [[CONCAT_VECTORS1:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>)
     ; GFX10-CONTRACT: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX10-CONTRACT: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX10-CONTRACT: [[CONCAT_VECTORS2:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>)
-    ; GFX10-CONTRACT: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-CONTRACT: [[FMUL:%[0-9]+]]:_(<4 x s16>) = G_FMUL [[CONCAT_VECTORS]], [[CONCAT_VECTORS1]]
     ; GFX10-CONTRACT: [[FADD:%[0-9]+]]:_(<4 x s16>) = G_FADD [[FMUL]], [[CONCAT_VECTORS2]]
     ; GFX10-CONTRACT: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[FADD]](<4 x s16>)
     ; GFX10-CONTRACT: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX10-CONTRACT: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX10-CONTRACT: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10-CONTRACT: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-DENORM-LABEL: name: test_4xhalf_add_mul
     ; GFX10-DENORM: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX10-DENORM: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX10-DENORM: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>)
     ; GFX10-DENORM: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX10-DENORM: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX10-DENORM: [[CONCAT_VECTORS1:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>)
     ; GFX10-DENORM: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX10-DENORM: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX10-DENORM: [[CONCAT_VECTORS2:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>)
-    ; GFX10-DENORM: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-DENORM: [[FMUL:%[0-9]+]]:_(<4 x s16>) = G_FMUL [[CONCAT_VECTORS]], [[CONCAT_VECTORS1]]
     ; GFX10-DENORM: [[FADD:%[0-9]+]]:_(<4 x s16>) = G_FADD [[FMUL]], [[CONCAT_VECTORS2]]
     ; GFX10-DENORM: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[FADD]](<4 x s16>)
     ; GFX10-DENORM: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX10-DENORM: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX10-DENORM: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10-DENORM: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-UNSAFE-LABEL: name: test_4xhalf_add_mul
     ; GFX10-UNSAFE: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX10-UNSAFE: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX10-UNSAFE: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>)
     ; GFX10-UNSAFE: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX10-UNSAFE: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX10-UNSAFE: [[CONCAT_VECTORS1:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>)
     ; GFX10-UNSAFE: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX10-UNSAFE: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX10-UNSAFE: [[CONCAT_VECTORS2:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>)
-    ; GFX10-UNSAFE: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-UNSAFE: [[FMUL:%[0-9]+]]:_(<4 x s16>) = G_FMUL [[CONCAT_VECTORS]], [[CONCAT_VECTORS1]]
     ; GFX10-UNSAFE: [[FADD:%[0-9]+]]:_(<4 x s16>) = G_FADD [[FMUL]], [[CONCAT_VECTORS2]]
     ; GFX10-UNSAFE: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[FADD]](<4 x s16>)
     ; GFX10-UNSAFE: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX10-UNSAFE: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX10-UNSAFE: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10-UNSAFE: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     %4:_(<2 x s16>) = COPY $vgpr0
     %5:_(<2 x s16>) = COPY $vgpr1
     %0:_(<4 x s16>) = G_CONCAT_VECTORS %4(<2 x s16>), %5(<2 x s16>)
     %6:_(<2 x s16>) = COPY $vgpr2
     %7:_(<2 x s16>) = COPY $vgpr3
     %1:_(<4 x s16>) = G_CONCAT_VECTORS %6(<2 x s16>), %7(<2 x s16>)
     %8:_(<2 x s16>) = COPY $vgpr4
     %9:_(<2 x s16>) = COPY $vgpr5
     %2:_(<4 x s16>) = G_CONCAT_VECTORS %8(<2 x s16>), %9(<2 x s16>)
-    %3:sgpr_64 = COPY $sgpr30_sgpr31
     %10:_(<4 x s16>) = G_FMUL %0, %1
     %11:_(<4 x s16>) = G_FADD %10, %2
     %13:_(<2 x s16>), %14:_(<2 x s16>) = G_UNMERGE_VALUES %11(<4 x s16>)
     $vgpr0 = COPY %13(<2 x s16>)
     $vgpr1 = COPY %14(<2 x s16>)
-    %12:ccr_sgpr_64 = COPY %3
-    S_SETPC_B64_return %12, implicit $vgpr0, implicit $vgpr1
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
 ...
 
 ---
 name:            test_3xhalf_add_mul_rhs
 body:             |
   bb.1.entry:
-    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $sgpr30_sgpr31
+    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
 
     ; GFX9-LABEL: name: test_3xhalf_add_mul_rhs
     ; GFX9: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX9: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX9: [[DEF:%[0-9]+]]:_(<2 x s16>) = G_IMPLICIT_DEF
     ; GFX9: [[CONCAT_VECTORS:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX9: [[BITCAST:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS]](<6 x s16>)
     ; GFX9: [[TRUNC:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST]](s96)
     ; GFX9: [[BITCAST1:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC]](s48)
     ; GFX9: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX9: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX9: [[CONCAT_VECTORS1:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX9: [[BITCAST2:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS1]](<6 x s16>)
     ; GFX9: [[TRUNC1:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST2]](s96)
     ; GFX9: [[BITCAST3:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC1]](s48)
     ; GFX9: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX9: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX9: [[CONCAT_VECTORS2:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX9: [[BITCAST4:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS2]](<6 x s16>)
     ; GFX9: [[TRUNC2:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST4]](s96)
     ; GFX9: [[BITCAST5:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC2]](s48)
-    ; GFX9: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9: [[FMUL:%[0-9]+]]:_(<3 x s16>) = G_FMUL [[BITCAST1]], [[BITCAST3]]
     ; GFX9: [[FADD:%[0-9]+]]:_(<3 x s16>) = G_FADD [[BITCAST5]], [[FMUL]]
     ; GFX9: [[DEF1:%[0-9]+]]:_(<3 x s16>) = G_IMPLICIT_DEF
     ; GFX9: [[CONCAT_VECTORS3:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[FADD]](<3 x s16>), [[DEF1]](<3 x s16>)
     ; GFX9: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>), [[UV2:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS3]](<6 x s16>)
     ; GFX9: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX9: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX9: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX9-CONTRACT-LABEL: name: test_3xhalf_add_mul_rhs
     ; GFX9-CONTRACT: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX9-CONTRACT: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX9-CONTRACT: [[DEF:%[0-9]+]]:_(<2 x s16>) = G_IMPLICIT_DEF
     ; GFX9-CONTRACT: [[CONCAT_VECTORS:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX9-CONTRACT: [[BITCAST:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS]](<6 x s16>)
     ; GFX9-CONTRACT: [[TRUNC:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST]](s96)
     ; GFX9-CONTRACT: [[BITCAST1:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC]](s48)
     ; GFX9-CONTRACT: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX9-CONTRACT: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX9-CONTRACT: [[CONCAT_VECTORS1:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX9-CONTRACT: [[BITCAST2:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS1]](<6 x s16>)
     ; GFX9-CONTRACT: [[TRUNC1:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST2]](s96)
     ; GFX9-CONTRACT: [[BITCAST3:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC1]](s48)
     ; GFX9-CONTRACT: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX9-CONTRACT: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX9-CONTRACT: [[CONCAT_VECTORS2:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX9-CONTRACT: [[BITCAST4:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS2]](<6 x s16>)
     ; GFX9-CONTRACT: [[TRUNC2:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST4]](s96)
     ; GFX9-CONTRACT: [[BITCAST5:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC2]](s48)
-    ; GFX9-CONTRACT: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-CONTRACT: [[FMUL:%[0-9]+]]:_(<3 x s16>) = G_FMUL [[BITCAST1]], [[BITCAST3]]
     ; GFX9-CONTRACT: [[FADD:%[0-9]+]]:_(<3 x s16>) = G_FADD [[BITCAST5]], [[FMUL]]
     ; GFX9-CONTRACT: [[DEF1:%[0-9]+]]:_(<3 x s16>) = G_IMPLICIT_DEF
     ; GFX9-CONTRACT: [[CONCAT_VECTORS3:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[FADD]](<3 x s16>), [[DEF1]](<3 x s16>)
     ; GFX9-CONTRACT: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>), [[UV2:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS3]](<6 x s16>)
     ; GFX9-CONTRACT: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX9-CONTRACT: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX9-CONTRACT: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9-CONTRACT: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX9-DENORM-LABEL: name: test_3xhalf_add_mul_rhs
     ; GFX9-DENORM: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX9-DENORM: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX9-DENORM: [[DEF:%[0-9]+]]:_(<2 x s16>) = G_IMPLICIT_DEF
     ; GFX9-DENORM: [[CONCAT_VECTORS:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX9-DENORM: [[BITCAST:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS]](<6 x s16>)
     ; GFX9-DENORM: [[TRUNC:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST]](s96)
     ; GFX9-DENORM: [[BITCAST1:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC]](s48)
     ; GFX9-DENORM: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX9-DENORM: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX9-DENORM: [[CONCAT_VECTORS1:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX9-DENORM: [[BITCAST2:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS1]](<6 x s16>)
     ; GFX9-DENORM: [[TRUNC1:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST2]](s96)
     ; GFX9-DENORM: [[BITCAST3:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC1]](s48)
     ; GFX9-DENORM: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX9-DENORM: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX9-DENORM: [[CONCAT_VECTORS2:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX9-DENORM: [[BITCAST4:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS2]](<6 x s16>)
     ; GFX9-DENORM: [[TRUNC2:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST4]](s96)
     ; GFX9-DENORM: [[BITCAST5:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC2]](s48)
-    ; GFX9-DENORM: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-DENORM: [[FMUL:%[0-9]+]]:_(<3 x s16>) = G_FMUL [[BITCAST1]], [[BITCAST3]]
     ; GFX9-DENORM: [[FADD:%[0-9]+]]:_(<3 x s16>) = G_FADD [[BITCAST5]], [[FMUL]]
     ; GFX9-DENORM: [[DEF1:%[0-9]+]]:_(<3 x s16>) = G_IMPLICIT_DEF
     ; GFX9-DENORM: [[CONCAT_VECTORS3:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[FADD]](<3 x s16>), [[DEF1]](<3 x s16>)
     ; GFX9-DENORM: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>), [[UV2:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS3]](<6 x s16>)
     ; GFX9-DENORM: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX9-DENORM: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX9-DENORM: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9-DENORM: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX9-UNSAFE-LABEL: name: test_3xhalf_add_mul_rhs
     ; GFX9-UNSAFE: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX9-UNSAFE: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX9-UNSAFE: [[DEF:%[0-9]+]]:_(<2 x s16>) = G_IMPLICIT_DEF
     ; GFX9-UNSAFE: [[CONCAT_VECTORS:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX9-UNSAFE: [[BITCAST:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS]](<6 x s16>)
     ; GFX9-UNSAFE: [[TRUNC:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST]](s96)
     ; GFX9-UNSAFE: [[BITCAST1:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC]](s48)
     ; GFX9-UNSAFE: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX9-UNSAFE: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX9-UNSAFE: [[CONCAT_VECTORS1:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX9-UNSAFE: [[BITCAST2:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS1]](<6 x s16>)
     ; GFX9-UNSAFE: [[TRUNC1:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST2]](s96)
     ; GFX9-UNSAFE: [[BITCAST3:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC1]](s48)
     ; GFX9-UNSAFE: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX9-UNSAFE: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX9-UNSAFE: [[CONCAT_VECTORS2:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX9-UNSAFE: [[BITCAST4:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS2]](<6 x s16>)
     ; GFX9-UNSAFE: [[TRUNC2:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST4]](s96)
     ; GFX9-UNSAFE: [[BITCAST5:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC2]](s48)
-    ; GFX9-UNSAFE: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-UNSAFE: [[FMUL:%[0-9]+]]:_(<3 x s16>) = G_FMUL [[BITCAST1]], [[BITCAST3]]
     ; GFX9-UNSAFE: [[FADD:%[0-9]+]]:_(<3 x s16>) = G_FADD [[BITCAST5]], [[FMUL]]
     ; GFX9-UNSAFE: [[DEF1:%[0-9]+]]:_(<3 x s16>) = G_IMPLICIT_DEF
     ; GFX9-UNSAFE: [[CONCAT_VECTORS3:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[FADD]](<3 x s16>), [[DEF1]](<3 x s16>)
     ; GFX9-UNSAFE: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>), [[UV2:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS3]](<6 x s16>)
     ; GFX9-UNSAFE: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX9-UNSAFE: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX9-UNSAFE: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9-UNSAFE: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-LABEL: name: test_3xhalf_add_mul_rhs
     ; GFX10: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX10: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX10: [[DEF:%[0-9]+]]:_(<2 x s16>) = G_IMPLICIT_DEF
     ; GFX10: [[CONCAT_VECTORS:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX10: [[BITCAST:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS]](<6 x s16>)
     ; GFX10: [[TRUNC:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST]](s96)
     ; GFX10: [[BITCAST1:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC]](s48)
     ; GFX10: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX10: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX10: [[CONCAT_VECTORS1:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX10: [[BITCAST2:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS1]](<6 x s16>)
     ; GFX10: [[TRUNC1:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST2]](s96)
     ; GFX10: [[BITCAST3:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC1]](s48)
     ; GFX10: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX10: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX10: [[CONCAT_VECTORS2:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX10: [[BITCAST4:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS2]](<6 x s16>)
     ; GFX10: [[TRUNC2:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST4]](s96)
     ; GFX10: [[BITCAST5:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC2]](s48)
-    ; GFX10: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10: [[FMUL:%[0-9]+]]:_(<3 x s16>) = G_FMUL [[BITCAST1]], [[BITCAST3]]
     ; GFX10: [[FADD:%[0-9]+]]:_(<3 x s16>) = G_FADD [[BITCAST5]], [[FMUL]]
     ; GFX10: [[DEF1:%[0-9]+]]:_(<3 x s16>) = G_IMPLICIT_DEF
     ; GFX10: [[CONCAT_VECTORS3:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[FADD]](<3 x s16>), [[DEF1]](<3 x s16>)
     ; GFX10: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>), [[UV2:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS3]](<6 x s16>)
     ; GFX10: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX10: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX10: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-CONTRACT-LABEL: name: test_3xhalf_add_mul_rhs
     ; GFX10-CONTRACT: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX10-CONTRACT: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX10-CONTRACT: [[DEF:%[0-9]+]]:_(<2 x s16>) = G_IMPLICIT_DEF
     ; GFX10-CONTRACT: [[CONCAT_VECTORS:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX10-CONTRACT: [[BITCAST:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS]](<6 x s16>)
     ; GFX10-CONTRACT: [[TRUNC:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST]](s96)
     ; GFX10-CONTRACT: [[BITCAST1:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC]](s48)
     ; GFX10-CONTRACT: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX10-CONTRACT: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX10-CONTRACT: [[CONCAT_VECTORS1:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX10-CONTRACT: [[BITCAST2:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS1]](<6 x s16>)
     ; GFX10-CONTRACT: [[TRUNC1:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST2]](s96)
     ; GFX10-CONTRACT: [[BITCAST3:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC1]](s48)
     ; GFX10-CONTRACT: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX10-CONTRACT: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX10-CONTRACT: [[CONCAT_VECTORS2:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX10-CONTRACT: [[BITCAST4:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS2]](<6 x s16>)
     ; GFX10-CONTRACT: [[TRUNC2:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST4]](s96)
     ; GFX10-CONTRACT: [[BITCAST5:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC2]](s48)
-    ; GFX10-CONTRACT: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-CONTRACT: [[FMUL:%[0-9]+]]:_(<3 x s16>) = G_FMUL [[BITCAST1]], [[BITCAST3]]
     ; GFX10-CONTRACT: [[FADD:%[0-9]+]]:_(<3 x s16>) = G_FADD [[BITCAST5]], [[FMUL]]
     ; GFX10-CONTRACT: [[DEF1:%[0-9]+]]:_(<3 x s16>) = G_IMPLICIT_DEF
     ; GFX10-CONTRACT: [[CONCAT_VECTORS3:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[FADD]](<3 x s16>), [[DEF1]](<3 x s16>)
     ; GFX10-CONTRACT: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>), [[UV2:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS3]](<6 x s16>)
     ; GFX10-CONTRACT: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX10-CONTRACT: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX10-CONTRACT: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10-CONTRACT: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-DENORM-LABEL: name: test_3xhalf_add_mul_rhs
     ; GFX10-DENORM: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX10-DENORM: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX10-DENORM: [[DEF:%[0-9]+]]:_(<2 x s16>) = G_IMPLICIT_DEF
     ; GFX10-DENORM: [[CONCAT_VECTORS:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX10-DENORM: [[BITCAST:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS]](<6 x s16>)
     ; GFX10-DENORM: [[TRUNC:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST]](s96)
     ; GFX10-DENORM: [[BITCAST1:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC]](s48)
     ; GFX10-DENORM: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX10-DENORM: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX10-DENORM: [[CONCAT_VECTORS1:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX10-DENORM: [[BITCAST2:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS1]](<6 x s16>)
     ; GFX10-DENORM: [[TRUNC1:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST2]](s96)
     ; GFX10-DENORM: [[BITCAST3:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC1]](s48)
     ; GFX10-DENORM: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX10-DENORM: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX10-DENORM: [[CONCAT_VECTORS2:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX10-DENORM: [[BITCAST4:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS2]](<6 x s16>)
     ; GFX10-DENORM: [[TRUNC2:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST4]](s96)
     ; GFX10-DENORM: [[BITCAST5:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC2]](s48)
-    ; GFX10-DENORM: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-DENORM: [[FMUL:%[0-9]+]]:_(<3 x s16>) = G_FMUL [[BITCAST1]], [[BITCAST3]]
     ; GFX10-DENORM: [[FADD:%[0-9]+]]:_(<3 x s16>) = G_FADD [[BITCAST5]], [[FMUL]]
     ; GFX10-DENORM: [[DEF1:%[0-9]+]]:_(<3 x s16>) = G_IMPLICIT_DEF
     ; GFX10-DENORM: [[CONCAT_VECTORS3:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[FADD]](<3 x s16>), [[DEF1]](<3 x s16>)
     ; GFX10-DENORM: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>), [[UV2:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS3]](<6 x s16>)
     ; GFX10-DENORM: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX10-DENORM: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX10-DENORM: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10-DENORM: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-UNSAFE-LABEL: name: test_3xhalf_add_mul_rhs
     ; GFX10-UNSAFE: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX10-UNSAFE: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX10-UNSAFE: [[DEF:%[0-9]+]]:_(<2 x s16>) = G_IMPLICIT_DEF
     ; GFX10-UNSAFE: [[CONCAT_VECTORS:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX10-UNSAFE: [[BITCAST:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS]](<6 x s16>)
     ; GFX10-UNSAFE: [[TRUNC:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST]](s96)
     ; GFX10-UNSAFE: [[BITCAST1:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC]](s48)
     ; GFX10-UNSAFE: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX10-UNSAFE: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX10-UNSAFE: [[CONCAT_VECTORS1:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX10-UNSAFE: [[BITCAST2:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS1]](<6 x s16>)
     ; GFX10-UNSAFE: [[TRUNC1:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST2]](s96)
     ; GFX10-UNSAFE: [[BITCAST3:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC1]](s48)
     ; GFX10-UNSAFE: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX10-UNSAFE: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX10-UNSAFE: [[CONCAT_VECTORS2:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX10-UNSAFE: [[BITCAST4:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS2]](<6 x s16>)
     ; GFX10-UNSAFE: [[TRUNC2:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST4]](s96)
     ; GFX10-UNSAFE: [[BITCAST5:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC2]](s48)
-    ; GFX10-UNSAFE: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-UNSAFE: [[FMUL:%[0-9]+]]:_(<3 x s16>) = G_FMUL [[BITCAST1]], [[BITCAST3]]
     ; GFX10-UNSAFE: [[FADD:%[0-9]+]]:_(<3 x s16>) = G_FADD [[BITCAST5]], [[FMUL]]
     ; GFX10-UNSAFE: [[DEF1:%[0-9]+]]:_(<3 x s16>) = G_IMPLICIT_DEF
     ; GFX10-UNSAFE: [[CONCAT_VECTORS3:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[FADD]](<3 x s16>), [[DEF1]](<3 x s16>)
     ; GFX10-UNSAFE: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>), [[UV2:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS3]](<6 x s16>)
     ; GFX10-UNSAFE: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX10-UNSAFE: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX10-UNSAFE: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10-UNSAFE: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     %4:_(<2 x s16>) = COPY $vgpr0
     %5:_(<2 x s16>) = COPY $vgpr1
     %10:_(<2 x s16>) = G_IMPLICIT_DEF
     %11:_(<6 x s16>) = G_CONCAT_VECTORS %4(<2 x s16>), %5(<2 x s16>), %10(<2 x s16>)
     %0:_(<3 x s16>), %12:_(<3 x s16>) = G_UNMERGE_VALUES %11(<6 x s16>)
     %6:_(<2 x s16>) = COPY $vgpr2
     %7:_(<2 x s16>) = COPY $vgpr3
     %13:_(<6 x s16>) = G_CONCAT_VECTORS %6(<2 x s16>), %7(<2 x s16>), %10(<2 x s16>)
     %1:_(<3 x s16>), %14:_(<3 x s16>) = G_UNMERGE_VALUES %13(<6 x s16>)
     %8:_(<2 x s16>) = COPY $vgpr4
     %9:_(<2 x s16>) = COPY $vgpr5
     %15:_(<6 x s16>) = G_CONCAT_VECTORS %8(<2 x s16>), %9(<2 x s16>), %10(<2 x s16>)
     %2:_(<3 x s16>), %16:_(<3 x s16>) = G_UNMERGE_VALUES %15(<6 x s16>)
-    %3:sgpr_64 = COPY $sgpr30_sgpr31
     %17:_(<3 x s16>) = G_FMUL %0, %1
     %18:_(<3 x s16>) = G_FADD %2, %17
     %22:_(<3 x s16>) = G_IMPLICIT_DEF
     %23:_(<6 x s16>) = G_CONCAT_VECTORS %18(<3 x s16>), %22(<3 x s16>)
     %20:_(<2 x s16>), %21:_(<2 x s16>), %24:_(<2 x s16>) = G_UNMERGE_VALUES %23(<6 x s16>)
     $vgpr0 = COPY %20(<2 x s16>)
     $vgpr1 = COPY %21(<2 x s16>)
-    %19:ccr_sgpr_64 = COPY %3
-    S_SETPC_B64_return %19, implicit $vgpr0, implicit $vgpr1
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
 ...
 
 ---
 name:            test_4xdouble_add_mul
 body:             |
   bb.1.entry:
-    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, $sgpr30_sgpr31
+    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
 
     ; GFX9-LABEL: name: test_4xdouble_add_mul
     ; GFX9: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
bb.1.entry:
     ; GFX9: [[COPY21:%[0-9]+]]:_(s32) = COPY $vgpr21
     ; GFX9: [[COPY22:%[0-9]+]]:_(s32) = COPY $vgpr22
     ; GFX9: [[COPY23:%[0-9]+]]:_(s32) = COPY $vgpr23
     ; GFX9: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX9: [[MV9:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY18]](s32), [[COPY19]](s32)
     ; GFX9: [[MV10:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY20]](s32), [[COPY21]](s32)
     ; GFX9: [[MV11:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY22]](s32), [[COPY23]](s32)
     ; GFX9: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s64>) = G_BUILD_VECTOR [[MV8]](s64), [[MV9]](s64), [[MV10]](s64), [[MV11]](s64)
-    ; GFX9: [[COPY24:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9: [[FMUL:%[0-9]+]]:_(<4 x s64>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX9: [[FADD:%[0-9]+]]:_(<4 x s64>) = G_FADD [[FMUL]], [[BUILD_VECTOR2]]
     ; GFX9: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<4 x s64>)
     ; GFX9: $vgpr0 = COPY [[UV]](s32)
     ; GFX9: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9: $vgpr2 = COPY [[UV2]](s32)
     ; GFX9: $vgpr3 = COPY [[UV3]](s32)
     ; GFX9: $vgpr4 = COPY [[UV4]](s32)
     ; GFX9: $vgpr5 = COPY [[UV5]](s32)
     ; GFX9: $vgpr6 = COPY [[UV6]](s32)
     ; GFX9: $vgpr7 = COPY [[UV7]](s32)
-    ; GFX9: [[COPY25:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY24]]
-    ; GFX9: S_SETPC_B64_return [[COPY25]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
+    ; GFX9: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
     ; GFX9-CONTRACT-LABEL: name: test_4xdouble_add_mul
     ; GFX9-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-CONTRACT: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-CONTRACT: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-CONTRACT: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-CONTRACT: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
bb.1.entry:
     ; GFX9-CONTRACT: [[COPY21:%[0-9]+]]:_(s32) = COPY $vgpr21
     ; GFX9-CONTRACT: [[COPY22:%[0-9]+]]:_(s32) = COPY $vgpr22
     ; GFX9-CONTRACT: [[COPY23:%[0-9]+]]:_(s32) = COPY $vgpr23
     ; GFX9-CONTRACT: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX9-CONTRACT: [[MV9:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY18]](s32), [[COPY19]](s32)
     ; GFX9-CONTRACT: [[MV10:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY20]](s32), [[COPY21]](s32)
     ; GFX9-CONTRACT: [[MV11:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY22]](s32), [[COPY23]](s32)
     ; GFX9-CONTRACT: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s64>) = G_BUILD_VECTOR [[MV8]](s64), [[MV9]](s64), [[MV10]](s64), [[MV11]](s64)
-    ; GFX9-CONTRACT: [[COPY24:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-CONTRACT: [[FMUL:%[0-9]+]]:_(<4 x s64>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX9-CONTRACT: [[FADD:%[0-9]+]]:_(<4 x s64>) = G_FADD [[FMUL]], [[BUILD_VECTOR2]]
     ; GFX9-CONTRACT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<4 x s64>)
     ; GFX9-CONTRACT: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-CONTRACT: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9-CONTRACT: $vgpr2 = COPY [[UV2]](s32)
     ; GFX9-CONTRACT: $vgpr3 = COPY [[UV3]](s32)
     ; GFX9-CONTRACT: $vgpr4 = COPY [[UV4]](s32)
     ; GFX9-CONTRACT: $vgpr5 = COPY [[UV5]](s32)
     ; GFX9-CONTRACT: $vgpr6 = COPY [[UV6]](s32)
     ; GFX9-CONTRACT: $vgpr7 = COPY [[UV7]](s32)
-    ; GFX9-CONTRACT: [[COPY25:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY24]]
-    ; GFX9-CONTRACT: S_SETPC_B64_return [[COPY25]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
+    ; GFX9-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
     ; GFX9-DENORM-LABEL: name: test_4xdouble_add_mul
     ; GFX9-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-DENORM: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-DENORM: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-DENORM: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-DENORM: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
bb.1.entry:
     ; GFX9-DENORM: [[COPY21:%[0-9]+]]:_(s32) = COPY $vgpr21
     ; GFX9-DENORM: [[COPY22:%[0-9]+]]:_(s32) = COPY $vgpr22
     ; GFX9-DENORM: [[COPY23:%[0-9]+]]:_(s32) = COPY $vgpr23
     ; GFX9-DENORM: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX9-DENORM: [[MV9:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY18]](s32), [[COPY19]](s32)
     ; GFX9-DENORM: [[MV10:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY20]](s32), [[COPY21]](s32)
     ; GFX9-DENORM: [[MV11:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY22]](s32), [[COPY23]](s32)
     ; GFX9-DENORM: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s64>) = G_BUILD_VECTOR [[MV8]](s64), [[MV9]](s64), [[MV10]](s64), [[MV11]](s64)
-    ; GFX9-DENORM: [[COPY24:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-DENORM: [[FMUL:%[0-9]+]]:_(<4 x s64>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX9-DENORM: [[FADD:%[0-9]+]]:_(<4 x s64>) = G_FADD [[FMUL]], [[BUILD_VECTOR2]]
     ; GFX9-DENORM: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<4 x s64>)
     ; GFX9-DENORM: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-DENORM: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9-DENORM: $vgpr2 = COPY [[UV2]](s32)
     ; GFX9-DENORM: $vgpr3 = COPY [[UV3]](s32)
     ; GFX9-DENORM: $vgpr4 = COPY [[UV4]](s32)
     ; GFX9-DENORM: $vgpr5 = COPY [[UV5]](s32)
     ; GFX9-DENORM: $vgpr6 = COPY [[UV6]](s32)
     ; GFX9-DENORM: $vgpr7 = COPY [[UV7]](s32)
-    ; GFX9-DENORM: [[COPY25:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY24]]
-    ; GFX9-DENORM: S_SETPC_B64_return [[COPY25]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
+    ; GFX9-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
     ; GFX9-UNSAFE-LABEL: name: test_4xdouble_add_mul
     ; GFX9-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-UNSAFE: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-UNSAFE: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-UNSAFE: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-UNSAFE: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
bb.1.entry:
     ; GFX9-UNSAFE: [[COPY21:%[0-9]+]]:_(s32) = COPY $vgpr21
     ; GFX9-UNSAFE: [[COPY22:%[0-9]+]]:_(s32) = COPY $vgpr22
     ; GFX9-UNSAFE: [[COPY23:%[0-9]+]]:_(s32) = COPY $vgpr23
     ; GFX9-UNSAFE: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX9-UNSAFE: [[MV9:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY18]](s32), [[COPY19]](s32)
     ; GFX9-UNSAFE: [[MV10:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY20]](s32), [[COPY21]](s32)
     ; GFX9-UNSAFE: [[MV11:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY22]](s32), [[COPY23]](s32)
     ; GFX9-UNSAFE: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s64>) = G_BUILD_VECTOR [[MV8]](s64), [[MV9]](s64), [[MV10]](s64), [[MV11]](s64)
-    ; GFX9-UNSAFE: [[COPY24:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-UNSAFE: [[FMUL:%[0-9]+]]:_(<4 x s64>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX9-UNSAFE: [[FADD:%[0-9]+]]:_(<4 x s64>) = G_FADD [[FMUL]], [[BUILD_VECTOR2]]
     ; GFX9-UNSAFE: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<4 x s64>)
     ; GFX9-UNSAFE: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-UNSAFE: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9-UNSAFE: $vgpr2 = COPY [[UV2]](s32)
     ; GFX9-UNSAFE: $vgpr3 = COPY [[UV3]](s32)
     ; GFX9-UNSAFE: $vgpr4 = COPY [[UV4]](s32)
     ; GFX9-UNSAFE: $vgpr5 = COPY [[UV5]](s32)
     ; GFX9-UNSAFE: $vgpr6 = COPY [[UV6]](s32)
     ; GFX9-UNSAFE: $vgpr7 = COPY [[UV7]](s32)
-    ; GFX9-UNSAFE: [[COPY25:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY24]]
-    ; GFX9-UNSAFE: S_SETPC_B64_return [[COPY25]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
+    ; GFX9-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
     ; GFX10-LABEL: name: test_4xdouble_add_mul
     ; GFX10: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
bb.1.entry:
     ; GFX10: [[COPY21:%[0-9]+]]:_(s32) = COPY $vgpr21
     ; GFX10: [[COPY22:%[0-9]+]]:_(s32) = COPY $vgpr22
     ; GFX10: [[COPY23:%[0-9]+]]:_(s32) = COPY $vgpr23
     ; GFX10: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX10: [[MV9:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY18]](s32), [[COPY19]](s32)
     ; GFX10: [[MV10:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY20]](s32), [[COPY21]](s32)
     ; GFX10: [[MV11:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY22]](s32), [[COPY23]](s32)
     ; GFX10: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s64>) = G_BUILD_VECTOR [[MV8]](s64), [[MV9]](s64), [[MV10]](s64), [[MV11]](s64)
-    ; GFX10: [[COPY24:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10: [[FMUL:%[0-9]+]]:_(<4 x s64>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX10: [[FADD:%[0-9]+]]:_(<4 x s64>) = G_FADD [[FMUL]], [[BUILD_VECTOR2]]
     ; GFX10: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<4 x s64>)
     ; GFX10: $vgpr0 = COPY [[UV]](s32)
     ; GFX10: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10: $vgpr2 = COPY [[UV2]](s32)
     ; GFX10: $vgpr3 = COPY [[UV3]](s32)
     ; GFX10: $vgpr4 = COPY [[UV4]](s32)
     ; GFX10: $vgpr5 = COPY [[UV5]](s32)
     ; GFX10: $vgpr6 = COPY [[UV6]](s32)
     ; GFX10: $vgpr7 = COPY [[UV7]](s32)
-    ; GFX10: [[COPY25:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY24]]
-    ; GFX10: S_SETPC_B64_return [[COPY25]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
+    ; GFX10: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
     ; GFX10-CONTRACT-LABEL: name: test_4xdouble_add_mul
     ; GFX10-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-CONTRACT: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-CONTRACT: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-CONTRACT: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-CONTRACT: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
bb.1.entry:
     ; GFX10-CONTRACT: [[COPY21:%[0-9]+]]:_(s32) = COPY $vgpr21
     ; GFX10-CONTRACT: [[COPY22:%[0-9]+]]:_(s32) = COPY $vgpr22
     ; GFX10-CONTRACT: [[COPY23:%[0-9]+]]:_(s32) = COPY $vgpr23
     ; GFX10-CONTRACT: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX10-CONTRACT: [[MV9:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY18]](s32), [[COPY19]](s32)
     ; GFX10-CONTRACT: [[MV10:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY20]](s32), [[COPY21]](s32)
     ; GFX10-CONTRACT: [[MV11:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY22]](s32), [[COPY23]](s32)
     ; GFX10-CONTRACT: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s64>) = G_BUILD_VECTOR [[MV8]](s64), [[MV9]](s64), [[MV10]](s64), [[MV11]](s64)
-    ; GFX10-CONTRACT: [[COPY24:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-CONTRACT: [[FMUL:%[0-9]+]]:_(<4 x s64>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX10-CONTRACT: [[FADD:%[0-9]+]]:_(<4 x s64>) = G_FADD [[FMUL]], [[BUILD_VECTOR2]]
     ; GFX10-CONTRACT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<4 x s64>)
     ; GFX10-CONTRACT: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-CONTRACT: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10-CONTRACT: $vgpr2 = COPY [[UV2]](s32)
     ; GFX10-CONTRACT: $vgpr3 = COPY [[UV3]](s32)
     ; GFX10-CONTRACT: $vgpr4 = COPY [[UV4]](s32)
     ; GFX10-CONTRACT: $vgpr5 = COPY [[UV5]](s32)
     ; GFX10-CONTRACT: $vgpr6 = COPY [[UV6]](s32)
     ; GFX10-CONTRACT: $vgpr7 = COPY [[UV7]](s32)
-    ; GFX10-CONTRACT: [[COPY25:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY24]]
-    ; GFX10-CONTRACT: S_SETPC_B64_return [[COPY25]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
+    ; GFX10-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
     ; GFX10-DENORM-LABEL: name: test_4xdouble_add_mul
     ; GFX10-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-DENORM: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-DENORM: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-DENORM: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-DENORM: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
bb.1.entry:
     ; GFX10-DENORM: [[COPY21:%[0-9]+]]:_(s32) = COPY $vgpr21
     ; GFX10-DENORM: [[COPY22:%[0-9]+]]:_(s32) = COPY $vgpr22
     ; GFX10-DENORM: [[COPY23:%[0-9]+]]:_(s32) = COPY $vgpr23
     ; GFX10-DENORM: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX10-DENORM: [[MV9:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY18]](s32), [[COPY19]](s32)
     ; GFX10-DENORM: [[MV10:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY20]](s32), [[COPY21]](s32)
     ; GFX10-DENORM: [[MV11:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY22]](s32), [[COPY23]](s32)
     ; GFX10-DENORM: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s64>) = G_BUILD_VECTOR [[MV8]](s64), [[MV9]](s64), [[MV10]](s64), [[MV11]](s64)
-    ; GFX10-DENORM: [[COPY24:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-DENORM: [[FMUL:%[0-9]+]]:_(<4 x s64>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX10-DENORM: [[FADD:%[0-9]+]]:_(<4 x s64>) = G_FADD [[FMUL]], [[BUILD_VECTOR2]]
     ; GFX10-DENORM: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<4 x s64>)
     ; GFX10-DENORM: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-DENORM: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10-DENORM: $vgpr2 = COPY [[UV2]](s32)
     ; GFX10-DENORM: $vgpr3 = COPY [[UV3]](s32)
     ; GFX10-DENORM: $vgpr4 = COPY [[UV4]](s32)
     ; GFX10-DENORM: $vgpr5 = COPY [[UV5]](s32)
     ; GFX10-DENORM: $vgpr6 = COPY [[UV6]](s32)
     ; GFX10-DENORM: $vgpr7 = COPY [[UV7]](s32)
-    ; GFX10-DENORM: [[COPY25:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY24]]
-    ; GFX10-DENORM: S_SETPC_B64_return [[COPY25]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
+    ; GFX10-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
     ; GFX10-UNSAFE-LABEL: name: test_4xdouble_add_mul
     ; GFX10-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-UNSAFE: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-UNSAFE: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-UNSAFE: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-UNSAFE: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
bb.1.entry:
     ; GFX10-UNSAFE: [[COPY21:%[0-9]+]]:_(s32) = COPY $vgpr21
     ; GFX10-UNSAFE: [[COPY22:%[0-9]+]]:_(s32) = COPY $vgpr22
     ; GFX10-UNSAFE: [[COPY23:%[0-9]+]]:_(s32) = COPY $vgpr23
     ; GFX10-UNSAFE: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX10-UNSAFE: [[MV9:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY18]](s32), [[COPY19]](s32)
     ; GFX10-UNSAFE: [[MV10:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY20]](s32), [[COPY21]](s32)
     ; GFX10-UNSAFE: [[MV11:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY22]](s32), [[COPY23]](s32)
     ; GFX10-UNSAFE: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s64>) = G_BUILD_VECTOR [[MV8]](s64), [[MV9]](s64), [[MV10]](s64), [[MV11]](s64)
-    ; GFX10-UNSAFE: [[COPY24:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-UNSAFE: [[FMUL:%[0-9]+]]:_(<4 x s64>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX10-UNSAFE: [[FADD:%[0-9]+]]:_(<4 x s64>) = G_FADD [[FMUL]], [[BUILD_VECTOR2]]
     ; GFX10-UNSAFE: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<4 x s64>)
     ; GFX10-UNSAFE: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-UNSAFE: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10-UNSAFE: $vgpr2 = COPY [[UV2]](s32)
     ; GFX10-UNSAFE: $vgpr3 = COPY [[UV3]](s32)
     ; GFX10-UNSAFE: $vgpr4 = COPY [[UV4]](s32)
     ; GFX10-UNSAFE: $vgpr5 = COPY [[UV5]](s32)
     ; GFX10-UNSAFE: $vgpr6 = COPY [[UV6]](s32)
     ; GFX10-UNSAFE: $vgpr7 = COPY [[UV7]](s32)
-    ; GFX10-UNSAFE: [[COPY25:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY24]]
-    ; GFX10-UNSAFE: S_SETPC_B64_return [[COPY25]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
+    ; GFX10-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
     %4:_(s32) = COPY $vgpr0
     %5:_(s32) = COPY $vgpr1
     %6:_(s32) = COPY $vgpr2
     %7:_(s32) = COPY $vgpr3
     %8:_(s32) = COPY $vgpr4
     %9:_(s32) = COPY $vgpr5
     %10:_(s32) = COPY $vgpr6
     %11:_(s32) = COPY $vgpr7
bb.1.entry:
     %25:_(s32) = COPY $vgpr21
     %26:_(s32) = COPY $vgpr22
     %27:_(s32) = COPY $vgpr23
     %36:_(s64) = G_MERGE_VALUES %20(s32), %21(s32)
     %37:_(s64) = G_MERGE_VALUES %22(s32), %23(s32)
     %38:_(s64) = G_MERGE_VALUES %24(s32), %25(s32)
     %39:_(s64) = G_MERGE_VALUES %26(s32), %27(s32)
     %2:_(<4 x s64>) = G_BUILD_VECTOR %36(s64), %37(s64), %38(s64), %39(s64)
-    %3:sgpr_64 = COPY $sgpr30_sgpr31
     %40:_(<4 x s64>) = G_FMUL %0, %1
     %41:_(<4 x s64>) = G_FADD %40, %2
     %43:_(s32), %44:_(s32), %45:_(s32), %46:_(s32), %47:_(s32), %48:_(s32), %49:_(s32), %50:_(s32) = G_UNMERGE_VALUES %41(<4 x s64>)
     $vgpr0 = COPY %43(s32)
     $vgpr1 = COPY %44(s32)
     $vgpr2 = COPY %45(s32)
     $vgpr3 = COPY %46(s32)
     $vgpr4 = COPY %47(s32)
     $vgpr5 = COPY %48(s32)
     $vgpr6 = COPY %49(s32)
     $vgpr7 = COPY %50(s32)
-    %42:ccr_sgpr_64 = COPY %3
-    S_SETPC_B64_return %42, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
 ...
 
 ---
 name:            test_3xdouble_add_mul_rhs
 body:             |
   bb.1.entry:
-    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7, $vgpr8, $vgpr9, $vgpr10, $vgpr11, $vgpr12, $vgpr13, $vgpr14, $vgpr15, $vgpr16, $vgpr17, $sgpr30_sgpr31
+    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7, $vgpr8, $vgpr9, $vgpr10, $vgpr11, $vgpr12, $vgpr13, $vgpr14, $vgpr15, $vgpr16, $vgpr17
 
     ; GFX9-LABEL: name: test_3xdouble_add_mul_rhs
     ; GFX9: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
bb.1.entry:
     ; GFX9: [[COPY14:%[0-9]+]]:_(s32) = COPY $vgpr14
     ; GFX9: [[COPY15:%[0-9]+]]:_(s32) = COPY $vgpr15
     ; GFX9: [[COPY16:%[0-9]+]]:_(s32) = COPY $vgpr16
     ; GFX9: [[COPY17:%[0-9]+]]:_(s32) = COPY $vgpr17
     ; GFX9: [[MV6:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY12]](s32), [[COPY13]](s32)
     ; GFX9: [[MV7:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY14]](s32), [[COPY15]](s32)
     ; GFX9: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX9: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s64>) = G_BUILD_VECTOR [[MV6]](s64), [[MV7]](s64), [[MV8]](s64)
-    ; GFX9: [[COPY18:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9: [[FMUL:%[0-9]+]]:_(<3 x s64>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX9: [[FADD:%[0-9]+]]:_(<3 x s64>) = G_FADD [[BUILD_VECTOR2]], [[FMUL]]
     ; GFX9: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<3 x s64>)
     ; GFX9: $vgpr0 = COPY [[UV]](s32)
     ; GFX9: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9: $vgpr2 = COPY [[UV2]](s32)
     ; GFX9: $vgpr3 = COPY [[UV3]](s32)
     ; GFX9: $vgpr4 = COPY [[UV4]](s32)
     ; GFX9: $vgpr5 = COPY [[UV5]](s32)
-    ; GFX9: [[COPY19:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY18]]
-    ; GFX9: S_SETPC_B64_return [[COPY19]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
+    ; GFX9: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
     ; GFX9-CONTRACT-LABEL: name: test_3xdouble_add_mul_rhs
     ; GFX9-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-CONTRACT: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-CONTRACT: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-CONTRACT: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-CONTRACT: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
bb.1.entry:
     ; GFX9-CONTRACT: [[COPY14:%[0-9]+]]:_(s32) = COPY $vgpr14
     ; GFX9-CONTRACT: [[COPY15:%[0-9]+]]:_(s32) = COPY $vgpr15
     ; GFX9-CONTRACT: [[COPY16:%[0-9]+]]:_(s32) = COPY $vgpr16
     ; GFX9-CONTRACT: [[COPY17:%[0-9]+]]:_(s32) = COPY $vgpr17
     ; GFX9-CONTRACT: [[MV6:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY12]](s32), [[COPY13]](s32)
     ; GFX9-CONTRACT: [[MV7:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY14]](s32), [[COPY15]](s32)
     ; GFX9-CONTRACT: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX9-CONTRACT: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s64>) = G_BUILD_VECTOR [[MV6]](s64), [[MV7]](s64), [[MV8]](s64)
-    ; GFX9-CONTRACT: [[COPY18:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-CONTRACT: [[FMUL:%[0-9]+]]:_(<3 x s64>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX9-CONTRACT: [[FADD:%[0-9]+]]:_(<3 x s64>) = G_FADD [[BUILD_VECTOR2]], [[FMUL]]
     ; GFX9-CONTRACT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<3 x s64>)
     ; GFX9-CONTRACT: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-CONTRACT: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9-CONTRACT: $vgpr2 = COPY [[UV2]](s32)
     ; GFX9-CONTRACT: $vgpr3 = COPY [[UV3]](s32)
     ; GFX9-CONTRACT: $vgpr4 = COPY [[UV4]](s32)
     ; GFX9-CONTRACT: $vgpr5 = COPY [[UV5]](s32)
-    ; GFX9-CONTRACT: [[COPY19:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY18]]
-    ; GFX9-CONTRACT: S_SETPC_B64_return [[COPY19]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
+    ; GFX9-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
     ; GFX9-DENORM-LABEL: name: test_3xdouble_add_mul_rhs
     ; GFX9-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-DENORM: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-DENORM: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-DENORM: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-DENORM: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
bb.1.entry:
     ; GFX9-DENORM: [[COPY14:%[0-9]+]]:_(s32) = COPY $vgpr14
     ; GFX9-DENORM: [[COPY15:%[0-9]+]]:_(s32) = COPY $vgpr15
     ; GFX9-DENORM: [[COPY16:%[0-9]+]]:_(s32) = COPY $vgpr16
     ; GFX9-DENORM: [[COPY17:%[0-9]+]]:_(s32) = COPY $vgpr17
     ; GFX9-DENORM: [[MV6:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY12]](s32), [[COPY13]](s32)
     ; GFX9-DENORM: [[MV7:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY14]](s32), [[COPY15]](s32)
     ; GFX9-DENORM: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX9-DENORM: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s64>) = G_BUILD_VECTOR [[MV6]](s64), [[MV7]](s64), [[MV8]](s64)
-    ; GFX9-DENORM: [[COPY18:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-DENORM: [[FMUL:%[0-9]+]]:_(<3 x s64>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX9-DENORM: [[FADD:%[0-9]+]]:_(<3 x s64>) = G_FADD [[BUILD_VECTOR2]], [[FMUL]]
     ; GFX9-DENORM: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<3 x s64>)
     ; GFX9-DENORM: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-DENORM: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9-DENORM: $vgpr2 = COPY [[UV2]](s32)
     ; GFX9-DENORM: $vgpr3 = COPY [[UV3]](s32)
     ; GFX9-DENORM: $vgpr4 = COPY [[UV4]](s32)
     ; GFX9-DENORM: $vgpr5 = COPY [[UV5]](s32)
-    ; GFX9-DENORM: [[COPY19:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY18]]
-    ; GFX9-DENORM: S_SETPC_B64_return [[COPY19]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
+    ; GFX9-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
     ; GFX9-UNSAFE-LABEL: name: test_3xdouble_add_mul_rhs
     ; GFX9-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-UNSAFE: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-UNSAFE: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-UNSAFE: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-UNSAFE: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
bb.1.entry:
     ; GFX9-UNSAFE: [[COPY14:%[0-9]+]]:_(s32) = COPY $vgpr14
     ; GFX9-UNSAFE: [[COPY15:%[0-9]+]]:_(s32) = COPY $vgpr15
     ; GFX9-UNSAFE: [[COPY16:%[0-9]+]]:_(s32) = COPY $vgpr16
     ; GFX9-UNSAFE: [[COPY17:%[0-9]+]]:_(s32) = COPY $vgpr17
     ; GFX9-UNSAFE: [[MV6:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY12]](s32), [[COPY13]](s32)
     ; GFX9-UNSAFE: [[MV7:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY14]](s32), [[COPY15]](s32)
     ; GFX9-UNSAFE: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX9-UNSAFE: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s64>) = G_BUILD_VECTOR [[MV6]](s64), [[MV7]](s64), [[MV8]](s64)
-    ; GFX9-UNSAFE: [[COPY18:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-UNSAFE: [[FMUL:%[0-9]+]]:_(<3 x s64>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX9-UNSAFE: [[FADD:%[0-9]+]]:_(<3 x s64>) = G_FADD [[BUILD_VECTOR2]], [[FMUL]]
     ; GFX9-UNSAFE: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<3 x s64>)
     ; GFX9-UNSAFE: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-UNSAFE: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9-UNSAFE: $vgpr2 = COPY [[UV2]](s32)
     ; GFX9-UNSAFE: $vgpr3 = COPY [[UV3]](s32)
     ; GFX9-UNSAFE: $vgpr4 = COPY [[UV4]](s32)
     ; GFX9-UNSAFE: $vgpr5 = COPY [[UV5]](s32)
-    ; GFX9-UNSAFE: [[COPY19:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY18]]
-    ; GFX9-UNSAFE: S_SETPC_B64_return [[COPY19]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
+    ; GFX9-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
     ; GFX10-LABEL: name: test_3xdouble_add_mul_rhs
     ; GFX10: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
bb.1.entry:
     ; GFX10: [[COPY14:%[0-9]+]]:_(s32) = COPY $vgpr14
     ; GFX10: [[COPY15:%[0-9]+]]:_(s32) = COPY $vgpr15
     ; GFX10: [[COPY16:%[0-9]+]]:_(s32) = COPY $vgpr16
     ; GFX10: [[COPY17:%[0-9]+]]:_(s32) = COPY $vgpr17
     ; GFX10: [[MV6:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY12]](s32), [[COPY13]](s32)
     ; GFX10: [[MV7:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY14]](s32), [[COPY15]](s32)
     ; GFX10: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX10: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s64>) = G_BUILD_VECTOR [[MV6]](s64), [[MV7]](s64), [[MV8]](s64)
-    ; GFX10: [[COPY18:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10: [[FMUL:%[0-9]+]]:_(<3 x s64>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX10: [[FADD:%[0-9]+]]:_(<3 x s64>) = G_FADD [[BUILD_VECTOR2]], [[FMUL]]
     ; GFX10: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<3 x s64>)
     ; GFX10: $vgpr0 = COPY [[UV]](s32)
     ; GFX10: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10: $vgpr2 = COPY [[UV2]](s32)
     ; GFX10: $vgpr3 = COPY [[UV3]](s32)
     ; GFX10: $vgpr4 = COPY [[UV4]](s32)
     ; GFX10: $vgpr5 = COPY [[UV5]](s32)
-    ; GFX10: [[COPY19:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY18]]
-    ; GFX10: S_SETPC_B64_return [[COPY19]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
+    ; GFX10: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
     ; GFX10-CONTRACT-LABEL: name: test_3xdouble_add_mul_rhs
     ; GFX10-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-CONTRACT: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-CONTRACT: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-CONTRACT: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-CONTRACT: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
bb.1.entry:
     ; GFX10-CONTRACT: [[COPY14:%[0-9]+]]:_(s32) = COPY $vgpr14
     ; GFX10-CONTRACT: [[COPY15:%[0-9]+]]:_(s32) = COPY $vgpr15
     ; GFX10-CONTRACT: [[COPY16:%[0-9]+]]:_(s32) = COPY $vgpr16
     ; GFX10-CONTRACT: [[COPY17:%[0-9]+]]:_(s32) = COPY $vgpr17
     ; GFX10-CONTRACT: [[MV6:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY12]](s32), [[COPY13]](s32)
     ; GFX10-CONTRACT: [[MV7:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY14]](s32), [[COPY15]](s32)
     ; GFX10-CONTRACT: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX10-CONTRACT: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s64>) = G_BUILD_VECTOR [[MV6]](s64), [[MV7]](s64), [[MV8]](s64)
-    ; GFX10-CONTRACT: [[COPY18:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-CONTRACT: [[FMUL:%[0-9]+]]:_(<3 x s64>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX10-CONTRACT: [[FADD:%[0-9]+]]:_(<3 x s64>) = G_FADD [[BUILD_VECTOR2]], [[FMUL]]
     ; GFX10-CONTRACT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<3 x s64>)
     ; GFX10-CONTRACT: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-CONTRACT: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10-CONTRACT: $vgpr2 = COPY [[UV2]](s32)
     ; GFX10-CONTRACT: $vgpr3 = COPY [[UV3]](s32)
     ; GFX10-CONTRACT: $vgpr4 = COPY [[UV4]](s32)
     ; GFX10-CONTRACT: $vgpr5 = COPY [[UV5]](s32)
-    ; GFX10-CONTRACT: [[COPY19:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY18]]
-    ; GFX10-CONTRACT: S_SETPC_B64_return [[COPY19]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
+    ; GFX10-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
     ; GFX10-DENORM-LABEL: name: test_3xdouble_add_mul_rhs
     ; GFX10-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-DENORM: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-DENORM: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-DENORM: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-DENORM: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
bb.1.entry:
     ; GFX10-DENORM: [[COPY14:%[0-9]+]]:_(s32) = COPY $vgpr14
     ; GFX10-DENORM: [[COPY15:%[0-9]+]]:_(s32) = COPY $vgpr15
     ; GFX10-DENORM: [[COPY16:%[0-9]+]]:_(s32) = COPY $vgpr16
     ; GFX10-DENORM: [[COPY17:%[0-9]+]]:_(s32) = COPY $vgpr17
     ; GFX10-DENORM: [[MV6:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY12]](s32), [[COPY13]](s32)
     ; GFX10-DENORM: [[MV7:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY14]](s32), [[COPY15]](s32)
     ; GFX10-DENORM: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX10-DENORM: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s64>) = G_BUILD_VECTOR [[MV6]](s64), [[MV7]](s64), [[MV8]](s64)
-    ; GFX10-DENORM: [[COPY18:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-DENORM: [[FMUL:%[0-9]+]]:_(<3 x s64>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX10-DENORM: [[FADD:%[0-9]+]]:_(<3 x s64>) = G_FADD [[BUILD_VECTOR2]], [[FMUL]]
     ; GFX10-DENORM: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<3 x s64>)
     ; GFX10-DENORM: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-DENORM: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10-DENORM: $vgpr2 = COPY [[UV2]](s32)
     ; GFX10-DENORM: $vgpr3 = COPY [[UV3]](s32)
     ; GFX10-DENORM: $vgpr4 = COPY [[UV4]](s32)
     ; GFX10-DENORM: $vgpr5 = COPY [[UV5]](s32)
-    ; GFX10-DENORM: [[COPY19:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY18]]
-    ; GFX10-DENORM: S_SETPC_B64_return [[COPY19]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
+    ; GFX10-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
     ; GFX10-UNSAFE-LABEL: name: test_3xdouble_add_mul_rhs
     ; GFX10-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-UNSAFE: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-UNSAFE: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-UNSAFE: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-UNSAFE: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
bb.1.entry:
     ; GFX10-UNSAFE: [[COPY14:%[0-9]+]]:_(s32) = COPY $vgpr14
     ; GFX10-UNSAFE: [[COPY15:%[0-9]+]]:_(s32) = COPY $vgpr15
     ; GFX10-UNSAFE: [[COPY16:%[0-9]+]]:_(s32) = COPY $vgpr16
     ; GFX10-UNSAFE: [[COPY17:%[0-9]+]]:_(s32) = COPY $vgpr17
     ; GFX10-UNSAFE: [[MV6:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY12]](s32), [[COPY13]](s32)
     ; GFX10-UNSAFE: [[MV7:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY14]](s32), [[COPY15]](s32)
     ; GFX10-UNSAFE: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX10-UNSAFE: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s64>) = G_BUILD_VECTOR [[MV6]](s64), [[MV7]](s64), [[MV8]](s64)
-    ; GFX10-UNSAFE: [[COPY18:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-UNSAFE: [[FMUL:%[0-9]+]]:_(<3 x s64>) = G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX10-UNSAFE: [[FADD:%[0-9]+]]:_(<3 x s64>) = G_FADD [[BUILD_VECTOR2]], [[FMUL]]
     ; GFX10-UNSAFE: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<3 x s64>)
     ; GFX10-UNSAFE: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-UNSAFE: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10-UNSAFE: $vgpr2 = COPY [[UV2]](s32)
     ; GFX10-UNSAFE: $vgpr3 = COPY [[UV3]](s32)
     ; GFX10-UNSAFE: $vgpr4 = COPY [[UV4]](s32)
     ; GFX10-UNSAFE: $vgpr5 = COPY [[UV5]](s32)
-    ; GFX10-UNSAFE: [[COPY19:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY18]]
-    ; GFX10-UNSAFE: S_SETPC_B64_return [[COPY19]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
+    ; GFX10-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
     %4:_(s32) = COPY $vgpr0
     %5:_(s32) = COPY $vgpr1
     %6:_(s32) = COPY $vgpr2
     %7:_(s32) = COPY $vgpr3
     %8:_(s32) = COPY $vgpr4
     %9:_(s32) = COPY $vgpr5
     %22:_(s64) = G_MERGE_VALUES %4(s32), %5(s32)
     %23:_(s64) = G_MERGE_VALUES %6(s32), %7(s32)
bb.1.entry:
     %18:_(s32) = COPY $vgpr14
     %19:_(s32) = COPY $vgpr15
     %20:_(s32) = COPY $vgpr16
     %21:_(s32) = COPY $vgpr17
     %28:_(s64) = G_MERGE_VALUES %16(s32), %17(s32)
     %29:_(s64) = G_MERGE_VALUES %18(s32), %19(s32)
     %30:_(s64) = G_MERGE_VALUES %20(s32), %21(s32)
     %2:_(<3 x s64>) = G_BUILD_VECTOR %28(s64), %29(s64), %30(s64)
-    %3:sgpr_64 = COPY $sgpr30_sgpr31
     %31:_(<3 x s64>) = G_FMUL %0, %1
     %32:_(<3 x s64>) = G_FADD %2, %31
     %34:_(s32), %35:_(s32), %36:_(s32), %37:_(s32), %38:_(s32), %39:_(s32) = G_UNMERGE_VALUES %32(<3 x s64>)
     $vgpr0 = COPY %34(s32)
     $vgpr1 = COPY %35(s32)
     $vgpr2 = COPY %36(s32)
     $vgpr3 = COPY %37(s32)
     $vgpr4 = COPY %38(s32)
     $vgpr5 = COPY %39(s32)
-    %33:ccr_sgpr_64 = COPY %3
-    S_SETPC_B64_return %33, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
 ...

llvm/test/CodeGen/AMDGPU/GlobalISel/combine-fma-add-mul-pre-legalize.mir

 # NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
 # RUN: llc -global-isel -march=amdgcn -mcpu=gfx900 -run-pass=amdgpu-prelegalizer-combiner %s -o - | FileCheck -check-prefix=GFX9 %s
 # RUN: llc -global-isel -march=amdgcn -mcpu=gfx900 -run-pass=amdgpu-prelegalizer-combiner -fp-contract=fast %s -o - | FileCheck -check-prefix=GFX9-CONTRACT %s
 # RUN: llc -global-isel -march=amdgcn -mcpu=gfx900 -run-pass=amdgpu-prelegalizer-combiner --denormal-fp-math=preserve-sign %s -o - | FileCheck -check-prefix=GFX9-DENORM %s
 # RUN: llc -global-isel -march=amdgcn -mcpu=gfx900 -run-pass=amdgpu-prelegalizer-combiner -enable-unsafe-fp-math %s -o - | FileCheck -check-prefix=GFX9-UNSAFE %s
 # RUN: llc -global-isel -march=amdgcn -mcpu=gfx1010 -run-pass=amdgpu-prelegalizer-combiner %s -o - | FileCheck -check-prefix=GFX10 %s
 # RUN: llc -global-isel -march=amdgcn -mcpu=gfx1010 -run-pass=amdgpu-prelegalizer-combiner -fp-contract=fast %s -o - | FileCheck -check-prefix=GFX10-CONTRACT %s
 # RUN: llc -global-isel -march=amdgcn -mcpu=gfx1010 -run-pass=amdgpu-prelegalizer-combiner --denormal-fp-math=preserve-sign %s -o - | FileCheck -check-prefix=GFX10-DENORM %s
 # RUN: llc -global-isel -march=amdgcn -mcpu=gfx1010 -run-pass=amdgpu-prelegalizer-combiner -enable-unsafe-fp-math %s -o - | FileCheck -check-prefix=GFX10-UNSAFE %s
 
 ---
 name:            test_f32_add_mul
 body:             |
   bb.1.entry:
-    liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
+    liveins: $vgpr0, $vgpr1, $vgpr2
 
     ; GFX9-LABEL: name: test_f32_add_mul
     ; GFX9: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX9: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9: [[FMUL:%[0-9]+]]:_(s32) = reassoc G_FMUL [[COPY]], [[COPY1]]
     ; GFX9: [[FADD:%[0-9]+]]:_(s32) = reassoc G_FADD [[FMUL]], [[COPY2]]
     ; GFX9: $vgpr0 = COPY [[FADD]](s32)
-    ; GFX9: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX9-CONTRACT-LABEL: name: test_f32_add_mul
     ; GFX9-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX9-CONTRACT: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-CONTRACT: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[COPY]], [[COPY1]], [[COPY2]]
     ; GFX9-CONTRACT: $vgpr0 = COPY [[FMA]](s32)
-    ; GFX9-CONTRACT: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9-CONTRACT: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX9-DENORM-LABEL: name: test_f32_add_mul
     ; GFX9-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX9-DENORM: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-DENORM: [[FMUL:%[0-9]+]]:_(s32) = reassoc G_FMUL [[COPY]], [[COPY1]]
     ; GFX9-DENORM: [[FADD:%[0-9]+]]:_(s32) = reassoc G_FADD [[FMUL]], [[COPY2]]
     ; GFX9-DENORM: $vgpr0 = COPY [[FADD]](s32)
-    ; GFX9-DENORM: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9-DENORM: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX9-UNSAFE-LABEL: name: test_f32_add_mul
     ; GFX9-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX9-UNSAFE: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-UNSAFE: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[COPY]], [[COPY1]], [[COPY2]]
     ; GFX9-UNSAFE: $vgpr0 = COPY [[FMA]](s32)
-    ; GFX9-UNSAFE: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9-UNSAFE: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-LABEL: name: test_f32_add_mul
     ; GFX10: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX10: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10: [[FMUL:%[0-9]+]]:_(s32) = reassoc G_FMUL [[COPY]], [[COPY1]]
     ; GFX10: [[FADD:%[0-9]+]]:_(s32) = reassoc G_FADD [[FMUL]], [[COPY2]]
     ; GFX10: $vgpr0 = COPY [[FADD]](s32)
-    ; GFX10: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-CONTRACT-LABEL: name: test_f32_add_mul
     ; GFX10-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX10-CONTRACT: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-CONTRACT: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[COPY]], [[COPY1]], [[COPY2]]
     ; GFX10-CONTRACT: $vgpr0 = COPY [[FMA]](s32)
-    ; GFX10-CONTRACT: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10-CONTRACT: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-DENORM-LABEL: name: test_f32_add_mul
     ; GFX10-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX10-DENORM: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-DENORM: [[FMUL:%[0-9]+]]:_(s32) = reassoc G_FMUL [[COPY]], [[COPY1]]
     ; GFX10-DENORM: [[FADD:%[0-9]+]]:_(s32) = reassoc G_FADD [[FMUL]], [[COPY2]]
     ; GFX10-DENORM: $vgpr0 = COPY [[FADD]](s32)
-    ; GFX10-DENORM: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10-DENORM: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-UNSAFE-LABEL: name: test_f32_add_mul
     ; GFX10-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX10-UNSAFE: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-UNSAFE: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[COPY]], [[COPY1]], [[COPY2]]
     ; GFX10-UNSAFE: $vgpr0 = COPY [[FMA]](s32)
-    ; GFX10-UNSAFE: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10-UNSAFE: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     %0:_(s32) = COPY $vgpr0
     %1:_(s32) = COPY $vgpr1
     %2:_(s32) = COPY $vgpr2
-    %3:sgpr_64 = COPY $sgpr30_sgpr31
     %4:_(s32) = reassoc G_FMUL %0, %1
     %5:_(s32) = reassoc G_FADD %4, %2
     $vgpr0 = COPY %5(s32)
-    %6:ccr_sgpr_64 = COPY %3
-    S_SETPC_B64_return %6, implicit $vgpr0
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
 ...
 
 ---
 name:            test_f32_add_mul_rhs
 body:             |
   bb.1.entry:
-    liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
+    liveins: $vgpr0, $vgpr1, $vgpr2
 
     ; GFX9-LABEL: name: test_f32_add_mul_rhs
     ; GFX9: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX9: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9: [[FMUL:%[0-9]+]]:_(s32) = reassoc G_FMUL [[COPY]], [[COPY1]]
     ; GFX9: [[FADD:%[0-9]+]]:_(s32) = reassoc G_FADD [[COPY2]], [[FMUL]]
     ; GFX9: $vgpr0 = COPY [[FADD]](s32)
-    ; GFX9: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX9-CONTRACT-LABEL: name: test_f32_add_mul_rhs
     ; GFX9-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX9-CONTRACT: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-CONTRACT: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[COPY]], [[COPY1]], [[COPY2]]
     ; GFX9-CONTRACT: $vgpr0 = COPY [[FMA]](s32)
-    ; GFX9-CONTRACT: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9-CONTRACT: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX9-DENORM-LABEL: name: test_f32_add_mul_rhs
     ; GFX9-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX9-DENORM: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-DENORM: [[FMUL:%[0-9]+]]:_(s32) = reassoc G_FMUL [[COPY]], [[COPY1]]
     ; GFX9-DENORM: [[FADD:%[0-9]+]]:_(s32) = reassoc G_FADD [[COPY2]], [[FMUL]]
     ; GFX9-DENORM: $vgpr0 = COPY [[FADD]](s32)
-    ; GFX9-DENORM: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9-DENORM: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX9-UNSAFE-LABEL: name: test_f32_add_mul_rhs
     ; GFX9-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX9-UNSAFE: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-UNSAFE: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[COPY]], [[COPY1]], [[COPY2]]
     ; GFX9-UNSAFE: $vgpr0 = COPY [[FMA]](s32)
-    ; GFX9-UNSAFE: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9-UNSAFE: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-LABEL: name: test_f32_add_mul_rhs
     ; GFX10: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX10: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10: [[FMUL:%[0-9]+]]:_(s32) = reassoc G_FMUL [[COPY]], [[COPY1]]
     ; GFX10: [[FADD:%[0-9]+]]:_(s32) = reassoc G_FADD [[COPY2]], [[FMUL]]
     ; GFX10: $vgpr0 = COPY [[FADD]](s32)
-    ; GFX10: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-CONTRACT-LABEL: name: test_f32_add_mul_rhs
     ; GFX10-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX10-CONTRACT: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-CONTRACT: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[COPY]], [[COPY1]], [[COPY2]]
     ; GFX10-CONTRACT: $vgpr0 = COPY [[FMA]](s32)
-    ; GFX10-CONTRACT: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10-CONTRACT: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-DENORM-LABEL: name: test_f32_add_mul_rhs
     ; GFX10-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX10-DENORM: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-DENORM: [[FMUL:%[0-9]+]]:_(s32) = reassoc G_FMUL [[COPY]], [[COPY1]]
     ; GFX10-DENORM: [[FADD:%[0-9]+]]:_(s32) = reassoc G_FADD [[COPY2]], [[FMUL]]
     ; GFX10-DENORM: $vgpr0 = COPY [[FADD]](s32)
-    ; GFX10-DENORM: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10-DENORM: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-UNSAFE-LABEL: name: test_f32_add_mul_rhs
     ; GFX10-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-    ; GFX10-UNSAFE: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-UNSAFE: [[FMA:%[0-9]+]]:_(s32) = G_FMA [[COPY]], [[COPY1]], [[COPY2]]
     ; GFX10-UNSAFE: $vgpr0 = COPY [[FMA]](s32)
-    ; GFX10-UNSAFE: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10-UNSAFE: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     %0:_(s32) = COPY $vgpr0
     %1:_(s32) = COPY $vgpr1
     %2:_(s32) = COPY $vgpr2
-    %3:sgpr_64 = COPY $sgpr30_sgpr31
     %4:_(s32) = reassoc G_FMUL %0, %1
     %5:_(s32) = reassoc G_FADD %2, %4
     $vgpr0 = COPY %5(s32)
-    %6:ccr_sgpr_64 = COPY %3
-    S_SETPC_B64_return %6, implicit $vgpr0
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
 ...
 
 ---
 name:            test_half_add_mul
 body:             |
   bb.1.entry:
-    liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
+    liveins: $vgpr0, $vgpr1, $vgpr2
 
     ; GFX9-LABEL: name: test_half_add_mul
     ; GFX9: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX9: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX9: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX9: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9: [[FMUL:%[0-9]+]]:_(s16) = reassoc G_FMUL [[TRUNC]], [[TRUNC1]]
     ; GFX9: [[FADD:%[0-9]+]]:_(s16) = reassoc G_FADD [[FMUL]], [[TRUNC2]]
     ; GFX9: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD]](s16)
     ; GFX9: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX9: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX9-CONTRACT-LABEL: name: test_half_add_mul
     ; GFX9-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-CONTRACT: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX9-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-CONTRACT: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX9-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-CONTRACT: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX9-CONTRACT: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-CONTRACT: [[FMA:%[0-9]+]]:_(s16) = G_FMA [[TRUNC]], [[TRUNC1]], [[TRUNC2]]
     ; GFX9-CONTRACT: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FMA]](s16)
     ; GFX9-CONTRACT: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX9-CONTRACT: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9-CONTRACT: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX9-DENORM-LABEL: name: test_half_add_mul
     ; GFX9-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-DENORM: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX9-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-DENORM: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX9-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-DENORM: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX9-DENORM: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-DENORM: [[FMUL:%[0-9]+]]:_(s16) = reassoc G_FMUL [[TRUNC]], [[TRUNC1]]
     ; GFX9-DENORM: [[FADD:%[0-9]+]]:_(s16) = reassoc G_FADD [[FMUL]], [[TRUNC2]]
     ; GFX9-DENORM: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD]](s16)
     ; GFX9-DENORM: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX9-DENORM: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9-DENORM: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX9-UNSAFE-LABEL: name: test_half_add_mul
     ; GFX9-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-UNSAFE: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX9-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-UNSAFE: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX9-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-UNSAFE: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX9-UNSAFE: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-UNSAFE: [[FMA:%[0-9]+]]:_(s16) = G_FMA [[TRUNC]], [[TRUNC1]], [[TRUNC2]]
     ; GFX9-UNSAFE: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FMA]](s16)
     ; GFX9-UNSAFE: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX9-UNSAFE: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9-UNSAFE: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-LABEL: name: test_half_add_mul
     ; GFX10: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX10: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX10: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX10: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10: [[FMUL:%[0-9]+]]:_(s16) = reassoc G_FMUL [[TRUNC]], [[TRUNC1]]
     ; GFX10: [[FADD:%[0-9]+]]:_(s16) = reassoc G_FADD [[FMUL]], [[TRUNC2]]
     ; GFX10: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD]](s16)
     ; GFX10: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX10: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-CONTRACT-LABEL: name: test_half_add_mul
     ; GFX10-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-CONTRACT: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX10-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-CONTRACT: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX10-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-CONTRACT: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX10-CONTRACT: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-CONTRACT: [[FMA:%[0-9]+]]:_(s16) = G_FMA [[TRUNC]], [[TRUNC1]], [[TRUNC2]]
     ; GFX10-CONTRACT: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FMA]](s16)
     ; GFX10-CONTRACT: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX10-CONTRACT: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10-CONTRACT: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-DENORM-LABEL: name: test_half_add_mul
     ; GFX10-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-DENORM: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX10-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-DENORM: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX10-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-DENORM: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX10-DENORM: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-DENORM: [[FMUL:%[0-9]+]]:_(s16) = reassoc G_FMUL [[TRUNC]], [[TRUNC1]]
     ; GFX10-DENORM: [[FADD:%[0-9]+]]:_(s16) = reassoc G_FADD [[FMUL]], [[TRUNC2]]
     ; GFX10-DENORM: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD]](s16)
     ; GFX10-DENORM: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX10-DENORM: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10-DENORM: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-UNSAFE-LABEL: name: test_half_add_mul
     ; GFX10-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-UNSAFE: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX10-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-UNSAFE: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX10-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-UNSAFE: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX10-UNSAFE: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-UNSAFE: [[FMA:%[0-9]+]]:_(s16) = G_FMA [[TRUNC]], [[TRUNC1]], [[TRUNC2]]
     ; GFX10-UNSAFE: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FMA]](s16)
     ; GFX10-UNSAFE: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX10-UNSAFE: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10-UNSAFE: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     %4:_(s32) = COPY $vgpr0
     %0:_(s16) = G_TRUNC %4(s32)
     %5:_(s32) = COPY $vgpr1
     %1:_(s16) = G_TRUNC %5(s32)
     %6:_(s32) = COPY $vgpr2
     %2:_(s16) = G_TRUNC %6(s32)
-    %3:sgpr_64 = COPY $sgpr30_sgpr31
     %7:_(s16) = reassoc G_FMUL %0, %1
     %8:_(s16) = reassoc G_FADD %7, %2
     %10:_(s32) = G_ANYEXT %8(s16)
     $vgpr0 = COPY %10(s32)
-    %9:ccr_sgpr_64 = COPY %3
-    S_SETPC_B64_return %9, implicit $vgpr0
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
 ...
 
 ---
 name:            test_half_add_mul_rhs
 body:             |
   bb.1.entry:
-    liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
+    liveins: $vgpr0, $vgpr1, $vgpr2
 
     ; GFX9-LABEL: name: test_half_add_mul_rhs
     ; GFX9: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX9: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX9: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX9: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9: [[FMUL:%[0-9]+]]:_(s16) = reassoc G_FMUL [[TRUNC]], [[TRUNC1]]
     ; GFX9: [[FADD:%[0-9]+]]:_(s16) = reassoc G_FADD [[TRUNC2]], [[FMUL]]
     ; GFX9: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD]](s16)
     ; GFX9: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX9: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX9-CONTRACT-LABEL: name: test_half_add_mul_rhs
     ; GFX9-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-CONTRACT: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX9-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-CONTRACT: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX9-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-CONTRACT: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX9-CONTRACT: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-CONTRACT: [[FMA:%[0-9]+]]:_(s16) = G_FMA [[TRUNC]], [[TRUNC1]], [[TRUNC2]]
     ; GFX9-CONTRACT: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FMA]](s16)
     ; GFX9-CONTRACT: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX9-CONTRACT: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9-CONTRACT: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX9-DENORM-LABEL: name: test_half_add_mul_rhs
     ; GFX9-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-DENORM: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX9-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-DENORM: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX9-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-DENORM: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX9-DENORM: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-DENORM: [[FMUL:%[0-9]+]]:_(s16) = reassoc G_FMUL [[TRUNC]], [[TRUNC1]]
     ; GFX9-DENORM: [[FADD:%[0-9]+]]:_(s16) = reassoc G_FADD [[TRUNC2]], [[FMUL]]
     ; GFX9-DENORM: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD]](s16)
     ; GFX9-DENORM: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX9-DENORM: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9-DENORM: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX9-UNSAFE-LABEL: name: test_half_add_mul_rhs
     ; GFX9-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-UNSAFE: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX9-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-UNSAFE: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX9-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-UNSAFE: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX9-UNSAFE: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-UNSAFE: [[FMA:%[0-9]+]]:_(s16) = G_FMA [[TRUNC]], [[TRUNC1]], [[TRUNC2]]
     ; GFX9-UNSAFE: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FMA]](s16)
     ; GFX9-UNSAFE: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX9-UNSAFE: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX9-UNSAFE: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX9-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-LABEL: name: test_half_add_mul_rhs
     ; GFX10: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX10: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX10: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX10: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10: [[FMUL:%[0-9]+]]:_(s16) = reassoc G_FMUL [[TRUNC]], [[TRUNC1]]
     ; GFX10: [[FADD:%[0-9]+]]:_(s16) = reassoc G_FADD [[TRUNC2]], [[FMUL]]
     ; GFX10: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD]](s16)
     ; GFX10: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX10: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-CONTRACT-LABEL: name: test_half_add_mul_rhs
     ; GFX10-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-CONTRACT: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX10-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-CONTRACT: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX10-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-CONTRACT: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX10-CONTRACT: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-CONTRACT: [[FMA:%[0-9]+]]:_(s16) = G_FMA [[TRUNC]], [[TRUNC1]], [[TRUNC2]]
     ; GFX10-CONTRACT: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FMA]](s16)
     ; GFX10-CONTRACT: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX10-CONTRACT: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10-CONTRACT: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-DENORM-LABEL: name: test_half_add_mul_rhs
     ; GFX10-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-DENORM: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX10-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-DENORM: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX10-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-DENORM: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX10-DENORM: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-DENORM: [[FMUL:%[0-9]+]]:_(s16) = reassoc G_FMUL [[TRUNC]], [[TRUNC1]]
     ; GFX10-DENORM: [[FADD:%[0-9]+]]:_(s16) = reassoc G_FADD [[TRUNC2]], [[FMUL]]
     ; GFX10-DENORM: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FADD]](s16)
     ; GFX10-DENORM: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX10-DENORM: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10-DENORM: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     ; GFX10-UNSAFE-LABEL: name: test_half_add_mul_rhs
     ; GFX10-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-UNSAFE: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
     ; GFX10-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-UNSAFE: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
     ; GFX10-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-UNSAFE: [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-    ; GFX10-UNSAFE: [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-UNSAFE: [[FMA:%[0-9]+]]:_(s16) = G_FMA [[TRUNC]], [[TRUNC1]], [[TRUNC2]]
     ; GFX10-UNSAFE: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[FMA]](s16)
     ; GFX10-UNSAFE: $vgpr0 = COPY [[ANYEXT]](s32)
-    ; GFX10-UNSAFE: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-    ; GFX10-UNSAFE: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; GFX10-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     %4:_(s32) = COPY $vgpr0
     %0:_(s16) = G_TRUNC %4(s32)
     %5:_(s32) = COPY $vgpr1
     %1:_(s16) = G_TRUNC %5(s32)
     %6:_(s32) = COPY $vgpr2
     %2:_(s16) = G_TRUNC %6(s32)
-    %3:sgpr_64 = COPY $sgpr30_sgpr31
     %7:_(s16) = reassoc G_FMUL %0, %1
     %8:_(s16) = reassoc G_FADD %2, %7
     %10:_(s32) = G_ANYEXT %8(s16)
     $vgpr0 = COPY %10(s32)
-    %9:ccr_sgpr_64 = COPY %3
-    S_SETPC_B64_return %9, implicit $vgpr0
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
 ...
 
 ---
 name:            test_double_add_mul
 body:             |
   bb.1.entry:
-    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $sgpr30_sgpr31
+    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
 
     ; GFX9-LABEL: name: test_double_add_mul
     ; GFX9: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX9: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX9: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX9: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9: [[FMUL:%[0-9]+]]:_(s64) = reassoc G_FMUL [[MV]], [[MV1]]
     ; GFX9: [[FADD:%[0-9]+]]:_(s64) = reassoc G_FADD [[FMUL]], [[MV2]]
     ; GFX9: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](s64)
     ; GFX9: $vgpr0 = COPY [[UV]](s32)
     ; GFX9: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX9: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX9-CONTRACT-LABEL: name: test_double_add_mul
     ; GFX9-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-CONTRACT: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX9-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-CONTRACT: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-CONTRACT: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX9-CONTRACT: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-CONTRACT: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-CONTRACT: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX9-CONTRACT: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-CONTRACT: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[MV]], [[MV1]], [[MV2]]
     ; GFX9-CONTRACT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](s64)
     ; GFX9-CONTRACT: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-CONTRACT: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX9-CONTRACT: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9-CONTRACT: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX9-DENORM-LABEL: name: test_double_add_mul
     ; GFX9-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-DENORM: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX9-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-DENORM: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-DENORM: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX9-DENORM: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-DENORM: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-DENORM: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX9-DENORM: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-DENORM: [[FMUL:%[0-9]+]]:_(s64) = reassoc G_FMUL [[MV]], [[MV1]]
     ; GFX9-DENORM: [[FADD:%[0-9]+]]:_(s64) = reassoc G_FADD [[FMUL]], [[MV2]]
     ; GFX9-DENORM: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](s64)
     ; GFX9-DENORM: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-DENORM: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX9-DENORM: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9-DENORM: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX9-UNSAFE-LABEL: name: test_double_add_mul
     ; GFX9-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-UNSAFE: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX9-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-UNSAFE: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-UNSAFE: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX9-UNSAFE: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-UNSAFE: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-UNSAFE: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX9-UNSAFE: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-UNSAFE: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[MV]], [[MV1]], [[MV2]]
     ; GFX9-UNSAFE: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](s64)
     ; GFX9-UNSAFE: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-UNSAFE: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX9-UNSAFE: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9-UNSAFE: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-LABEL: name: test_double_add_mul
     ; GFX10: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX10: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX10: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX10: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10: [[FMUL:%[0-9]+]]:_(s64) = reassoc G_FMUL [[MV]], [[MV1]]
     ; GFX10: [[FADD:%[0-9]+]]:_(s64) = reassoc G_FADD [[FMUL]], [[MV2]]
     ; GFX10: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](s64)
     ; GFX10: $vgpr0 = COPY [[UV]](s32)
     ; GFX10: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX10: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-CONTRACT-LABEL: name: test_double_add_mul
     ; GFX10-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-CONTRACT: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX10-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-CONTRACT: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-CONTRACT: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX10-CONTRACT: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-CONTRACT: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-CONTRACT: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX10-CONTRACT: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-CONTRACT: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[MV]], [[MV1]], [[MV2]]
     ; GFX10-CONTRACT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](s64)
     ; GFX10-CONTRACT: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-CONTRACT: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX10-CONTRACT: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10-CONTRACT: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-DENORM-LABEL: name: test_double_add_mul
     ; GFX10-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-DENORM: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX10-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-DENORM: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-DENORM: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX10-DENORM: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-DENORM: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-DENORM: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX10-DENORM: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-DENORM: [[FMUL:%[0-9]+]]:_(s64) = reassoc G_FMUL [[MV]], [[MV1]]
     ; GFX10-DENORM: [[FADD:%[0-9]+]]:_(s64) = reassoc G_FADD [[FMUL]], [[MV2]]
     ; GFX10-DENORM: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](s64)
     ; GFX10-DENORM: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-DENORM: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX10-DENORM: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10-DENORM: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-UNSAFE-LABEL: name: test_double_add_mul
     ; GFX10-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-UNSAFE: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX10-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-UNSAFE: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-UNSAFE: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX10-UNSAFE: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-UNSAFE: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-UNSAFE: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX10-UNSAFE: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-UNSAFE: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[MV]], [[MV1]], [[MV2]]
     ; GFX10-UNSAFE: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](s64)
     ; GFX10-UNSAFE: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-UNSAFE: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX10-UNSAFE: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10-UNSAFE: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     %4:_(s32) = COPY $vgpr0
     %5:_(s32) = COPY $vgpr1
     %0:_(s64) = G_MERGE_VALUES %4(s32), %5(s32)
     %6:_(s32) = COPY $vgpr2
     %7:_(s32) = COPY $vgpr3
     %1:_(s64) = G_MERGE_VALUES %6(s32), %7(s32)
     %8:_(s32) = COPY $vgpr4
     %9:_(s32) = COPY $vgpr5
     %2:_(s64) = G_MERGE_VALUES %8(s32), %9(s32)
-    %3:sgpr_64 = COPY $sgpr30_sgpr31
     %10:_(s64) = reassoc G_FMUL %0, %1
     %11:_(s64) = reassoc G_FADD %10, %2
     %13:_(s32), %14:_(s32) = G_UNMERGE_VALUES %11(s64)
     $vgpr0 = COPY %13(s32)
     $vgpr1 = COPY %14(s32)
-    %12:ccr_sgpr_64 = COPY %3
-    S_SETPC_B64_return %12, implicit $vgpr0, implicit $vgpr1
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
 ...
 
 ---
 name:            test_double_add_mul_rhs
 body:             |
   bb.1.entry:
-    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $sgpr30_sgpr31
+    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
 
     ; GFX9-LABEL: name: test_double_add_mul_rhs
     ; GFX9: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX9: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX9: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX9: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9: [[FMUL:%[0-9]+]]:_(s64) = reassoc G_FMUL [[MV]], [[MV1]]
     ; GFX9: [[FADD:%[0-9]+]]:_(s64) = reassoc G_FADD [[MV2]], [[FMUL]]
     ; GFX9: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](s64)
     ; GFX9: $vgpr0 = COPY [[UV]](s32)
     ; GFX9: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX9: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX9-CONTRACT-LABEL: name: test_double_add_mul_rhs
     ; GFX9-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-CONTRACT: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX9-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-CONTRACT: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-CONTRACT: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX9-CONTRACT: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-CONTRACT: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-CONTRACT: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX9-CONTRACT: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-CONTRACT: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[MV]], [[MV1]], [[MV2]]
     ; GFX9-CONTRACT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](s64)
     ; GFX9-CONTRACT: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-CONTRACT: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX9-CONTRACT: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9-CONTRACT: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX9-DENORM-LABEL: name: test_double_add_mul_rhs
     ; GFX9-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-DENORM: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX9-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-DENORM: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-DENORM: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX9-DENORM: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-DENORM: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-DENORM: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX9-DENORM: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-DENORM: [[FMUL:%[0-9]+]]:_(s64) = reassoc G_FMUL [[MV]], [[MV1]]
     ; GFX9-DENORM: [[FADD:%[0-9]+]]:_(s64) = reassoc G_FADD [[MV2]], [[FMUL]]
     ; GFX9-DENORM: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](s64)
     ; GFX9-DENORM: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-DENORM: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX9-DENORM: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9-DENORM: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX9-UNSAFE-LABEL: name: test_double_add_mul_rhs
     ; GFX9-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-UNSAFE: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX9-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-UNSAFE: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-UNSAFE: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX9-UNSAFE: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-UNSAFE: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-UNSAFE: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX9-UNSAFE: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-UNSAFE: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[MV]], [[MV1]], [[MV2]]
     ; GFX9-UNSAFE: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](s64)
     ; GFX9-UNSAFE: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-UNSAFE: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX9-UNSAFE: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9-UNSAFE: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-LABEL: name: test_double_add_mul_rhs
     ; GFX10: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX10: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX10: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX10: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10: [[FMUL:%[0-9]+]]:_(s64) = reassoc G_FMUL [[MV]], [[MV1]]
     ; GFX10: [[FADD:%[0-9]+]]:_(s64) = reassoc G_FADD [[MV2]], [[FMUL]]
     ; GFX10: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](s64)
     ; GFX10: $vgpr0 = COPY [[UV]](s32)
     ; GFX10: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX10: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-CONTRACT-LABEL: name: test_double_add_mul_rhs
     ; GFX10-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-CONTRACT: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX10-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-CONTRACT: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-CONTRACT: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX10-CONTRACT: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-CONTRACT: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-CONTRACT: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX10-CONTRACT: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-CONTRACT: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[MV]], [[MV1]], [[MV2]]
     ; GFX10-CONTRACT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](s64)
     ; GFX10-CONTRACT: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-CONTRACT: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX10-CONTRACT: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10-CONTRACT: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-DENORM-LABEL: name: test_double_add_mul_rhs
     ; GFX10-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-DENORM: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX10-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-DENORM: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-DENORM: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX10-DENORM: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-DENORM: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-DENORM: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX10-DENORM: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-DENORM: [[FMUL:%[0-9]+]]:_(s64) = reassoc G_FMUL [[MV]], [[MV1]]
     ; GFX10-DENORM: [[FADD:%[0-9]+]]:_(s64) = reassoc G_FADD [[MV2]], [[FMUL]]
     ; GFX10-DENORM: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](s64)
     ; GFX10-DENORM: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-DENORM: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX10-DENORM: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10-DENORM: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-UNSAFE-LABEL: name: test_double_add_mul_rhs
     ; GFX10-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-UNSAFE: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; GFX10-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-UNSAFE: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-UNSAFE: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
     ; GFX10-UNSAFE: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-UNSAFE: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-UNSAFE: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-    ; GFX10-UNSAFE: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-UNSAFE: [[FMA:%[0-9]+]]:_(s64) = G_FMA [[MV]], [[MV1]], [[MV2]]
     ; GFX10-UNSAFE: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](s64)
     ; GFX10-UNSAFE: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-UNSAFE: $vgpr1 = COPY [[UV1]](s32)
-    ; GFX10-UNSAFE: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10-UNSAFE: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     %4:_(s32) = COPY $vgpr0
     %5:_(s32) = COPY $vgpr1
     %0:_(s64) = G_MERGE_VALUES %4(s32), %5(s32)
     %6:_(s32) = COPY $vgpr2
     %7:_(s32) = COPY $vgpr3
     %1:_(s64) = G_MERGE_VALUES %6(s32), %7(s32)
     %8:_(s32) = COPY $vgpr4
     %9:_(s32) = COPY $vgpr5
     %2:_(s64) = G_MERGE_VALUES %8(s32), %9(s32)
-    %3:sgpr_64 = COPY $sgpr30_sgpr31
     %10:_(s64) = reassoc G_FMUL %0, %1
     %11:_(s64) = reassoc G_FADD %2, %10
     %13:_(s32), %14:_(s32) = G_UNMERGE_VALUES %11(s64)
     $vgpr0 = COPY %13(s32)
     $vgpr1 = COPY %14(s32)
-    %12:ccr_sgpr_64 = COPY %3
-    S_SETPC_B64_return %12, implicit $vgpr0, implicit $vgpr1
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
 ...
 
 ---
 name:            test_4xfloat_add_mul
 body:             |
   bb.1.entry:
-    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7, $vgpr8, $vgpr9, $vgpr10, $vgpr11, $sgpr30_sgpr31
+    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7, $vgpr8, $vgpr9, $vgpr10, $vgpr11
 
     ; GFX9-LABEL: name: test_4xfloat_add_mul
     ; GFX9: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9: [[BUILD_VECTOR:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
     ; GFX9: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX9: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX9: [[BUILD_VECTOR1:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY4]](s32), [[COPY5]](s32), [[COPY6]](s32), [[COPY7]](s32)
     ; GFX9: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX9: [[COPY9:%[0-9]+]]:_(s32) = COPY $vgpr9
     ; GFX9: [[COPY10:%[0-9]+]]:_(s32) = COPY $vgpr10
     ; GFX9: [[COPY11:%[0-9]+]]:_(s32) = COPY $vgpr11
     ; GFX9: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY8]](s32), [[COPY9]](s32), [[COPY10]](s32), [[COPY11]](s32)
-    ; GFX9: [[COPY12:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9: [[FMUL:%[0-9]+]]:_(<4 x s32>) = reassoc G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX9: [[FADD:%[0-9]+]]:_(<4 x s32>) = reassoc G_FADD [[FMUL]], [[BUILD_VECTOR2]]
     ; GFX9: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<4 x s32>)
     ; GFX9: $vgpr0 = COPY [[UV]](s32)
     ; GFX9: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9: $vgpr2 = COPY [[UV2]](s32)
     ; GFX9: $vgpr3 = COPY [[UV3]](s32)
-    ; GFX9: [[COPY13:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY12]]
-    ; GFX9: S_SETPC_B64_return [[COPY13]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+    ; GFX9: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
     ; GFX9-CONTRACT-LABEL: name: test_4xfloat_add_mul
     ; GFX9-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-CONTRACT: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-CONTRACT: [[BUILD_VECTOR:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
     ; GFX9-CONTRACT: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-CONTRACT: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-CONTRACT: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX9-CONTRACT: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX9-CONTRACT: [[BUILD_VECTOR1:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY4]](s32), [[COPY5]](s32), [[COPY6]](s32), [[COPY7]](s32)
     ; GFX9-CONTRACT: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX9-CONTRACT: [[COPY9:%[0-9]+]]:_(s32) = COPY $vgpr9
     ; GFX9-CONTRACT: [[COPY10:%[0-9]+]]:_(s32) = COPY $vgpr10
     ; GFX9-CONTRACT: [[COPY11:%[0-9]+]]:_(s32) = COPY $vgpr11
     ; GFX9-CONTRACT: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY8]](s32), [[COPY9]](s32), [[COPY10]](s32), [[COPY11]](s32)
-    ; GFX9-CONTRACT: [[COPY12:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-CONTRACT: [[FMA:%[0-9]+]]:_(<4 x s32>) = G_FMA [[BUILD_VECTOR]], [[BUILD_VECTOR1]], [[BUILD_VECTOR2]]
     ; GFX9-CONTRACT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](<4 x s32>)
     ; GFX9-CONTRACT: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-CONTRACT: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9-CONTRACT: $vgpr2 = COPY [[UV2]](s32)
     ; GFX9-CONTRACT: $vgpr3 = COPY [[UV3]](s32)
-    ; GFX9-CONTRACT: [[COPY13:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY12]]
-    ; GFX9-CONTRACT: S_SETPC_B64_return [[COPY13]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+    ; GFX9-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
     ; GFX9-DENORM-LABEL: name: test_4xfloat_add_mul
     ; GFX9-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-DENORM: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-DENORM: [[BUILD_VECTOR:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
     ; GFX9-DENORM: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-DENORM: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-DENORM: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX9-DENORM: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX9-DENORM: [[BUILD_VECTOR1:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY4]](s32), [[COPY5]](s32), [[COPY6]](s32), [[COPY7]](s32)
     ; GFX9-DENORM: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX9-DENORM: [[COPY9:%[0-9]+]]:_(s32) = COPY $vgpr9
     ; GFX9-DENORM: [[COPY10:%[0-9]+]]:_(s32) = COPY $vgpr10
     ; GFX9-DENORM: [[COPY11:%[0-9]+]]:_(s32) = COPY $vgpr11
     ; GFX9-DENORM: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY8]](s32), [[COPY9]](s32), [[COPY10]](s32), [[COPY11]](s32)
-    ; GFX9-DENORM: [[COPY12:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-DENORM: [[FMUL:%[0-9]+]]:_(<4 x s32>) = reassoc G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX9-DENORM: [[FADD:%[0-9]+]]:_(<4 x s32>) = reassoc G_FADD [[FMUL]], [[BUILD_VECTOR2]]
     ; GFX9-DENORM: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<4 x s32>)
     ; GFX9-DENORM: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-DENORM: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9-DENORM: $vgpr2 = COPY [[UV2]](s32)
     ; GFX9-DENORM: $vgpr3 = COPY [[UV3]](s32)
-    ; GFX9-DENORM: [[COPY13:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY12]]
-    ; GFX9-DENORM: S_SETPC_B64_return [[COPY13]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+    ; GFX9-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
     ; GFX9-UNSAFE-LABEL: name: test_4xfloat_add_mul
     ; GFX9-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-UNSAFE: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-UNSAFE: [[BUILD_VECTOR:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
     ; GFX9-UNSAFE: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-UNSAFE: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-UNSAFE: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX9-UNSAFE: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX9-UNSAFE: [[BUILD_VECTOR1:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY4]](s32), [[COPY5]](s32), [[COPY6]](s32), [[COPY7]](s32)
     ; GFX9-UNSAFE: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX9-UNSAFE: [[COPY9:%[0-9]+]]:_(s32) = COPY $vgpr9
     ; GFX9-UNSAFE: [[COPY10:%[0-9]+]]:_(s32) = COPY $vgpr10
     ; GFX9-UNSAFE: [[COPY11:%[0-9]+]]:_(s32) = COPY $vgpr11
     ; GFX9-UNSAFE: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY8]](s32), [[COPY9]](s32), [[COPY10]](s32), [[COPY11]](s32)
-    ; GFX9-UNSAFE: [[COPY12:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-UNSAFE: [[FMA:%[0-9]+]]:_(<4 x s32>) = G_FMA [[BUILD_VECTOR]], [[BUILD_VECTOR1]], [[BUILD_VECTOR2]]
     ; GFX9-UNSAFE: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](<4 x s32>)
     ; GFX9-UNSAFE: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-UNSAFE: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9-UNSAFE: $vgpr2 = COPY [[UV2]](s32)
     ; GFX9-UNSAFE: $vgpr3 = COPY [[UV3]](s32)
-    ; GFX9-UNSAFE: [[COPY13:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY12]]
-    ; GFX9-UNSAFE: S_SETPC_B64_return [[COPY13]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+    ; GFX9-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
     ; GFX10-LABEL: name: test_4xfloat_add_mul
     ; GFX10: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10: [[BUILD_VECTOR:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
     ; GFX10: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX10: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX10: [[BUILD_VECTOR1:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY4]](s32), [[COPY5]](s32), [[COPY6]](s32), [[COPY7]](s32)
     ; GFX10: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX10: [[COPY9:%[0-9]+]]:_(s32) = COPY $vgpr9
     ; GFX10: [[COPY10:%[0-9]+]]:_(s32) = COPY $vgpr10
     ; GFX10: [[COPY11:%[0-9]+]]:_(s32) = COPY $vgpr11
     ; GFX10: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY8]](s32), [[COPY9]](s32), [[COPY10]](s32), [[COPY11]](s32)
-    ; GFX10: [[COPY12:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10: [[FMUL:%[0-9]+]]:_(<4 x s32>) = reassoc G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX10: [[FADD:%[0-9]+]]:_(<4 x s32>) = reassoc G_FADD [[FMUL]], [[BUILD_VECTOR2]]
     ; GFX10: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<4 x s32>)
     ; GFX10: $vgpr0 = COPY [[UV]](s32)
     ; GFX10: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10: $vgpr2 = COPY [[UV2]](s32)
     ; GFX10: $vgpr3 = COPY [[UV3]](s32)
-    ; GFX10: [[COPY13:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY12]]
-    ; GFX10: S_SETPC_B64_return [[COPY13]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+    ; GFX10: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
     ; GFX10-CONTRACT-LABEL: name: test_4xfloat_add_mul
     ; GFX10-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-CONTRACT: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-CONTRACT: [[BUILD_VECTOR:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
     ; GFX10-CONTRACT: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-CONTRACT: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-CONTRACT: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX10-CONTRACT: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX10-CONTRACT: [[BUILD_VECTOR1:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY4]](s32), [[COPY5]](s32), [[COPY6]](s32), [[COPY7]](s32)
     ; GFX10-CONTRACT: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX10-CONTRACT: [[COPY9:%[0-9]+]]:_(s32) = COPY $vgpr9
     ; GFX10-CONTRACT: [[COPY10:%[0-9]+]]:_(s32) = COPY $vgpr10
     ; GFX10-CONTRACT: [[COPY11:%[0-9]+]]:_(s32) = COPY $vgpr11
     ; GFX10-CONTRACT: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY8]](s32), [[COPY9]](s32), [[COPY10]](s32), [[COPY11]](s32)
-    ; GFX10-CONTRACT: [[COPY12:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-CONTRACT: [[FMA:%[0-9]+]]:_(<4 x s32>) = G_FMA [[BUILD_VECTOR]], [[BUILD_VECTOR1]], [[BUILD_VECTOR2]]
     ; GFX10-CONTRACT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](<4 x s32>)
     ; GFX10-CONTRACT: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-CONTRACT: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10-CONTRACT: $vgpr2 = COPY [[UV2]](s32)
     ; GFX10-CONTRACT: $vgpr3 = COPY [[UV3]](s32)
-    ; GFX10-CONTRACT: [[COPY13:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY12]]
-    ; GFX10-CONTRACT: S_SETPC_B64_return [[COPY13]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+    ; GFX10-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
     ; GFX10-DENORM-LABEL: name: test_4xfloat_add_mul
     ; GFX10-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-DENORM: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-DENORM: [[BUILD_VECTOR:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
     ; GFX10-DENORM: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-DENORM: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-DENORM: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX10-DENORM: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX10-DENORM: [[BUILD_VECTOR1:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY4]](s32), [[COPY5]](s32), [[COPY6]](s32), [[COPY7]](s32)
     ; GFX10-DENORM: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX10-DENORM: [[COPY9:%[0-9]+]]:_(s32) = COPY $vgpr9
     ; GFX10-DENORM: [[COPY10:%[0-9]+]]:_(s32) = COPY $vgpr10
     ; GFX10-DENORM: [[COPY11:%[0-9]+]]:_(s32) = COPY $vgpr11
     ; GFX10-DENORM: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY8]](s32), [[COPY9]](s32), [[COPY10]](s32), [[COPY11]](s32)
-    ; GFX10-DENORM: [[COPY12:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-DENORM: [[FMUL:%[0-9]+]]:_(<4 x s32>) = reassoc G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX10-DENORM: [[FADD:%[0-9]+]]:_(<4 x s32>) = reassoc G_FADD [[FMUL]], [[BUILD_VECTOR2]]
     ; GFX10-DENORM: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<4 x s32>)
     ; GFX10-DENORM: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-DENORM: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10-DENORM: $vgpr2 = COPY [[UV2]](s32)
     ; GFX10-DENORM: $vgpr3 = COPY [[UV3]](s32)
-    ; GFX10-DENORM: [[COPY13:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY12]]
-    ; GFX10-DENORM: S_SETPC_B64_return [[COPY13]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+    ; GFX10-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
     ; GFX10-UNSAFE-LABEL: name: test_4xfloat_add_mul
     ; GFX10-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-UNSAFE: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-UNSAFE: [[BUILD_VECTOR:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32), [[COPY3]](s32)
     ; GFX10-UNSAFE: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-UNSAFE: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-UNSAFE: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX10-UNSAFE: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX10-UNSAFE: [[BUILD_VECTOR1:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY4]](s32), [[COPY5]](s32), [[COPY6]](s32), [[COPY7]](s32)
     ; GFX10-UNSAFE: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX10-UNSAFE: [[COPY9:%[0-9]+]]:_(s32) = COPY $vgpr9
     ; GFX10-UNSAFE: [[COPY10:%[0-9]+]]:_(s32) = COPY $vgpr10
     ; GFX10-UNSAFE: [[COPY11:%[0-9]+]]:_(s32) = COPY $vgpr11
     ; GFX10-UNSAFE: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s32>) = G_BUILD_VECTOR [[COPY8]](s32), [[COPY9]](s32), [[COPY10]](s32), [[COPY11]](s32)
-    ; GFX10-UNSAFE: [[COPY12:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-UNSAFE: [[FMA:%[0-9]+]]:_(<4 x s32>) = G_FMA [[BUILD_VECTOR]], [[BUILD_VECTOR1]], [[BUILD_VECTOR2]]
     ; GFX10-UNSAFE: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](<4 x s32>)
     ; GFX10-UNSAFE: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-UNSAFE: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10-UNSAFE: $vgpr2 = COPY [[UV2]](s32)
     ; GFX10-UNSAFE: $vgpr3 = COPY [[UV3]](s32)
-    ; GFX10-UNSAFE: [[COPY13:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY12]]
-    ; GFX10-UNSAFE: S_SETPC_B64_return [[COPY13]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+    ; GFX10-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
     %4:_(s32) = COPY $vgpr0
     %5:_(s32) = COPY $vgpr1
     %6:_(s32) = COPY $vgpr2
     %7:_(s32) = COPY $vgpr3
     %0:_(<4 x s32>) = G_BUILD_VECTOR %4(s32), %5(s32), %6(s32), %7(s32)
     %8:_(s32) = COPY $vgpr4
     %9:_(s32) = COPY $vgpr5
     %10:_(s32) = COPY $vgpr6
     %11:_(s32) = COPY $vgpr7
     %1:_(<4 x s32>) = G_BUILD_VECTOR %8(s32), %9(s32), %10(s32), %11(s32)
     %12:_(s32) = COPY $vgpr8
     %13:_(s32) = COPY $vgpr9
     %14:_(s32) = COPY $vgpr10
     %15:_(s32) = COPY $vgpr11
     %2:_(<4 x s32>) = G_BUILD_VECTOR %12(s32), %13(s32), %14(s32), %15(s32)
-    %3:sgpr_64 = COPY $sgpr30_sgpr31
     %16:_(<4 x s32>) = reassoc G_FMUL %0, %1
     %17:_(<4 x s32>) = reassoc G_FADD %16, %2
     %19:_(s32), %20:_(s32), %21:_(s32), %22:_(s32) = G_UNMERGE_VALUES %17(<4 x s32>)
     $vgpr0 = COPY %19(s32)
     $vgpr1 = COPY %20(s32)
     $vgpr2 = COPY %21(s32)
     $vgpr3 = COPY %22(s32)
-    %18:ccr_sgpr_64 = COPY %3
-    S_SETPC_B64_return %18, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
 ...
 
 ---
 name:            test_3xfloat_add_mul_rhs
 body:             |
   bb.1.entry:
-    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7, $vgpr8, $sgpr30_sgpr31
+    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7, $vgpr8
 
     ; GFX9-LABEL: name: test_3xfloat_add_mul_rhs
     ; GFX9: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9: [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32)
     ; GFX9: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9: [[BUILD_VECTOR1:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY3]](s32), [[COPY4]](s32), [[COPY5]](s32)
     ; GFX9: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX9: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX9: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX9: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY6]](s32), [[COPY7]](s32), [[COPY8]](s32)
-    ; GFX9: [[COPY9:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9: [[FMUL:%[0-9]+]]:_(<3 x s32>) = reassoc G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX9: [[FADD:%[0-9]+]]:_(<3 x s32>) = reassoc G_FADD [[BUILD_VECTOR2]], [[FMUL]]
     ; GFX9: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<3 x s32>)
     ; GFX9: $vgpr0 = COPY [[UV]](s32)
     ; GFX9: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9: $vgpr2 = COPY [[UV2]](s32)
-    ; GFX9: [[COPY10:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY9]]
-    ; GFX9: S_SETPC_B64_return [[COPY10]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
+    ; GFX9: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
     ; GFX9-CONTRACT-LABEL: name: test_3xfloat_add_mul_rhs
     ; GFX9-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-CONTRACT: [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32)
     ; GFX9-CONTRACT: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-CONTRACT: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-CONTRACT: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-CONTRACT: [[BUILD_VECTOR1:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY3]](s32), [[COPY4]](s32), [[COPY5]](s32)
     ; GFX9-CONTRACT: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX9-CONTRACT: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX9-CONTRACT: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX9-CONTRACT: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY6]](s32), [[COPY7]](s32), [[COPY8]](s32)
-    ; GFX9-CONTRACT: [[COPY9:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-CONTRACT: [[FMA:%[0-9]+]]:_(<3 x s32>) = G_FMA [[BUILD_VECTOR]], [[BUILD_VECTOR1]], [[BUILD_VECTOR2]]
     ; GFX9-CONTRACT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](<3 x s32>)
     ; GFX9-CONTRACT: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-CONTRACT: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9-CONTRACT: $vgpr2 = COPY [[UV2]](s32)
-    ; GFX9-CONTRACT: [[COPY10:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY9]]
-    ; GFX9-CONTRACT: S_SETPC_B64_return [[COPY10]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
+    ; GFX9-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
     ; GFX9-DENORM-LABEL: name: test_3xfloat_add_mul_rhs
     ; GFX9-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-DENORM: [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32)
     ; GFX9-DENORM: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-DENORM: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-DENORM: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-DENORM: [[BUILD_VECTOR1:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY3]](s32), [[COPY4]](s32), [[COPY5]](s32)
     ; GFX9-DENORM: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX9-DENORM: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX9-DENORM: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX9-DENORM: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY6]](s32), [[COPY7]](s32), [[COPY8]](s32)
-    ; GFX9-DENORM: [[COPY9:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-DENORM: [[FMUL:%[0-9]+]]:_(<3 x s32>) = reassoc G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX9-DENORM: [[FADD:%[0-9]+]]:_(<3 x s32>) = reassoc G_FADD [[BUILD_VECTOR2]], [[FMUL]]
     ; GFX9-DENORM: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<3 x s32>)
     ; GFX9-DENORM: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-DENORM: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9-DENORM: $vgpr2 = COPY [[UV2]](s32)
-    ; GFX9-DENORM: [[COPY10:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY9]]
-    ; GFX9-DENORM: S_SETPC_B64_return [[COPY10]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
+    ; GFX9-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
     ; GFX9-UNSAFE-LABEL: name: test_3xfloat_add_mul_rhs
     ; GFX9-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-UNSAFE: [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32)
     ; GFX9-UNSAFE: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-UNSAFE: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-UNSAFE: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-UNSAFE: [[BUILD_VECTOR1:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY3]](s32), [[COPY4]](s32), [[COPY5]](s32)
     ; GFX9-UNSAFE: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX9-UNSAFE: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX9-UNSAFE: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX9-UNSAFE: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY6]](s32), [[COPY7]](s32), [[COPY8]](s32)
-    ; GFX9-UNSAFE: [[COPY9:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-UNSAFE: [[FMA:%[0-9]+]]:_(<3 x s32>) = G_FMA [[BUILD_VECTOR]], [[BUILD_VECTOR1]], [[BUILD_VECTOR2]]
     ; GFX9-UNSAFE: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](<3 x s32>)
     ; GFX9-UNSAFE: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-UNSAFE: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9-UNSAFE: $vgpr2 = COPY [[UV2]](s32)
-    ; GFX9-UNSAFE: [[COPY10:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY9]]
-    ; GFX9-UNSAFE: S_SETPC_B64_return [[COPY10]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
+    ; GFX9-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
     ; GFX10-LABEL: name: test_3xfloat_add_mul_rhs
     ; GFX10: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10: [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32)
     ; GFX10: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10: [[BUILD_VECTOR1:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY3]](s32), [[COPY4]](s32), [[COPY5]](s32)
     ; GFX10: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX10: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX10: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX10: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY6]](s32), [[COPY7]](s32), [[COPY8]](s32)
-    ; GFX10: [[COPY9:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10: [[FMUL:%[0-9]+]]:_(<3 x s32>) = reassoc G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX10: [[FADD:%[0-9]+]]:_(<3 x s32>) = reassoc G_FADD [[BUILD_VECTOR2]], [[FMUL]]
     ; GFX10: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<3 x s32>)
     ; GFX10: $vgpr0 = COPY [[UV]](s32)
     ; GFX10: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10: $vgpr2 = COPY [[UV2]](s32)
-    ; GFX10: [[COPY10:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY9]]
-    ; GFX10: S_SETPC_B64_return [[COPY10]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
+    ; GFX10: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
     ; GFX10-CONTRACT-LABEL: name: test_3xfloat_add_mul_rhs
     ; GFX10-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-CONTRACT: [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32)
     ; GFX10-CONTRACT: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-CONTRACT: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-CONTRACT: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-CONTRACT: [[BUILD_VECTOR1:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY3]](s32), [[COPY4]](s32), [[COPY5]](s32)
     ; GFX10-CONTRACT: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX10-CONTRACT: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX10-CONTRACT: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX10-CONTRACT: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY6]](s32), [[COPY7]](s32), [[COPY8]](s32)
-    ; GFX10-CONTRACT: [[COPY9:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-CONTRACT: [[FMA:%[0-9]+]]:_(<3 x s32>) = G_FMA [[BUILD_VECTOR]], [[BUILD_VECTOR1]], [[BUILD_VECTOR2]]
     ; GFX10-CONTRACT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](<3 x s32>)
     ; GFX10-CONTRACT: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-CONTRACT: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10-CONTRACT: $vgpr2 = COPY [[UV2]](s32)
-    ; GFX10-CONTRACT: [[COPY10:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY9]]
-    ; GFX10-CONTRACT: S_SETPC_B64_return [[COPY10]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
+    ; GFX10-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
     ; GFX10-DENORM-LABEL: name: test_3xfloat_add_mul_rhs
     ; GFX10-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-DENORM: [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32)
     ; GFX10-DENORM: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-DENORM: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-DENORM: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-DENORM: [[BUILD_VECTOR1:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY3]](s32), [[COPY4]](s32), [[COPY5]](s32)
     ; GFX10-DENORM: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX10-DENORM: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX10-DENORM: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX10-DENORM: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY6]](s32), [[COPY7]](s32), [[COPY8]](s32)
-    ; GFX10-DENORM: [[COPY9:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-DENORM: [[FMUL:%[0-9]+]]:_(<3 x s32>) = reassoc G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX10-DENORM: [[FADD:%[0-9]+]]:_(<3 x s32>) = reassoc G_FADD [[BUILD_VECTOR2]], [[FMUL]]
     ; GFX10-DENORM: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<3 x s32>)
     ; GFX10-DENORM: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-DENORM: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10-DENORM: $vgpr2 = COPY [[UV2]](s32)
-    ; GFX10-DENORM: [[COPY10:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY9]]
-    ; GFX10-DENORM: S_SETPC_B64_return [[COPY10]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
+    ; GFX10-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
     ; GFX10-UNSAFE-LABEL: name: test_3xfloat_add_mul_rhs
     ; GFX10-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-UNSAFE: [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32), [[COPY2]](s32)
     ; GFX10-UNSAFE: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-UNSAFE: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-UNSAFE: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-UNSAFE: [[BUILD_VECTOR1:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY3]](s32), [[COPY4]](s32), [[COPY5]](s32)
     ; GFX10-UNSAFE: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
     ; GFX10-UNSAFE: [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
     ; GFX10-UNSAFE: [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
     ; GFX10-UNSAFE: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[COPY6]](s32), [[COPY7]](s32), [[COPY8]](s32)
-    ; GFX10-UNSAFE: [[COPY9:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-UNSAFE: [[FMA:%[0-9]+]]:_(<3 x s32>) = G_FMA [[BUILD_VECTOR]], [[BUILD_VECTOR1]], [[BUILD_VECTOR2]]
     ; GFX10-UNSAFE: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](<3 x s32>)
     ; GFX10-UNSAFE: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-UNSAFE: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10-UNSAFE: $vgpr2 = COPY [[UV2]](s32)
-    ; GFX10-UNSAFE: [[COPY10:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY9]]
-    ; GFX10-UNSAFE: S_SETPC_B64_return [[COPY10]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
+    ; GFX10-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
     %4:_(s32) = COPY $vgpr0
     %5:_(s32) = COPY $vgpr1
     %6:_(s32) = COPY $vgpr2
     %0:_(<3 x s32>) = G_BUILD_VECTOR %4(s32), %5(s32), %6(s32)
     %7:_(s32) = COPY $vgpr3
     %8:_(s32) = COPY $vgpr4
     %9:_(s32) = COPY $vgpr5
     %1:_(<3 x s32>) = G_BUILD_VECTOR %7(s32), %8(s32), %9(s32)
     %10:_(s32) = COPY $vgpr6
     %11:_(s32) = COPY $vgpr7
     %12:_(s32) = COPY $vgpr8
     %2:_(<3 x s32>) = G_BUILD_VECTOR %10(s32), %11(s32), %12(s32)
-    %3:sgpr_64 = COPY $sgpr30_sgpr31
     %13:_(<3 x s32>) = reassoc G_FMUL %0, %1
     %14:_(<3 x s32>) = reassoc G_FADD %2, %13
     %16:_(s32), %17:_(s32), %18:_(s32) = G_UNMERGE_VALUES %14(<3 x s32>)
     $vgpr0 = COPY %16(s32)
     $vgpr1 = COPY %17(s32)
     $vgpr2 = COPY %18(s32)
-    %15:ccr_sgpr_64 = COPY %3
-    S_SETPC_B64_return %15, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
 ...
 
 ---
 name:            test_4xhalf_add_mul
 body:             |
   bb.1.entry:
-    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $sgpr30_sgpr31
+    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
 
     ; GFX9-LABEL: name: test_4xhalf_add_mul
     ; GFX9: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX9: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX9: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>)
     ; GFX9: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX9: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX9: [[CONCAT_VECTORS1:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>)
     ; GFX9: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX9: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX9: [[CONCAT_VECTORS2:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>)
-    ; GFX9: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9: [[FMUL:%[0-9]+]]:_(<4 x s16>) = reassoc G_FMUL [[CONCAT_VECTORS]], [[CONCAT_VECTORS1]]
     ; GFX9: [[FADD:%[0-9]+]]:_(<4 x s16>) = reassoc G_FADD [[FMUL]], [[CONCAT_VECTORS2]]
     ; GFX9: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[FADD]](<4 x s16>)
     ; GFX9: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX9: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX9: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX9-CONTRACT-LABEL: name: test_4xhalf_add_mul
     ; GFX9-CONTRACT: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX9-CONTRACT: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX9-CONTRACT: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>)
     ; GFX9-CONTRACT: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX9-CONTRACT: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX9-CONTRACT: [[CONCAT_VECTORS1:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>)
     ; GFX9-CONTRACT: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX9-CONTRACT: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX9-CONTRACT: [[CONCAT_VECTORS2:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>)
-    ; GFX9-CONTRACT: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-CONTRACT: [[FMA:%[0-9]+]]:_(<4 x s16>) = G_FMA [[CONCAT_VECTORS]], [[CONCAT_VECTORS1]], [[CONCAT_VECTORS2]]
     ; GFX9-CONTRACT: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[FMA]](<4 x s16>)
     ; GFX9-CONTRACT: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX9-CONTRACT: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX9-CONTRACT: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9-CONTRACT: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX9-DENORM-LABEL: name: test_4xhalf_add_mul
     ; GFX9-DENORM: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX9-DENORM: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX9-DENORM: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>)
     ; GFX9-DENORM: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX9-DENORM: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX9-DENORM: [[CONCAT_VECTORS1:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>)
     ; GFX9-DENORM: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX9-DENORM: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX9-DENORM: [[CONCAT_VECTORS2:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>)
-    ; GFX9-DENORM: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-DENORM: [[FMUL:%[0-9]+]]:_(<4 x s16>) = reassoc G_FMUL [[CONCAT_VECTORS]], [[CONCAT_VECTORS1]]
     ; GFX9-DENORM: [[FADD:%[0-9]+]]:_(<4 x s16>) = reassoc G_FADD [[FMUL]], [[CONCAT_VECTORS2]]
     ; GFX9-DENORM: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[FADD]](<4 x s16>)
     ; GFX9-DENORM: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX9-DENORM: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX9-DENORM: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9-DENORM: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX9-UNSAFE-LABEL: name: test_4xhalf_add_mul
     ; GFX9-UNSAFE: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX9-UNSAFE: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX9-UNSAFE: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>)
     ; GFX9-UNSAFE: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX9-UNSAFE: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX9-UNSAFE: [[CONCAT_VECTORS1:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>)
     ; GFX9-UNSAFE: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX9-UNSAFE: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX9-UNSAFE: [[CONCAT_VECTORS2:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>)
-    ; GFX9-UNSAFE: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-UNSAFE: [[FMA:%[0-9]+]]:_(<4 x s16>) = G_FMA [[CONCAT_VECTORS]], [[CONCAT_VECTORS1]], [[CONCAT_VECTORS2]]
     ; GFX9-UNSAFE: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[FMA]](<4 x s16>)
     ; GFX9-UNSAFE: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX9-UNSAFE: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX9-UNSAFE: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9-UNSAFE: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-LABEL: name: test_4xhalf_add_mul
     ; GFX10: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX10: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX10: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>)
     ; GFX10: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX10: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX10: [[CONCAT_VECTORS1:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>)
     ; GFX10: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX10: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX10: [[CONCAT_VECTORS2:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>)
-    ; GFX10: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10: [[FMUL:%[0-9]+]]:_(<4 x s16>) = reassoc G_FMUL [[CONCAT_VECTORS]], [[CONCAT_VECTORS1]]
     ; GFX10: [[FADD:%[0-9]+]]:_(<4 x s16>) = reassoc G_FADD [[FMUL]], [[CONCAT_VECTORS2]]
     ; GFX10: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[FADD]](<4 x s16>)
     ; GFX10: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX10: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX10: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-CONTRACT-LABEL: name: test_4xhalf_add_mul
     ; GFX10-CONTRACT: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX10-CONTRACT: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX10-CONTRACT: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>)
     ; GFX10-CONTRACT: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX10-CONTRACT: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX10-CONTRACT: [[CONCAT_VECTORS1:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>)
     ; GFX10-CONTRACT: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX10-CONTRACT: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX10-CONTRACT: [[CONCAT_VECTORS2:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>)
-    ; GFX10-CONTRACT: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-CONTRACT: [[FMA:%[0-9]+]]:_(<4 x s16>) = G_FMA [[CONCAT_VECTORS]], [[CONCAT_VECTORS1]], [[CONCAT_VECTORS2]]
     ; GFX10-CONTRACT: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[FMA]](<4 x s16>)
     ; GFX10-CONTRACT: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX10-CONTRACT: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX10-CONTRACT: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10-CONTRACT: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-DENORM-LABEL: name: test_4xhalf_add_mul
     ; GFX10-DENORM: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX10-DENORM: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX10-DENORM: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>)
     ; GFX10-DENORM: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX10-DENORM: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX10-DENORM: [[CONCAT_VECTORS1:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>)
     ; GFX10-DENORM: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX10-DENORM: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX10-DENORM: [[CONCAT_VECTORS2:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>)
-    ; GFX10-DENORM: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-DENORM: [[FMUL:%[0-9]+]]:_(<4 x s16>) = reassoc G_FMUL [[CONCAT_VECTORS]], [[CONCAT_VECTORS1]]
     ; GFX10-DENORM: [[FADD:%[0-9]+]]:_(<4 x s16>) = reassoc G_FADD [[FMUL]], [[CONCAT_VECTORS2]]
     ; GFX10-DENORM: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[FADD]](<4 x s16>)
     ; GFX10-DENORM: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX10-DENORM: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX10-DENORM: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10-DENORM: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-UNSAFE-LABEL: name: test_4xhalf_add_mul
     ; GFX10-UNSAFE: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX10-UNSAFE: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX10-UNSAFE: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>)
     ; GFX10-UNSAFE: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX10-UNSAFE: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX10-UNSAFE: [[CONCAT_VECTORS1:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>)
     ; GFX10-UNSAFE: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX10-UNSAFE: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX10-UNSAFE: [[CONCAT_VECTORS2:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>)
-    ; GFX10-UNSAFE: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-UNSAFE: [[FMA:%[0-9]+]]:_(<4 x s16>) = G_FMA [[CONCAT_VECTORS]], [[CONCAT_VECTORS1]], [[CONCAT_VECTORS2]]
     ; GFX10-UNSAFE: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[FMA]](<4 x s16>)
     ; GFX10-UNSAFE: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX10-UNSAFE: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX10-UNSAFE: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10-UNSAFE: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     %4:_(<2 x s16>) = COPY $vgpr0
     %5:_(<2 x s16>) = COPY $vgpr1
     %0:_(<4 x s16>) = G_CONCAT_VECTORS %4(<2 x s16>), %5(<2 x s16>)
     %6:_(<2 x s16>) = COPY $vgpr2
     %7:_(<2 x s16>) = COPY $vgpr3
     %1:_(<4 x s16>) = G_CONCAT_VECTORS %6(<2 x s16>), %7(<2 x s16>)
     %8:_(<2 x s16>) = COPY $vgpr4
     %9:_(<2 x s16>) = COPY $vgpr5
     %2:_(<4 x s16>) = G_CONCAT_VECTORS %8(<2 x s16>), %9(<2 x s16>)
-    %3:sgpr_64 = COPY $sgpr30_sgpr31
     %10:_(<4 x s16>) = reassoc G_FMUL %0, %1
     %11:_(<4 x s16>) = reassoc G_FADD %10, %2
     %13:_(<2 x s16>), %14:_(<2 x s16>) = G_UNMERGE_VALUES %11(<4 x s16>)
     $vgpr0 = COPY %13(<2 x s16>)
     $vgpr1 = COPY %14(<2 x s16>)
-    %12:ccr_sgpr_64 = COPY %3
-    S_SETPC_B64_return %12, implicit $vgpr0, implicit $vgpr1
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
 ...
 
 ---
 name:            test_3xhalf_add_mul_rhs
 body:             |
   bb.1.entry:
-    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $sgpr30_sgpr31
+    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
 
     ; GFX9-LABEL: name: test_3xhalf_add_mul_rhs
     ; GFX9: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX9: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX9: [[DEF:%[0-9]+]]:_(<2 x s16>) = G_IMPLICIT_DEF
     ; GFX9: [[CONCAT_VECTORS:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX9: [[BITCAST:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS]](<6 x s16>)
     ; GFX9: [[TRUNC:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST]](s96)
     ; GFX9: [[BITCAST1:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC]](s48)
     ; GFX9: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX9: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX9: [[CONCAT_VECTORS1:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX9: [[BITCAST2:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS1]](<6 x s16>)
     ; GFX9: [[TRUNC1:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST2]](s96)
     ; GFX9: [[BITCAST3:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC1]](s48)
     ; GFX9: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX9: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX9: [[CONCAT_VECTORS2:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX9: [[BITCAST4:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS2]](<6 x s16>)
     ; GFX9: [[TRUNC2:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST4]](s96)
     ; GFX9: [[BITCAST5:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC2]](s48)
-    ; GFX9: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9: [[FMUL:%[0-9]+]]:_(<3 x s16>) = reassoc G_FMUL [[BITCAST1]], [[BITCAST3]]
     ; GFX9: [[FADD:%[0-9]+]]:_(<3 x s16>) = reassoc G_FADD [[BITCAST5]], [[FMUL]]
     ; GFX9: [[DEF1:%[0-9]+]]:_(<3 x s16>) = G_IMPLICIT_DEF
     ; GFX9: [[CONCAT_VECTORS3:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[FADD]](<3 x s16>), [[DEF1]](<3 x s16>)
     ; GFX9: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>), [[UV2:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS3]](<6 x s16>)
     ; GFX9: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX9: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX9: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX9-CONTRACT-LABEL: name: test_3xhalf_add_mul_rhs
     ; GFX9-CONTRACT: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX9-CONTRACT: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX9-CONTRACT: [[DEF:%[0-9]+]]:_(<2 x s16>) = G_IMPLICIT_DEF
     ; GFX9-CONTRACT: [[CONCAT_VECTORS:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX9-CONTRACT: [[BITCAST:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS]](<6 x s16>)
     ; GFX9-CONTRACT: [[TRUNC:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST]](s96)
     ; GFX9-CONTRACT: [[BITCAST1:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC]](s48)
     ; GFX9-CONTRACT: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX9-CONTRACT: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX9-CONTRACT: [[CONCAT_VECTORS1:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX9-CONTRACT: [[BITCAST2:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS1]](<6 x s16>)
     ; GFX9-CONTRACT: [[TRUNC1:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST2]](s96)
     ; GFX9-CONTRACT: [[BITCAST3:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC1]](s48)
     ; GFX9-CONTRACT: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX9-CONTRACT: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX9-CONTRACT: [[CONCAT_VECTORS2:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX9-CONTRACT: [[BITCAST4:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS2]](<6 x s16>)
     ; GFX9-CONTRACT: [[TRUNC2:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST4]](s96)
     ; GFX9-CONTRACT: [[BITCAST5:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC2]](s48)
-    ; GFX9-CONTRACT: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-CONTRACT: [[FMA:%[0-9]+]]:_(<3 x s16>) = G_FMA [[BITCAST1]], [[BITCAST3]], [[BITCAST5]]
     ; GFX9-CONTRACT: [[DEF1:%[0-9]+]]:_(<3 x s16>) = G_IMPLICIT_DEF
     ; GFX9-CONTRACT: [[CONCAT_VECTORS3:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[FMA]](<3 x s16>), [[DEF1]](<3 x s16>)
     ; GFX9-CONTRACT: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>), [[UV2:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS3]](<6 x s16>)
     ; GFX9-CONTRACT: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX9-CONTRACT: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX9-CONTRACT: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9-CONTRACT: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX9-DENORM-LABEL: name: test_3xhalf_add_mul_rhs
     ; GFX9-DENORM: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX9-DENORM: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX9-DENORM: [[DEF:%[0-9]+]]:_(<2 x s16>) = G_IMPLICIT_DEF
     ; GFX9-DENORM: [[CONCAT_VECTORS:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX9-DENORM: [[BITCAST:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS]](<6 x s16>)
     ; GFX9-DENORM: [[TRUNC:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST]](s96)
     ; GFX9-DENORM: [[BITCAST1:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC]](s48)
     ; GFX9-DENORM: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX9-DENORM: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX9-DENORM: [[CONCAT_VECTORS1:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX9-DENORM: [[BITCAST2:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS1]](<6 x s16>)
     ; GFX9-DENORM: [[TRUNC1:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST2]](s96)
     ; GFX9-DENORM: [[BITCAST3:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC1]](s48)
     ; GFX9-DENORM: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX9-DENORM: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX9-DENORM: [[CONCAT_VECTORS2:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX9-DENORM: [[BITCAST4:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS2]](<6 x s16>)
     ; GFX9-DENORM: [[TRUNC2:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST4]](s96)
     ; GFX9-DENORM: [[BITCAST5:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC2]](s48)
-    ; GFX9-DENORM: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-DENORM: [[FMUL:%[0-9]+]]:_(<3 x s16>) = reassoc G_FMUL [[BITCAST1]], [[BITCAST3]]
     ; GFX9-DENORM: [[FADD:%[0-9]+]]:_(<3 x s16>) = reassoc G_FADD [[BITCAST5]], [[FMUL]]
     ; GFX9-DENORM: [[DEF1:%[0-9]+]]:_(<3 x s16>) = G_IMPLICIT_DEF
     ; GFX9-DENORM: [[CONCAT_VECTORS3:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[FADD]](<3 x s16>), [[DEF1]](<3 x s16>)
     ; GFX9-DENORM: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>), [[UV2:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS3]](<6 x s16>)
     ; GFX9-DENORM: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX9-DENORM: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX9-DENORM: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9-DENORM: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX9-UNSAFE-LABEL: name: test_3xhalf_add_mul_rhs
     ; GFX9-UNSAFE: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX9-UNSAFE: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX9-UNSAFE: [[DEF:%[0-9]+]]:_(<2 x s16>) = G_IMPLICIT_DEF
     ; GFX9-UNSAFE: [[CONCAT_VECTORS:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX9-UNSAFE: [[BITCAST:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS]](<6 x s16>)
     ; GFX9-UNSAFE: [[TRUNC:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST]](s96)
     ; GFX9-UNSAFE: [[BITCAST1:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC]](s48)
     ; GFX9-UNSAFE: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX9-UNSAFE: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX9-UNSAFE: [[CONCAT_VECTORS1:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX9-UNSAFE: [[BITCAST2:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS1]](<6 x s16>)
     ; GFX9-UNSAFE: [[TRUNC1:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST2]](s96)
     ; GFX9-UNSAFE: [[BITCAST3:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC1]](s48)
     ; GFX9-UNSAFE: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX9-UNSAFE: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX9-UNSAFE: [[CONCAT_VECTORS2:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX9-UNSAFE: [[BITCAST4:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS2]](<6 x s16>)
     ; GFX9-UNSAFE: [[TRUNC2:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST4]](s96)
     ; GFX9-UNSAFE: [[BITCAST5:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC2]](s48)
-    ; GFX9-UNSAFE: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-UNSAFE: [[FMA:%[0-9]+]]:_(<3 x s16>) = G_FMA [[BITCAST1]], [[BITCAST3]], [[BITCAST5]]
     ; GFX9-UNSAFE: [[DEF1:%[0-9]+]]:_(<3 x s16>) = G_IMPLICIT_DEF
     ; GFX9-UNSAFE: [[CONCAT_VECTORS3:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[FMA]](<3 x s16>), [[DEF1]](<3 x s16>)
     ; GFX9-UNSAFE: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>), [[UV2:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS3]](<6 x s16>)
     ; GFX9-UNSAFE: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX9-UNSAFE: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX9-UNSAFE: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX9-UNSAFE: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX9-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-LABEL: name: test_3xhalf_add_mul_rhs
     ; GFX10: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX10: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX10: [[DEF:%[0-9]+]]:_(<2 x s16>) = G_IMPLICIT_DEF
     ; GFX10: [[CONCAT_VECTORS:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX10: [[BITCAST:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS]](<6 x s16>)
     ; GFX10: [[TRUNC:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST]](s96)
     ; GFX10: [[BITCAST1:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC]](s48)
     ; GFX10: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX10: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX10: [[CONCAT_VECTORS1:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX10: [[BITCAST2:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS1]](<6 x s16>)
     ; GFX10: [[TRUNC1:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST2]](s96)
     ; GFX10: [[BITCAST3:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC1]](s48)
     ; GFX10: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX10: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX10: [[CONCAT_VECTORS2:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX10: [[BITCAST4:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS2]](<6 x s16>)
     ; GFX10: [[TRUNC2:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST4]](s96)
     ; GFX10: [[BITCAST5:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC2]](s48)
-    ; GFX10: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10: [[FMUL:%[0-9]+]]:_(<3 x s16>) = reassoc G_FMUL [[BITCAST1]], [[BITCAST3]]
     ; GFX10: [[FADD:%[0-9]+]]:_(<3 x s16>) = reassoc G_FADD [[BITCAST5]], [[FMUL]]
     ; GFX10: [[DEF1:%[0-9]+]]:_(<3 x s16>) = G_IMPLICIT_DEF
     ; GFX10: [[CONCAT_VECTORS3:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[FADD]](<3 x s16>), [[DEF1]](<3 x s16>)
     ; GFX10: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>), [[UV2:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS3]](<6 x s16>)
     ; GFX10: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX10: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX10: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-CONTRACT-LABEL: name: test_3xhalf_add_mul_rhs
     ; GFX10-CONTRACT: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX10-CONTRACT: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX10-CONTRACT: [[DEF:%[0-9]+]]:_(<2 x s16>) = G_IMPLICIT_DEF
     ; GFX10-CONTRACT: [[CONCAT_VECTORS:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX10-CONTRACT: [[BITCAST:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS]](<6 x s16>)
     ; GFX10-CONTRACT: [[TRUNC:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST]](s96)
     ; GFX10-CONTRACT: [[BITCAST1:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC]](s48)
     ; GFX10-CONTRACT: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX10-CONTRACT: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX10-CONTRACT: [[CONCAT_VECTORS1:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX10-CONTRACT: [[BITCAST2:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS1]](<6 x s16>)
     ; GFX10-CONTRACT: [[TRUNC1:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST2]](s96)
     ; GFX10-CONTRACT: [[BITCAST3:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC1]](s48)
     ; GFX10-CONTRACT: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX10-CONTRACT: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX10-CONTRACT: [[CONCAT_VECTORS2:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX10-CONTRACT: [[BITCAST4:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS2]](<6 x s16>)
     ; GFX10-CONTRACT: [[TRUNC2:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST4]](s96)
     ; GFX10-CONTRACT: [[BITCAST5:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC2]](s48)
-    ; GFX10-CONTRACT: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-CONTRACT: [[FMA:%[0-9]+]]:_(<3 x s16>) = G_FMA [[BITCAST1]], [[BITCAST3]], [[BITCAST5]]
     ; GFX10-CONTRACT: [[DEF1:%[0-9]+]]:_(<3 x s16>) = G_IMPLICIT_DEF
     ; GFX10-CONTRACT: [[CONCAT_VECTORS3:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[FMA]](<3 x s16>), [[DEF1]](<3 x s16>)
     ; GFX10-CONTRACT: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>), [[UV2:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS3]](<6 x s16>)
     ; GFX10-CONTRACT: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX10-CONTRACT: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX10-CONTRACT: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10-CONTRACT: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-DENORM-LABEL: name: test_3xhalf_add_mul_rhs
     ; GFX10-DENORM: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX10-DENORM: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX10-DENORM: [[DEF:%[0-9]+]]:_(<2 x s16>) = G_IMPLICIT_DEF
     ; GFX10-DENORM: [[CONCAT_VECTORS:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX10-DENORM: [[BITCAST:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS]](<6 x s16>)
     ; GFX10-DENORM: [[TRUNC:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST]](s96)
     ; GFX10-DENORM: [[BITCAST1:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC]](s48)
     ; GFX10-DENORM: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX10-DENORM: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX10-DENORM: [[CONCAT_VECTORS1:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX10-DENORM: [[BITCAST2:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS1]](<6 x s16>)
     ; GFX10-DENORM: [[TRUNC1:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST2]](s96)
     ; GFX10-DENORM: [[BITCAST3:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC1]](s48)
     ; GFX10-DENORM: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX10-DENORM: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX10-DENORM: [[CONCAT_VECTORS2:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX10-DENORM: [[BITCAST4:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS2]](<6 x s16>)
     ; GFX10-DENORM: [[TRUNC2:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST4]](s96)
     ; GFX10-DENORM: [[BITCAST5:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC2]](s48)
-    ; GFX10-DENORM: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-DENORM: [[FMUL:%[0-9]+]]:_(<3 x s16>) = reassoc G_FMUL [[BITCAST1]], [[BITCAST3]]
     ; GFX10-DENORM: [[FADD:%[0-9]+]]:_(<3 x s16>) = reassoc G_FADD [[BITCAST5]], [[FMUL]]
     ; GFX10-DENORM: [[DEF1:%[0-9]+]]:_(<3 x s16>) = G_IMPLICIT_DEF
     ; GFX10-DENORM: [[CONCAT_VECTORS3:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[FADD]](<3 x s16>), [[DEF1]](<3 x s16>)
     ; GFX10-DENORM: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>), [[UV2:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS3]](<6 x s16>)
     ; GFX10-DENORM: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX10-DENORM: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX10-DENORM: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10-DENORM: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     ; GFX10-UNSAFE-LABEL: name: test_3xhalf_add_mul_rhs
     ; GFX10-UNSAFE: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
     ; GFX10-UNSAFE: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
     ; GFX10-UNSAFE: [[DEF:%[0-9]+]]:_(<2 x s16>) = G_IMPLICIT_DEF
     ; GFX10-UNSAFE: [[CONCAT_VECTORS:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX10-UNSAFE: [[BITCAST:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS]](<6 x s16>)
     ; GFX10-UNSAFE: [[TRUNC:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST]](s96)
     ; GFX10-UNSAFE: [[BITCAST1:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC]](s48)
     ; GFX10-UNSAFE: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
     ; GFX10-UNSAFE: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
     ; GFX10-UNSAFE: [[CONCAT_VECTORS1:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX10-UNSAFE: [[BITCAST2:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS1]](<6 x s16>)
     ; GFX10-UNSAFE: [[TRUNC1:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST2]](s96)
     ; GFX10-UNSAFE: [[BITCAST3:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC1]](s48)
     ; GFX10-UNSAFE: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
     ; GFX10-UNSAFE: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
     ; GFX10-UNSAFE: [[CONCAT_VECTORS2:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>), [[DEF]](<2 x s16>)
     ; GFX10-UNSAFE: [[BITCAST4:%[0-9]+]]:_(s96) = G_BITCAST [[CONCAT_VECTORS2]](<6 x s16>)
     ; GFX10-UNSAFE: [[TRUNC2:%[0-9]+]]:_(s48) = G_TRUNC [[BITCAST4]](s96)
     ; GFX10-UNSAFE: [[BITCAST5:%[0-9]+]]:_(<3 x s16>) = G_BITCAST [[TRUNC2]](s48)
-    ; GFX10-UNSAFE: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-UNSAFE: [[FMA:%[0-9]+]]:_(<3 x s16>) = G_FMA [[BITCAST1]], [[BITCAST3]], [[BITCAST5]]
     ; GFX10-UNSAFE: [[DEF1:%[0-9]+]]:_(<3 x s16>) = G_IMPLICIT_DEF
     ; GFX10-UNSAFE: [[CONCAT_VECTORS3:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[FMA]](<3 x s16>), [[DEF1]](<3 x s16>)
     ; GFX10-UNSAFE: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>), [[UV2:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS3]](<6 x s16>)
     ; GFX10-UNSAFE: $vgpr0 = COPY [[UV]](<2 x s16>)
     ; GFX10-UNSAFE: $vgpr1 = COPY [[UV1]](<2 x s16>)
-    ; GFX10-UNSAFE: [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-    ; GFX10-UNSAFE: S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1
+    ; GFX10-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     %4:_(<2 x s16>) = COPY $vgpr0
     %5:_(<2 x s16>) = COPY $vgpr1
     %10:_(<2 x s16>) = G_IMPLICIT_DEF
     %11:_(<6 x s16>) = G_CONCAT_VECTORS %4(<2 x s16>), %5(<2 x s16>), %10(<2 x s16>)
     %0:_(<3 x s16>), %12:_(<3 x s16>) = G_UNMERGE_VALUES %11(<6 x s16>)
     %6:_(<2 x s16>) = COPY $vgpr2
     %7:_(<2 x s16>) = COPY $vgpr3
     %13:_(<6 x s16>) = G_CONCAT_VECTORS %6(<2 x s16>), %7(<2 x s16>), %10(<2 x s16>)
     %1:_(<3 x s16>), %14:_(<3 x s16>) = G_UNMERGE_VALUES %13(<6 x s16>)
     %8:_(<2 x s16>) = COPY $vgpr4
     %9:_(<2 x s16>) = COPY $vgpr5
     %15:_(<6 x s16>) = G_CONCAT_VECTORS %8(<2 x s16>), %9(<2 x s16>), %10(<2 x s16>)
     %2:_(<3 x s16>), %16:_(<3 x s16>) = G_UNMERGE_VALUES %15(<6 x s16>)
-    %3:sgpr_64 = COPY $sgpr30_sgpr31
     %17:_(<3 x s16>) = reassoc G_FMUL %0, %1
     %18:_(<3 x s16>) = reassoc G_FADD %2, %17
     %22:_(<3 x s16>) = G_IMPLICIT_DEF
     %23:_(<6 x s16>) = G_CONCAT_VECTORS %18(<3 x s16>), %22(<3 x s16>)
     %20:_(<2 x s16>), %21:_(<2 x s16>), %24:_(<2 x s16>) = G_UNMERGE_VALUES %23(<6 x s16>)
     $vgpr0 = COPY %20(<2 x s16>)
     $vgpr1 = COPY %21(<2 x s16>)
-    %19:ccr_sgpr_64 = COPY %3
-    S_SETPC_B64_return %19, implicit $vgpr0, implicit $vgpr1
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
 ...
 
 ---
 name:            test_4xdouble_add_mul
 body:             |
   bb.1.entry:
-    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, $sgpr30_sgpr31
+    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
 
     ; GFX9-LABEL: name: test_4xdouble_add_mul
     ; GFX9: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
bb.1.entry:
     ; GFX9: [[COPY21:%[0-9]+]]:_(s32) = COPY $vgpr21
     ; GFX9: [[COPY22:%[0-9]+]]:_(s32) = COPY $vgpr22
     ; GFX9: [[COPY23:%[0-9]+]]:_(s32) = COPY $vgpr23
     ; GFX9: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX9: [[MV9:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY18]](s32), [[COPY19]](s32)
     ; GFX9: [[MV10:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY20]](s32), [[COPY21]](s32)
     ; GFX9: [[MV11:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY22]](s32), [[COPY23]](s32)
     ; GFX9: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s64>) = G_BUILD_VECTOR [[MV8]](s64), [[MV9]](s64), [[MV10]](s64), [[MV11]](s64)
-    ; GFX9: [[COPY24:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9: [[FMUL:%[0-9]+]]:_(<4 x s64>) = reassoc G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX9: [[FADD:%[0-9]+]]:_(<4 x s64>) = reassoc G_FADD [[FMUL]], [[BUILD_VECTOR2]]
     ; GFX9: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<4 x s64>)
     ; GFX9: $vgpr0 = COPY [[UV]](s32)
     ; GFX9: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9: $vgpr2 = COPY [[UV2]](s32)
     ; GFX9: $vgpr3 = COPY [[UV3]](s32)
     ; GFX9: $vgpr4 = COPY [[UV4]](s32)
     ; GFX9: $vgpr5 = COPY [[UV5]](s32)
     ; GFX9: $vgpr6 = COPY [[UV6]](s32)
     ; GFX9: $vgpr7 = COPY [[UV7]](s32)
-    ; GFX9: [[COPY25:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY24]]
-    ; GFX9: S_SETPC_B64_return [[COPY25]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
+    ; GFX9: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
     ; GFX9-CONTRACT-LABEL: name: test_4xdouble_add_mul
     ; GFX9-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-CONTRACT: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-CONTRACT: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-CONTRACT: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-CONTRACT: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
bb.1.entry:
     ; GFX9-CONTRACT: [[COPY21:%[0-9]+]]:_(s32) = COPY $vgpr21
     ; GFX9-CONTRACT: [[COPY22:%[0-9]+]]:_(s32) = COPY $vgpr22
     ; GFX9-CONTRACT: [[COPY23:%[0-9]+]]:_(s32) = COPY $vgpr23
     ; GFX9-CONTRACT: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX9-CONTRACT: [[MV9:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY18]](s32), [[COPY19]](s32)
     ; GFX9-CONTRACT: [[MV10:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY20]](s32), [[COPY21]](s32)
     ; GFX9-CONTRACT: [[MV11:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY22]](s32), [[COPY23]](s32)
     ; GFX9-CONTRACT: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s64>) = G_BUILD_VECTOR [[MV8]](s64), [[MV9]](s64), [[MV10]](s64), [[MV11]](s64)
-    ; GFX9-CONTRACT: [[COPY24:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-CONTRACT: [[FMA:%[0-9]+]]:_(<4 x s64>) = G_FMA [[BUILD_VECTOR]], [[BUILD_VECTOR1]], [[BUILD_VECTOR2]]
     ; GFX9-CONTRACT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](<4 x s64>)
     ; GFX9-CONTRACT: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-CONTRACT: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9-CONTRACT: $vgpr2 = COPY [[UV2]](s32)
     ; GFX9-CONTRACT: $vgpr3 = COPY [[UV3]](s32)
     ; GFX9-CONTRACT: $vgpr4 = COPY [[UV4]](s32)
     ; GFX9-CONTRACT: $vgpr5 = COPY [[UV5]](s32)
     ; GFX9-CONTRACT: $vgpr6 = COPY [[UV6]](s32)
     ; GFX9-CONTRACT: $vgpr7 = COPY [[UV7]](s32)
-    ; GFX9-CONTRACT: [[COPY25:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY24]]
-    ; GFX9-CONTRACT: S_SETPC_B64_return [[COPY25]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
+    ; GFX9-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
     ; GFX9-DENORM-LABEL: name: test_4xdouble_add_mul
     ; GFX9-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-DENORM: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-DENORM: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-DENORM: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-DENORM: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
bb.1.entry:
     ; GFX9-DENORM: [[COPY21:%[0-9]+]]:_(s32) = COPY $vgpr21
     ; GFX9-DENORM: [[COPY22:%[0-9]+]]:_(s32) = COPY $vgpr22
     ; GFX9-DENORM: [[COPY23:%[0-9]+]]:_(s32) = COPY $vgpr23
     ; GFX9-DENORM: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX9-DENORM: [[MV9:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY18]](s32), [[COPY19]](s32)
     ; GFX9-DENORM: [[MV10:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY20]](s32), [[COPY21]](s32)
     ; GFX9-DENORM: [[MV11:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY22]](s32), [[COPY23]](s32)
     ; GFX9-DENORM: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s64>) = G_BUILD_VECTOR [[MV8]](s64), [[MV9]](s64), [[MV10]](s64), [[MV11]](s64)
-    ; GFX9-DENORM: [[COPY24:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-DENORM: [[FMUL:%[0-9]+]]:_(<4 x s64>) = reassoc G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX9-DENORM: [[FADD:%[0-9]+]]:_(<4 x s64>) = reassoc G_FADD [[FMUL]], [[BUILD_VECTOR2]]
     ; GFX9-DENORM: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<4 x s64>)
     ; GFX9-DENORM: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-DENORM: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9-DENORM: $vgpr2 = COPY [[UV2]](s32)
     ; GFX9-DENORM: $vgpr3 = COPY [[UV3]](s32)
     ; GFX9-DENORM: $vgpr4 = COPY [[UV4]](s32)
     ; GFX9-DENORM: $vgpr5 = COPY [[UV5]](s32)
     ; GFX9-DENORM: $vgpr6 = COPY [[UV6]](s32)
     ; GFX9-DENORM: $vgpr7 = COPY [[UV7]](s32)
-    ; GFX9-DENORM: [[COPY25:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY24]]
-    ; GFX9-DENORM: S_SETPC_B64_return [[COPY25]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
+    ; GFX9-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
     ; GFX9-UNSAFE-LABEL: name: test_4xdouble_add_mul
     ; GFX9-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-UNSAFE: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-UNSAFE: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-UNSAFE: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-UNSAFE: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
bb.1.entry:
     ; GFX9-UNSAFE: [[COPY21:%[0-9]+]]:_(s32) = COPY $vgpr21
     ; GFX9-UNSAFE: [[COPY22:%[0-9]+]]:_(s32) = COPY $vgpr22
     ; GFX9-UNSAFE: [[COPY23:%[0-9]+]]:_(s32) = COPY $vgpr23
     ; GFX9-UNSAFE: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX9-UNSAFE: [[MV9:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY18]](s32), [[COPY19]](s32)
     ; GFX9-UNSAFE: [[MV10:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY20]](s32), [[COPY21]](s32)
     ; GFX9-UNSAFE: [[MV11:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY22]](s32), [[COPY23]](s32)
     ; GFX9-UNSAFE: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s64>) = G_BUILD_VECTOR [[MV8]](s64), [[MV9]](s64), [[MV10]](s64), [[MV11]](s64)
-    ; GFX9-UNSAFE: [[COPY24:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-UNSAFE: [[FMA:%[0-9]+]]:_(<4 x s64>) = G_FMA [[BUILD_VECTOR]], [[BUILD_VECTOR1]], [[BUILD_VECTOR2]]
     ; GFX9-UNSAFE: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](<4 x s64>)
     ; GFX9-UNSAFE: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-UNSAFE: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9-UNSAFE: $vgpr2 = COPY [[UV2]](s32)
     ; GFX9-UNSAFE: $vgpr3 = COPY [[UV3]](s32)
     ; GFX9-UNSAFE: $vgpr4 = COPY [[UV4]](s32)
     ; GFX9-UNSAFE: $vgpr5 = COPY [[UV5]](s32)
     ; GFX9-UNSAFE: $vgpr6 = COPY [[UV6]](s32)
     ; GFX9-UNSAFE: $vgpr7 = COPY [[UV7]](s32)
-    ; GFX9-UNSAFE: [[COPY25:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY24]]
-    ; GFX9-UNSAFE: S_SETPC_B64_return [[COPY25]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
+    ; GFX9-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
     ; GFX10-LABEL: name: test_4xdouble_add_mul
     ; GFX10: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
bb.1.entry:
     ; GFX10: [[COPY21:%[0-9]+]]:_(s32) = COPY $vgpr21
     ; GFX10: [[COPY22:%[0-9]+]]:_(s32) = COPY $vgpr22
     ; GFX10: [[COPY23:%[0-9]+]]:_(s32) = COPY $vgpr23
     ; GFX10: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX10: [[MV9:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY18]](s32), [[COPY19]](s32)
     ; GFX10: [[MV10:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY20]](s32), [[COPY21]](s32)
     ; GFX10: [[MV11:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY22]](s32), [[COPY23]](s32)
     ; GFX10: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s64>) = G_BUILD_VECTOR [[MV8]](s64), [[MV9]](s64), [[MV10]](s64), [[MV11]](s64)
-    ; GFX10: [[COPY24:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10: [[FMUL:%[0-9]+]]:_(<4 x s64>) = reassoc G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX10: [[FADD:%[0-9]+]]:_(<4 x s64>) = reassoc G_FADD [[FMUL]], [[BUILD_VECTOR2]]
     ; GFX10: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<4 x s64>)
     ; GFX10: $vgpr0 = COPY [[UV]](s32)
     ; GFX10: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10: $vgpr2 = COPY [[UV2]](s32)
     ; GFX10: $vgpr3 = COPY [[UV3]](s32)
     ; GFX10: $vgpr4 = COPY [[UV4]](s32)
     ; GFX10: $vgpr5 = COPY [[UV5]](s32)
     ; GFX10: $vgpr6 = COPY [[UV6]](s32)
     ; GFX10: $vgpr7 = COPY [[UV7]](s32)
-    ; GFX10: [[COPY25:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY24]]
-    ; GFX10: S_SETPC_B64_return [[COPY25]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
+    ; GFX10: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
     ; GFX10-CONTRACT-LABEL: name: test_4xdouble_add_mul
     ; GFX10-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-CONTRACT: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-CONTRACT: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-CONTRACT: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-CONTRACT: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
bb.1.entry:
     ; GFX10-CONTRACT: [[COPY21:%[0-9]+]]:_(s32) = COPY $vgpr21
     ; GFX10-CONTRACT: [[COPY22:%[0-9]+]]:_(s32) = COPY $vgpr22
     ; GFX10-CONTRACT: [[COPY23:%[0-9]+]]:_(s32) = COPY $vgpr23
     ; GFX10-CONTRACT: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX10-CONTRACT: [[MV9:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY18]](s32), [[COPY19]](s32)
     ; GFX10-CONTRACT: [[MV10:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY20]](s32), [[COPY21]](s32)
     ; GFX10-CONTRACT: [[MV11:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY22]](s32), [[COPY23]](s32)
     ; GFX10-CONTRACT: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s64>) = G_BUILD_VECTOR [[MV8]](s64), [[MV9]](s64), [[MV10]](s64), [[MV11]](s64)
-    ; GFX10-CONTRACT: [[COPY24:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-CONTRACT: [[FMA:%[0-9]+]]:_(<4 x s64>) = G_FMA [[BUILD_VECTOR]], [[BUILD_VECTOR1]], [[BUILD_VECTOR2]]
     ; GFX10-CONTRACT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](<4 x s64>)
     ; GFX10-CONTRACT: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-CONTRACT: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10-CONTRACT: $vgpr2 = COPY [[UV2]](s32)
     ; GFX10-CONTRACT: $vgpr3 = COPY [[UV3]](s32)
     ; GFX10-CONTRACT: $vgpr4 = COPY [[UV4]](s32)
     ; GFX10-CONTRACT: $vgpr5 = COPY [[UV5]](s32)
     ; GFX10-CONTRACT: $vgpr6 = COPY [[UV6]](s32)
     ; GFX10-CONTRACT: $vgpr7 = COPY [[UV7]](s32)
-    ; GFX10-CONTRACT: [[COPY25:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY24]]
-    ; GFX10-CONTRACT: S_SETPC_B64_return [[COPY25]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
+    ; GFX10-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
     ; GFX10-DENORM-LABEL: name: test_4xdouble_add_mul
     ; GFX10-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-DENORM: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-DENORM: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-DENORM: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-DENORM: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
bb.1.entry:
     ; GFX10-DENORM: [[COPY21:%[0-9]+]]:_(s32) = COPY $vgpr21
     ; GFX10-DENORM: [[COPY22:%[0-9]+]]:_(s32) = COPY $vgpr22
     ; GFX10-DENORM: [[COPY23:%[0-9]+]]:_(s32) = COPY $vgpr23
     ; GFX10-DENORM: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX10-DENORM: [[MV9:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY18]](s32), [[COPY19]](s32)
     ; GFX10-DENORM: [[MV10:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY20]](s32), [[COPY21]](s32)
     ; GFX10-DENORM: [[MV11:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY22]](s32), [[COPY23]](s32)
     ; GFX10-DENORM: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s64>) = G_BUILD_VECTOR [[MV8]](s64), [[MV9]](s64), [[MV10]](s64), [[MV11]](s64)
-    ; GFX10-DENORM: [[COPY24:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-DENORM: [[FMUL:%[0-9]+]]:_(<4 x s64>) = reassoc G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX10-DENORM: [[FADD:%[0-9]+]]:_(<4 x s64>) = reassoc G_FADD [[FMUL]], [[BUILD_VECTOR2]]
     ; GFX10-DENORM: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<4 x s64>)
     ; GFX10-DENORM: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-DENORM: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10-DENORM: $vgpr2 = COPY [[UV2]](s32)
     ; GFX10-DENORM: $vgpr3 = COPY [[UV3]](s32)
     ; GFX10-DENORM: $vgpr4 = COPY [[UV4]](s32)
     ; GFX10-DENORM: $vgpr5 = COPY [[UV5]](s32)
     ; GFX10-DENORM: $vgpr6 = COPY [[UV6]](s32)
     ; GFX10-DENORM: $vgpr7 = COPY [[UV7]](s32)
-    ; GFX10-DENORM: [[COPY25:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY24]]
-    ; GFX10-DENORM: S_SETPC_B64_return [[COPY25]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
+    ; GFX10-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
     ; GFX10-UNSAFE-LABEL: name: test_4xdouble_add_mul
     ; GFX10-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-UNSAFE: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-UNSAFE: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-UNSAFE: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-UNSAFE: [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
bb.1.entry:
     ; GFX10-UNSAFE: [[COPY21:%[0-9]+]]:_(s32) = COPY $vgpr21
     ; GFX10-UNSAFE: [[COPY22:%[0-9]+]]:_(s32) = COPY $vgpr22
     ; GFX10-UNSAFE: [[COPY23:%[0-9]+]]:_(s32) = COPY $vgpr23
     ; GFX10-UNSAFE: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX10-UNSAFE: [[MV9:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY18]](s32), [[COPY19]](s32)
     ; GFX10-UNSAFE: [[MV10:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY20]](s32), [[COPY21]](s32)
     ; GFX10-UNSAFE: [[MV11:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY22]](s32), [[COPY23]](s32)
     ; GFX10-UNSAFE: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s64>) = G_BUILD_VECTOR [[MV8]](s64), [[MV9]](s64), [[MV10]](s64), [[MV11]](s64)
-    ; GFX10-UNSAFE: [[COPY24:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-UNSAFE: [[FMA:%[0-9]+]]:_(<4 x s64>) = G_FMA [[BUILD_VECTOR]], [[BUILD_VECTOR1]], [[BUILD_VECTOR2]]
     ; GFX10-UNSAFE: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](<4 x s64>)
     ; GFX10-UNSAFE: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-UNSAFE: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10-UNSAFE: $vgpr2 = COPY [[UV2]](s32)
     ; GFX10-UNSAFE: $vgpr3 = COPY [[UV3]](s32)
     ; GFX10-UNSAFE: $vgpr4 = COPY [[UV4]](s32)
     ; GFX10-UNSAFE: $vgpr5 = COPY [[UV5]](s32)
     ; GFX10-UNSAFE: $vgpr6 = COPY [[UV6]](s32)
     ; GFX10-UNSAFE: $vgpr7 = COPY [[UV7]](s32)
-    ; GFX10-UNSAFE: [[COPY25:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY24]]
-    ; GFX10-UNSAFE: S_SETPC_B64_return [[COPY25]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
+    ; GFX10-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
     %4:_(s32) = COPY $vgpr0
     %5:_(s32) = COPY $vgpr1
     %6:_(s32) = COPY $vgpr2
     %7:_(s32) = COPY $vgpr3
     %8:_(s32) = COPY $vgpr4
     %9:_(s32) = COPY $vgpr5
     %10:_(s32) = COPY $vgpr6
     %11:_(s32) = COPY $vgpr7
bb.1.entry:
     %25:_(s32) = COPY $vgpr21
     %26:_(s32) = COPY $vgpr22
     %27:_(s32) = COPY $vgpr23
     %36:_(s64) = G_MERGE_VALUES %20(s32), %21(s32)
     %37:_(s64) = G_MERGE_VALUES %22(s32), %23(s32)
     %38:_(s64) = G_MERGE_VALUES %24(s32), %25(s32)
     %39:_(s64) = G_MERGE_VALUES %26(s32), %27(s32)
     %2:_(<4 x s64>) = G_BUILD_VECTOR %36(s64), %37(s64), %38(s64), %39(s64)
-    %3:sgpr_64 = COPY $sgpr30_sgpr31
     %40:_(<4 x s64>) = reassoc G_FMUL %0, %1
     %41:_(<4 x s64>) = reassoc G_FADD %40, %2
     %43:_(s32), %44:_(s32), %45:_(s32), %46:_(s32), %47:_(s32), %48:_(s32), %49:_(s32), %50:_(s32) = G_UNMERGE_VALUES %41(<4 x s64>)
     $vgpr0 = COPY %43(s32)
     $vgpr1 = COPY %44(s32)
     $vgpr2 = COPY %45(s32)
     $vgpr3 = COPY %46(s32)
     $vgpr4 = COPY %47(s32)
     $vgpr5 = COPY %48(s32)
     $vgpr6 = COPY %49(s32)
     $vgpr7 = COPY %50(s32)
-    %42:ccr_sgpr_64 = COPY %3
-    S_SETPC_B64_return %42, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
 ...
 
 ---
 name:            test_3xdouble_add_mul_rhs
 body:             |
   bb.1.entry:
-    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7, $vgpr8, $vgpr9, $vgpr10, $vgpr11, $vgpr12, $vgpr13, $vgpr14, $vgpr15, $vgpr16, $vgpr17, $sgpr30_sgpr31
+    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7, $vgpr8, $vgpr9, $vgpr10, $vgpr11, $vgpr12, $vgpr13, $vgpr14, $vgpr15, $vgpr16, $vgpr17
 
     ; GFX9-LABEL: name: test_3xdouble_add_mul_rhs
     ; GFX9: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
bb.1.entry:
     ; GFX9: [[COPY14:%[0-9]+]]:_(s32) = COPY $vgpr14
     ; GFX9: [[COPY15:%[0-9]+]]:_(s32) = COPY $vgpr15
     ; GFX9: [[COPY16:%[0-9]+]]:_(s32) = COPY $vgpr16
     ; GFX9: [[COPY17:%[0-9]+]]:_(s32) = COPY $vgpr17
     ; GFX9: [[MV6:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY12]](s32), [[COPY13]](s32)
     ; GFX9: [[MV7:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY14]](s32), [[COPY15]](s32)
     ; GFX9: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX9: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s64>) = G_BUILD_VECTOR [[MV6]](s64), [[MV7]](s64), [[MV8]](s64)
-    ; GFX9: [[COPY18:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9: [[FMUL:%[0-9]+]]:_(<3 x s64>) = reassoc G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX9: [[FADD:%[0-9]+]]:_(<3 x s64>) = reassoc G_FADD [[BUILD_VECTOR2]], [[FMUL]]
     ; GFX9: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<3 x s64>)
     ; GFX9: $vgpr0 = COPY [[UV]](s32)
     ; GFX9: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9: $vgpr2 = COPY [[UV2]](s32)
     ; GFX9: $vgpr3 = COPY [[UV3]](s32)
     ; GFX9: $vgpr4 = COPY [[UV4]](s32)
     ; GFX9: $vgpr5 = COPY [[UV5]](s32)
-    ; GFX9: [[COPY19:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY18]]
-    ; GFX9: S_SETPC_B64_return [[COPY19]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
+    ; GFX9: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
     ; GFX9-CONTRACT-LABEL: name: test_3xdouble_add_mul_rhs
     ; GFX9-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-CONTRACT: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-CONTRACT: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-CONTRACT: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-CONTRACT: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
bb.1.entry:
     ; GFX9-CONTRACT: [[COPY14:%[0-9]+]]:_(s32) = COPY $vgpr14
     ; GFX9-CONTRACT: [[COPY15:%[0-9]+]]:_(s32) = COPY $vgpr15
     ; GFX9-CONTRACT: [[COPY16:%[0-9]+]]:_(s32) = COPY $vgpr16
     ; GFX9-CONTRACT: [[COPY17:%[0-9]+]]:_(s32) = COPY $vgpr17
     ; GFX9-CONTRACT: [[MV6:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY12]](s32), [[COPY13]](s32)
     ; GFX9-CONTRACT: [[MV7:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY14]](s32), [[COPY15]](s32)
     ; GFX9-CONTRACT: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX9-CONTRACT: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s64>) = G_BUILD_VECTOR [[MV6]](s64), [[MV7]](s64), [[MV8]](s64)
-    ; GFX9-CONTRACT: [[COPY18:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-CONTRACT: [[FMA:%[0-9]+]]:_(<3 x s64>) = G_FMA [[BUILD_VECTOR]], [[BUILD_VECTOR1]], [[BUILD_VECTOR2]]
     ; GFX9-CONTRACT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](<3 x s64>)
     ; GFX9-CONTRACT: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-CONTRACT: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9-CONTRACT: $vgpr2 = COPY [[UV2]](s32)
     ; GFX9-CONTRACT: $vgpr3 = COPY [[UV3]](s32)
     ; GFX9-CONTRACT: $vgpr4 = COPY [[UV4]](s32)
     ; GFX9-CONTRACT: $vgpr5 = COPY [[UV5]](s32)
-    ; GFX9-CONTRACT: [[COPY19:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY18]]
-    ; GFX9-CONTRACT: S_SETPC_B64_return [[COPY19]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
+    ; GFX9-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
     ; GFX9-DENORM-LABEL: name: test_3xdouble_add_mul_rhs
     ; GFX9-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-DENORM: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-DENORM: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-DENORM: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-DENORM: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
bb.1.entry:
     ; GFX9-DENORM: [[COPY14:%[0-9]+]]:_(s32) = COPY $vgpr14
     ; GFX9-DENORM: [[COPY15:%[0-9]+]]:_(s32) = COPY $vgpr15
     ; GFX9-DENORM: [[COPY16:%[0-9]+]]:_(s32) = COPY $vgpr16
     ; GFX9-DENORM: [[COPY17:%[0-9]+]]:_(s32) = COPY $vgpr17
     ; GFX9-DENORM: [[MV6:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY12]](s32), [[COPY13]](s32)
     ; GFX9-DENORM: [[MV7:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY14]](s32), [[COPY15]](s32)
     ; GFX9-DENORM: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX9-DENORM: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s64>) = G_BUILD_VECTOR [[MV6]](s64), [[MV7]](s64), [[MV8]](s64)
-    ; GFX9-DENORM: [[COPY18:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-DENORM: [[FMUL:%[0-9]+]]:_(<3 x s64>) = reassoc G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX9-DENORM: [[FADD:%[0-9]+]]:_(<3 x s64>) = reassoc G_FADD [[BUILD_VECTOR2]], [[FMUL]]
     ; GFX9-DENORM: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<3 x s64>)
     ; GFX9-DENORM: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-DENORM: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9-DENORM: $vgpr2 = COPY [[UV2]](s32)
     ; GFX9-DENORM: $vgpr3 = COPY [[UV3]](s32)
     ; GFX9-DENORM: $vgpr4 = COPY [[UV4]](s32)
     ; GFX9-DENORM: $vgpr5 = COPY [[UV5]](s32)
-    ; GFX9-DENORM: [[COPY19:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY18]]
-    ; GFX9-DENORM: S_SETPC_B64_return [[COPY19]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
+    ; GFX9-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
     ; GFX9-UNSAFE-LABEL: name: test_3xdouble_add_mul_rhs
     ; GFX9-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX9-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX9-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX9-UNSAFE: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX9-UNSAFE: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX9-UNSAFE: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX9-UNSAFE: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
bb.1.entry:
     ; GFX9-UNSAFE: [[COPY14:%[0-9]+]]:_(s32) = COPY $vgpr14
     ; GFX9-UNSAFE: [[COPY15:%[0-9]+]]:_(s32) = COPY $vgpr15
     ; GFX9-UNSAFE: [[COPY16:%[0-9]+]]:_(s32) = COPY $vgpr16
     ; GFX9-UNSAFE: [[COPY17:%[0-9]+]]:_(s32) = COPY $vgpr17
     ; GFX9-UNSAFE: [[MV6:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY12]](s32), [[COPY13]](s32)
     ; GFX9-UNSAFE: [[MV7:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY14]](s32), [[COPY15]](s32)
     ; GFX9-UNSAFE: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX9-UNSAFE: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s64>) = G_BUILD_VECTOR [[MV6]](s64), [[MV7]](s64), [[MV8]](s64)
-    ; GFX9-UNSAFE: [[COPY18:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX9-UNSAFE: [[FMA:%[0-9]+]]:_(<3 x s64>) = G_FMA [[BUILD_VECTOR]], [[BUILD_VECTOR1]], [[BUILD_VECTOR2]]
     ; GFX9-UNSAFE: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](<3 x s64>)
     ; GFX9-UNSAFE: $vgpr0 = COPY [[UV]](s32)
     ; GFX9-UNSAFE: $vgpr1 = COPY [[UV1]](s32)
     ; GFX9-UNSAFE: $vgpr2 = COPY [[UV2]](s32)
     ; GFX9-UNSAFE: $vgpr3 = COPY [[UV3]](s32)
     ; GFX9-UNSAFE: $vgpr4 = COPY [[UV4]](s32)
     ; GFX9-UNSAFE: $vgpr5 = COPY [[UV5]](s32)
-    ; GFX9-UNSAFE: [[COPY19:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY18]]
-    ; GFX9-UNSAFE: S_SETPC_B64_return [[COPY19]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
+    ; GFX9-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
     ; GFX10-LABEL: name: test_3xdouble_add_mul_rhs
     ; GFX10: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
bb.1.entry:
     ; GFX10: [[COPY14:%[0-9]+]]:_(s32) = COPY $vgpr14
     ; GFX10: [[COPY15:%[0-9]+]]:_(s32) = COPY $vgpr15
     ; GFX10: [[COPY16:%[0-9]+]]:_(s32) = COPY $vgpr16
     ; GFX10: [[COPY17:%[0-9]+]]:_(s32) = COPY $vgpr17
     ; GFX10: [[MV6:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY12]](s32), [[COPY13]](s32)
     ; GFX10: [[MV7:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY14]](s32), [[COPY15]](s32)
     ; GFX10: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX10: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s64>) = G_BUILD_VECTOR [[MV6]](s64), [[MV7]](s64), [[MV8]](s64)
-    ; GFX10: [[COPY18:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10: [[FMUL:%[0-9]+]]:_(<3 x s64>) = reassoc G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX10: [[FADD:%[0-9]+]]:_(<3 x s64>) = reassoc G_FADD [[BUILD_VECTOR2]], [[FMUL]]
     ; GFX10: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<3 x s64>)
     ; GFX10: $vgpr0 = COPY [[UV]](s32)
     ; GFX10: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10: $vgpr2 = COPY [[UV2]](s32)
     ; GFX10: $vgpr3 = COPY [[UV3]](s32)
     ; GFX10: $vgpr4 = COPY [[UV4]](s32)
     ; GFX10: $vgpr5 = COPY [[UV5]](s32)
-    ; GFX10: [[COPY19:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY18]]
-    ; GFX10: S_SETPC_B64_return [[COPY19]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
+    ; GFX10: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
     ; GFX10-CONTRACT-LABEL: name: test_3xdouble_add_mul_rhs
     ; GFX10-CONTRACT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-CONTRACT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-CONTRACT: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-CONTRACT: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-CONTRACT: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-CONTRACT: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-CONTRACT: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
bb.1.entry:
     ; GFX10-CONTRACT: [[COPY14:%[0-9]+]]:_(s32) = COPY $vgpr14
     ; GFX10-CONTRACT: [[COPY15:%[0-9]+]]:_(s32) = COPY $vgpr15
     ; GFX10-CONTRACT: [[COPY16:%[0-9]+]]:_(s32) = COPY $vgpr16
     ; GFX10-CONTRACT: [[COPY17:%[0-9]+]]:_(s32) = COPY $vgpr17
     ; GFX10-CONTRACT: [[MV6:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY12]](s32), [[COPY13]](s32)
     ; GFX10-CONTRACT: [[MV7:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY14]](s32), [[COPY15]](s32)
     ; GFX10-CONTRACT: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX10-CONTRACT: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s64>) = G_BUILD_VECTOR [[MV6]](s64), [[MV7]](s64), [[MV8]](s64)
-    ; GFX10-CONTRACT: [[COPY18:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-CONTRACT: [[FMA:%[0-9]+]]:_(<3 x s64>) = G_FMA [[BUILD_VECTOR]], [[BUILD_VECTOR1]], [[BUILD_VECTOR2]]
     ; GFX10-CONTRACT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](<3 x s64>)
     ; GFX10-CONTRACT: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-CONTRACT: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10-CONTRACT: $vgpr2 = COPY [[UV2]](s32)
     ; GFX10-CONTRACT: $vgpr3 = COPY [[UV3]](s32)
     ; GFX10-CONTRACT: $vgpr4 = COPY [[UV4]](s32)
     ; GFX10-CONTRACT: $vgpr5 = COPY [[UV5]](s32)
-    ; GFX10-CONTRACT: [[COPY19:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY18]]
-    ; GFX10-CONTRACT: S_SETPC_B64_return [[COPY19]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
+    ; GFX10-CONTRACT: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
     ; GFX10-DENORM-LABEL: name: test_3xdouble_add_mul_rhs
     ; GFX10-DENORM: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-DENORM: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-DENORM: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-DENORM: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-DENORM: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-DENORM: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-DENORM: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
bb.1.entry:
     ; GFX10-DENORM: [[COPY14:%[0-9]+]]:_(s32) = COPY $vgpr14
     ; GFX10-DENORM: [[COPY15:%[0-9]+]]:_(s32) = COPY $vgpr15
     ; GFX10-DENORM: [[COPY16:%[0-9]+]]:_(s32) = COPY $vgpr16
     ; GFX10-DENORM: [[COPY17:%[0-9]+]]:_(s32) = COPY $vgpr17
     ; GFX10-DENORM: [[MV6:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY12]](s32), [[COPY13]](s32)
     ; GFX10-DENORM: [[MV7:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY14]](s32), [[COPY15]](s32)
     ; GFX10-DENORM: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX10-DENORM: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s64>) = G_BUILD_VECTOR [[MV6]](s64), [[MV7]](s64), [[MV8]](s64)
-    ; GFX10-DENORM: [[COPY18:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-DENORM: [[FMUL:%[0-9]+]]:_(<3 x s64>) = reassoc G_FMUL [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
     ; GFX10-DENORM: [[FADD:%[0-9]+]]:_(<3 x s64>) = reassoc G_FADD [[BUILD_VECTOR2]], [[FMUL]]
     ; GFX10-DENORM: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FADD]](<3 x s64>)
     ; GFX10-DENORM: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-DENORM: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10-DENORM: $vgpr2 = COPY [[UV2]](s32)
     ; GFX10-DENORM: $vgpr3 = COPY [[UV3]](s32)
     ; GFX10-DENORM: $vgpr4 = COPY [[UV4]](s32)
     ; GFX10-DENORM: $vgpr5 = COPY [[UV5]](s32)
-    ; GFX10-DENORM: [[COPY19:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY18]]
-    ; GFX10-DENORM: S_SETPC_B64_return [[COPY19]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
+    ; GFX10-DENORM: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
     ; GFX10-UNSAFE-LABEL: name: test_3xdouble_add_mul_rhs
     ; GFX10-UNSAFE: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; GFX10-UNSAFE: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
     ; GFX10-UNSAFE: [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
     ; GFX10-UNSAFE: [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
     ; GFX10-UNSAFE: [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
     ; GFX10-UNSAFE: [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
     ; GFX10-UNSAFE: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
bb.1.entry:
     ; GFX10-UNSAFE: [[COPY14:%[0-9]+]]:_(s32) = COPY $vgpr14
     ; GFX10-UNSAFE: [[COPY15:%[0-9]+]]:_(s32) = COPY $vgpr15
     ; GFX10-UNSAFE: [[COPY16:%[0-9]+]]:_(s32) = COPY $vgpr16
     ; GFX10-UNSAFE: [[COPY17:%[0-9]+]]:_(s32) = COPY $vgpr17
     ; GFX10-UNSAFE: [[MV6:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY12]](s32), [[COPY13]](s32)
     ; GFX10-UNSAFE: [[MV7:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY14]](s32), [[COPY15]](s32)
     ; GFX10-UNSAFE: [[MV8:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY16]](s32), [[COPY17]](s32)
     ; GFX10-UNSAFE: [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s64>) = G_BUILD_VECTOR [[MV6]](s64), [[MV7]](s64), [[MV8]](s64)
-    ; GFX10-UNSAFE: [[COPY18:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; GFX10-UNSAFE: [[FMA:%[0-9]+]]:_(<3 x s64>) = G_FMA [[BUILD_VECTOR]], [[BUILD_VECTOR1]], [[BUILD_VECTOR2]]
     ; GFX10-UNSAFE: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[FMA]](<3 x s64>)
     ; GFX10-UNSAFE: $vgpr0 = COPY [[UV]](s32)
     ; GFX10-UNSAFE: $vgpr1 = COPY [[UV1]](s32)
     ; GFX10-UNSAFE: $vgpr2 = COPY [[UV2]](s32)
     ; GFX10-UNSAFE: $vgpr3 = COPY [[UV3]](s32)
     ; GFX10-UNSAFE: $vgpr4 = COPY [[UV4]](s32)
     ; GFX10-UNSAFE: $vgpr5 = COPY [[UV5]](s32)
-    ; GFX10-UNSAFE: [[COPY19:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY18]]
-    ; GFX10-UNSAFE: S_SETPC_B64_return [[COPY19]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
+    ; GFX10-UNSAFE: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
     %4:_(s32) = COPY $vgpr0
     %5:_(s32) = COPY $vgpr1
     %6:_(s32) = COPY $vgpr2
     %7:_(s32) = COPY $vgpr3
     %8:_(s32) = COPY $vgpr4
     %9:_(s32) = COPY $vgpr5
     %22:_(s64) = G_MERGE_VALUES %4(s32), %5(s32)
     %23:_(s64) = G_MERGE_VALUES %6(s32), %7(s32)
bb.1.entry:
     %18:_(s32) = COPY $vgpr14
     %19:_(s32) = COPY $vgpr15
     %20:_(s32) = COPY $vgpr16
     %21:_(s32) = COPY $vgpr17
     %28:_(s64) = G_MERGE_VALUES %16(s32), %17(s32)
     %29:_(s64) = G_MERGE_VALUES %18(s32), %19(s32)
     %30:_(s64) = G_MERGE_VALUES %20(s32), %21(s32)
     %2:_(<3 x s64>) = G_BUILD_VECTOR %28(s64), %29(s64), %30(s64)
-    %3:sgpr_64 = COPY $sgpr30_sgpr31
     %31:_(<3 x s64>) = reassoc G_FMUL %0, %1
     %32:_(<3 x s64>) = reassoc G_FADD %2, %31
     %34:_(s32), %35:_(s32), %36:_(s32), %37:_(s32), %38:_(s32), %39:_(s32) = G_UNMERGE_VALUES %32(<3 x s64>)
     $vgpr0 = COPY %34(s32)
     $vgpr1 = COPY %35(s32)
     $vgpr2 = COPY %36(s32)
     $vgpr3 = COPY %37(s32)
     $vgpr4 = COPY %38(s32)
     $vgpr5 = COPY %39(s32)
-    %33:ccr_sgpr_64 = COPY %3
-    S_SETPC_B64_return %33, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
 ...

llvm/test/CodeGen/AMDGPU/GlobalISel/combine-shift-imm-chain-illegal-types.mir

 ---
 name:            ushlsat_i44
 tracksRegLiveness: true
 body:             |
   bb.0:
     liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
 
     ; CHECK-LABEL: name: ushlsat_i44
-    ; CHECK: liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+    ; CHECK: liveins: $vgpr0, $vgpr1
     ; CHECK: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; CHECK: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-    ; CHECK: [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; CHECK: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; CHECK: [[TRUNC:%[0-9]+]]:_(s44) = G_TRUNC [[MV]](s64)
     ; CHECK: [[C:%[0-9]+]]:_(s44) = G_CONSTANT i44 22
     ; CHECK: [[USHLSAT:%[0-9]+]]:_(s44) = G_USHLSAT [[TRUNC]], [[C]](s44)
     ; CHECK: [[USHLSAT1:%[0-9]+]]:_(s44) = G_USHLSAT [[USHLSAT]], [[C]](s44)
     ; CHECK: [[ANYEXT:%[0-9]+]]:_(s64) = G_ANYEXT [[USHLSAT1]](s44)
     ; CHECK: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[ANYEXT]](s64)
     ; CHECK: $vgpr0 = COPY [[UV]](s32)
     ; CHECK: $vgpr1 = COPY [[UV1]](s32)
-    ; CHECK: [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-    ; CHECK: S_SETPC_B64_return [[COPY3]], implicit $vgpr0, implicit $vgpr1
+    ; CHECK: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     %2:_(s32) = COPY $vgpr0
     %3:_(s32) = COPY $vgpr1
-    %1:sgpr_64 = COPY $sgpr30_sgpr31
     %4:_(s64) = G_MERGE_VALUES %2(s32), %3(s32)
     %0:_(s44) = G_TRUNC %4(s64)
     %5:_(s44) = G_CONSTANT i44 22
     %6:_(s44) = G_USHLSAT %0, %5(s44)
     %7:_(s44) = G_USHLSAT %6, %5(s44)
     %9:_(s64) = G_ANYEXT %7(s44)
     %10:_(s32), %11:_(s32) = G_UNMERGE_VALUES %9(s64)
     $vgpr0 = COPY %10(s32)
     $vgpr1 = COPY %11(s32)
-    %8:ccr_sgpr_64 = COPY %1
-    S_SETPC_B64_return %8, implicit $vgpr0, implicit $vgpr1
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
 
 ...
 ---
 name:            ushlsat_i55
 tracksRegLiveness: true
 body:             |
   bb.0:
-    liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+    liveins: $vgpr0, $vgpr1
 
     ; CHECK-LABEL: name: ushlsat_i55
-    ; CHECK: liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+    ; CHECK: liveins: $vgpr0, $vgpr1
     ; CHECK: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
     ; CHECK: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-    ; CHECK: [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; CHECK: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
     ; CHECK: [[TRUNC:%[0-9]+]]:_(s55) = G_TRUNC [[MV]](s64)
     ; CHECK: [[C:%[0-9]+]]:_(s55) = G_CONSTANT i55 53
     ; CHECK: [[USHLSAT:%[0-9]+]]:_(s55) = G_USHLSAT [[TRUNC]], [[C]](s55)
     ; CHECK: [[ANYEXT:%[0-9]+]]:_(s64) = G_ANYEXT [[USHLSAT]](s55)
     ; CHECK: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[ANYEXT]](s64)
     ; CHECK: $vgpr0 = COPY [[UV]](s32)
     ; CHECK: $vgpr1 = COPY [[UV1]](s32)
-    ; CHECK: [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-    ; CHECK: S_SETPC_B64_return [[COPY3]], implicit $vgpr0, implicit $vgpr1
+    ; CHECK: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
     %2:_(s32) = COPY $vgpr0
     %3:_(s32) = COPY $vgpr1
-    %1:sgpr_64 = COPY $sgpr30_sgpr31
     %4:_(s64) = G_MERGE_VALUES %2(s32), %3(s32)
     %0:_(s55) = G_TRUNC %4(s64)
     %5:_(s55) = G_CONSTANT i55 50
     %7:_(s55) = G_CONSTANT i55 3
     %6:_(s55) = G_USHLSAT %0, %5(s55)
     %8:_(s55) = G_USHLSAT %6, %7(s55)
     %10:_(s64) = G_ANYEXT %8(s55)
     %11:_(s32), %12:_(s32) = G_UNMERGE_VALUES %10(s64)
     $vgpr0 = COPY %11(s32)
     $vgpr1 = COPY %12(s32)
-    %9:ccr_sgpr_64 = COPY %1
-    S_SETPC_B64_return %9, implicit $vgpr0, implicit $vgpr1
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit $vgpr1
 
 ...

llvm/test/CodeGen/AMDGPU/GlobalISel/dummy-target.ll

 ; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
 ; RUN: llc -global-isel -mtriple=amdgcn-amd-amdhsa -stop-after=legalizer -o - %s | FileCheck %s
 
 ; Make sure legalizer info doesn't assert on dummy targets
 
 define i16 @vop3p_add_i16(i16 %arg0) #0 {
   ; CHECK-LABEL: name: vop3p_add_i16
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[ADD:%[0-9]+]]:_(s16) = G_ADD [[TRUNC]], [[TRUNC]]
-  ; CHECK:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[ADD]](s16)
-  ; CHECK:   $vgpr0 = COPY [[ANYEXT]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-  ; CHECK:   S_SETPC_B64_return [[COPY2]], implicit $vgpr0
+  ; CHECK-NEXT:   liveins: $vgpr0
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
+  ; CHECK-NEXT:   [[ADD:%[0-9]+]]:_(s16) = G_ADD [[TRUNC]], [[TRUNC]]
+  ; CHECK-NEXT:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[ADD]](s16)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[ANYEXT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %add = add i16 %arg0, %arg0
   ret i16 %add
 }
 
 define <2 x i16> @vop3p_add_v2i16(<2 x i16> %arg0) #0 {
   ; CHECK-LABEL: name: vop3p_add_v2i16
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>)
-  ; CHECK:   [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32)
-  ; CHECK:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
-  ; CHECK:   [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
-  ; CHECK:   [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
-  ; CHECK:   [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>)
-  ; CHECK:   [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32)
-  ; CHECK:   [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
-  ; CHECK:   [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32)
-  ; CHECK:   [[ADD:%[0-9]+]]:_(s16) = G_ADD [[TRUNC]], [[TRUNC2]]
-  ; CHECK:   [[ADD1:%[0-9]+]]:_(s16) = G_ADD [[TRUNC1]], [[TRUNC3]]
-  ; CHECK:   [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[ADD]](s16)
-  ; CHECK:   [[ZEXT1:%[0-9]+]]:_(s32) = G_ZEXT [[ADD1]](s16)
-  ; CHECK:   [[SHL:%[0-9]+]]:_(s32) = G_SHL [[ZEXT1]], [[C]](s32)
-  ; CHECK:   [[OR:%[0-9]+]]:_(s32) = G_OR [[ZEXT]], [[SHL]]
-  ; CHECK:   [[BITCAST2:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32)
-  ; CHECK:   $vgpr0 = COPY [[BITCAST2]](<2 x s16>)
-  ; CHECK:   [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-  ; CHECK:   S_SETPC_B64_return [[COPY2]], implicit $vgpr0
+  ; CHECK-NEXT:   liveins: $vgpr0
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
+  ; CHECK-NEXT:   [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>)
+  ; CHECK-NEXT:   [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST]](s32)
+  ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
+  ; CHECK-NEXT:   [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
+  ; CHECK-NEXT:   [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
+  ; CHECK-NEXT:   [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>)
+  ; CHECK-NEXT:   [[TRUNC2:%[0-9]+]]:_(s16) = G_TRUNC [[BITCAST1]](s32)
+  ; CHECK-NEXT:   [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
+  ; CHECK-NEXT:   [[TRUNC3:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32)
+  ; CHECK-NEXT:   [[ADD:%[0-9]+]]:_(s16) = G_ADD [[TRUNC]], [[TRUNC2]]
+  ; CHECK-NEXT:   [[ADD1:%[0-9]+]]:_(s16) = G_ADD [[TRUNC1]], [[TRUNC3]]
+  ; CHECK-NEXT:   [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[ADD]](s16)
+  ; CHECK-NEXT:   [[ZEXT1:%[0-9]+]]:_(s32) = G_ZEXT [[ADD1]](s16)
+  ; CHECK-NEXT:   [[SHL:%[0-9]+]]:_(s32) = G_SHL [[ZEXT1]], [[C]](s32)
+  ; CHECK-NEXT:   [[OR:%[0-9]+]]:_(s32) = G_OR [[ZEXT]], [[SHL]]
+  ; CHECK-NEXT:   [[BITCAST2:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[BITCAST2]](<2 x s16>)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %add = add <2 x i16> %arg0, %arg0
   ret <2 x i16> %add
 }
 
 define i16 @halfinsts_add_i16(i16 %arg0) #1 {
   ; CHECK-LABEL: name: halfinsts_add_i16
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[ADD:%[0-9]+]]:_(s32) = G_ADD [[COPY]], [[COPY]]
-  ; CHECK:   $vgpr0 = COPY [[ADD]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-  ; CHECK:   S_SETPC_B64_return [[COPY2]], implicit $vgpr0
+  ; CHECK-NEXT:   liveins: $vgpr0
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[ADD:%[0-9]+]]:_(s32) = G_ADD [[COPY]], [[COPY]]
+  ; CHECK-NEXT:   $vgpr0 = COPY [[ADD]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %add = add i16 %arg0, %arg0
   ret i16 %add
 }
 
 define <2 x i16> @halfinsts_add_v2i16(<2 x i16> %arg0) #1 {
   ; CHECK-LABEL: name: halfinsts_add_v2i16
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[ADD:%[0-9]+]]:_(s32) = G_ADD [[COPY]], [[COPY]]
-  ; CHECK:   [[ADD1:%[0-9]+]]:_(s32) = G_ADD [[COPY1]], [[COPY1]]
-  ; CHECK:   $vgpr0 = COPY [[ADD]](s32)
-  ; CHECK:   $vgpr1 = COPY [[ADD1]](s32)
-  ; CHECK:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[ADD:%[0-9]+]]:_(s32) = G_ADD [[COPY]], [[COPY]]
+  ; CHECK-NEXT:   [[ADD1:%[0-9]+]]:_(s32) = G_ADD [[COPY1]], [[COPY1]]
+  ; CHECK-NEXT:   $vgpr0 = COPY [[ADD]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[ADD1]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %add = add <2 x i16> %arg0, %arg0
   ret <2 x i16> %add
 }
 
 attributes #0 = { "target-features"="+vop3p" }
 attributes #0 = { "target-features"="+16-bit-insts" }

llvm/test/CodeGen/AMDGPU/GlobalISel/fdiv.f32.ll

 ; GFX89-IEEE-NEXT:    v_fma_f32 v9, v11, v9, v9
 ; GFX89-IEEE-NEXT:    v_mul_f32_e32 v10, v6, v8
 ; GFX89-IEEE-NEXT:    v_mul_f32_e32 v11, v7, v9
 ; GFX89-IEEE-NEXT:    v_fma_f32 v12, -v4, v10, v6
 ; GFX89-IEEE-NEXT:    v_fma_f32 v13, -v5, v11, v7
 ; GFX89-IEEE-NEXT:    v_fma_f32 v10, v12, v8, v10
 ; GFX89-IEEE-NEXT:    v_fma_f32 v11, v13, v9, v11
 ; GFX89-IEEE-NEXT:    v_fma_f32 v4, -v4, v10, v6
-; GFX89-IEEE-NEXT:    v_div_fmas_f32 v4, v4, v8, v10
 ; GFX89-IEEE-NEXT:    v_fma_f32 v5, -v5, v11, v7
+; GFX89-IEEE-NEXT:    v_div_fmas_f32 v4, v4, v8, v10
 ; GFX89-IEEE-NEXT:    s_mov_b64 vcc, s[4:5]
 ; GFX89-IEEE-NEXT:    v_div_fmas_f32 v5, v5, v9, v11
 ; GFX89-IEEE-NEXT:    v_div_fixup_f32 v0, v4, v2, v0
 ; GFX89-IEEE-NEXT:    v_div_fixup_f32 v1, v5, v3, v1
 ; GFX89-IEEE-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX89-FLUSH-LABEL: v_fdiv_v2f32:
 ; GFX89-FLUSH:       ; %bb.0:

 ; GFX89-IEEE-NEXT:    v_fma_f32 v9, v11, v9, v9
 ; GFX89-IEEE-NEXT:    v_mul_f32_e32 v10, v6, v8
 ; GFX89-IEEE-NEXT:    v_mul_f32_e32 v11, v7, v9
 ; GFX89-IEEE-NEXT:    v_fma_f32 v12, -v4, v10, v6
 ; GFX89-IEEE-NEXT:    v_fma_f32 v13, -v5, v11, v7
 ; GFX89-IEEE-NEXT:    v_fma_f32 v10, v12, v8, v10
 ; GFX89-IEEE-NEXT:    v_fma_f32 v11, v13, v9, v11
 ; GFX89-IEEE-NEXT:    v_fma_f32 v4, -v4, v10, v6
-; GFX89-IEEE-NEXT:    v_div_fmas_f32 v4, v4, v8, v10
 ; GFX89-IEEE-NEXT:    v_fma_f32 v5, -v5, v11, v7
+; GFX89-IEEE-NEXT:    v_div_fmas_f32 v4, v4, v8, v10
 ; GFX89-IEEE-NEXT:    s_mov_b64 vcc, s[4:5]
 ; GFX89-IEEE-NEXT:    v_div_fmas_f32 v5, v5, v9, v11
 ; GFX89-IEEE-NEXT:    v_div_fixup_f32 v0, v4, v2, v0
 ; GFX89-IEEE-NEXT:    v_div_fixup_f32 v1, v5, v3, v1
 ; GFX89-IEEE-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-IEEE-LABEL: v_fdiv_v2f32_ulp25:
 ; GFX10-IEEE:       ; %bb.0:

 ; GFX89-IEEE-NEXT:    v_fma_f32 v9, v11, v9, v9
 ; GFX89-IEEE-NEXT:    v_mul_f32_e32 v10, v6, v8
 ; GFX89-IEEE-NEXT:    v_mul_f32_e32 v11, v7, v9
 ; GFX89-IEEE-NEXT:    v_fma_f32 v12, -v4, v10, v6
 ; GFX89-IEEE-NEXT:    v_fma_f32 v13, -v5, v11, v7
 ; GFX89-IEEE-NEXT:    v_fma_f32 v10, v12, v8, v10
 ; GFX89-IEEE-NEXT:    v_fma_f32 v11, v13, v9, v11
 ; GFX89-IEEE-NEXT:    v_fma_f32 v4, -v4, v10, v6
-; GFX89-IEEE-NEXT:    v_div_fmas_f32 v4, v4, v8, v10
 ; GFX89-IEEE-NEXT:    v_fma_f32 v5, -v5, v11, v7
+; GFX89-IEEE-NEXT:    v_div_fmas_f32 v4, v4, v8, v10
 ; GFX89-IEEE-NEXT:    s_mov_b64 vcc, s[4:5]
 ; GFX89-IEEE-NEXT:    v_div_fmas_f32 v5, v5, v9, v11
 ; GFX89-IEEE-NEXT:    v_div_fixup_f32 v0, v4, v2, v0
 ; GFX89-IEEE-NEXT:    v_div_fixup_f32 v1, v5, v3, v1
 ; GFX89-IEEE-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-IEEE-LABEL: v_fdiv_v2f32_arcp_ulp25:
 ; GFX10-IEEE:       ; %bb.0:

llvm/test/CodeGen/AMDGPU/GlobalISel/fdiv.f64.ll

; GFX10-NEXT:    s_setpc_b64 s[30:31]
   ret <2 x double> %fdiv
 }
 
 define <2 x double> @v_rcp_v2f64(<2 x double> %x) {
 ; GFX6-LABEL: v_rcp_v2f64:
 ; GFX6:       ; %bb.0:
 ; GFX6-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX6-NEXT:    v_div_scale_f64 v[4:5], s[4:5], v[0:1], v[0:1], 1.0
-; GFX6-NEXT:    v_div_scale_f64 v[10:11], s[4:5], 1.0, v[0:1], 1.0
+; GFX6-NEXT:    v_div_scale_f64 v[10:11], s[4:5], v[2:3], v[2:3], 1.0
 ; GFX6-NEXT:    v_rcp_f64_e32 v[6:7], v[4:5]
 ; GFX6-NEXT:    v_mov_b32_e32 v18, 0x3ff00000
-; GFX6-NEXT:    v_cmp_eq_u32_e32 vcc, v18, v11
+; GFX6-NEXT:    v_rcp_f64_e32 v[14:15], v[10:11]
 ; GFX6-NEXT:    v_fma_f64 v[8:9], -v[4:5], v[6:7], 1.0
+; GFX6-NEXT:    v_fma_f64 v[16:17], -v[10:11], v[14:15], 1.0
 ; GFX6-NEXT:    v_fma_f64 v[6:7], v[6:7], v[8:9], v[6:7]
-; GFX6-NEXT:    v_div_scale_f64 v[8:9], s[4:5], v[2:3], v[2:3], 1.0
+; GFX6-NEXT:    v_div_scale_f64 v[8:9], s[4:5], 1.0, v[0:1], 1.0
 ; GFX6-NEXT:    v_fma_f64 v[12:13], -v[4:5], v[6:7], 1.0
-; GFX6-NEXT:    v_cmp_eq_u32_e64 s[4:5], v1, v5
+; GFX6-NEXT:    v_fma_f64 v[14:15], v[14:15], v[16:17], v[14:15]
 ; GFX6-NEXT:    v_fma_f64 v[6:7], v[6:7], v[12:13], v[6:7]
-; GFX6-NEXT:    v_rcp_f64_e32 v[12:13], v[8:9]
-; GFX6-NEXT:    v_mul_f64 v[14:15], v[10:11], v[6:7]
-; GFX6-NEXT:    s_xor_b64 vcc, vcc, s[4:5]
-; GFX6-NEXT:    v_fma_f64 v[10:11], -v[4:5], v[14:15], v[10:11]
-; GFX6-NEXT:    v_fma_f64 v[16:17], -v[8:9], v[12:13], 1.0
-; GFX6-NEXT:    v_div_fmas_f64 v[6:7], v[10:11], v[6:7], v[14:15]
-; GFX6-NEXT:    v_fma_f64 v[12:13], v[12:13], v[16:17], v[12:13]
+; GFX6-NEXT:    v_cmp_eq_u32_e32 vcc, v18, v9
+; GFX6-NEXT:    v_mul_f64 v[12:13], v[8:9], v[6:7]
+; GFX6-NEXT:    v_cmp_eq_u32_e64 s[4:5], v1, v5
+; GFX6-NEXT:    v_fma_f64 v[8:9], -v[4:5], v[12:13], v[8:9]
+; GFX6-NEXT:    v_fma_f64 v[4:5], -v[10:11], v[14:15], 1.0
 ; GFX6-NEXT:    v_div_scale_f64 v[16:17], s[6:7], 1.0, v[2:3], 1.0
-; GFX6-NEXT:    v_fma_f64 v[4:5], -v[8:9], v[12:13], 1.0
-; GFX6-NEXT:    v_cmp_eq_u32_e64 s[4:5], v3, v9
-; GFX6-NEXT:    v_fma_f64 v[4:5], v[12:13], v[4:5], v[12:13]
+; GFX6-NEXT:    v_fma_f64 v[4:5], v[14:15], v[4:5], v[14:15]
+; GFX6-NEXT:    s_xor_b64 vcc, vcc, s[4:5]
+; GFX6-NEXT:    v_mul_f64 v[14:15], v[16:17], v[4:5]
+; GFX6-NEXT:    v_div_fmas_f64 v[6:7], v[8:9], v[6:7], v[12:13]
+; GFX6-NEXT:    v_fma_f64 v[8:9], -v[10:11], v[14:15], v[16:17]
 ; GFX6-NEXT:    v_cmp_eq_u32_e32 vcc, v18, v17
-; GFX6-NEXT:    v_mul_f64 v[12:13], v[16:17], v[4:5]
+; GFX6-NEXT:    v_cmp_eq_u32_e64 s[4:5], v3, v11
 ; GFX6-NEXT:    s_xor_b64 vcc, vcc, s[4:5]
-; GFX6-NEXT:    v_fma_f64 v[10:11], -v[8:9], v[12:13], v[16:17]
 ; GFX6-NEXT:    v_div_fixup_f64 v[0:1], v[6:7], v[0:1], 1.0
-; GFX6-NEXT:    v_div_fmas_f64 v[4:5], v[10:11], v[4:5], v[12:13]
+; GFX6-NEXT:    s_nop 0
+; GFX6-NEXT:    v_div_fmas_f64 v[4:5], v[8:9], v[4:5], v[14:15]
 ; GFX6-NEXT:    v_div_fixup_f64 v[2:3], v[4:5], v[2:3], 1.0
 ; GFX6-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX8-LABEL: v_rcp_v2f64:
 ; GFX8:       ; %bb.0:
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX8-NEXT:    v_div_scale_f64 v[4:5], s[4:5], v[0:1], v[0:1], 1.0
 ; GFX8-NEXT:    v_div_scale_f64 v[6:7], s[4:5], v[2:3], v[2:3], 1.0
; GFX10-NEXT:    s_setpc_b64 s[30:31]
   ret <2 x double> %fdiv
 }
 
 define <2 x double> @v_rcp_v2f64_arcp(<2 x double> %x) {
 ; GFX6-LABEL: v_rcp_v2f64_arcp:
 ; GFX6:       ; %bb.0:
 ; GFX6-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX6-NEXT:    v_div_scale_f64 v[4:5], s[4:5], v[0:1], v[0:1], 1.0
-; GFX6-NEXT:    v_div_scale_f64 v[10:11], s[4:5], 1.0, v[0:1], 1.0
+; GFX6-NEXT:    v_div_scale_f64 v[10:11], s[4:5], v[2:3], v[2:3], 1.0
 ; GFX6-NEXT:    v_rcp_f64_e32 v[6:7], v[4:5]
 ; GFX6-NEXT:    v_mov_b32_e32 v18, 0x3ff00000
-; GFX6-NEXT:    v_cmp_eq_u32_e32 vcc, v18, v11
+; GFX6-NEXT:    v_rcp_f64_e32 v[14:15], v[10:11]
 ; GFX6-NEXT:    v_fma_f64 v[8:9], -v[4:5], v[6:7], 1.0
+; GFX6-NEXT:    v_fma_f64 v[16:17], -v[10:11], v[14:15], 1.0
 ; GFX6-NEXT:    v_fma_f64 v[6:7], v[6:7], v[8:9], v[6:7]
-; GFX6-NEXT:    v_div_scale_f64 v[8:9], s[4:5], v[2:3], v[2:3], 1.0
+; GFX6-NEXT:    v_div_scale_f64 v[8:9], s[4:5], 1.0, v[0:1], 1.0
 ; GFX6-NEXT:    v_fma_f64 v[12:13], -v[4:5], v[6:7], 1.0
-; GFX6-NEXT:    v_cmp_eq_u32_e64 s[4:5], v1, v5
+; GFX6-NEXT:    v_fma_f64 v[14:15], v[14:15], v[16:17], v[14:15]
 ; GFX6-NEXT:    v_fma_f64 v[6:7], v[6:7], v[12:13], v[6:7]
-; GFX6-NEXT:    v_rcp_f64_e32 v[12:13], v[8:9]
-; GFX6-NEXT:    v_mul_f64 v[14:15], v[10:11], v[6:7]
-; GFX6-NEXT:    s_xor_b64 vcc, vcc, s[4:5]
-; GFX6-NEXT:    v_fma_f64 v[10:11], -v[4:5], v[14:15], v[10:11]
-; GFX6-NEXT:    v_fma_f64 v[16:17], -v[8:9], v[12:13], 1.0
-; GFX6-NEXT:    v_div_fmas_f64 v[6:7], v[10:11], v[6:7], v[14:15]
-; GFX6-NEXT:    v_fma_f64 v[12:13], v[12:13], v[16:17], v[12:13]
+; GFX6-NEXT:    v_cmp_eq_u32_e32 vcc, v18, v9
+; GFX6-NEXT:    v_mul_f64 v[12:13], v[8:9], v[6:7]
+; GFX6-NEXT:    v_cmp_eq_u32_e64 s[4:5], v1, v5
+; GFX6-NEXT:    v_fma_f64 v[8:9], -v[4:5], v[12:13], v[8:9]
+; GFX6-NEXT:    v_fma_f64 v[4:5], -v[10:11], v[14:15], 1.0
 ; GFX6-NEXT:    v_div_scale_f64 v[16:17], s[6:7], 1.0, v[2:3], 1.0
-; GFX6-NEXT:    v_fma_f64 v[4:5], -v[8:9], v[12:13], 1.0
-; GFX6-NEXT:    v_cmp_eq_u32_e64 s[4:5], v3, v9
-; GFX6-NEXT:    v_fma_f64 v[4:5], v[12:13], v[4:5], v[12:13]
+; GFX6-NEXT:    v_fma_f64 v[4:5], v[14:15], v[4:5], v[14:15]
+; GFX6-NEXT:    s_xor_b64 vcc, vcc, s[4:5]
+; GFX6-NEXT:    v_mul_f64 v[14:15], v[16:17], v[4:5]
+; GFX6-NEXT:    v_div_fmas_f64 v[6:7], v[8:9], v[6:7], v[12:13]
+; GFX6-NEXT:    v_fma_f64 v[8:9], -v[10:11], v[14:15], v[16:17]
 ; GFX6-NEXT:    v_cmp_eq_u32_e32 vcc, v18, v17
-; GFX6-NEXT:    v_mul_f64 v[12:13], v[16:17], v[4:5]
+; GFX6-NEXT:    v_cmp_eq_u32_e64 s[4:5], v3, v11
 ; GFX6-NEXT:    s_xor_b64 vcc, vcc, s[4:5]
-; GFX6-NEXT:    v_fma_f64 v[10:11], -v[8:9], v[12:13], v[16:17]
 ; GFX6-NEXT:    v_div_fixup_f64 v[0:1], v[6:7], v[0:1], 1.0
-; GFX6-NEXT:    v_div_fmas_f64 v[4:5], v[10:11], v[4:5], v[12:13]
+; GFX6-NEXT:    s_nop 0
+; GFX6-NEXT:    v_div_fmas_f64 v[4:5], v[8:9], v[4:5], v[14:15]
 ; GFX6-NEXT:    v_div_fixup_f64 v[2:3], v[4:5], v[2:3], 1.0
 ; GFX6-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX8-LABEL: v_rcp_v2f64_arcp:
 ; GFX8:       ; %bb.0:
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX8-NEXT:    v_div_scale_f64 v[4:5], s[4:5], v[0:1], v[0:1], 1.0
 ; GFX8-NEXT:    v_div_scale_f64 v[6:7], s[4:5], v[2:3], v[2:3], 1.0
; GFX10-NEXT:    s_setpc_b64 s[30:31]
   ret <2 x double> %fdiv
 }
 
 define <2 x double> @v_rcp_v2f64_ulp25(<2 x double> %x) {
 ; GFX6-LABEL: v_rcp_v2f64_ulp25:
 ; GFX6:       ; %bb.0:
 ; GFX6-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX6-NEXT:    v_div_scale_f64 v[4:5], s[4:5], v[0:1], v[0:1], 1.0
-; GFX6-NEXT:    v_div_scale_f64 v[10:11], s[4:5], 1.0, v[0:1], 1.0
+; GFX6-NEXT:    v_div_scale_f64 v[10:11], s[4:5], v[2:3], v[2:3], 1.0
 ; GFX6-NEXT:    v_rcp_f64_e32 v[6:7], v[4:5]
 ; GFX6-NEXT:    v_mov_b32_e32 v18, 0x3ff00000
-; GFX6-NEXT:    v_cmp_eq_u32_e32 vcc, v18, v11
+; GFX6-NEXT:    v_rcp_f64_e32 v[14:15], v[10:11]
 ; GFX6-NEXT:    v_fma_f64 v[8:9], -v[4:5], v[6:7], 1.0
+; GFX6-NEXT:    v_fma_f64 v[16:17], -v[10:11], v[14:15], 1.0
 ; GFX6-NEXT:    v_fma_f64 v[6:7], v[6:7], v[8:9], v[6:7]
-; GFX6-NEXT:    v_div_scale_f64 v[8:9], s[4:5], v[2:3], v[2:3], 1.0
+; GFX6-NEXT:    v_div_scale_f64 v[8:9], s[4:5], 1.0, v[0:1], 1.0
 ; GFX6-NEXT:    v_fma_f64 v[12:13], -v[4:5], v[6:7], 1.0
-; GFX6-NEXT:    v_cmp_eq_u32_e64 s[4:5], v1, v5
+; GFX6-NEXT:    v_fma_f64 v[14:15], v[14:15], v[16:17], v[14:15]
 ; GFX6-NEXT:    v_fma_f64 v[6:7], v[6:7], v[12:13], v[6:7]
-; GFX6-NEXT:    v_rcp_f64_e32 v[12:13], v[8:9]
-; GFX6-NEXT:    v_mul_f64 v[14:15], v[10:11], v[6:7]
-; GFX6-NEXT:    s_xor_b64 vcc, vcc, s[4:5]
-; GFX6-NEXT:    v_fma_f64 v[10:11], -v[4:5], v[14:15], v[10:11]
-; GFX6-NEXT:    v_fma_f64 v[16:17], -v[8:9], v[12:13], 1.0
-; GFX6-NEXT:    v_div_fmas_f64 v[6:7], v[10:11], v[6:7], v[14:15]
-; GFX6-NEXT:    v_fma_f64 v[12:13], v[12:13], v[16:17], v[12:13]
+; GFX6-NEXT:    v_cmp_eq_u32_e32 vcc, v18, v9
+; GFX6-NEXT:    v_mul_f64 v[12:13], v[8:9], v[6:7]
+; GFX6-NEXT:    v_cmp_eq_u32_e64 s[4:5], v1, v5
+; GFX6-NEXT:    v_fma_f64 v[8:9], -v[4:5], v[12:13], v[8:9]
+; GFX6-NEXT:    v_fma_f64 v[4:5], -v[10:11], v[14:15], 1.0
 ; GFX6-NEXT:    v_div_scale_f64 v[16:17], s[6:7], 1.0, v[2:3], 1.0
-; GFX6-NEXT:    v_fma_f64 v[4:5], -v[8:9], v[12:13], 1.0
-; GFX6-NEXT:    v_cmp_eq_u32_e64 s[4:5], v3, v9
-; GFX6-NEXT:    v_fma_f64 v[4:5], v[12:13], v[4:5], v[12:13]
+; GFX6-NEXT:    v_fma_f64 v[4:5], v[14:15], v[4:5], v[14:15]
+; GFX6-NEXT:    s_xor_b64 vcc, vcc, s[4:5]
+; GFX6-NEXT:    v_mul_f64 v[14:15], v[16:17], v[4:5]
+; GFX6-NEXT:    v_div_fmas_f64 v[6:7], v[8:9], v[6:7], v[12:13]
+; GFX6-NEXT:    v_fma_f64 v[8:9], -v[10:11], v[14:15], v[16:17]
 ; GFX6-NEXT:    v_cmp_eq_u32_e32 vcc, v18, v17
-; GFX6-NEXT:    v_mul_f64 v[12:13], v[16:17], v[4:5]
+; GFX6-NEXT:    v_cmp_eq_u32_e64 s[4:5], v3, v11
 ; GFX6-NEXT:    s_xor_b64 vcc, vcc, s[4:5]
-; GFX6-NEXT:    v_fma_f64 v[10:11], -v[8:9], v[12:13], v[16:17]
 ; GFX6-NEXT:    v_div_fixup_f64 v[0:1], v[6:7], v[0:1], 1.0
-; GFX6-NEXT:    v_div_fmas_f64 v[4:5], v[10:11], v[4:5], v[12:13]
+; GFX6-NEXT:    s_nop 0
+; GFX6-NEXT:    v_div_fmas_f64 v[4:5], v[8:9], v[4:5], v[14:15]
 ; GFX6-NEXT:    v_div_fixup_f64 v[2:3], v[4:5], v[2:3], 1.0
 ; GFX6-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX8-LABEL: v_rcp_v2f64_ulp25:
 ; GFX8:       ; %bb.0:
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX8-NEXT:    v_div_scale_f64 v[4:5], s[4:5], v[0:1], v[0:1], 1.0
 ; GFX8-NEXT:    v_div_scale_f64 v[6:7], s[4:5], v[2:3], v[2:3], 1.0

llvm/test/CodeGen/AMDGPU/GlobalISel/fpow.ll

 ; GFX9-NEXT:    v_log_f16_e32 v2, v0
 ; GFX9-NEXT:    v_log_f16_sdwa v0, v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX9-NEXT:    v_cvt_f32_f16_e32 v3, v1
 ; GFX9-NEXT:    v_cvt_f32_f16_sdwa v1, v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX9-NEXT:    v_cvt_f32_f16_e32 v2, v2
 ; GFX9-NEXT:    v_cvt_f32_f16_e32 v0, v0
 ; GFX9-NEXT:    v_mul_legacy_f32_e32 v2, v2, v3
 ; GFX9-NEXT:    v_mul_legacy_f32_e32 v0, v0, v1
-; GFX9-NEXT:    v_cvt_f16_f32_e32 v2, v2
+; GFX9-NEXT:    v_cvt_f16_f32_e32 v1, v2
 ; GFX9-NEXT:    v_cvt_f16_f32_e32 v0, v0
-; GFX9-NEXT:    v_exp_f16_e32 v1, v2
-; GFX9-NEXT:    v_exp_f16_sdwa v0, v0 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD
 ; GFX9-NEXT:    v_mov_b32_e32 v2, 0xffff
+; GFX9-NEXT:    v_exp_f16_e32 v1, v1
+; GFX9-NEXT:    v_exp_f16_sdwa v0, v0 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD
 ; GFX9-NEXT:    v_and_or_b32 v0, v1, v2, v0
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: v_pow_v2f16:
 ; GFX10:       ; %bb.0:
 ; GFX10-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    v_log_f16_e32 v2, v0

 ; GFX9-NEXT:    v_log_f16_e32 v2, v0
 ; GFX9-NEXT:    v_log_f16_sdwa v0, v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX9-NEXT:    v_cvt_f32_f16_e32 v3, v1
 ; GFX9-NEXT:    v_cvt_f32_f16_sdwa v1, v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX9-NEXT:    v_cvt_f32_f16_e32 v2, v2
 ; GFX9-NEXT:    v_cvt_f32_f16_e32 v0, v0
 ; GFX9-NEXT:    v_mul_legacy_f32_e32 v2, v2, v3
 ; GFX9-NEXT:    v_mul_legacy_f32_e32 v0, v0, v1
-; GFX9-NEXT:    v_cvt_f16_f32_e32 v2, v2
+; GFX9-NEXT:    v_cvt_f16_f32_e32 v1, v2
 ; GFX9-NEXT:    v_cvt_f16_f32_e32 v0, v0
-; GFX9-NEXT:    v_exp_f16_e32 v1, v2
-; GFX9-NEXT:    v_exp_f16_sdwa v0, v0 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD
 ; GFX9-NEXT:    v_mov_b32_e32 v2, 0xffff
+; GFX9-NEXT:    v_exp_f16_e32 v1, v1
+; GFX9-NEXT:    v_exp_f16_sdwa v0, v0 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD
 ; GFX9-NEXT:    v_and_or_b32 v0, v1, v2, v0
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: v_pow_v2f16_fneg_lhs:
 ; GFX10:       ; %bb.0:
 ; GFX10-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    v_xor_b32_e32 v0, 0x80008000, v0

 ; GFX9-LABEL: v_pow_v2f16_fneg_rhs:
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    v_log_f16_e32 v2, v0
 ; GFX9-NEXT:    v_log_f16_sdwa v0, v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX9-NEXT:    v_xor_b32_e32 v1, 0x80008000, v1
 ; GFX9-NEXT:    v_cvt_f32_f16_e32 v3, v1
 ; GFX9-NEXT:    v_cvt_f32_f16_e32 v2, v2
-; GFX9-NEXT:    v_cvt_f32_f16_e32 v0, v0
 ; GFX9-NEXT:    v_cvt_f32_f16_sdwa v1, v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
+; GFX9-NEXT:    v_cvt_f32_f16_e32 v0, v0
 ; GFX9-NEXT:    v_mul_legacy_f32_e32 v2, v2, v3
 ; GFX9-NEXT:    v_cvt_f16_f32_e32 v2, v2
 ; GFX9-NEXT:    v_mul_legacy_f32_e32 v0, v0, v1
 ; GFX9-NEXT:    v_cvt_f16_f32_e32 v0, v0
 ; GFX9-NEXT:    v_exp_f16_e32 v1, v2
 ; GFX9-NEXT:    v_mov_b32_e32 v2, 0xffff
 ; GFX9-NEXT:    v_exp_f16_sdwa v0, v0 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD
 ; GFX9-NEXT:    v_and_or_b32 v0, v1, v2, v0

llvm/test/CodeGen/AMDGPU/GlobalISel/fshl.ll

 ; GFX9-NEXT:    v_rcp_iflag_f32_e32 v6, v6
 ; GFX9-NEXT:    v_mov_b32_e32 v7, 0xffffffe8
 ; GFX9-NEXT:    v_cvt_f32_ubyte0_e32 v9, 24
 ; GFX9-NEXT:    v_rcp_iflag_f32_e32 v9, v9
 ; GFX9-NEXT:    v_mul_f32_e32 v6, 0x4f7ffffe, v6
 ; GFX9-NEXT:    v_cvt_u32_f32_e32 v6, v6
 ; GFX9-NEXT:    v_and_b32_e32 v4, 0xffffff, v4
 ; GFX9-NEXT:    v_mul_f32_e32 v9, 0x4f7ffffe, v9
-; GFX9-NEXT:    v_bfe_u32 v2, v2, 1, 23
+; GFX9-NEXT:    v_cvt_u32_f32_e32 v9, v9
 ; GFX9-NEXT:    v_mul_lo_u32 v8, v7, v6
+; GFX9-NEXT:    v_bfe_u32 v2, v2, 1, 23
 ; GFX9-NEXT:    v_bfe_u32 v3, v3, 1, 23
+; GFX9-NEXT:    v_mul_lo_u32 v7, v7, v9
 ; GFX9-NEXT:    v_mul_hi_u32 v8, v6, v8
+; GFX9-NEXT:    v_mul_hi_u32 v7, v9, v7
 ; GFX9-NEXT:    v_add_u32_e32 v6, v6, v8
-; GFX9-NEXT:    v_cvt_u32_f32_e32 v8, v9
 ; GFX9-NEXT:    v_mul_hi_u32 v6, v4, v6
-; GFX9-NEXT:    v_mov_b32_e32 v9, 0xffffff
-; GFX9-NEXT:    v_and_b32_e32 v5, v5, v9
-; GFX9-NEXT:    v_mul_lo_u32 v7, v7, v8
+; GFX9-NEXT:    v_mov_b32_e32 v8, 0xffffff
+; GFX9-NEXT:    v_and_b32_e32 v5, v5, v8
 ; GFX9-NEXT:    v_mul_lo_u32 v6, v6, 24
-; GFX9-NEXT:    v_mul_hi_u32 v7, v8, v7
 ; GFX9-NEXT:    v_sub_u32_e32 v4, v4, v6
 ; GFX9-NEXT:    v_subrev_u32_e32 v6, 24, v4
 ; GFX9-NEXT:    v_cmp_le_u32_e32 vcc, 24, v4
 ; GFX9-NEXT:    v_cndmask_b32_e32 v4, v4, v6, vcc
 ; GFX9-NEXT:    v_subrev_u32_e32 v6, 24, v4
 ; GFX9-NEXT:    v_cmp_le_u32_e32 vcc, 24, v4
 ; GFX9-NEXT:    v_cndmask_b32_e32 v4, v4, v6, vcc
-; GFX9-NEXT:    v_add_u32_e32 v6, v8, v7
+; GFX9-NEXT:    v_add_u32_e32 v6, v9, v7
 ; GFX9-NEXT:    v_mul_hi_u32 v6, v5, v6
 ; GFX9-NEXT:    v_sub_u32_e32 v7, 23, v4
-; GFX9-NEXT:    v_and_b32_e32 v7, v7, v9
-; GFX9-NEXT:    v_and_b32_e32 v4, v4, v9
+; GFX9-NEXT:    v_and_b32_e32 v7, v7, v8
+; GFX9-NEXT:    v_and_b32_e32 v4, v4, v8
 ; GFX9-NEXT:    v_mul_lo_u32 v6, v6, 24
 ; GFX9-NEXT:    v_lshrrev_b32_e32 v2, v7, v2
 ; GFX9-NEXT:    v_lshl_or_b32 v0, v0, v4, v2
 ; GFX9-NEXT:    v_sub_u32_e32 v2, v5, v6
 ; GFX9-NEXT:    v_subrev_u32_e32 v4, 24, v2
 ; GFX9-NEXT:    v_cmp_le_u32_e32 vcc, 24, v2
 ; GFX9-NEXT:    v_cndmask_b32_e32 v2, v2, v4, vcc
 ; GFX9-NEXT:    v_subrev_u32_e32 v4, 24, v2
 ; GFX9-NEXT:    v_cmp_le_u32_e32 vcc, 24, v2
 ; GFX9-NEXT:    v_cndmask_b32_e32 v2, v2, v4, vcc
 ; GFX9-NEXT:    v_sub_u32_e32 v4, 23, v2
-; GFX9-NEXT:    v_and_b32_e32 v4, v4, v9
-; GFX9-NEXT:    v_and_b32_e32 v2, v2, v9
+; GFX9-NEXT:    v_and_b32_e32 v4, v4, v8
+; GFX9-NEXT:    v_and_b32_e32 v2, v2, v8
 ; GFX9-NEXT:    v_lshrrev_b32_e32 v3, v4, v3
 ; GFX9-NEXT:    v_lshl_or_b32 v1, v1, v2, v3
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: v_fshl_v2i24:
 ; GFX10:       ; %bb.0:
 ; GFX10-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0

llvm/test/CodeGen/AMDGPU/GlobalISel/fshr.ll

 ; GFX8-NEXT:    v_or_b32_e32 v1, v1, v2
 ; GFX8-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX9-LABEL: v_fshr_v2i24:
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    v_cvt_f32_ubyte0_e32 v6, 24
 ; GFX9-NEXT:    v_rcp_iflag_f32_e32 v6, v6
-; GFX9-NEXT:    v_mov_b32_e32 v7, 0xffffffe8
 ; GFX9-NEXT:    v_cvt_f32_ubyte0_e32 v9, 24
 ; GFX9-NEXT:    v_rcp_iflag_f32_e32 v9, v9
+; GFX9-NEXT:    v_mov_b32_e32 v7, 0xffffffe8
 ; GFX9-NEXT:    v_mul_f32_e32 v6, 0x4f7ffffe, v6
 ; GFX9-NEXT:    v_cvt_u32_f32_e32 v6, v6
-; GFX9-NEXT:    v_and_b32_e32 v4, 0xffffff, v4
 ; GFX9-NEXT:    v_mul_f32_e32 v9, 0x4f7ffffe, v9
-; GFX9-NEXT:    v_lshlrev_b32_e32 v0, 1, v0
+; GFX9-NEXT:    v_cvt_u32_f32_e32 v9, v9
+; GFX9-NEXT:    v_and_b32_e32 v4, 0xffffff, v4
 ; GFX9-NEXT:    v_mul_lo_u32 v8, v7, v6
+; GFX9-NEXT:    v_lshlrev_b32_e32 v0, 1, v0
+; GFX9-NEXT:    v_mul_lo_u32 v7, v7, v9
 ; GFX9-NEXT:    v_lshlrev_b32_e32 v1, 1, v1
 ; GFX9-NEXT:    v_mul_hi_u32 v8, v6, v8
+; GFX9-NEXT:    v_mul_hi_u32 v7, v9, v7
 ; GFX9-NEXT:    v_add_u32_e32 v6, v6, v8
-; GFX9-NEXT:    v_cvt_u32_f32_e32 v8, v9
 ; GFX9-NEXT:    v_mul_hi_u32 v6, v4, v6
-; GFX9-NEXT:    v_mov_b32_e32 v9, 0xffffff
-; GFX9-NEXT:    v_and_b32_e32 v5, v5, v9
-; GFX9-NEXT:    v_mul_lo_u32 v7, v7, v8
+; GFX9-NEXT:    v_mov_b32_e32 v8, 0xffffff
+; GFX9-NEXT:    v_and_b32_e32 v5, v5, v8
+; GFX9-NEXT:    v_add_u32_e32 v7, v9, v7
 ; GFX9-NEXT:    v_mul_lo_u32 v6, v6, 24
-; GFX9-NEXT:    v_and_b32_e32 v2, v2, v9
-; GFX9-NEXT:    v_and_b32_e32 v3, v3, v9
-; GFX9-NEXT:    v_mul_hi_u32 v7, v8, v7
+; GFX9-NEXT:    v_mul_hi_u32 v7, v5, v7
+; GFX9-NEXT:    v_and_b32_e32 v2, v2, v8
+; GFX9-NEXT:    v_and_b32_e32 v3, v3, v8
 ; GFX9-NEXT:    v_sub_u32_e32 v4, v4, v6
 ; GFX9-NEXT:    v_subrev_u32_e32 v6, 24, v4
 ; GFX9-NEXT:    v_cmp_le_u32_e32 vcc, 24, v4
-; GFX9-NEXT:    v_add_u32_e32 v7, v8, v7
-; GFX9-NEXT:    v_mul_hi_u32 v7, v5, v7
 ; GFX9-NEXT:    v_cndmask_b32_e32 v4, v4, v6, vcc
 ; GFX9-NEXT:    v_subrev_u32_e32 v6, 24, v4
 ; GFX9-NEXT:    v_cmp_le_u32_e32 vcc, 24, v4
 ; GFX9-NEXT:    v_mul_lo_u32 v7, v7, 24
 ; GFX9-NEXT:    v_cndmask_b32_e32 v4, v4, v6, vcc
 ; GFX9-NEXT:    v_sub_u32_e32 v6, 23, v4
-; GFX9-NEXT:    v_and_b32_e32 v4, v4, v9
-; GFX9-NEXT:    v_and_b32_e32 v6, v6, v9
+; GFX9-NEXT:    v_and_b32_e32 v4, v4, v8
+; GFX9-NEXT:    v_and_b32_e32 v6, v6, v8
 ; GFX9-NEXT:    v_lshrrev_b32_e32 v2, v4, v2
 ; GFX9-NEXT:    v_lshl_or_b32 v0, v0, v6, v2
 ; GFX9-NEXT:    v_sub_u32_e32 v2, v5, v7
 ; GFX9-NEXT:    v_subrev_u32_e32 v4, 24, v2
 ; GFX9-NEXT:    v_cmp_le_u32_e32 vcc, 24, v2
 ; GFX9-NEXT:    v_cndmask_b32_e32 v2, v2, v4, vcc
 ; GFX9-NEXT:    v_subrev_u32_e32 v4, 24, v2
 ; GFX9-NEXT:    v_cmp_le_u32_e32 vcc, 24, v2
 ; GFX9-NEXT:    v_cndmask_b32_e32 v2, v2, v4, vcc
 ; GFX9-NEXT:    v_sub_u32_e32 v4, 23, v2
-; GFX9-NEXT:    v_and_b32_e32 v2, v2, v9
-; GFX9-NEXT:    v_and_b32_e32 v4, v4, v9
+; GFX9-NEXT:    v_and_b32_e32 v2, v2, v8
+; GFX9-NEXT:    v_and_b32_e32 v4, v4, v8
 ; GFX9-NEXT:    v_lshrrev_b32_e32 v2, v2, v3
 ; GFX9-NEXT:    v_lshl_or_b32 v1, v1, v4, v2
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: v_fshr_v2i24:
 ; GFX10:       ; %bb.0:
 ; GFX10-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0

 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    v_xor_b32_e32 v6, -1, v4
 ; GFX10-NEXT:    v_xor_b32_e32 v7, -1, v5
 ; GFX10-NEXT:    s_mov_b32 s4, 0xf000f
 ; GFX10-NEXT:    v_pk_lshlrev_b16 v0, 1, v0 op_sel_hi:[0,1]
 ; GFX10-NEXT:    v_and_b32_e32 v4, s4, v4
 ; GFX10-NEXT:    v_and_b32_e32 v6, s4, v6
 ; GFX10-NEXT:    v_and_b32_e32 v5, s4, v5
-; GFX10-NEXT:    v_and_b32_e32 v7, s4, v7
 ; GFX10-NEXT:    v_pk_lshlrev_b16 v1, 1, v1 op_sel_hi:[0,1]
+; GFX10-NEXT:    v_and_b32_e32 v7, s4, v7
 ; GFX10-NEXT:    v_pk_lshrrev_b16 v2, v4, v2
 ; GFX10-NEXT:    v_pk_lshlrev_b16 v0, v6, v0
 ; GFX10-NEXT:    v_pk_lshrrev_b16 v3, v5, v3
 ; GFX10-NEXT:    v_pk_lshlrev_b16 v1, v7, v1
 ; GFX10-NEXT:    v_or_b32_e32 v0, v0, v2
 ; GFX10-NEXT:    v_or_b32_e32 v1, v1, v3
 ; GFX10-NEXT:    s_setpc_b64 s[30:31]
   %result = call <4 x i16> @llvm.fshr.v4i16(<4 x i16> %lhs, <4 x i16> %rhs, <4 x i16> %amt)

llvm/test/CodeGen/AMDGPU/GlobalISel/function-returns.ll

 ; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
 ; RUN: llc -global-isel -global-isel-abort=0 -stop-after=irtranslator -mtriple=amdgcn-mesa-mesa3d -mcpu=fiji -verify-machineinstrs -o - %s | FileCheck -enable-var-scope %s
 
 ; FIXME: Also test with a pre-gfx8 target.
 
 define i1 @i1_func_void() #0 {
   ; CHECK-LABEL: name: i1_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s1) = G_LOAD [[DEF]](p1) :: (load (s1) from `i1 addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[LOAD]](s1)
-  ; CHECK:   $vgpr0 = COPY [[ANYEXT]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s1) = G_LOAD [[DEF]](p1) :: (load (s1) from `i1 addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[LOAD]](s1)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[ANYEXT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = load i1, i1 addrspace(1)* undef
   ret i1 %val
 }
 
 define zeroext i1 @i1_zeroext_func_void() #0 {
   ; CHECK-LABEL: name: i1_zeroext_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s1) = G_LOAD [[DEF]](p1) :: (load (s1) from `i1 addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[LOAD]](s1)
-  ; CHECK:   $vgpr0 = COPY [[ZEXT]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s1) = G_LOAD [[DEF]](p1) :: (load (s1) from `i1 addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[LOAD]](s1)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[ZEXT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = load i1, i1 addrspace(1)* undef
   ret i1 %val
 }
 
 define signext i1 @i1_signext_func_void() #0 {
   ; CHECK-LABEL: name: i1_signext_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s1) = G_LOAD [[DEF]](p1) :: (load (s1) from `i1 addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   [[SEXT:%[0-9]+]]:_(s32) = G_SEXT [[LOAD]](s1)
-  ; CHECK:   $vgpr0 = COPY [[SEXT]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s1) = G_LOAD [[DEF]](p1) :: (load (s1) from `i1 addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   [[SEXT:%[0-9]+]]:_(s32) = G_SEXT [[LOAD]](s1)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[SEXT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = load i1, i1 addrspace(1)* undef
   ret i1 %val
 }
 
 define i7 @i7_func_void() #0 {
   ; CHECK-LABEL: name: i7_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s7) = G_LOAD [[DEF]](p1) :: (load (s7) from `i7 addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[LOAD]](s7)
-  ; CHECK:   $vgpr0 = COPY [[ANYEXT]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s7) = G_LOAD [[DEF]](p1) :: (load (s7) from `i7 addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[LOAD]](s7)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[ANYEXT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = load i7, i7 addrspace(1)* undef
   ret i7 %val
 }
 
 define zeroext i7 @i7_zeroext_func_void() #0 {
   ; CHECK-LABEL: name: i7_zeroext_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s7) = G_LOAD [[DEF]](p1) :: (load (s7) from `i7 addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[LOAD]](s7)
-  ; CHECK:   $vgpr0 = COPY [[ZEXT]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s7) = G_LOAD [[DEF]](p1) :: (load (s7) from `i7 addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[LOAD]](s7)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[ZEXT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = load i7, i7 addrspace(1)* undef
   ret i7 %val
 }
 
 define signext i7 @i7_signext_func_void() #0 {
   ; CHECK-LABEL: name: i7_signext_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s7) = G_LOAD [[DEF]](p1) :: (load (s7) from `i7 addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   [[SEXT:%[0-9]+]]:_(s32) = G_SEXT [[LOAD]](s7)
-  ; CHECK:   $vgpr0 = COPY [[SEXT]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s7) = G_LOAD [[DEF]](p1) :: (load (s7) from `i7 addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   [[SEXT:%[0-9]+]]:_(s32) = G_SEXT [[LOAD]](s7)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[SEXT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = load i7, i7 addrspace(1)* undef
   ret i7 %val
 }
 
 define i8 @i8_func_void() #0 {
   ; CHECK-LABEL: name: i8_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s8) = G_LOAD [[DEF]](p1) :: (load (s8) from `i8 addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[LOAD]](s8)
-  ; CHECK:   $vgpr0 = COPY [[ANYEXT]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s8) = G_LOAD [[DEF]](p1) :: (load (s8) from `i8 addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[LOAD]](s8)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[ANYEXT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = load i8, i8 addrspace(1)* undef
   ret i8 %val
 }
 
 define zeroext i8 @i8_zeroext_func_void() #0 {
   ; CHECK-LABEL: name: i8_zeroext_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s8) = G_LOAD [[DEF]](p1) :: (load (s8) from `i8 addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[LOAD]](s8)
-  ; CHECK:   $vgpr0 = COPY [[ZEXT]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s8) = G_LOAD [[DEF]](p1) :: (load (s8) from `i8 addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[LOAD]](s8)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[ZEXT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = load i8, i8 addrspace(1)* undef
   ret i8 %val
 }
 
 define signext i8 @i8_signext_func_void() #0 {
   ; CHECK-LABEL: name: i8_signext_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s8) = G_LOAD [[DEF]](p1) :: (load (s8) from `i8 addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   [[SEXT:%[0-9]+]]:_(s32) = G_SEXT [[LOAD]](s8)
-  ; CHECK:   $vgpr0 = COPY [[SEXT]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s8) = G_LOAD [[DEF]](p1) :: (load (s8) from `i8 addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   [[SEXT:%[0-9]+]]:_(s32) = G_SEXT [[LOAD]](s8)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[SEXT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = load i8, i8 addrspace(1)* undef
   ret i8 %val
 }
 
 define i16 @i16_func_void() #0 {
   ; CHECK-LABEL: name: i16_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s16) = G_LOAD [[DEF]](p1) :: (load (s16) from `i16 addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[LOAD]](s16)
-  ; CHECK:   $vgpr0 = COPY [[ANYEXT]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s16) = G_LOAD [[DEF]](p1) :: (load (s16) from `i16 addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[LOAD]](s16)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[ANYEXT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = load i16, i16 addrspace(1)* undef
   ret i16 %val
 }
 
 define zeroext i16 @i16_zeroext_func_void() #0 {
   ; CHECK-LABEL: name: i16_zeroext_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s16) = G_LOAD [[DEF]](p1) :: (load (s16) from `i16 addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[LOAD]](s16)
-  ; CHECK:   $vgpr0 = COPY [[ZEXT]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s16) = G_LOAD [[DEF]](p1) :: (load (s16) from `i16 addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[LOAD]](s16)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[ZEXT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = load i16, i16 addrspace(1)* undef
   ret i16 %val
 }
 
 define signext i16 @i16_signext_func_void() #0 {
   ; CHECK-LABEL: name: i16_signext_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s16) = G_LOAD [[DEF]](p1) :: (load (s16) from `i16 addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   [[SEXT:%[0-9]+]]:_(s32) = G_SEXT [[LOAD]](s16)
-  ; CHECK:   $vgpr0 = COPY [[SEXT]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s16) = G_LOAD [[DEF]](p1) :: (load (s16) from `i16 addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   [[SEXT:%[0-9]+]]:_(s32) = G_SEXT [[LOAD]](s16)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[SEXT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = load i16, i16 addrspace(1)* undef
   ret i16 %val
 }
 
 define half @f16_func_void() #0 {
   ; CHECK-LABEL: name: f16_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s16) = G_LOAD [[DEF]](p1) :: (load (s16) from `half addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[LOAD]](s16)
-  ; CHECK:   $vgpr0 = COPY [[ANYEXT]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s16) = G_LOAD [[DEF]](p1) :: (load (s16) from `half addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[LOAD]](s16)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[ANYEXT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = load half, half addrspace(1)* undef
   ret half %val
 }
 
 define i24 @i24_func_void() #0 {
   ; CHECK-LABEL: name: i24_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s24) = G_LOAD [[DEF]](p1) :: (load (s24) from `i24 addrspace(1)* undef`, align 4, addrspace 1)
-  ; CHECK:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[LOAD]](s24)
-  ; CHECK:   $vgpr0 = COPY [[ANYEXT]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s24) = G_LOAD [[DEF]](p1) :: (load (s24) from `i24 addrspace(1)* undef`, align 4, addrspace 1)
+  ; CHECK-NEXT:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[LOAD]](s24)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[ANYEXT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = load i24, i24 addrspace(1)* undef
   ret i24 %val
 }
 
 define zeroext i24 @i24_zeroext_func_void() #0 {
   ; CHECK-LABEL: name: i24_zeroext_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s24) = G_LOAD [[DEF]](p1) :: (load (s24) from `i24 addrspace(1)* undef`, align 4, addrspace 1)
-  ; CHECK:   [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[LOAD]](s24)
-  ; CHECK:   $vgpr0 = COPY [[ZEXT]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s24) = G_LOAD [[DEF]](p1) :: (load (s24) from `i24 addrspace(1)* undef`, align 4, addrspace 1)
+  ; CHECK-NEXT:   [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[LOAD]](s24)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[ZEXT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = load i24, i24 addrspace(1)* undef
   ret i24 %val
 }
 
 define signext i24 @i24_signext_func_void() #0 {
   ; CHECK-LABEL: name: i24_signext_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s24) = G_LOAD [[DEF]](p1) :: (load (s24) from `i24 addrspace(1)* undef`, align 4, addrspace 1)
-  ; CHECK:   [[SEXT:%[0-9]+]]:_(s32) = G_SEXT [[LOAD]](s24)
-  ; CHECK:   $vgpr0 = COPY [[SEXT]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s24) = G_LOAD [[DEF]](p1) :: (load (s24) from `i24 addrspace(1)* undef`, align 4, addrspace 1)
+  ; CHECK-NEXT:   [[SEXT:%[0-9]+]]:_(s32) = G_SEXT [[LOAD]](s24)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[SEXT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = load i24, i24 addrspace(1)* undef
   ret i24 %val
 }
 
 define <2 x i24> @v2i24_func_void() #0 {
   ; CHECK-LABEL: name: v2i24_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(<2 x s24>) = G_LOAD [[DEF]](p1) :: (load (<2 x s24>) from `<2 x i24> addrspace(1)* undef`, align 8, addrspace 1)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s24), [[UV1:%[0-9]+]]:_(s24) = G_UNMERGE_VALUES [[LOAD]](<2 x s24>)
-  ; CHECK:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[UV]](s24)
-  ; CHECK:   [[ANYEXT1:%[0-9]+]]:_(s32) = G_ANYEXT [[UV1]](s24)
-  ; CHECK:   $vgpr0 = COPY [[ANYEXT]](s32)
-  ; CHECK:   $vgpr1 = COPY [[ANYEXT1]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(<2 x s24>) = G_LOAD [[DEF]](p1) :: (load (<2 x s24>) from `<2 x i24> addrspace(1)* undef`, align 8, addrspace 1)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s24), [[UV1:%[0-9]+]]:_(s24) = G_UNMERGE_VALUES [[LOAD]](<2 x s24>)
+  ; CHECK-NEXT:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[UV]](s24)
+  ; CHECK-NEXT:   [[ANYEXT1:%[0-9]+]]:_(s32) = G_ANYEXT [[UV1]](s24)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[ANYEXT]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[ANYEXT1]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %val = load <2 x i24>, <2 x i24> addrspace(1)* undef
   ret <2 x i24> %val
 }
 
 define <3 x i24> @v3i24_func_void() #0 {
   ; CHECK-LABEL: name: v3i24_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(<3 x s24>) = G_LOAD [[DEF]](p1) :: (load (<3 x s24>) from `<3 x i24> addrspace(1)* undef`, align 16, addrspace 1)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s24), [[UV1:%[0-9]+]]:_(s24), [[UV2:%[0-9]+]]:_(s24) = G_UNMERGE_VALUES [[LOAD]](<3 x s24>)
-  ; CHECK:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[UV]](s24)
-  ; CHECK:   [[ANYEXT1:%[0-9]+]]:_(s32) = G_ANYEXT [[UV1]](s24)
-  ; CHECK:   [[ANYEXT2:%[0-9]+]]:_(s32) = G_ANYEXT [[UV2]](s24)
-  ; CHECK:   $vgpr0 = COPY [[ANYEXT]](s32)
-  ; CHECK:   $vgpr1 = COPY [[ANYEXT1]](s32)
-  ; CHECK:   $vgpr2 = COPY [[ANYEXT2]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(<3 x s24>) = G_LOAD [[DEF]](p1) :: (load (<3 x s24>) from `<3 x i24> addrspace(1)* undef`, align 16, addrspace 1)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s24), [[UV1:%[0-9]+]]:_(s24), [[UV2:%[0-9]+]]:_(s24) = G_UNMERGE_VALUES [[LOAD]](<3 x s24>)
+  ; CHECK-NEXT:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[UV]](s24)
+  ; CHECK-NEXT:   [[ANYEXT1:%[0-9]+]]:_(s32) = G_ANYEXT [[UV1]](s24)
+  ; CHECK-NEXT:   [[ANYEXT2:%[0-9]+]]:_(s32) = G_ANYEXT [[UV2]](s24)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[ANYEXT]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[ANYEXT1]](s32)
+  ; CHECK-NEXT:   $vgpr2 = COPY [[ANYEXT2]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
   %val = load <3 x i24>, <3 x i24> addrspace(1)* undef
   ret <3 x i24> %val
 }
 
 define i32 @i32_func_void() #0 {
   ; CHECK-LABEL: name: i32_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s32) = G_LOAD [[DEF]](p1) :: (load (s32) from `i32 addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   $vgpr0 = COPY [[LOAD]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s32) = G_LOAD [[DEF]](p1) :: (load (s32) from `i32 addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[LOAD]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = load i32, i32 addrspace(1)* undef
   ret i32 %val
 }
 
 define i48 @i48_func_void() #0 {
   ; CHECK-LABEL: name: i48_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s48) = G_LOAD [[DEF]](p1) :: (load (s48) from `i48 addrspace(1)* undef`, align 8, addrspace 1)
-  ; CHECK:   [[ANYEXT:%[0-9]+]]:_(s64) = G_ANYEXT [[LOAD]](s48)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[ANYEXT]](s64)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s48) = G_LOAD [[DEF]](p1) :: (load (s48) from `i48 addrspace(1)* undef`, align 8, addrspace 1)
+  ; CHECK-NEXT:   [[ANYEXT:%[0-9]+]]:_(s64) = G_ANYEXT [[LOAD]](s48)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[ANYEXT]](s64)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %val = load i48, i48 addrspace(1)* undef, align 8
   ret i48 %val
 }
 
 define signext i48 @i48_signext_func_void() #0 {
   ; CHECK-LABEL: name: i48_signext_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s48) = G_LOAD [[DEF]](p1) :: (load (s48) from `i48 addrspace(1)* undef`, align 8, addrspace 1)
-  ; CHECK:   [[SEXT:%[0-9]+]]:_(s64) = G_SEXT [[LOAD]](s48)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[SEXT]](s64)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s48) = G_LOAD [[DEF]](p1) :: (load (s48) from `i48 addrspace(1)* undef`, align 8, addrspace 1)
+  ; CHECK-NEXT:   [[SEXT:%[0-9]+]]:_(s64) = G_SEXT [[LOAD]](s48)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[SEXT]](s64)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %val = load i48, i48 addrspace(1)* undef, align 8
   ret i48 %val
 }
 
 define zeroext i48 @i48_zeroext_func_void() #0 {
   ; CHECK-LABEL: name: i48_zeroext_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s48) = G_LOAD [[DEF]](p1) :: (load (s48) from `i48 addrspace(1)* undef`, align 8, addrspace 1)
-  ; CHECK:   [[ZEXT:%[0-9]+]]:_(s64) = G_ZEXT [[LOAD]](s48)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[ZEXT]](s64)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s48) = G_LOAD [[DEF]](p1) :: (load (s48) from `i48 addrspace(1)* undef`, align 8, addrspace 1)
+  ; CHECK-NEXT:   [[ZEXT:%[0-9]+]]:_(s64) = G_ZEXT [[LOAD]](s48)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[ZEXT]](s64)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %val = load i48, i48 addrspace(1)* undef, align 8
   ret i48 %val
 }
 
 define i64 @i64_func_void() #0 {
   ; CHECK-LABEL: name: i64_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s64) = G_LOAD [[DEF]](p1) :: (load (s64) from `i64 addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD]](s64)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s64) = G_LOAD [[DEF]](p1) :: (load (s64) from `i64 addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD]](s64)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %val = load i64, i64 addrspace(1)* undef
   ret i64 %val
 }
 
 define i65 @i65_func_void() #0 {
   ; CHECK-LABEL: name: i65_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s65) = G_LOAD [[DEF]](p1) :: (load (s65) from `i65 addrspace(1)* undef`, align 8, addrspace 1)
-  ; CHECK:   [[ANYEXT:%[0-9]+]]:_(s96) = G_ANYEXT [[LOAD]](s65)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[ANYEXT]](s96)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   $vgpr2 = COPY [[UV2]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s65) = G_LOAD [[DEF]](p1) :: (load (s65) from `i65 addrspace(1)* undef`, align 8, addrspace 1)
+  ; CHECK-NEXT:   [[ANYEXT:%[0-9]+]]:_(s96) = G_ANYEXT [[LOAD]](s65)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[ANYEXT]](s96)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
   %val = load i65, i65 addrspace(1)* undef
   ret i65 %val
 }
 
 define signext i65 @i65_signext_func_void() #0 {
   ; CHECK-LABEL: name: i65_signext_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s65) = G_LOAD [[DEF]](p1) :: (load (s65) from `i65 addrspace(1)* undef`, align 8, addrspace 1)
-  ; CHECK:   [[SEXT:%[0-9]+]]:_(s96) = G_SEXT [[LOAD]](s65)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[SEXT]](s96)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   $vgpr2 = COPY [[UV2]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s65) = G_LOAD [[DEF]](p1) :: (load (s65) from `i65 addrspace(1)* undef`, align 8, addrspace 1)
+  ; CHECK-NEXT:   [[SEXT:%[0-9]+]]:_(s96) = G_SEXT [[LOAD]](s65)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[SEXT]](s96)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
   %val = load i65, i65 addrspace(1)* undef
   ret i65 %val
 }
 
 define zeroext i65 @i65_zeroext_func_void() #0 {
   ; CHECK-LABEL: name: i65_zeroext_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s65) = G_LOAD [[DEF]](p1) :: (load (s65) from `i65 addrspace(1)* undef`, align 8, addrspace 1)
-  ; CHECK:   [[ZEXT:%[0-9]+]]:_(s96) = G_ZEXT [[LOAD]](s65)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[ZEXT]](s96)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   $vgpr2 = COPY [[UV2]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s65) = G_LOAD [[DEF]](p1) :: (load (s65) from `i65 addrspace(1)* undef`, align 8, addrspace 1)
+  ; CHECK-NEXT:   [[ZEXT:%[0-9]+]]:_(s96) = G_ZEXT [[LOAD]](s65)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[ZEXT]](s96)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
   %val = load i65, i65 addrspace(1)* undef
   ret i65 %val
 }
 
 define float @f32_func_void() #0 {
   ; CHECK-LABEL: name: f32_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s32) = G_LOAD [[DEF]](p1) :: (load (s32) from `float addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   $vgpr0 = COPY [[LOAD]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s32) = G_LOAD [[DEF]](p1) :: (load (s32) from `float addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[LOAD]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = load float, float addrspace(1)* undef
   ret float %val
 }
 
 define double @f64_func_void() #0 {
   ; CHECK-LABEL: name: f64_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s64) = G_LOAD [[DEF]](p1) :: (load (s64) from `double addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD]](s64)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s64) = G_LOAD [[DEF]](p1) :: (load (s64) from `double addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD]](s64)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %val = load double, double addrspace(1)* undef
   ret double %val
 }
 
 define <2 x double> @v2f64_func_void() #0 {
   ; CHECK-LABEL: name: v2f64_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(<2 x s64>) = G_LOAD [[DEF]](p1) :: (load (<2 x s64>) from `<2 x double> addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD]](<2 x s64>)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   $vgpr2 = COPY [[UV2]](s32)
-  ; CHECK:   $vgpr3 = COPY [[UV3]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(<2 x s64>) = G_LOAD [[DEF]](p1) :: (load (<2 x s64>) from `<2 x double> addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD]](<2 x s64>)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
+  ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
   %val = load <2 x double>, <2 x double> addrspace(1)* undef
   ret <2 x double> %val
 }
 
 define <2 x i32> @v2i32_func_void() #0 {
   ; CHECK-LABEL: name: v2i32_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(<2 x s32>) = G_LOAD [[DEF]](p1) :: (load (<2 x s32>) from `<2 x i32> addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD]](<2 x s32>)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(<2 x s32>) = G_LOAD [[DEF]](p1) :: (load (<2 x s32>) from `<2 x i32> addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD]](<2 x s32>)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %val = load <2 x i32>, <2 x i32> addrspace(1)* undef
   ret <2 x i32> %val
 }
 
 define <3 x i32> @v3i32_func_void() #0 {
   ; CHECK-LABEL: name: v3i32_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(<3 x s32>) = G_LOAD [[DEF]](p1) :: (load (<3 x s32>) from `<3 x i32> addrspace(1)* undef`, align 16, addrspace 1)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD]](<3 x s32>)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   $vgpr2 = COPY [[UV2]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(<3 x s32>) = G_LOAD [[DEF]](p1) :: (load (<3 x s32>) from `<3 x i32> addrspace(1)* undef`, align 16, addrspace 1)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD]](<3 x s32>)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
   %val = load <3 x i32>, <3 x i32> addrspace(1)* undef
   ret <3 x i32> %val
 }
 
 define <4 x i32> @v4i32_func_void() #0 {
   ; CHECK-LABEL: name: v4i32_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(<4 x s32>) = G_LOAD [[DEF]](p1) :: (load (<4 x s32>) from `<4 x i32> addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD]](<4 x s32>)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   $vgpr2 = COPY [[UV2]](s32)
-  ; CHECK:   $vgpr3 = COPY [[UV3]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(<4 x s32>) = G_LOAD [[DEF]](p1) :: (load (<4 x s32>) from `<4 x i32> addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD]](<4 x s32>)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
+  ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
   %val = load <4 x i32>, <4 x i32> addrspace(1)* undef
   ret <4 x i32> %val
 }
 
 define <5 x i32> @v5i32_func_void() #0 {
   ; CHECK-LABEL: name: v5i32_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(<5 x s32>) = G_LOAD [[DEF]](p1) :: (volatile load (<5 x s32>) from `<5 x i32> addrspace(1)* undef`, align 32, addrspace 1)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD]](<5 x s32>)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   $vgpr2 = COPY [[UV2]](s32)
-  ; CHECK:   $vgpr3 = COPY [[UV3]](s32)
-  ; CHECK:   $vgpr4 = COPY [[UV4]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(<5 x s32>) = G_LOAD [[DEF]](p1) :: (volatile load (<5 x s32>) from `<5 x i32> addrspace(1)* undef`, align 32, addrspace 1)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD]](<5 x s32>)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
+  ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
+  ; CHECK-NEXT:   $vgpr4 = COPY [[UV4]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4
   %val = load volatile <5 x i32>, <5 x i32> addrspace(1)* undef
   ret <5 x i32> %val
 }
 
 define <8 x i32> @v8i32_func_void() #0 {
   ; CHECK-LABEL: name: v8i32_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `<8 x i32> addrspace(1)* addrspace(4)* undef`, addrspace 4)
-  ; CHECK:   [[LOAD1:%[0-9]+]]:_(<8 x s32>) = G_LOAD [[LOAD]](p1) :: (load (<8 x s32>) from %ir.ptr, addrspace 1)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD1]](<8 x s32>)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   $vgpr2 = COPY [[UV2]](s32)
-  ; CHECK:   $vgpr3 = COPY [[UV3]](s32)
-  ; CHECK:   $vgpr4 = COPY [[UV4]](s32)
-  ; CHECK:   $vgpr5 = COPY [[UV5]](s32)
-  ; CHECK:   $vgpr6 = COPY [[UV6]](s32)
-  ; CHECK:   $vgpr7 = COPY [[UV7]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `<8 x i32> addrspace(1)* addrspace(4)* undef`, addrspace 4)
+  ; CHECK-NEXT:   [[LOAD1:%[0-9]+]]:_(<8 x s32>) = G_LOAD [[LOAD]](p1) :: (load (<8 x s32>) from %ir.ptr, addrspace 1)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD1]](<8 x s32>)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
+  ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
+  ; CHECK-NEXT:   $vgpr4 = COPY [[UV4]](s32)
+  ; CHECK-NEXT:   $vgpr5 = COPY [[UV5]](s32)
+  ; CHECK-NEXT:   $vgpr6 = COPY [[UV6]](s32)
+  ; CHECK-NEXT:   $vgpr7 = COPY [[UV7]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
   %ptr = load volatile <8 x i32> addrspace(1)*, <8 x i32> addrspace(1)* addrspace(4)* undef
   %val = load <8 x i32>, <8 x i32> addrspace(1)* %ptr
   ret <8 x i32> %val
 }
 
 define <16 x i32> @v16i32_func_void() #0 {
   ; CHECK-LABEL: name: v16i32_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `<16 x i32> addrspace(1)* addrspace(4)* undef`, addrspace 4)
-  ; CHECK:   [[LOAD1:%[0-9]+]]:_(<16 x s32>) = G_LOAD [[LOAD]](p1) :: (load (<16 x s32>) from %ir.ptr, addrspace 1)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32), [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32), [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32), [[UV12:%[0-9]+]]:_(s32), [[UV13:%[0-9]+]]:_(s32), [[UV14:%[0-9]+]]:_(s32), [[UV15:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD1]](<16 x s32>)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   $vgpr2 = COPY [[UV2]](s32)
-  ; CHECK:   $vgpr3 = COPY [[UV3]](s32)
-  ; CHECK:   $vgpr4 = COPY [[UV4]](s32)
-  ; CHECK:   $vgpr5 = COPY [[UV5]](s32)
-  ; CHECK:   $vgpr6 = COPY [[UV6]](s32)
-  ; CHECK:   $vgpr7 = COPY [[UV7]](s32)
-  ; CHECK:   $vgpr8 = COPY [[UV8]](s32)
-  ; CHECK:   $vgpr9 = COPY [[UV9]](s32)
-  ; CHECK:   $vgpr10 = COPY [[UV10]](s32)
-  ; CHECK:   $vgpr11 = COPY [[UV11]](s32)
-  ; CHECK:   $vgpr12 = COPY [[UV12]](s32)
-  ; CHECK:   $vgpr13 = COPY [[UV13]](s32)
-  ; CHECK:   $vgpr14 = COPY [[UV14]](s32)
-  ; CHECK:   $vgpr15 = COPY [[UV15]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `<16 x i32> addrspace(1)* addrspace(4)* undef`, addrspace 4)
+  ; CHECK-NEXT:   [[LOAD1:%[0-9]+]]:_(<16 x s32>) = G_LOAD [[LOAD]](p1) :: (load (<16 x s32>) from %ir.ptr, addrspace 1)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32), [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32), [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32), [[UV12:%[0-9]+]]:_(s32), [[UV13:%[0-9]+]]:_(s32), [[UV14:%[0-9]+]]:_(s32), [[UV15:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD1]](<16 x s32>)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
+  ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
+  ; CHECK-NEXT:   $vgpr4 = COPY [[UV4]](s32)
+  ; CHECK-NEXT:   $vgpr5 = COPY [[UV5]](s32)
+  ; CHECK-NEXT:   $vgpr6 = COPY [[UV6]](s32)
+  ; CHECK-NEXT:   $vgpr7 = COPY [[UV7]](s32)
+  ; CHECK-NEXT:   $vgpr8 = COPY [[UV8]](s32)
+  ; CHECK-NEXT:   $vgpr9 = COPY [[UV9]](s32)
+  ; CHECK-NEXT:   $vgpr10 = COPY [[UV10]](s32)
+  ; CHECK-NEXT:   $vgpr11 = COPY [[UV11]](s32)
+  ; CHECK-NEXT:   $vgpr12 = COPY [[UV12]](s32)
+  ; CHECK-NEXT:   $vgpr13 = COPY [[UV13]](s32)
+  ; CHECK-NEXT:   $vgpr14 = COPY [[UV14]](s32)
+  ; CHECK-NEXT:   $vgpr15 = COPY [[UV15]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15
   %ptr = load volatile <16 x i32> addrspace(1)*, <16 x i32> addrspace(1)* addrspace(4)* undef
   %val = load <16 x i32>, <16 x i32> addrspace(1)* %ptr
   ret <16 x i32> %val
 }
 
 define <32 x i32> @v32i32_func_void() #0 {
   ; CHECK-LABEL: name: v32i32_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `<32 x i32> addrspace(1)* addrspace(4)* undef`, addrspace 4)
-  ; CHECK:   [[LOAD1:%[0-9]+]]:_(<32 x s32>) = G_LOAD [[LOAD]](p1) :: (load (<32 x s32>) from %ir.ptr, addrspace 1)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32), [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32), [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32), [[UV12:%[0-9]+]]:_(s32), [[UV13:%[0-9]+]]:_(s32), [[UV14:%[0-9]+]]:_(s32), [[UV15:%[0-9]+]]:_(s32), [[UV16:%[0-9]+]]:_(s32), [[UV17:%[0-9]+]]:_(s32), [[UV18:%[0-9]+]]:_(s32), [[UV19:%[0-9]+]]:_(s32), [[UV20:%[0-9]+]]:_(s32), [[UV21:%[0-9]+]]:_(s32), [[UV22:%[0-9]+]]:_(s32), [[UV23:%[0-9]+]]:_(s32), [[UV24:%[0-9]+]]:_(s32), [[UV25:%[0-9]+]]:_(s32), [[UV26:%[0-9]+]]:_(s32), [[UV27:%[0-9]+]]:_(s32), [[UV28:%[0-9]+]]:_(s32), [[UV29:%[0-9]+]]:_(s32), [[UV30:%[0-9]+]]:_(s32), [[UV31:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD1]](<32 x s32>)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   $vgpr2 = COPY [[UV2]](s32)
-  ; CHECK:   $vgpr3 = COPY [[UV3]](s32)
-  ; CHECK:   $vgpr4 = COPY [[UV4]](s32)
-  ; CHECK:   $vgpr5 = COPY [[UV5]](s32)
-  ; CHECK:   $vgpr6 = COPY [[UV6]](s32)
-  ; CHECK:   $vgpr7 = COPY [[UV7]](s32)
-  ; CHECK:   $vgpr8 = COPY [[UV8]](s32)
-  ; CHECK:   $vgpr9 = COPY [[UV9]](s32)
-  ; CHECK:   $vgpr10 = COPY [[UV10]](s32)
-  ; CHECK:   $vgpr11 = COPY [[UV11]](s32)
-  ; CHECK:   $vgpr12 = COPY [[UV12]](s32)
-  ; CHECK:   $vgpr13 = COPY [[UV13]](s32)
-  ; CHECK:   $vgpr14 = COPY [[UV14]](s32)
-  ; CHECK:   $vgpr15 = COPY [[UV15]](s32)
-  ; CHECK:   $vgpr16 = COPY [[UV16]](s32)
-  ; CHECK:   $vgpr17 = COPY [[UV17]](s32)
-  ; CHECK:   $vgpr18 = COPY [[UV18]](s32)
-  ; CHECK:   $vgpr19 = COPY [[UV19]](s32)
-  ; CHECK:   $vgpr20 = COPY [[UV20]](s32)
-  ; CHECK:   $vgpr21 = COPY [[UV21]](s32)
-  ; CHECK:   $vgpr22 = COPY [[UV22]](s32)
-  ; CHECK:   $vgpr23 = COPY [[UV23]](s32)
-  ; CHECK:   $vgpr24 = COPY [[UV24]](s32)
-  ; CHECK:   $vgpr25 = COPY [[UV25]](s32)
-  ; CHECK:   $vgpr26 = COPY [[UV26]](s32)
-  ; CHECK:   $vgpr27 = COPY [[UV27]](s32)
-  ; CHECK:   $vgpr28 = COPY [[UV28]](s32)
-  ; CHECK:   $vgpr29 = COPY [[UV29]](s32)
-  ; CHECK:   $vgpr30 = COPY [[UV30]](s32)
-  ; CHECK:   $vgpr31 = COPY [[UV31]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15, implicit $vgpr16, implicit $vgpr17, implicit $vgpr18, implicit $vgpr19, implicit $vgpr20, implicit $vgpr21, implicit $vgpr22, implicit $vgpr23, implicit $vgpr24, implicit $vgpr25, implicit $vgpr26, implicit $vgpr27, implicit $vgpr28, implicit $vgpr29, implicit $vgpr30, implicit $vgpr31
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `<32 x i32> addrspace(1)* addrspace(4)* undef`, addrspace 4)
+  ; CHECK-NEXT:   [[LOAD1:%[0-9]+]]:_(<32 x s32>) = G_LOAD [[LOAD]](p1) :: (load (<32 x s32>) from %ir.ptr, addrspace 1)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32), [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32), [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32), [[UV12:%[0-9]+]]:_(s32), [[UV13:%[0-9]+]]:_(s32), [[UV14:%[0-9]+]]:_(s32), [[UV15:%[0-9]+]]:_(s32), [[UV16:%[0-9]+]]:_(s32), [[UV17:%[0-9]+]]:_(s32), [[UV18:%[0-9]+]]:_(s32), [[UV19:%[0-9]+]]:_(s32), [[UV20:%[0-9]+]]:_(s32), [[UV21:%[0-9]+]]:_(s32), [[UV22:%[0-9]+]]:_(s32), [[UV23:%[0-9]+]]:_(s32), [[UV24:%[0-9]+]]:_(s32), [[UV25:%[0-9]+]]:_(s32), [[UV26:%[0-9]+]]:_(s32), [[UV27:%[0-9]+]]:_(s32), [[UV28:%[0-9]+]]:_(s32), [[UV29:%[0-9]+]]:_(s32), [[UV30:%[0-9]+]]:_(s32), [[UV31:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD1]](<32 x s32>)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
+  ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
+  ; CHECK-NEXT:   $vgpr4 = COPY [[UV4]](s32)
+  ; CHECK-NEXT:   $vgpr5 = COPY [[UV5]](s32)
+  ; CHECK-NEXT:   $vgpr6 = COPY [[UV6]](s32)
+  ; CHECK-NEXT:   $vgpr7 = COPY [[UV7]](s32)
+  ; CHECK-NEXT:   $vgpr8 = COPY [[UV8]](s32)
+  ; CHECK-NEXT:   $vgpr9 = COPY [[UV9]](s32)
+  ; CHECK-NEXT:   $vgpr10 = COPY [[UV10]](s32)
+  ; CHECK-NEXT:   $vgpr11 = COPY [[UV11]](s32)
+  ; CHECK-NEXT:   $vgpr12 = COPY [[UV12]](s32)
+  ; CHECK-NEXT:   $vgpr13 = COPY [[UV13]](s32)
+  ; CHECK-NEXT:   $vgpr14 = COPY [[UV14]](s32)
+  ; CHECK-NEXT:   $vgpr15 = COPY [[UV15]](s32)
+  ; CHECK-NEXT:   $vgpr16 = COPY [[UV16]](s32)
+  ; CHECK-NEXT:   $vgpr17 = COPY [[UV17]](s32)
+  ; CHECK-NEXT:   $vgpr18 = COPY [[UV18]](s32)
+  ; CHECK-NEXT:   $vgpr19 = COPY [[UV19]](s32)
+  ; CHECK-NEXT:   $vgpr20 = COPY [[UV20]](s32)
+  ; CHECK-NEXT:   $vgpr21 = COPY [[UV21]](s32)
+  ; CHECK-NEXT:   $vgpr22 = COPY [[UV22]](s32)
+  ; CHECK-NEXT:   $vgpr23 = COPY [[UV23]](s32)
+  ; CHECK-NEXT:   $vgpr24 = COPY [[UV24]](s32)
+  ; CHECK-NEXT:   $vgpr25 = COPY [[UV25]](s32)
+  ; CHECK-NEXT:   $vgpr26 = COPY [[UV26]](s32)
+  ; CHECK-NEXT:   $vgpr27 = COPY [[UV27]](s32)
+  ; CHECK-NEXT:   $vgpr28 = COPY [[UV28]](s32)
+  ; CHECK-NEXT:   $vgpr29 = COPY [[UV29]](s32)
+  ; CHECK-NEXT:   $vgpr30 = COPY [[UV30]](s32)
+  ; CHECK-NEXT:   $vgpr31 = COPY [[UV31]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15, implicit $vgpr16, implicit $vgpr17, implicit $vgpr18, implicit $vgpr19, implicit $vgpr20, implicit $vgpr21, implicit $vgpr22, implicit $vgpr23, implicit $vgpr24, implicit $vgpr25, implicit $vgpr26, implicit $vgpr27, implicit $vgpr28, implicit $vgpr29, implicit $vgpr30, implicit $vgpr31
   %ptr = load volatile <32 x i32> addrspace(1)*, <32 x i32> addrspace(1)* addrspace(4)* undef
   %val = load <32 x i32>, <32 x i32> addrspace(1)* %ptr
   ret <32 x i32> %val
 }
 
 define <2 x i64> @v2i64_func_void() #0 {
   ; CHECK-LABEL: name: v2i64_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(<2 x s64>) = G_LOAD [[DEF]](p1) :: (load (<2 x s64>) from `<2 x i64> addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD]](<2 x s64>)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   $vgpr2 = COPY [[UV2]](s32)
-  ; CHECK:   $vgpr3 = COPY [[UV3]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(<2 x s64>) = G_LOAD [[DEF]](p1) :: (load (<2 x s64>) from `<2 x i64> addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD]](<2 x s64>)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
+  ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
   %val = load <2 x i64>, <2 x i64> addrspace(1)* undef
   ret <2 x i64> %val
 }
 
 define <3 x i64> @v3i64_func_void() #0 {
   ; CHECK-LABEL: name: v3i64_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `<3 x i64> addrspace(1)* addrspace(4)* undef`, addrspace 4)
-  ; CHECK:   [[LOAD1:%[0-9]+]]:_(<3 x s64>) = G_LOAD [[LOAD]](p1) :: (load (<3 x s64>) from %ir.ptr, align 32, addrspace 1)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD1]](<3 x s64>)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   $vgpr2 = COPY [[UV2]](s32)
-  ; CHECK:   $vgpr3 = COPY [[UV3]](s32)
-  ; CHECK:   $vgpr4 = COPY [[UV4]](s32)
-  ; CHECK:   $vgpr5 = COPY [[UV5]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `<3 x i64> addrspace(1)* addrspace(4)* undef`, addrspace 4)
+  ; CHECK-NEXT:   [[LOAD1:%[0-9]+]]:_(<3 x s64>) = G_LOAD [[LOAD]](p1) :: (load (<3 x s64>) from %ir.ptr, align 32, addrspace 1)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD1]](<3 x s64>)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
+  ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
+  ; CHECK-NEXT:   $vgpr4 = COPY [[UV4]](s32)
+  ; CHECK-NEXT:   $vgpr5 = COPY [[UV5]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5
   %ptr = load volatile <3 x i64> addrspace(1)*, <3 x i64> addrspace(1)* addrspace(4)* undef
   %val = load <3 x i64>, <3 x i64> addrspace(1)* %ptr
   ret <3 x i64> %val
 }
 
 define <4 x i64> @v4i64_func_void() #0 {
   ; CHECK-LABEL: name: v4i64_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `<4 x i64> addrspace(1)* addrspace(4)* undef`, addrspace 4)
-  ; CHECK:   [[LOAD1:%[0-9]+]]:_(<4 x s64>) = G_LOAD [[LOAD]](p1) :: (load (<4 x s64>) from %ir.ptr, addrspace 1)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD1]](<4 x s64>)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   $vgpr2 = COPY [[UV2]](s32)
-  ; CHECK:   $vgpr3 = COPY [[UV3]](s32)
-  ; CHECK:   $vgpr4 = COPY [[UV4]](s32)
-  ; CHECK:   $vgpr5 = COPY [[UV5]](s32)
-  ; CHECK:   $vgpr6 = COPY [[UV6]](s32)
-  ; CHECK:   $vgpr7 = COPY [[UV7]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `<4 x i64> addrspace(1)* addrspace(4)* undef`, addrspace 4)
+  ; CHECK-NEXT:   [[LOAD1:%[0-9]+]]:_(<4 x s64>) = G_LOAD [[LOAD]](p1) :: (load (<4 x s64>) from %ir.ptr, addrspace 1)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD1]](<4 x s64>)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
+  ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
+  ; CHECK-NEXT:   $vgpr4 = COPY [[UV4]](s32)
+  ; CHECK-NEXT:   $vgpr5 = COPY [[UV5]](s32)
+  ; CHECK-NEXT:   $vgpr6 = COPY [[UV6]](s32)
+  ; CHECK-NEXT:   $vgpr7 = COPY [[UV7]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
   %ptr = load volatile <4 x i64> addrspace(1)*, <4 x i64> addrspace(1)* addrspace(4)* undef
   %val = load <4 x i64>, <4 x i64> addrspace(1)* %ptr
   ret <4 x i64> %val
 }
 
 define <5 x i64> @v5i64_func_void() #0 {
   ; CHECK-LABEL: name: v5i64_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `<5 x i64> addrspace(1)* addrspace(4)* undef`, addrspace 4)
-  ; CHECK:   [[LOAD1:%[0-9]+]]:_(<5 x s64>) = G_LOAD [[LOAD]](p1) :: (load (<5 x s64>) from %ir.ptr, align 64, addrspace 1)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32), [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD1]](<5 x s64>)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   $vgpr2 = COPY [[UV2]](s32)
-  ; CHECK:   $vgpr3 = COPY [[UV3]](s32)
-  ; CHECK:   $vgpr4 = COPY [[UV4]](s32)
-  ; CHECK:   $vgpr5 = COPY [[UV5]](s32)
-  ; CHECK:   $vgpr6 = COPY [[UV6]](s32)
-  ; CHECK:   $vgpr7 = COPY [[UV7]](s32)
-  ; CHECK:   $vgpr8 = COPY [[UV8]](s32)
-  ; CHECK:   $vgpr9 = COPY [[UV9]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `<5 x i64> addrspace(1)* addrspace(4)* undef`, addrspace 4)
+  ; CHECK-NEXT:   [[LOAD1:%[0-9]+]]:_(<5 x s64>) = G_LOAD [[LOAD]](p1) :: (load (<5 x s64>) from %ir.ptr, align 64, addrspace 1)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32), [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD1]](<5 x s64>)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
+  ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
+  ; CHECK-NEXT:   $vgpr4 = COPY [[UV4]](s32)
+  ; CHECK-NEXT:   $vgpr5 = COPY [[UV5]](s32)
+  ; CHECK-NEXT:   $vgpr6 = COPY [[UV6]](s32)
+  ; CHECK-NEXT:   $vgpr7 = COPY [[UV7]](s32)
+  ; CHECK-NEXT:   $vgpr8 = COPY [[UV8]](s32)
+  ; CHECK-NEXT:   $vgpr9 = COPY [[UV9]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9
   %ptr = load volatile <5 x i64> addrspace(1)*, <5 x i64> addrspace(1)* addrspace(4)* undef
   %val = load <5 x i64>, <5 x i64> addrspace(1)* %ptr
   ret <5 x i64> %val
 }
 
 define <8 x i64> @v8i64_func_void() #0 {
   ; CHECK-LABEL: name: v8i64_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `<8 x i64> addrspace(1)* addrspace(4)* undef`, addrspace 4)
-  ; CHECK:   [[LOAD1:%[0-9]+]]:_(<8 x s64>) = G_LOAD [[LOAD]](p1) :: (load (<8 x s64>) from %ir.ptr, addrspace 1)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32), [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32), [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32), [[UV12:%[0-9]+]]:_(s32), [[UV13:%[0-9]+]]:_(s32), [[UV14:%[0-9]+]]:_(s32), [[UV15:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD1]](<8 x s64>)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   $vgpr2 = COPY [[UV2]](s32)
-  ; CHECK:   $vgpr3 = COPY [[UV3]](s32)
-  ; CHECK:   $vgpr4 = COPY [[UV4]](s32)
-  ; CHECK:   $vgpr5 = COPY [[UV5]](s32)
-  ; CHECK:   $vgpr6 = COPY [[UV6]](s32)
-  ; CHECK:   $vgpr7 = COPY [[UV7]](s32)
-  ; CHECK:   $vgpr8 = COPY [[UV8]](s32)
-  ; CHECK:   $vgpr9 = COPY [[UV9]](s32)
-  ; CHECK:   $vgpr10 = COPY [[UV10]](s32)
-  ; CHECK:   $vgpr11 = COPY [[UV11]](s32)
-  ; CHECK:   $vgpr12 = COPY [[UV12]](s32)
-  ; CHECK:   $vgpr13 = COPY [[UV13]](s32)
-  ; CHECK:   $vgpr14 = COPY [[UV14]](s32)
-  ; CHECK:   $vgpr15 = COPY [[UV15]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `<8 x i64> addrspace(1)* addrspace(4)* undef`, addrspace 4)
+  ; CHECK-NEXT:   [[LOAD1:%[0-9]+]]:_(<8 x s64>) = G_LOAD [[LOAD]](p1) :: (load (<8 x s64>) from %ir.ptr, addrspace 1)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32), [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32), [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32), [[UV12:%[0-9]+]]:_(s32), [[UV13:%[0-9]+]]:_(s32), [[UV14:%[0-9]+]]:_(s32), [[UV15:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD1]](<8 x s64>)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
+  ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
+  ; CHECK-NEXT:   $vgpr4 = COPY [[UV4]](s32)
+  ; CHECK-NEXT:   $vgpr5 = COPY [[UV5]](s32)
+  ; CHECK-NEXT:   $vgpr6 = COPY [[UV6]](s32)
+  ; CHECK-NEXT:   $vgpr7 = COPY [[UV7]](s32)
+  ; CHECK-NEXT:   $vgpr8 = COPY [[UV8]](s32)
+  ; CHECK-NEXT:   $vgpr9 = COPY [[UV9]](s32)
+  ; CHECK-NEXT:   $vgpr10 = COPY [[UV10]](s32)
+  ; CHECK-NEXT:   $vgpr11 = COPY [[UV11]](s32)
+  ; CHECK-NEXT:   $vgpr12 = COPY [[UV12]](s32)
+  ; CHECK-NEXT:   $vgpr13 = COPY [[UV13]](s32)
+  ; CHECK-NEXT:   $vgpr14 = COPY [[UV14]](s32)
+  ; CHECK-NEXT:   $vgpr15 = COPY [[UV15]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15
   %ptr = load volatile <8 x i64> addrspace(1)*, <8 x i64> addrspace(1)* addrspace(4)* undef
   %val = load <8 x i64>, <8 x i64> addrspace(1)* %ptr
   ret <8 x i64> %val
 }
 
 define <16 x i64> @v16i64_func_void() #0 {
   ; CHECK-LABEL: name: v16i64_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `<16 x i64> addrspace(1)* addrspace(4)* undef`, addrspace 4)
-  ; CHECK:   [[LOAD1:%[0-9]+]]:_(<16 x s64>) = G_LOAD [[LOAD]](p1) :: (load (<16 x s64>) from %ir.ptr, addrspace 1)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32), [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32), [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32), [[UV12:%[0-9]+]]:_(s32), [[UV13:%[0-9]+]]:_(s32), [[UV14:%[0-9]+]]:_(s32), [[UV15:%[0-9]+]]:_(s32), [[UV16:%[0-9]+]]:_(s32), [[UV17:%[0-9]+]]:_(s32), [[UV18:%[0-9]+]]:_(s32), [[UV19:%[0-9]+]]:_(s32), [[UV20:%[0-9]+]]:_(s32), [[UV21:%[0-9]+]]:_(s32), [[UV22:%[0-9]+]]:_(s32), [[UV23:%[0-9]+]]:_(s32), [[UV24:%[0-9]+]]:_(s32), [[UV25:%[0-9]+]]:_(s32), [[UV26:%[0-9]+]]:_(s32), [[UV27:%[0-9]+]]:_(s32), [[UV28:%[0-9]+]]:_(s32), [[UV29:%[0-9]+]]:_(s32), [[UV30:%[0-9]+]]:_(s32), [[UV31:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD1]](<16 x s64>)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   $vgpr2 = COPY [[UV2]](s32)
-  ; CHECK:   $vgpr3 = COPY [[UV3]](s32)
-  ; CHECK:   $vgpr4 = COPY [[UV4]](s32)
-  ; CHECK:   $vgpr5 = COPY [[UV5]](s32)
-  ; CHECK:   $vgpr6 = COPY [[UV6]](s32)
-  ; CHECK:   $vgpr7 = COPY [[UV7]](s32)
-  ; CHECK:   $vgpr8 = COPY [[UV8]](s32)
-  ; CHECK:   $vgpr9 = COPY [[UV9]](s32)
-  ; CHECK:   $vgpr10 = COPY [[UV10]](s32)
-  ; CHECK:   $vgpr11 = COPY [[UV11]](s32)
-  ; CHECK:   $vgpr12 = COPY [[UV12]](s32)
-  ; CHECK:   $vgpr13 = COPY [[UV13]](s32)
-  ; CHECK:   $vgpr14 = COPY [[UV14]](s32)
-  ; CHECK:   $vgpr15 = COPY [[UV15]](s32)
-  ; CHECK:   $vgpr16 = COPY [[UV16]](s32)
-  ; CHECK:   $vgpr17 = COPY [[UV17]](s32)
-  ; CHECK:   $vgpr18 = COPY [[UV18]](s32)
-  ; CHECK:   $vgpr19 = COPY [[UV19]](s32)
-  ; CHECK:   $vgpr20 = COPY [[UV20]](s32)
-  ; CHECK:   $vgpr21 = COPY [[UV21]](s32)
-  ; CHECK:   $vgpr22 = COPY [[UV22]](s32)
-  ; CHECK:   $vgpr23 = COPY [[UV23]](s32)
-  ; CHECK:   $vgpr24 = COPY [[UV24]](s32)
-  ; CHECK:   $vgpr25 = COPY [[UV25]](s32)
-  ; CHECK:   $vgpr26 = COPY [[UV26]](s32)
-  ; CHECK:   $vgpr27 = COPY [[UV27]](s32)
-  ; CHECK:   $vgpr28 = COPY [[UV28]](s32)
-  ; CHECK:   $vgpr29 = COPY [[UV29]](s32)
-  ; CHECK:   $vgpr30 = COPY [[UV30]](s32)
-  ; CHECK:   $vgpr31 = COPY [[UV31]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15, implicit $vgpr16, implicit $vgpr17, implicit $vgpr18, implicit $vgpr19, implicit $vgpr20, implicit $vgpr21, implicit $vgpr22, implicit $vgpr23, implicit $vgpr24, implicit $vgpr25, implicit $vgpr26, implicit $vgpr27, implicit $vgpr28, implicit $vgpr29, implicit $vgpr30, implicit $vgpr31
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `<16 x i64> addrspace(1)* addrspace(4)* undef`, addrspace 4)
+  ; CHECK-NEXT:   [[LOAD1:%[0-9]+]]:_(<16 x s64>) = G_LOAD [[LOAD]](p1) :: (load (<16 x s64>) from %ir.ptr, addrspace 1)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32), [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32), [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32), [[UV12:%[0-9]+]]:_(s32), [[UV13:%[0-9]+]]:_(s32), [[UV14:%[0-9]+]]:_(s32), [[UV15:%[0-9]+]]:_(s32), [[UV16:%[0-9]+]]:_(s32), [[UV17:%[0-9]+]]:_(s32), [[UV18:%[0-9]+]]:_(s32), [[UV19:%[0-9]+]]:_(s32), [[UV20:%[0-9]+]]:_(s32), [[UV21:%[0-9]+]]:_(s32), [[UV22:%[0-9]+]]:_(s32), [[UV23:%[0-9]+]]:_(s32), [[UV24:%[0-9]+]]:_(s32), [[UV25:%[0-9]+]]:_(s32), [[UV26:%[0-9]+]]:_(s32), [[UV27:%[0-9]+]]:_(s32), [[UV28:%[0-9]+]]:_(s32), [[UV29:%[0-9]+]]:_(s32), [[UV30:%[0-9]+]]:_(s32), [[UV31:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD1]](<16 x s64>)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
+  ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
+  ; CHECK-NEXT:   $vgpr4 = COPY [[UV4]](s32)
+  ; CHECK-NEXT:   $vgpr5 = COPY [[UV5]](s32)
+  ; CHECK-NEXT:   $vgpr6 = COPY [[UV6]](s32)
+  ; CHECK-NEXT:   $vgpr7 = COPY [[UV7]](s32)
+  ; CHECK-NEXT:   $vgpr8 = COPY [[UV8]](s32)
+  ; CHECK-NEXT:   $vgpr9 = COPY [[UV9]](s32)
+  ; CHECK-NEXT:   $vgpr10 = COPY [[UV10]](s32)
+  ; CHECK-NEXT:   $vgpr11 = COPY [[UV11]](s32)
+  ; CHECK-NEXT:   $vgpr12 = COPY [[UV12]](s32)
+  ; CHECK-NEXT:   $vgpr13 = COPY [[UV13]](s32)
+  ; CHECK-NEXT:   $vgpr14 = COPY [[UV14]](s32)
+  ; CHECK-NEXT:   $vgpr15 = COPY [[UV15]](s32)
+  ; CHECK-NEXT:   $vgpr16 = COPY [[UV16]](s32)
+  ; CHECK-NEXT:   $vgpr17 = COPY [[UV17]](s32)
+  ; CHECK-NEXT:   $vgpr18 = COPY [[UV18]](s32)
+  ; CHECK-NEXT:   $vgpr19 = COPY [[UV19]](s32)
+  ; CHECK-NEXT:   $vgpr20 = COPY [[UV20]](s32)
+  ; CHECK-NEXT:   $vgpr21 = COPY [[UV21]](s32)
+  ; CHECK-NEXT:   $vgpr22 = COPY [[UV22]](s32)
+  ; CHECK-NEXT:   $vgpr23 = COPY [[UV23]](s32)
+  ; CHECK-NEXT:   $vgpr24 = COPY [[UV24]](s32)
+  ; CHECK-NEXT:   $vgpr25 = COPY [[UV25]](s32)
+  ; CHECK-NEXT:   $vgpr26 = COPY [[UV26]](s32)
+  ; CHECK-NEXT:   $vgpr27 = COPY [[UV27]](s32)
+  ; CHECK-NEXT:   $vgpr28 = COPY [[UV28]](s32)
+  ; CHECK-NEXT:   $vgpr29 = COPY [[UV29]](s32)
+  ; CHECK-NEXT:   $vgpr30 = COPY [[UV30]](s32)
+  ; CHECK-NEXT:   $vgpr31 = COPY [[UV31]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15, implicit $vgpr16, implicit $vgpr17, implicit $vgpr18, implicit $vgpr19, implicit $vgpr20, implicit $vgpr21, implicit $vgpr22, implicit $vgpr23, implicit $vgpr24, implicit $vgpr25, implicit $vgpr26, implicit $vgpr27, implicit $vgpr28, implicit $vgpr29, implicit $vgpr30, implicit $vgpr31
   %ptr = load volatile <16 x i64> addrspace(1)*, <16 x i64> addrspace(1)* addrspace(4)* undef
   %val = load <16 x i64>, <16 x i64> addrspace(1)* %ptr
   ret <16 x i64> %val
 }
 
 define <2 x i16> @v2i16_func_void() #0 {
   ; CHECK-LABEL: name: v2i16_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(<2 x s16>) = G_LOAD [[DEF]](p1) :: (load (<2 x s16>) from `<2 x i16> addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   $vgpr0 = COPY [[LOAD]](<2 x s16>)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(<2 x s16>) = G_LOAD [[DEF]](p1) :: (load (<2 x s16>) from `<2 x i16> addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[LOAD]](<2 x s16>)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = load <2 x i16>, <2 x i16> addrspace(1)* undef
   ret <2 x i16> %val
 }
 
 define <2 x half> @v2f16_func_void() #0 {
   ; CHECK-LABEL: name: v2f16_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(<2 x s16>) = G_LOAD [[DEF]](p1) :: (load (<2 x s16>) from `<2 x half> addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   $vgpr0 = COPY [[LOAD]](<2 x s16>)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(<2 x s16>) = G_LOAD [[DEF]](p1) :: (load (<2 x s16>) from `<2 x half> addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[LOAD]](<2 x s16>)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = load <2 x half>, <2 x half> addrspace(1)* undef
   ret <2 x half> %val
 }
 
 define <3 x i16> @v3i16_func_void() #0 {
   ; CHECK-LABEL: name: v3i16_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(<3 x s16>) = G_LOAD [[DEF]](p1) :: (load (<3 x s16>) from `<3 x i16> addrspace(1)* undef`, align 8, addrspace 1)
-  ; CHECK:   [[DEF1:%[0-9]+]]:_(<3 x s16>) = G_IMPLICIT_DEF
-  ; CHECK:   [[CONCAT_VECTORS:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[LOAD]](<3 x s16>), [[DEF1]](<3 x s16>)
-  ; CHECK:   [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>), [[UV2:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<6 x s16>)
-  ; CHECK:   $vgpr0 = COPY [[UV]](<2 x s16>)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](<2 x s16>)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(<3 x s16>) = G_LOAD [[DEF]](p1) :: (load (<3 x s16>) from `<3 x i16> addrspace(1)* undef`, align 8, addrspace 1)
+  ; CHECK-NEXT:   [[DEF1:%[0-9]+]]:_(<3 x s16>) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[LOAD]](<3 x s16>), [[DEF1]](<3 x s16>)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>), [[UV2:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<6 x s16>)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](<2 x s16>)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](<2 x s16>)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %val = load <3 x i16>, <3 x i16> addrspace(1)* undef
   ret <3 x i16> %val
 }
 
 define <4 x i16> @v4i16_func_void() #0 {
   ; CHECK-LABEL: name: v4i16_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(<4 x s16>) = G_LOAD [[DEF]](p1) :: (load (<4 x s16>) from `<4 x i16> addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[LOAD]](<4 x s16>)
-  ; CHECK:   $vgpr0 = COPY [[UV]](<2 x s16>)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](<2 x s16>)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(<4 x s16>) = G_LOAD [[DEF]](p1) :: (load (<4 x s16>) from `<4 x i16> addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[LOAD]](<4 x s16>)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](<2 x s16>)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](<2 x s16>)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %val = load <4 x i16>, <4 x i16> addrspace(1)* undef
   ret <4 x i16> %val
 }
 
 define <4 x half> @v4f16_func_void() #0 {
   ; CHECK-LABEL: name: v4f16_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(<4 x s16>) = G_LOAD [[DEF]](p1) :: (load (<4 x s16>) from `<4 x half> addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[LOAD]](<4 x s16>)
-  ; CHECK:   $vgpr0 = COPY [[UV]](<2 x s16>)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](<2 x s16>)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(<4 x s16>) = G_LOAD [[DEF]](p1) :: (load (<4 x s16>) from `<4 x half> addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[LOAD]](<4 x s16>)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](<2 x s16>)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](<2 x s16>)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %val = load <4 x half>, <4 x half> addrspace(1)* undef
   ret <4 x half> %val
 }
 
 define <5 x i16> @v5i16_func_void() #0 {
   ; CHECK-LABEL: name: v5i16_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `<5 x i16> addrspace(1)* addrspace(4)* undef`, addrspace 4)
-  ; CHECK:   [[LOAD1:%[0-9]+]]:_(<5 x s16>) = G_LOAD [[LOAD]](p1) :: (load (<5 x s16>) from %ir.ptr, align 16, addrspace 1)
-  ; CHECK:   [[DEF1:%[0-9]+]]:_(<5 x s16>) = G_IMPLICIT_DEF
-  ; CHECK:   [[CONCAT_VECTORS:%[0-9]+]]:_(<10 x s16>) = G_CONCAT_VECTORS [[LOAD1]](<5 x s16>), [[DEF1]](<5 x s16>)
-  ; CHECK:   [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>), [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>), [[UV4:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<10 x s16>)
-  ; CHECK:   $vgpr0 = COPY [[UV]](<2 x s16>)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](<2 x s16>)
-  ; CHECK:   $vgpr2 = COPY [[UV2]](<2 x s16>)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `<5 x i16> addrspace(1)* addrspace(4)* undef`, addrspace 4)
+  ; CHECK-NEXT:   [[LOAD1:%[0-9]+]]:_(<5 x s16>) = G_LOAD [[LOAD]](p1) :: (load (<5 x s16>) from %ir.ptr, align 16, addrspace 1)
+  ; CHECK-NEXT:   [[DEF1:%[0-9]+]]:_(<5 x s16>) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:_(<10 x s16>) = G_CONCAT_VECTORS [[LOAD1]](<5 x s16>), [[DEF1]](<5 x s16>)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>), [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>), [[UV4:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<10 x s16>)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](<2 x s16>)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](<2 x s16>)
+  ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](<2 x s16>)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
   %ptr = load volatile <5 x i16> addrspace(1)*, <5 x i16> addrspace(1)* addrspace(4)* undef
   %val = load <5 x i16>, <5 x i16> addrspace(1)* %ptr
   ret <5 x i16> %val
 }
 
 define <8 x i16> @v8i16_func_void() #0 {
   ; CHECK-LABEL: name: v8i16_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `<8 x i16> addrspace(1)* addrspace(4)* undef`, addrspace 4)
-  ; CHECK:   [[LOAD1:%[0-9]+]]:_(<8 x s16>) = G_LOAD [[LOAD]](p1) :: (load (<8 x s16>) from %ir.ptr, addrspace 1)
-  ; CHECK:   [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>), [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[LOAD1]](<8 x s16>)
-  ; CHECK:   $vgpr0 = COPY [[UV]](<2 x s16>)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](<2 x s16>)
-  ; CHECK:   $vgpr2 = COPY [[UV2]](<2 x s16>)
-  ; CHECK:   $vgpr3 = COPY [[UV3]](<2 x s16>)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `<8 x i16> addrspace(1)* addrspace(4)* undef`, addrspace 4)
+  ; CHECK-NEXT:   [[LOAD1:%[0-9]+]]:_(<8 x s16>) = G_LOAD [[LOAD]](p1) :: (load (<8 x s16>) from %ir.ptr, addrspace 1)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>), [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[LOAD1]](<8 x s16>)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](<2 x s16>)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](<2 x s16>)
+  ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](<2 x s16>)
+  ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](<2 x s16>)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
   %ptr = load volatile <8 x i16> addrspace(1)*, <8 x i16> addrspace(1)* addrspace(4)* undef
   %val = load <8 x i16>, <8 x i16> addrspace(1)* %ptr
   ret <8 x i16> %val
 }
 
 define <16 x i16> @v16i16_func_void() #0 {
   ; CHECK-LABEL: name: v16i16_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `<16 x i16> addrspace(1)* addrspace(4)* undef`, addrspace 4)
-  ; CHECK:   [[LOAD1:%[0-9]+]]:_(<16 x s16>) = G_LOAD [[LOAD]](p1) :: (load (<16 x s16>) from %ir.ptr, addrspace 1)
-  ; CHECK:   [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>), [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>), [[UV4:%[0-9]+]]:_(<2 x s16>), [[UV5:%[0-9]+]]:_(<2 x s16>), [[UV6:%[0-9]+]]:_(<2 x s16>), [[UV7:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[LOAD1]](<16 x s16>)
-  ; CHECK:   $vgpr0 = COPY [[UV]](<2 x s16>)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](<2 x s16>)
-  ; CHECK:   $vgpr2 = COPY [[UV2]](<2 x s16>)
-  ; CHECK:   $vgpr3 = COPY [[UV3]](<2 x s16>)
-  ; CHECK:   $vgpr4 = COPY [[UV4]](<2 x s16>)
-  ; CHECK:   $vgpr5 = COPY [[UV5]](<2 x s16>)
-  ; CHECK:   $vgpr6 = COPY [[UV6]](<2 x s16>)
-  ; CHECK:   $vgpr7 = COPY [[UV7]](<2 x s16>)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `<16 x i16> addrspace(1)* addrspace(4)* undef`, addrspace 4)
+  ; CHECK-NEXT:   [[LOAD1:%[0-9]+]]:_(<16 x s16>) = G_LOAD [[LOAD]](p1) :: (load (<16 x s16>) from %ir.ptr, addrspace 1)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>), [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>), [[UV4:%[0-9]+]]:_(<2 x s16>), [[UV5:%[0-9]+]]:_(<2 x s16>), [[UV6:%[0-9]+]]:_(<2 x s16>), [[UV7:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[LOAD1]](<16 x s16>)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](<2 x s16>)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](<2 x s16>)
+  ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](<2 x s16>)
+  ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](<2 x s16>)
+  ; CHECK-NEXT:   $vgpr4 = COPY [[UV4]](<2 x s16>)
+  ; CHECK-NEXT:   $vgpr5 = COPY [[UV5]](<2 x s16>)
+  ; CHECK-NEXT:   $vgpr6 = COPY [[UV6]](<2 x s16>)
+  ; CHECK-NEXT:   $vgpr7 = COPY [[UV7]](<2 x s16>)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
   %ptr = load volatile <16 x i16> addrspace(1)*, <16 x i16> addrspace(1)* addrspace(4)* undef
   %val = load <16 x i16>, <16 x i16> addrspace(1)* %ptr
   ret <16 x i16> %val
 }
 
 define <16 x i8> @v16i8_func_void() #0 {
   ; CHECK-LABEL: name: v16i8_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `<16 x i8> addrspace(1)* addrspace(4)* undef`, addrspace 4)
-  ; CHECK:   [[LOAD1:%[0-9]+]]:_(<16 x s8>) = G_LOAD [[LOAD]](p1) :: (load (<16 x s8>) from %ir.ptr, addrspace 1)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s8), [[UV1:%[0-9]+]]:_(s8), [[UV2:%[0-9]+]]:_(s8), [[UV3:%[0-9]+]]:_(s8), [[UV4:%[0-9]+]]:_(s8), [[UV5:%[0-9]+]]:_(s8), [[UV6:%[0-9]+]]:_(s8), [[UV7:%[0-9]+]]:_(s8), [[UV8:%[0-9]+]]:_(s8), [[UV9:%[0-9]+]]:_(s8), [[UV10:%[0-9]+]]:_(s8), [[UV11:%[0-9]+]]:_(s8), [[UV12:%[0-9]+]]:_(s8), [[UV13:%[0-9]+]]:_(s8), [[UV14:%[0-9]+]]:_(s8), [[UV15:%[0-9]+]]:_(s8) = G_UNMERGE_VALUES [[LOAD1]](<16 x s8>)
-  ; CHECK:   [[ANYEXT:%[0-9]+]]:_(s16) = G_ANYEXT [[UV]](s8)
-  ; CHECK:   [[ANYEXT1:%[0-9]+]]:_(s16) = G_ANYEXT [[UV1]](s8)
-  ; CHECK:   [[ANYEXT2:%[0-9]+]]:_(s16) = G_ANYEXT [[UV2]](s8)
-  ; CHECK:   [[ANYEXT3:%[0-9]+]]:_(s16) = G_ANYEXT [[UV3]](s8)
-  ; CHECK:   [[ANYEXT4:%[0-9]+]]:_(s16) = G_ANYEXT [[UV4]](s8)
-  ; CHECK:   [[ANYEXT5:%[0-9]+]]:_(s16) = G_ANYEXT [[UV5]](s8)
-  ; CHECK:   [[ANYEXT6:%[0-9]+]]:_(s16) = G_ANYEXT [[UV6]](s8)
-  ; CHECK:   [[ANYEXT7:%[0-9]+]]:_(s16) = G_ANYEXT [[UV7]](s8)
-  ; CHECK:   [[ANYEXT8:%[0-9]+]]:_(s16) = G_ANYEXT [[UV8]](s8)
-  ; CHECK:   [[ANYEXT9:%[0-9]+]]:_(s16) = G_ANYEXT [[UV9]](s8)
-  ; CHECK:   [[ANYEXT10:%[0-9]+]]:_(s16) = G_ANYEXT [[UV10]](s8)
-  ; CHECK:   [[ANYEXT11:%[0-9]+]]:_(s16) = G_ANYEXT [[UV11]](s8)
-  ; CHECK:   [[ANYEXT12:%[0-9]+]]:_(s16) = G_ANYEXT [[UV12]](s8)
-  ; CHECK:   [[ANYEXT13:%[0-9]+]]:_(s16) = G_ANYEXT [[UV13]](s8)
-  ; CHECK:   [[ANYEXT14:%[0-9]+]]:_(s16) = G_ANYEXT [[UV14]](s8)
-  ; CHECK:   [[ANYEXT15:%[0-9]+]]:_(s16) = G_ANYEXT [[UV15]](s8)
-  ; CHECK:   [[ANYEXT16:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT]](s16)
-  ; CHECK:   $vgpr0 = COPY [[ANYEXT16]](s32)
-  ; CHECK:   [[ANYEXT17:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT1]](s16)
-  ; CHECK:   $vgpr1 = COPY [[ANYEXT17]](s32)
-  ; CHECK:   [[ANYEXT18:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT2]](s16)
-  ; CHECK:   $vgpr2 = COPY [[ANYEXT18]](s32)
-  ; CHECK:   [[ANYEXT19:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT3]](s16)
-  ; CHECK:   $vgpr3 = COPY [[ANYEXT19]](s32)
-  ; CHECK:   [[ANYEXT20:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT4]](s16)
-  ; CHECK:   $vgpr4 = COPY [[ANYEXT20]](s32)
-  ; CHECK:   [[ANYEXT21:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT5]](s16)
-  ; CHECK:   $vgpr5 = COPY [[ANYEXT21]](s32)
-  ; CHECK:   [[ANYEXT22:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT6]](s16)
-  ; CHECK:   $vgpr6 = COPY [[ANYEXT22]](s32)
-  ; CHECK:   [[ANYEXT23:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT7]](s16)
-  ; CHECK:   $vgpr7 = COPY [[ANYEXT23]](s32)
-  ; CHECK:   [[ANYEXT24:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT8]](s16)
-  ; CHECK:   $vgpr8 = COPY [[ANYEXT24]](s32)
-  ; CHECK:   [[ANYEXT25:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT9]](s16)
-  ; CHECK:   $vgpr9 = COPY [[ANYEXT25]](s32)
-  ; CHECK:   [[ANYEXT26:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT10]](s16)
-  ; CHECK:   $vgpr10 = COPY [[ANYEXT26]](s32)
-  ; CHECK:   [[ANYEXT27:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT11]](s16)
-  ; CHECK:   $vgpr11 = COPY [[ANYEXT27]](s32)
-  ; CHECK:   [[ANYEXT28:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT12]](s16)
-  ; CHECK:   $vgpr12 = COPY [[ANYEXT28]](s32)
-  ; CHECK:   [[ANYEXT29:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT13]](s16)
-  ; CHECK:   $vgpr13 = COPY [[ANYEXT29]](s32)
-  ; CHECK:   [[ANYEXT30:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT14]](s16)
-  ; CHECK:   $vgpr14 = COPY [[ANYEXT30]](s32)
-  ; CHECK:   [[ANYEXT31:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT15]](s16)
-  ; CHECK:   $vgpr15 = COPY [[ANYEXT31]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `<16 x i8> addrspace(1)* addrspace(4)* undef`, addrspace 4)
+  ; CHECK-NEXT:   [[LOAD1:%[0-9]+]]:_(<16 x s8>) = G_LOAD [[LOAD]](p1) :: (load (<16 x s8>) from %ir.ptr, addrspace 1)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s8), [[UV1:%[0-9]+]]:_(s8), [[UV2:%[0-9]+]]:_(s8), [[UV3:%[0-9]+]]:_(s8), [[UV4:%[0-9]+]]:_(s8), [[UV5:%[0-9]+]]:_(s8), [[UV6:%[0-9]+]]:_(s8), [[UV7:%[0-9]+]]:_(s8), [[UV8:%[0-9]+]]:_(s8), [[UV9:%[0-9]+]]:_(s8), [[UV10:%[0-9]+]]:_(s8), [[UV11:%[0-9]+]]:_(s8), [[UV12:%[0-9]+]]:_(s8), [[UV13:%[0-9]+]]:_(s8), [[UV14:%[0-9]+]]:_(s8), [[UV15:%[0-9]+]]:_(s8) = G_UNMERGE_VALUES [[LOAD1]](<16 x s8>)
+  ; CHECK-NEXT:   [[ANYEXT:%[0-9]+]]:_(s16) = G_ANYEXT [[UV]](s8)
+  ; CHECK-NEXT:   [[ANYEXT1:%[0-9]+]]:_(s16) = G_ANYEXT [[UV1]](s8)
+  ; CHECK-NEXT:   [[ANYEXT2:%[0-9]+]]:_(s16) = G_ANYEXT [[UV2]](s8)
+  ; CHECK-NEXT:   [[ANYEXT3:%[0-9]+]]:_(s16) = G_ANYEXT [[UV3]](s8)
+  ; CHECK-NEXT:   [[ANYEXT4:%[0-9]+]]:_(s16) = G_ANYEXT [[UV4]](s8)
+  ; CHECK-NEXT:   [[ANYEXT5:%[0-9]+]]:_(s16) = G_ANYEXT [[UV5]](s8)
+  ; CHECK-NEXT:   [[ANYEXT6:%[0-9]+]]:_(s16) = G_ANYEXT [[UV6]](s8)
+  ; CHECK-NEXT:   [[ANYEXT7:%[0-9]+]]:_(s16) = G_ANYEXT [[UV7]](s8)
+  ; CHECK-NEXT:   [[ANYEXT8:%[0-9]+]]:_(s16) = G_ANYEXT [[UV8]](s8)
+  ; CHECK-NEXT:   [[ANYEXT9:%[0-9]+]]:_(s16) = G_ANYEXT [[UV9]](s8)
+  ; CHECK-NEXT:   [[ANYEXT10:%[0-9]+]]:_(s16) = G_ANYEXT [[UV10]](s8)
+  ; CHECK-NEXT:   [[ANYEXT11:%[0-9]+]]:_(s16) = G_ANYEXT [[UV11]](s8)
+  ; CHECK-NEXT:   [[ANYEXT12:%[0-9]+]]:_(s16) = G_ANYEXT [[UV12]](s8)
+  ; CHECK-NEXT:   [[ANYEXT13:%[0-9]+]]:_(s16) = G_ANYEXT [[UV13]](s8)
+  ; CHECK-NEXT:   [[ANYEXT14:%[0-9]+]]:_(s16) = G_ANYEXT [[UV14]](s8)
+  ; CHECK-NEXT:   [[ANYEXT15:%[0-9]+]]:_(s16) = G_ANYEXT [[UV15]](s8)
+  ; CHECK-NEXT:   [[ANYEXT16:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT]](s16)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[ANYEXT16]](s32)
+  ; CHECK-NEXT:   [[ANYEXT17:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT1]](s16)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[ANYEXT17]](s32)
+  ; CHECK-NEXT:   [[ANYEXT18:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT2]](s16)
+  ; CHECK-NEXT:   $vgpr2 = COPY [[ANYEXT18]](s32)
+  ; CHECK-NEXT:   [[ANYEXT19:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT3]](s16)
+  ; CHECK-NEXT:   $vgpr3 = COPY [[ANYEXT19]](s32)
+  ; CHECK-NEXT:   [[ANYEXT20:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT4]](s16)
+  ; CHECK-NEXT:   $vgpr4 = COPY [[ANYEXT20]](s32)
+  ; CHECK-NEXT:   [[ANYEXT21:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT5]](s16)
+  ; CHECK-NEXT:   $vgpr5 = COPY [[ANYEXT21]](s32)
+  ; CHECK-NEXT:   [[ANYEXT22:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT6]](s16)
+  ; CHECK-NEXT:   $vgpr6 = COPY [[ANYEXT22]](s32)
+  ; CHECK-NEXT:   [[ANYEXT23:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT7]](s16)
+  ; CHECK-NEXT:   $vgpr7 = COPY [[ANYEXT23]](s32)
+  ; CHECK-NEXT:   [[ANYEXT24:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT8]](s16)
+  ; CHECK-NEXT:   $vgpr8 = COPY [[ANYEXT24]](s32)
+  ; CHECK-NEXT:   [[ANYEXT25:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT9]](s16)
+  ; CHECK-NEXT:   $vgpr9 = COPY [[ANYEXT25]](s32)
+  ; CHECK-NEXT:   [[ANYEXT26:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT10]](s16)
+  ; CHECK-NEXT:   $vgpr10 = COPY [[ANYEXT26]](s32)
+  ; CHECK-NEXT:   [[ANYEXT27:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT11]](s16)
+  ; CHECK-NEXT:   $vgpr11 = COPY [[ANYEXT27]](s32)
+  ; CHECK-NEXT:   [[ANYEXT28:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT12]](s16)
+  ; CHECK-NEXT:   $vgpr12 = COPY [[ANYEXT28]](s32)
+  ; CHECK-NEXT:   [[ANYEXT29:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT13]](s16)
+  ; CHECK-NEXT:   $vgpr13 = COPY [[ANYEXT29]](s32)
+  ; CHECK-NEXT:   [[ANYEXT30:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT14]](s16)
+  ; CHECK-NEXT:   $vgpr14 = COPY [[ANYEXT30]](s32)
+  ; CHECK-NEXT:   [[ANYEXT31:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT15]](s16)
+  ; CHECK-NEXT:   $vgpr15 = COPY [[ANYEXT31]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15
   %ptr = load volatile <16 x i8> addrspace(1)*, <16 x i8> addrspace(1)* addrspace(4)* undef
   %val = load <16 x i8>, <16 x i8> addrspace(1)* %ptr
   ret <16 x i8> %val
 }
 
 define <2 x i8> @v2i8_func_void() #0 {
   ; CHECK-LABEL: name: v2i8_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(<2 x s8>) = G_LOAD [[DEF]](p1) :: (load (<2 x s8>) from `<2 x i8> addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s8), [[UV1:%[0-9]+]]:_(s8) = G_UNMERGE_VALUES [[LOAD]](<2 x s8>)
-  ; CHECK:   [[ANYEXT:%[0-9]+]]:_(s16) = G_ANYEXT [[UV]](s8)
-  ; CHECK:   [[ANYEXT1:%[0-9]+]]:_(s16) = G_ANYEXT [[UV1]](s8)
-  ; CHECK:   [[ANYEXT2:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT]](s16)
-  ; CHECK:   $vgpr0 = COPY [[ANYEXT2]](s32)
-  ; CHECK:   [[ANYEXT3:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT1]](s16)
-  ; CHECK:   $vgpr1 = COPY [[ANYEXT3]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(<2 x s8>) = G_LOAD [[DEF]](p1) :: (load (<2 x s8>) from `<2 x i8> addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s8), [[UV1:%[0-9]+]]:_(s8) = G_UNMERGE_VALUES [[LOAD]](<2 x s8>)
+  ; CHECK-NEXT:   [[ANYEXT:%[0-9]+]]:_(s16) = G_ANYEXT [[UV]](s8)
+  ; CHECK-NEXT:   [[ANYEXT1:%[0-9]+]]:_(s16) = G_ANYEXT [[UV1]](s8)
+  ; CHECK-NEXT:   [[ANYEXT2:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT]](s16)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[ANYEXT2]](s32)
+  ; CHECK-NEXT:   [[ANYEXT3:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT1]](s16)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[ANYEXT3]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %val = load <2 x i8>, <2 x i8> addrspace(1)* undef
   ret <2 x i8> %val
 }
 
 define <3 x i8> @v3i8_func_void() #0 {
   ; CHECK-LABEL: name: v3i8_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(<3 x s8>) = G_LOAD [[DEF]](p1) :: (load (<3 x s8>) from `<3 x i8> addrspace(1)* undef`, align 4, addrspace 1)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s8), [[UV1:%[0-9]+]]:_(s8), [[UV2:%[0-9]+]]:_(s8) = G_UNMERGE_VALUES [[LOAD]](<3 x s8>)
-  ; CHECK:   [[ANYEXT:%[0-9]+]]:_(s16) = G_ANYEXT [[UV]](s8)
-  ; CHECK:   [[ANYEXT1:%[0-9]+]]:_(s16) = G_ANYEXT [[UV1]](s8)
-  ; CHECK:   [[ANYEXT2:%[0-9]+]]:_(s16) = G_ANYEXT [[UV2]](s8)
-  ; CHECK:   [[ANYEXT3:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT]](s16)
-  ; CHECK:   $vgpr0 = COPY [[ANYEXT3]](s32)
-  ; CHECK:   [[ANYEXT4:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT1]](s16)
-  ; CHECK:   $vgpr1 = COPY [[ANYEXT4]](s32)
-  ; CHECK:   [[ANYEXT5:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT2]](s16)
-  ; CHECK:   $vgpr2 = COPY [[ANYEXT5]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(<3 x s8>) = G_LOAD [[DEF]](p1) :: (load (<3 x s8>) from `<3 x i8> addrspace(1)* undef`, align 4, addrspace 1)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s8), [[UV1:%[0-9]+]]:_(s8), [[UV2:%[0-9]+]]:_(s8) = G_UNMERGE_VALUES [[LOAD]](<3 x s8>)
+  ; CHECK-NEXT:   [[ANYEXT:%[0-9]+]]:_(s16) = G_ANYEXT [[UV]](s8)
+  ; CHECK-NEXT:   [[ANYEXT1:%[0-9]+]]:_(s16) = G_ANYEXT [[UV1]](s8)
+  ; CHECK-NEXT:   [[ANYEXT2:%[0-9]+]]:_(s16) = G_ANYEXT [[UV2]](s8)
+  ; CHECK-NEXT:   [[ANYEXT3:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT]](s16)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[ANYEXT3]](s32)
+  ; CHECK-NEXT:   [[ANYEXT4:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT1]](s16)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[ANYEXT4]](s32)
+  ; CHECK-NEXT:   [[ANYEXT5:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT2]](s16)
+  ; CHECK-NEXT:   $vgpr2 = COPY [[ANYEXT5]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
   %val = load <3 x i8>, <3 x i8> addrspace(1)* undef
   ret <3 x i8> %val
 }
 
 define <4  x i8> @v4i8_func_void() #0 {
   ; CHECK-LABEL: name: v4i8_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `<4 x i8> addrspace(1)* addrspace(4)* undef`, addrspace 4)
-  ; CHECK:   [[LOAD1:%[0-9]+]]:_(<4 x s8>) = G_LOAD [[LOAD]](p1) :: (load (<4 x s8>) from %ir.ptr, addrspace 1)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s8), [[UV1:%[0-9]+]]:_(s8), [[UV2:%[0-9]+]]:_(s8), [[UV3:%[0-9]+]]:_(s8) = G_UNMERGE_VALUES [[LOAD1]](<4 x s8>)
-  ; CHECK:   [[ANYEXT:%[0-9]+]]:_(s16) = G_ANYEXT [[UV]](s8)
-  ; CHECK:   [[ANYEXT1:%[0-9]+]]:_(s16) = G_ANYEXT [[UV1]](s8)
-  ; CHECK:   [[ANYEXT2:%[0-9]+]]:_(s16) = G_ANYEXT [[UV2]](s8)
-  ; CHECK:   [[ANYEXT3:%[0-9]+]]:_(s16) = G_ANYEXT [[UV3]](s8)
-  ; CHECK:   [[ANYEXT4:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT]](s16)
-  ; CHECK:   $vgpr0 = COPY [[ANYEXT4]](s32)
-  ; CHECK:   [[ANYEXT5:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT1]](s16)
-  ; CHECK:   $vgpr1 = COPY [[ANYEXT5]](s32)
-  ; CHECK:   [[ANYEXT6:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT2]](s16)
-  ; CHECK:   $vgpr2 = COPY [[ANYEXT6]](s32)
-  ; CHECK:   [[ANYEXT7:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT3]](s16)
-  ; CHECK:   $vgpr3 = COPY [[ANYEXT7]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `<4 x i8> addrspace(1)* addrspace(4)* undef`, addrspace 4)
+  ; CHECK-NEXT:   [[LOAD1:%[0-9]+]]:_(<4 x s8>) = G_LOAD [[LOAD]](p1) :: (load (<4 x s8>) from %ir.ptr, addrspace 1)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s8), [[UV1:%[0-9]+]]:_(s8), [[UV2:%[0-9]+]]:_(s8), [[UV3:%[0-9]+]]:_(s8) = G_UNMERGE_VALUES [[LOAD1]](<4 x s8>)
+  ; CHECK-NEXT:   [[ANYEXT:%[0-9]+]]:_(s16) = G_ANYEXT [[UV]](s8)
+  ; CHECK-NEXT:   [[ANYEXT1:%[0-9]+]]:_(s16) = G_ANYEXT [[UV1]](s8)
+  ; CHECK-NEXT:   [[ANYEXT2:%[0-9]+]]:_(s16) = G_ANYEXT [[UV2]](s8)
+  ; CHECK-NEXT:   [[ANYEXT3:%[0-9]+]]:_(s16) = G_ANYEXT [[UV3]](s8)
+  ; CHECK-NEXT:   [[ANYEXT4:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT]](s16)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[ANYEXT4]](s32)
+  ; CHECK-NEXT:   [[ANYEXT5:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT1]](s16)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[ANYEXT5]](s32)
+  ; CHECK-NEXT:   [[ANYEXT6:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT2]](s16)
+  ; CHECK-NEXT:   $vgpr2 = COPY [[ANYEXT6]](s32)
+  ; CHECK-NEXT:   [[ANYEXT7:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT3]](s16)
+  ; CHECK-NEXT:   $vgpr3 = COPY [[ANYEXT7]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
   %ptr = load volatile <4  x i8> addrspace(1)*, <4  x i8> addrspace(1)* addrspace(4)* undef
   %val = load <4  x i8>, <4  x i8> addrspace(1)* %ptr
   ret <4  x i8> %val
 }
 
 define {i8, i32} @struct_i8_i32_func_void() #0 {
   ; CHECK-LABEL: name: struct_i8_i32_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s8) = G_LOAD [[DEF]](p1) :: (load (s8) from `{ i8, i32 } addrspace(1)* undef`, align 4, addrspace 1)
-  ; CHECK:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 4
-  ; CHECK:   [[PTR_ADD:%[0-9]+]]:_(p1) = G_PTR_ADD [[DEF]], [[C]](s64)
-  ; CHECK:   [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD]](p1) :: (load (s32) from `{ i8, i32 } addrspace(1)* undef` + 4, addrspace 1)
-  ; CHECK:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[LOAD]](s8)
-  ; CHECK:   $vgpr0 = COPY [[ANYEXT]](s32)
-  ; CHECK:   $vgpr1 = COPY [[LOAD1]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s8) = G_LOAD [[DEF]](p1) :: (load (s8) from `{ i8, i32 } addrspace(1)* undef`, align 4, addrspace 1)
+  ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 4
+  ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p1) = G_PTR_ADD [[DEF]], [[C]](s64)
+  ; CHECK-NEXT:   [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD]](p1) :: (load (s32) from `{ i8, i32 } addrspace(1)* undef` + 4, addrspace 1)
+  ; CHECK-NEXT:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[LOAD]](s8)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[ANYEXT]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[LOAD1]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %val = load { i8, i32 }, { i8, i32 } addrspace(1)* undef
   ret { i8, i32 } %val
 }
 
 define void @void_func_sret_struct_i8_i32({ i8, i32 } addrspace(5)* sret({ i8, i32 }) %arg0) #0 {
   ; CHECK-LABEL: name: void_func_sret_struct_i8_i32
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(p5) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(p1) = COPY [[DEF]](p1)
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s8) = G_LOAD [[DEF]](p1) :: (volatile load (s8) from `i8 addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[COPY2]](p1) :: (volatile load (s32) from `i32 addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 4
-  ; CHECK:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY]], [[C]](s32)
-  ; CHECK:   G_STORE [[LOAD]](s8), [[COPY]](p5) :: (store (s8) into %ir.gep01, addrspace 5)
-  ; CHECK:   G_STORE [[LOAD1]](s32), [[PTR_ADD]](p5) :: (store (s32) into %ir.gep1, addrspace 5)
-  ; CHECK:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-  ; CHECK:   S_SETPC_B64_return [[COPY3]]
+  ; CHECK-NEXT:   liveins: $vgpr0
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(p5) = COPY $vgpr0
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(p1) = COPY [[DEF]](p1)
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s8) = G_LOAD [[DEF]](p1) :: (volatile load (s8) from `i8 addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[COPY1]](p1) :: (volatile load (s32) from `i32 addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 4
+  ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY]], [[C]](s32)
+  ; CHECK-NEXT:   G_STORE [[LOAD]](s8), [[COPY]](p5) :: (store (s8) into %ir.gep01, addrspace 5)
+  ; CHECK-NEXT:   G_STORE [[LOAD1]](s32), [[PTR_ADD]](p5) :: (store (s32) into %ir.gep1, addrspace 5)
+  ; CHECK-NEXT:   SI_RETURN
   %val0 = load volatile i8, i8 addrspace(1)* undef
   %val1 = load volatile i32, i32 addrspace(1)* undef
   %gep0 = getelementptr inbounds { i8, i32 }, { i8, i32 } addrspace(5)* %arg0, i32 0, i32 0
   %gep1 = getelementptr inbounds { i8, i32 }, { i8, i32 } addrspace(5)* %arg0, i32 0, i32 1
   store i8 %val0, i8 addrspace(5)* %gep0
   store i32 %val1, i32 addrspace(5)* %gep1
   ret void
 }
 
 ; FIXME: Should be able to fold offsets in all of these pre-gfx9. Call
 ; lowering introduces an extra CopyToReg/CopyFromReg obscuring the
 ; AssertZext inserted. Not using it introduces the spills.
 
 define <33 x i32> @v33i32_func_void() #0 {
   ; CHECK-LABEL: name: v33i32_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(p5) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `<33 x i32> addrspace(1)* addrspace(4)* undef`, addrspace 4)
-  ; CHECK:   [[LOAD1:%[0-9]+]]:_(<33 x s32>) = G_LOAD [[LOAD]](p1) :: (load (<33 x s32>) from %ir.ptr, align 256, addrspace 1)
-  ; CHECK:   G_STORE [[LOAD1]](<33 x s32>), [[COPY]](p5) :: (store (<33 x s32>), align 256, addrspace 5)
-  ; CHECK:   [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-  ; CHECK:   S_SETPC_B64_return [[COPY2]]
+  ; CHECK-NEXT:   liveins: $vgpr0
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(p5) = COPY $vgpr0
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `<33 x i32> addrspace(1)* addrspace(4)* undef`, addrspace 4)
+  ; CHECK-NEXT:   [[LOAD1:%[0-9]+]]:_(<33 x s32>) = G_LOAD [[LOAD]](p1) :: (load (<33 x s32>) from %ir.ptr, align 256, addrspace 1)
+  ; CHECK-NEXT:   G_STORE [[LOAD1]](<33 x s32>), [[COPY]](p5) :: (store (<33 x s32>), align 256, addrspace 5)
+  ; CHECK-NEXT:   SI_RETURN
   %ptr = load volatile <33 x i32> addrspace(1)*, <33 x i32> addrspace(1)* addrspace(4)* undef
   %val = load <33 x i32>, <33 x i32> addrspace(1)* %ptr
   ret <33 x i32> %val
 }
 
 define <33 x i32> @v33i32_func_v33i32_i32(<33 x i32> addrspace(1)* %p, i32 %idx) #0 {
   ; CHECK-LABEL: name: v33i32_func_v33i32_i32
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(p5) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; CHECK:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY1]](s32), [[COPY2]](s32)
-  ; CHECK:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
-  ; CHECK:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[SEXT:%[0-9]+]]:_(s64) = G_SEXT [[COPY3]](s32)
-  ; CHECK:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 256
-  ; CHECK:   [[MUL:%[0-9]+]]:_(s64) = G_MUL [[SEXT]], [[C]]
-  ; CHECK:   [[PTR_ADD:%[0-9]+]]:_(p1) = G_PTR_ADD [[MV]], [[MUL]](s64)
-  ; CHECK:   [[COPY5:%[0-9]+]]:_(p1) = COPY [[PTR_ADD]](p1)
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(<33 x s32>) = G_LOAD [[COPY5]](p1) :: (load (<33 x s32>) from %ir.gep, align 256, addrspace 1)
-  ; CHECK:   G_STORE [[LOAD]](<33 x s32>), [[COPY]](p5) :: (store (<33 x s32>), align 256, addrspace 5)
-  ; CHECK:   [[COPY6:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY4]]
-  ; CHECK:   S_SETPC_B64_return [[COPY6]]
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(p5) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
+  ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY1]](s32), [[COPY2]](s32)
+  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
+  ; CHECK-NEXT:   [[SEXT:%[0-9]+]]:_(s64) = G_SEXT [[COPY3]](s32)
+  ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 256
+  ; CHECK-NEXT:   [[MUL:%[0-9]+]]:_(s64) = G_MUL [[SEXT]], [[C]]
+  ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p1) = G_PTR_ADD [[MV]], [[MUL]](s64)
+  ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:_(p1) = COPY [[PTR_ADD]](p1)
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(<33 x s32>) = G_LOAD [[COPY4]](p1) :: (load (<33 x s32>) from %ir.gep, align 256, addrspace 1)
+  ; CHECK-NEXT:   G_STORE [[LOAD]](<33 x s32>), [[COPY]](p5) :: (store (<33 x s32>), align 256, addrspace 5)
+  ; CHECK-NEXT:   SI_RETURN
   %gep = getelementptr inbounds <33 x i32>, <33 x i32> addrspace(1)* %p, i32 %idx
   %val = load <33 x i32>, <33 x i32> addrspace(1)* %gep
   ret <33 x i32> %val
 }
 
 define { <32 x i32>, i32 } @struct_v32i32_i32_func_void() #0 {
   ; CHECK-LABEL: name: struct_v32i32_i32_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(p5) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `{ <32 x i32>, i32 } addrspace(1)* addrspace(4)* undef`, addrspace 4)
-  ; CHECK:   [[LOAD1:%[0-9]+]]:_(<32 x s32>) = G_LOAD [[LOAD]](p1) :: (load (<32 x s32>) from %ir.ptr, addrspace 1)
-  ; CHECK:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 128
-  ; CHECK:   [[PTR_ADD:%[0-9]+]]:_(p1) = G_PTR_ADD [[LOAD]], [[C]](s64)
-  ; CHECK:   [[LOAD2:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD]](p1) :: (load (s32) from %ir.ptr + 128, align 128, addrspace 1)
-  ; CHECK:   G_STORE [[LOAD1]](<32 x s32>), [[COPY]](p5) :: (store (<32 x s32>), addrspace 5)
-  ; CHECK:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 128
-  ; CHECK:   [[PTR_ADD1:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY]], [[C1]](s32)
-  ; CHECK:   G_STORE [[LOAD2]](s32), [[PTR_ADD1]](p5) :: (store (s32), align 128, addrspace 5)
-  ; CHECK:   [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-  ; CHECK:   S_SETPC_B64_return [[COPY2]]
+  ; CHECK-NEXT:   liveins: $vgpr0
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(p5) = COPY $vgpr0
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `{ <32 x i32>, i32 } addrspace(1)* addrspace(4)* undef`, addrspace 4)
+  ; CHECK-NEXT:   [[LOAD1:%[0-9]+]]:_(<32 x s32>) = G_LOAD [[LOAD]](p1) :: (load (<32 x s32>) from %ir.ptr, addrspace 1)
+  ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 128
+  ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p1) = G_PTR_ADD [[LOAD]], [[C]](s64)
+  ; CHECK-NEXT:   [[LOAD2:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD]](p1) :: (load (s32) from %ir.ptr + 128, align 128, addrspace 1)
+  ; CHECK-NEXT:   G_STORE [[LOAD1]](<32 x s32>), [[COPY]](p5) :: (store (<32 x s32>), addrspace 5)
+  ; CHECK-NEXT:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 128
+  ; CHECK-NEXT:   [[PTR_ADD1:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY]], [[C1]](s32)
+  ; CHECK-NEXT:   G_STORE [[LOAD2]](s32), [[PTR_ADD1]](p5) :: (store (s32), align 128, addrspace 5)
+  ; CHECK-NEXT:   SI_RETURN
   %ptr = load volatile { <32 x i32>, i32 } addrspace(1)*, { <32 x i32>, i32 } addrspace(1)* addrspace(4)* undef
   %val = load { <32 x i32>, i32 }, { <32 x i32>, i32 } addrspace(1)* %ptr
   ret { <32 x i32>, i32 }%val
 }
 
 define { i32, <32 x i32> } @struct_i32_v32i32_func_void() #0 {
   ; CHECK-LABEL: name: struct_i32_v32i32_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(p5) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `{ i32, <32 x i32> } addrspace(1)* addrspace(4)* undef`, addrspace 4)
-  ; CHECK:   [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[LOAD]](p1) :: (load (s32) from %ir.ptr, align 128, addrspace 1)
-  ; CHECK:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 128
-  ; CHECK:   [[PTR_ADD:%[0-9]+]]:_(p1) = G_PTR_ADD [[LOAD]], [[C]](s64)
-  ; CHECK:   [[LOAD2:%[0-9]+]]:_(<32 x s32>) = G_LOAD [[PTR_ADD]](p1) :: (load (<32 x s32>) from %ir.ptr + 128, addrspace 1)
-  ; CHECK:   G_STORE [[LOAD1]](s32), [[COPY]](p5) :: (store (s32), align 128, addrspace 5)
-  ; CHECK:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 128
-  ; CHECK:   [[PTR_ADD1:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY]], [[C1]](s32)
-  ; CHECK:   G_STORE [[LOAD2]](<32 x s32>), [[PTR_ADD1]](p5) :: (store (<32 x s32>), addrspace 5)
-  ; CHECK:   [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-  ; CHECK:   S_SETPC_B64_return [[COPY2]]
+  ; CHECK-NEXT:   liveins: $vgpr0
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(p5) = COPY $vgpr0
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (volatile load (p1) from `{ i32, <32 x i32> } addrspace(1)* addrspace(4)* undef`, addrspace 4)
+  ; CHECK-NEXT:   [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[LOAD]](p1) :: (load (s32) from %ir.ptr, align 128, addrspace 1)
+  ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 128
+  ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p1) = G_PTR_ADD [[LOAD]], [[C]](s64)
+  ; CHECK-NEXT:   [[LOAD2:%[0-9]+]]:_(<32 x s32>) = G_LOAD [[PTR_ADD]](p1) :: (load (<32 x s32>) from %ir.ptr + 128, addrspace 1)
+  ; CHECK-NEXT:   G_STORE [[LOAD1]](s32), [[COPY]](p5) :: (store (s32), align 128, addrspace 5)
+  ; CHECK-NEXT:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 128
+  ; CHECK-NEXT:   [[PTR_ADD1:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY]], [[C1]](s32)
+  ; CHECK-NEXT:   G_STORE [[LOAD2]](<32 x s32>), [[PTR_ADD1]](p5) :: (store (<32 x s32>), addrspace 5)
+  ; CHECK-NEXT:   SI_RETURN
   %ptr = load volatile { i32, <32 x i32> } addrspace(1)*, { i32, <32 x i32> } addrspace(1)* addrspace(4)* undef
   %val = load { i32, <32 x i32> }, { i32, <32 x i32> } addrspace(1)* %ptr
   ret { i32, <32 x i32> }%val
 }
 
 ; Make sure the last struct component is returned in v3, not v4.
 define { <3 x i32>, i32 } @v3i32_struct_func_void_wasted_reg() #0 {
   ; CHECK-LABEL: name: v3i32_struct_func_void_wasted_reg
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p3) = G_IMPLICIT_DEF
-  ; CHECK:   [[DEF1:%[0-9]+]]:_(<3 x s32>) = G_IMPLICIT_DEF
-  ; CHECK:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
-  ; CHECK:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 1
-  ; CHECK:   [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 2
-  ; CHECK:   [[DEF2:%[0-9]+]]:_(s32) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s32) = G_LOAD [[DEF]](p3) :: (volatile load (s32) from `i32 addrspace(3)* undef`, addrspace 3)
-  ; CHECK:   [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[DEF]](p3) :: (volatile load (s32) from `i32 addrspace(3)* undef`, addrspace 3)
-  ; CHECK:   [[LOAD2:%[0-9]+]]:_(s32) = G_LOAD [[DEF]](p3) :: (volatile load (s32) from `i32 addrspace(3)* undef`, addrspace 3)
-  ; CHECK:   [[LOAD3:%[0-9]+]]:_(s32) = G_LOAD [[DEF]](p3) :: (volatile load (s32) from `i32 addrspace(3)* undef`, addrspace 3)
-  ; CHECK:   [[IVEC:%[0-9]+]]:_(<3 x s32>) = G_INSERT_VECTOR_ELT [[DEF1]], [[LOAD]](s32), [[C]](s32)
-  ; CHECK:   [[IVEC1:%[0-9]+]]:_(<3 x s32>) = G_INSERT_VECTOR_ELT [[IVEC]], [[LOAD1]](s32), [[C1]](s32)
-  ; CHECK:   [[IVEC2:%[0-9]+]]:_(<3 x s32>) = G_INSERT_VECTOR_ELT [[IVEC1]], [[LOAD2]](s32), [[C2]](s32)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[IVEC2]](<3 x s32>)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   $vgpr2 = COPY [[UV2]](s32)
-  ; CHECK:   $vgpr3 = COPY [[LOAD3]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p3) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[DEF1:%[0-9]+]]:_(<3 x s32>) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
+  ; CHECK-NEXT:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 1
+  ; CHECK-NEXT:   [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 2
+  ; CHECK-NEXT:   [[DEF2:%[0-9]+]]:_(s32) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s32) = G_LOAD [[DEF]](p3) :: (volatile load (s32) from `i32 addrspace(3)* undef`, addrspace 3)
+  ; CHECK-NEXT:   [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[DEF]](p3) :: (volatile load (s32) from `i32 addrspace(3)* undef`, addrspace 3)
+  ; CHECK-NEXT:   [[LOAD2:%[0-9]+]]:_(s32) = G_LOAD [[DEF]](p3) :: (volatile load (s32) from `i32 addrspace(3)* undef`, addrspace 3)
+  ; CHECK-NEXT:   [[LOAD3:%[0-9]+]]:_(s32) = G_LOAD [[DEF]](p3) :: (volatile load (s32) from `i32 addrspace(3)* undef`, addrspace 3)
+  ; CHECK-NEXT:   [[IVEC:%[0-9]+]]:_(<3 x s32>) = G_INSERT_VECTOR_ELT [[DEF1]], [[LOAD]](s32), [[C]](s32)
+  ; CHECK-NEXT:   [[IVEC1:%[0-9]+]]:_(<3 x s32>) = G_INSERT_VECTOR_ELT [[IVEC]], [[LOAD1]](s32), [[C1]](s32)
+  ; CHECK-NEXT:   [[IVEC2:%[0-9]+]]:_(<3 x s32>) = G_INSERT_VECTOR_ELT [[IVEC1]], [[LOAD2]](s32), [[C2]](s32)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[IVEC2]](<3 x s32>)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
+  ; CHECK-NEXT:   $vgpr3 = COPY [[LOAD3]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
   %load0 = load volatile i32, i32 addrspace(3)* undef
   %load1 = load volatile i32, i32 addrspace(3)* undef
   %load2 = load volatile i32, i32 addrspace(3)* undef
   %load3 = load volatile i32, i32 addrspace(3)* undef
 
   %insert.0 = insertelement <3 x i32> undef, i32 %load0, i32 0
   %insert.1 = insertelement <3 x i32> %insert.0, i32 %load1, i32 1
   %insert.2 = insertelement <3 x i32> %insert.1, i32 %load2, i32 2
   %insert.3 = insertvalue { <3 x i32>, i32 } undef, <3 x i32> %insert.2, 0
   %insert.4 = insertvalue { <3 x i32>, i32 } %insert.3, i32 %load3, 1
   ret { <3 x i32>, i32 } %insert.4
 }
 
 define { <3 x float>, i32 } @v3f32_struct_func_void_wasted_reg() #0 {
   ; CHECK-LABEL: name: v3f32_struct_func_void_wasted_reg
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p3) = G_IMPLICIT_DEF
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(p3) = COPY [[DEF]](p3)
-  ; CHECK:   [[DEF1:%[0-9]+]]:_(<3 x s32>) = G_IMPLICIT_DEF
-  ; CHECK:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
-  ; CHECK:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 1
-  ; CHECK:   [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 2
-  ; CHECK:   [[DEF2:%[0-9]+]]:_(s32) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s32) = G_LOAD [[DEF]](p3) :: (volatile load (s32) from `float addrspace(3)* undef`, addrspace 3)
-  ; CHECK:   [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[DEF]](p3) :: (volatile load (s32) from `float addrspace(3)* undef`, addrspace 3)
-  ; CHECK:   [[LOAD2:%[0-9]+]]:_(s32) = G_LOAD [[DEF]](p3) :: (volatile load (s32) from `float addrspace(3)* undef`, addrspace 3)
-  ; CHECK:   [[LOAD3:%[0-9]+]]:_(s32) = G_LOAD [[COPY1]](p3) :: (volatile load (s32) from `i32 addrspace(3)* undef`, addrspace 3)
-  ; CHECK:   [[IVEC:%[0-9]+]]:_(<3 x s32>) = G_INSERT_VECTOR_ELT [[DEF1]], [[LOAD]](s32), [[C]](s32)
-  ; CHECK:   [[IVEC1:%[0-9]+]]:_(<3 x s32>) = G_INSERT_VECTOR_ELT [[IVEC]], [[LOAD1]](s32), [[C1]](s32)
-  ; CHECK:   [[IVEC2:%[0-9]+]]:_(<3 x s32>) = G_INSERT_VECTOR_ELT [[IVEC1]], [[LOAD2]](s32), [[C2]](s32)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[IVEC2]](<3 x s32>)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   $vgpr2 = COPY [[UV2]](s32)
-  ; CHECK:   $vgpr3 = COPY [[LOAD3]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY2]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p3) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(p3) = COPY [[DEF]](p3)
+  ; CHECK-NEXT:   [[DEF1:%[0-9]+]]:_(<3 x s32>) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
+  ; CHECK-NEXT:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 1
+  ; CHECK-NEXT:   [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 2
+  ; CHECK-NEXT:   [[DEF2:%[0-9]+]]:_(s32) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s32) = G_LOAD [[DEF]](p3) :: (volatile load (s32) from `float addrspace(3)* undef`, addrspace 3)
+  ; CHECK-NEXT:   [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[DEF]](p3) :: (volatile load (s32) from `float addrspace(3)* undef`, addrspace 3)
+  ; CHECK-NEXT:   [[LOAD2:%[0-9]+]]:_(s32) = G_LOAD [[DEF]](p3) :: (volatile load (s32) from `float addrspace(3)* undef`, addrspace 3)
+  ; CHECK-NEXT:   [[LOAD3:%[0-9]+]]:_(s32) = G_LOAD [[COPY]](p3) :: (volatile load (s32) from `i32 addrspace(3)* undef`, addrspace 3)
+  ; CHECK-NEXT:   [[IVEC:%[0-9]+]]:_(<3 x s32>) = G_INSERT_VECTOR_ELT [[DEF1]], [[LOAD]](s32), [[C]](s32)
+  ; CHECK-NEXT:   [[IVEC1:%[0-9]+]]:_(<3 x s32>) = G_INSERT_VECTOR_ELT [[IVEC]], [[LOAD1]](s32), [[C1]](s32)
+  ; CHECK-NEXT:   [[IVEC2:%[0-9]+]]:_(<3 x s32>) = G_INSERT_VECTOR_ELT [[IVEC1]], [[LOAD2]](s32), [[C2]](s32)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[IVEC2]](<3 x s32>)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
+  ; CHECK-NEXT:   $vgpr3 = COPY [[LOAD3]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
   %load0 = load volatile float, float addrspace(3)* undef
   %load1 = load volatile float, float addrspace(3)* undef
   %load2 = load volatile float, float addrspace(3)* undef
   %load3 = load volatile i32, i32 addrspace(3)* undef
 
   %insert.0 = insertelement <3 x float> undef, float %load0, i32 0
   %insert.1 = insertelement <3 x float> %insert.0, float %load1, i32 1
   %insert.2 = insertelement <3 x float> %insert.1, float %load2, i32 2
   %insert.3 = insertvalue { <3 x float>, i32 } undef, <3 x float> %insert.2, 0
   %insert.4 = insertvalue { <3 x float>, i32 } %insert.3, i32 %load3, 1
   ret { <3 x float>, i32 } %insert.4
 }
 
 define void @void_func_sret_max_known_zero_bits(i8 addrspace(5)* sret(i8) %arg0) #0 {
   ; CHECK-LABEL: name: void_func_sret_max_known_zero_bits
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(p5) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
-  ; CHECK:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 17
-  ; CHECK:   [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 18
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p3) = G_IMPLICIT_DEF
-  ; CHECK:   [[PTRTOINT:%[0-9]+]]:_(s32) = G_PTRTOINT [[COPY]](p5)
-  ; CHECK:   [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[PTRTOINT]], [[C]](s32)
-  ; CHECK:   [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[PTRTOINT]], [[C1]](s32)
-  ; CHECK:   [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[PTRTOINT]], [[C2]](s32)
-  ; CHECK:   G_STORE [[LSHR]](s32), [[DEF]](p3) :: (volatile store (s32) into `i32 addrspace(3)* undef`, addrspace 3)
-  ; CHECK:   G_STORE [[LSHR1]](s32), [[DEF]](p3) :: (volatile store (s32) into `i32 addrspace(3)* undef`, addrspace 3)
-  ; CHECK:   G_STORE [[LSHR2]](s32), [[DEF]](p3) :: (volatile store (s32) into `i32 addrspace(3)* undef`, addrspace 3)
-  ; CHECK:   [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-  ; CHECK:   S_SETPC_B64_return [[COPY2]]
+  ; CHECK-NEXT:   liveins: $vgpr0
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(p5) = COPY $vgpr0
+  ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
+  ; CHECK-NEXT:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 17
+  ; CHECK-NEXT:   [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 18
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p3) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[PTRTOINT:%[0-9]+]]:_(s32) = G_PTRTOINT [[COPY]](p5)
+  ; CHECK-NEXT:   [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[PTRTOINT]], [[C]](s32)
+  ; CHECK-NEXT:   [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[PTRTOINT]], [[C1]](s32)
+  ; CHECK-NEXT:   [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[PTRTOINT]], [[C2]](s32)
+  ; CHECK-NEXT:   G_STORE [[LSHR]](s32), [[DEF]](p3) :: (volatile store (s32) into `i32 addrspace(3)* undef`, addrspace 3)
+  ; CHECK-NEXT:   G_STORE [[LSHR1]](s32), [[DEF]](p3) :: (volatile store (s32) into `i32 addrspace(3)* undef`, addrspace 3)
+  ; CHECK-NEXT:   G_STORE [[LSHR2]](s32), [[DEF]](p3) :: (volatile store (s32) into `i32 addrspace(3)* undef`, addrspace 3)
+  ; CHECK-NEXT:   SI_RETURN
   %arg0.int = ptrtoint i8 addrspace(5)* %arg0 to i32
 
   %lshr0 = lshr i32 %arg0.int, 16
   %lshr1 = lshr i32 %arg0.int, 17
   %lshr2 = lshr i32 %arg0.int, 18
 
   store volatile i32 %lshr0, i32 addrspace(3)* undef
   store volatile i32 %lshr1, i32 addrspace(3)* undef
   store volatile i32 %lshr2, i32 addrspace(3)* undef
   ret void
 }
 
 define i1022 @i1022_func_void() #0 {
   ; CHECK-LABEL: name: i1022_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s1022) = G_LOAD [[DEF]](p1) :: (load (s1022) from `i1022 addrspace(1)* undef`, align 8, addrspace 1)
-  ; CHECK:   [[ANYEXT:%[0-9]+]]:_(s1024) = G_ANYEXT [[LOAD]](s1022)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32), [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32), [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32), [[UV12:%[0-9]+]]:_(s32), [[UV13:%[0-9]+]]:_(s32), [[UV14:%[0-9]+]]:_(s32), [[UV15:%[0-9]+]]:_(s32), [[UV16:%[0-9]+]]:_(s32), [[UV17:%[0-9]+]]:_(s32), [[UV18:%[0-9]+]]:_(s32), [[UV19:%[0-9]+]]:_(s32), [[UV20:%[0-9]+]]:_(s32), [[UV21:%[0-9]+]]:_(s32), [[UV22:%[0-9]+]]:_(s32), [[UV23:%[0-9]+]]:_(s32), [[UV24:%[0-9]+]]:_(s32), [[UV25:%[0-9]+]]:_(s32), [[UV26:%[0-9]+]]:_(s32), [[UV27:%[0-9]+]]:_(s32), [[UV28:%[0-9]+]]:_(s32), [[UV29:%[0-9]+]]:_(s32), [[UV30:%[0-9]+]]:_(s32), [[UV31:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[ANYEXT]](s1024)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   $vgpr2 = COPY [[UV2]](s32)
-  ; CHECK:   $vgpr3 = COPY [[UV3]](s32)
-  ; CHECK:   $vgpr4 = COPY [[UV4]](s32)
-  ; CHECK:   $vgpr5 = COPY [[UV5]](s32)
-  ; CHECK:   $vgpr6 = COPY [[UV6]](s32)
-  ; CHECK:   $vgpr7 = COPY [[UV7]](s32)
-  ; CHECK:   $vgpr8 = COPY [[UV8]](s32)
-  ; CHECK:   $vgpr9 = COPY [[UV9]](s32)
-  ; CHECK:   $vgpr10 = COPY [[UV10]](s32)
-  ; CHECK:   $vgpr11 = COPY [[UV11]](s32)
-  ; CHECK:   $vgpr12 = COPY [[UV12]](s32)
-  ; CHECK:   $vgpr13 = COPY [[UV13]](s32)
-  ; CHECK:   $vgpr14 = COPY [[UV14]](s32)
-  ; CHECK:   $vgpr15 = COPY [[UV15]](s32)
-  ; CHECK:   $vgpr16 = COPY [[UV16]](s32)
-  ; CHECK:   $vgpr17 = COPY [[UV17]](s32)
-  ; CHECK:   $vgpr18 = COPY [[UV18]](s32)
-  ; CHECK:   $vgpr19 = COPY [[UV19]](s32)
-  ; CHECK:   $vgpr20 = COPY [[UV20]](s32)
-  ; CHECK:   $vgpr21 = COPY [[UV21]](s32)
-  ; CHECK:   $vgpr22 = COPY [[UV22]](s32)
-  ; CHECK:   $vgpr23 = COPY [[UV23]](s32)
-  ; CHECK:   $vgpr24 = COPY [[UV24]](s32)
-  ; CHECK:   $vgpr25 = COPY [[UV25]](s32)
-  ; CHECK:   $vgpr26 = COPY [[UV26]](s32)
-  ; CHECK:   $vgpr27 = COPY [[UV27]](s32)
-  ; CHECK:   $vgpr28 = COPY [[UV28]](s32)
-  ; CHECK:   $vgpr29 = COPY [[UV29]](s32)
-  ; CHECK:   $vgpr30 = COPY [[UV30]](s32)
-  ; CHECK:   $vgpr31 = COPY [[UV31]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15, implicit $vgpr16, implicit $vgpr17, implicit $vgpr18, implicit $vgpr19, implicit $vgpr20, implicit $vgpr21, implicit $vgpr22, implicit $vgpr23, implicit $vgpr24, implicit $vgpr25, implicit $vgpr26, implicit $vgpr27, implicit $vgpr28, implicit $vgpr29, implicit $vgpr30, implicit $vgpr31
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s1022) = G_LOAD [[DEF]](p1) :: (load (s1022) from `i1022 addrspace(1)* undef`, align 8, addrspace 1)
+  ; CHECK-NEXT:   [[ANYEXT:%[0-9]+]]:_(s1024) = G_ANYEXT [[LOAD]](s1022)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32), [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32), [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32), [[UV12:%[0-9]+]]:_(s32), [[UV13:%[0-9]+]]:_(s32), [[UV14:%[0-9]+]]:_(s32), [[UV15:%[0-9]+]]:_(s32), [[UV16:%[0-9]+]]:_(s32), [[UV17:%[0-9]+]]:_(s32), [[UV18:%[0-9]+]]:_(s32), [[UV19:%[0-9]+]]:_(s32), [[UV20:%[0-9]+]]:_(s32), [[UV21:%[0-9]+]]:_(s32), [[UV22:%[0-9]+]]:_(s32), [[UV23:%[0-9]+]]:_(s32), [[UV24:%[0-9]+]]:_(s32), [[UV25:%[0-9]+]]:_(s32), [[UV26:%[0-9]+]]:_(s32), [[UV27:%[0-9]+]]:_(s32), [[UV28:%[0-9]+]]:_(s32), [[UV29:%[0-9]+]]:_(s32), [[UV30:%[0-9]+]]:_(s32), [[UV31:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[ANYEXT]](s1024)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
+  ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
+  ; CHECK-NEXT:   $vgpr4 = COPY [[UV4]](s32)
+  ; CHECK-NEXT:   $vgpr5 = COPY [[UV5]](s32)
+  ; CHECK-NEXT:   $vgpr6 = COPY [[UV6]](s32)
+  ; CHECK-NEXT:   $vgpr7 = COPY [[UV7]](s32)
+  ; CHECK-NEXT:   $vgpr8 = COPY [[UV8]](s32)
+  ; CHECK-NEXT:   $vgpr9 = COPY [[UV9]](s32)
+  ; CHECK-NEXT:   $vgpr10 = COPY [[UV10]](s32)
+  ; CHECK-NEXT:   $vgpr11 = COPY [[UV11]](s32)
+  ; CHECK-NEXT:   $vgpr12 = COPY [[UV12]](s32)
+  ; CHECK-NEXT:   $vgpr13 = COPY [[UV13]](s32)
+  ; CHECK-NEXT:   $vgpr14 = COPY [[UV14]](s32)
+  ; CHECK-NEXT:   $vgpr15 = COPY [[UV15]](s32)
+  ; CHECK-NEXT:   $vgpr16 = COPY [[UV16]](s32)
+  ; CHECK-NEXT:   $vgpr17 = COPY [[UV17]](s32)
+  ; CHECK-NEXT:   $vgpr18 = COPY [[UV18]](s32)
+  ; CHECK-NEXT:   $vgpr19 = COPY [[UV19]](s32)
+  ; CHECK-NEXT:   $vgpr20 = COPY [[UV20]](s32)
+  ; CHECK-NEXT:   $vgpr21 = COPY [[UV21]](s32)
+  ; CHECK-NEXT:   $vgpr22 = COPY [[UV22]](s32)
+  ; CHECK-NEXT:   $vgpr23 = COPY [[UV23]](s32)
+  ; CHECK-NEXT:   $vgpr24 = COPY [[UV24]](s32)
+  ; CHECK-NEXT:   $vgpr25 = COPY [[UV25]](s32)
+  ; CHECK-NEXT:   $vgpr26 = COPY [[UV26]](s32)
+  ; CHECK-NEXT:   $vgpr27 = COPY [[UV27]](s32)
+  ; CHECK-NEXT:   $vgpr28 = COPY [[UV28]](s32)
+  ; CHECK-NEXT:   $vgpr29 = COPY [[UV29]](s32)
+  ; CHECK-NEXT:   $vgpr30 = COPY [[UV30]](s32)
+  ; CHECK-NEXT:   $vgpr31 = COPY [[UV31]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15, implicit $vgpr16, implicit $vgpr17, implicit $vgpr18, implicit $vgpr19, implicit $vgpr20, implicit $vgpr21, implicit $vgpr22, implicit $vgpr23, implicit $vgpr24, implicit $vgpr25, implicit $vgpr26, implicit $vgpr27, implicit $vgpr28, implicit $vgpr29, implicit $vgpr30, implicit $vgpr31
   %val = load i1022, i1022 addrspace(1)* undef
   ret i1022 %val
 }
 
 define signext i1022 @i1022_signext_func_void() #0 {
   ; CHECK-LABEL: name: i1022_signext_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s1022) = G_LOAD [[DEF]](p1) :: (load (s1022) from `i1022 addrspace(1)* undef`, align 8, addrspace 1)
-  ; CHECK:   [[SEXT:%[0-9]+]]:_(s1024) = G_SEXT [[LOAD]](s1022)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32), [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32), [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32), [[UV12:%[0-9]+]]:_(s32), [[UV13:%[0-9]+]]:_(s32), [[UV14:%[0-9]+]]:_(s32), [[UV15:%[0-9]+]]:_(s32), [[UV16:%[0-9]+]]:_(s32), [[UV17:%[0-9]+]]:_(s32), [[UV18:%[0-9]+]]:_(s32), [[UV19:%[0-9]+]]:_(s32), [[UV20:%[0-9]+]]:_(s32), [[UV21:%[0-9]+]]:_(s32), [[UV22:%[0-9]+]]:_(s32), [[UV23:%[0-9]+]]:_(s32), [[UV24:%[0-9]+]]:_(s32), [[UV25:%[0-9]+]]:_(s32), [[UV26:%[0-9]+]]:_(s32), [[UV27:%[0-9]+]]:_(s32), [[UV28:%[0-9]+]]:_(s32), [[UV29:%[0-9]+]]:_(s32), [[UV30:%[0-9]+]]:_(s32), [[UV31:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[SEXT]](s1024)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   $vgpr2 = COPY [[UV2]](s32)
-  ; CHECK:   $vgpr3 = COPY [[UV3]](s32)
-  ; CHECK:   $vgpr4 = COPY [[UV4]](s32)
-  ; CHECK:   $vgpr5 = COPY [[UV5]](s32)
-  ; CHECK:   $vgpr6 = COPY [[UV6]](s32)
-  ; CHECK:   $vgpr7 = COPY [[UV7]](s32)
-  ; CHECK:   $vgpr8 = COPY [[UV8]](s32)
-  ; CHECK:   $vgpr9 = COPY [[UV9]](s32)
-  ; CHECK:   $vgpr10 = COPY [[UV10]](s32)
-  ; CHECK:   $vgpr11 = COPY [[UV11]](s32)
-  ; CHECK:   $vgpr12 = COPY [[UV12]](s32)
-  ; CHECK:   $vgpr13 = COPY [[UV13]](s32)
-  ; CHECK:   $vgpr14 = COPY [[UV14]](s32)
-  ; CHECK:   $vgpr15 = COPY [[UV15]](s32)
-  ; CHECK:   $vgpr16 = COPY [[UV16]](s32)
-  ; CHECK:   $vgpr17 = COPY [[UV17]](s32)
-  ; CHECK:   $vgpr18 = COPY [[UV18]](s32)
-  ; CHECK:   $vgpr19 = COPY [[UV19]](s32)
-  ; CHECK:   $vgpr20 = COPY [[UV20]](s32)
-  ; CHECK:   $vgpr21 = COPY [[UV21]](s32)
-  ; CHECK:   $vgpr22 = COPY [[UV22]](s32)
-  ; CHECK:   $vgpr23 = COPY [[UV23]](s32)
-  ; CHECK:   $vgpr24 = COPY [[UV24]](s32)
-  ; CHECK:   $vgpr25 = COPY [[UV25]](s32)
-  ; CHECK:   $vgpr26 = COPY [[UV26]](s32)
-  ; CHECK:   $vgpr27 = COPY [[UV27]](s32)
-  ; CHECK:   $vgpr28 = COPY [[UV28]](s32)
-  ; CHECK:   $vgpr29 = COPY [[UV29]](s32)
-  ; CHECK:   $vgpr30 = COPY [[UV30]](s32)
-  ; CHECK:   $vgpr31 = COPY [[UV31]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15, implicit $vgpr16, implicit $vgpr17, implicit $vgpr18, implicit $vgpr19, implicit $vgpr20, implicit $vgpr21, implicit $vgpr22, implicit $vgpr23, implicit $vgpr24, implicit $vgpr25, implicit $vgpr26, implicit $vgpr27, implicit $vgpr28, implicit $vgpr29, implicit $vgpr30, implicit $vgpr31
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s1022) = G_LOAD [[DEF]](p1) :: (load (s1022) from `i1022 addrspace(1)* undef`, align 8, addrspace 1)
+  ; CHECK-NEXT:   [[SEXT:%[0-9]+]]:_(s1024) = G_SEXT [[LOAD]](s1022)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32), [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32), [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32), [[UV12:%[0-9]+]]:_(s32), [[UV13:%[0-9]+]]:_(s32), [[UV14:%[0-9]+]]:_(s32), [[UV15:%[0-9]+]]:_(s32), [[UV16:%[0-9]+]]:_(s32), [[UV17:%[0-9]+]]:_(s32), [[UV18:%[0-9]+]]:_(s32), [[UV19:%[0-9]+]]:_(s32), [[UV20:%[0-9]+]]:_(s32), [[UV21:%[0-9]+]]:_(s32), [[UV22:%[0-9]+]]:_(s32), [[UV23:%[0-9]+]]:_(s32), [[UV24:%[0-9]+]]:_(s32), [[UV25:%[0-9]+]]:_(s32), [[UV26:%[0-9]+]]:_(s32), [[UV27:%[0-9]+]]:_(s32), [[UV28:%[0-9]+]]:_(s32), [[UV29:%[0-9]+]]:_(s32), [[UV30:%[0-9]+]]:_(s32), [[UV31:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[SEXT]](s1024)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
+  ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
+  ; CHECK-NEXT:   $vgpr4 = COPY [[UV4]](s32)
+  ; CHECK-NEXT:   $vgpr5 = COPY [[UV5]](s32)
+  ; CHECK-NEXT:   $vgpr6 = COPY [[UV6]](s32)
+  ; CHECK-NEXT:   $vgpr7 = COPY [[UV7]](s32)
+  ; CHECK-NEXT:   $vgpr8 = COPY [[UV8]](s32)
+  ; CHECK-NEXT:   $vgpr9 = COPY [[UV9]](s32)
+  ; CHECK-NEXT:   $vgpr10 = COPY [[UV10]](s32)
+  ; CHECK-NEXT:   $vgpr11 = COPY [[UV11]](s32)
+  ; CHECK-NEXT:   $vgpr12 = COPY [[UV12]](s32)
+  ; CHECK-NEXT:   $vgpr13 = COPY [[UV13]](s32)
+  ; CHECK-NEXT:   $vgpr14 = COPY [[UV14]](s32)
+  ; CHECK-NEXT:   $vgpr15 = COPY [[UV15]](s32)
+  ; CHECK-NEXT:   $vgpr16 = COPY [[UV16]](s32)
+  ; CHECK-NEXT:   $vgpr17 = COPY [[UV17]](s32)
+  ; CHECK-NEXT:   $vgpr18 = COPY [[UV18]](s32)
+  ; CHECK-NEXT:   $vgpr19 = COPY [[UV19]](s32)
+  ; CHECK-NEXT:   $vgpr20 = COPY [[UV20]](s32)
+  ; CHECK-NEXT:   $vgpr21 = COPY [[UV21]](s32)
+  ; CHECK-NEXT:   $vgpr22 = COPY [[UV22]](s32)
+  ; CHECK-NEXT:   $vgpr23 = COPY [[UV23]](s32)
+  ; CHECK-NEXT:   $vgpr24 = COPY [[UV24]](s32)
+  ; CHECK-NEXT:   $vgpr25 = COPY [[UV25]](s32)
+  ; CHECK-NEXT:   $vgpr26 = COPY [[UV26]](s32)
+  ; CHECK-NEXT:   $vgpr27 = COPY [[UV27]](s32)
+  ; CHECK-NEXT:   $vgpr28 = COPY [[UV28]](s32)
+  ; CHECK-NEXT:   $vgpr29 = COPY [[UV29]](s32)
+  ; CHECK-NEXT:   $vgpr30 = COPY [[UV30]](s32)
+  ; CHECK-NEXT:   $vgpr31 = COPY [[UV31]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15, implicit $vgpr16, implicit $vgpr17, implicit $vgpr18, implicit $vgpr19, implicit $vgpr20, implicit $vgpr21, implicit $vgpr22, implicit $vgpr23, implicit $vgpr24, implicit $vgpr25, implicit $vgpr26, implicit $vgpr27, implicit $vgpr28, implicit $vgpr29, implicit $vgpr30, implicit $vgpr31
   %val = load i1022, i1022 addrspace(1)* undef
   ret i1022 %val
 }
 
 define zeroext i1022 @i1022_zeroext_func_void() #0 {
   ; CHECK-LABEL: name: i1022_zeroext_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(s1022) = G_LOAD [[DEF]](p1) :: (load (s1022) from `i1022 addrspace(1)* undef`, align 8, addrspace 1)
-  ; CHECK:   [[ZEXT:%[0-9]+]]:_(s1024) = G_ZEXT [[LOAD]](s1022)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32), [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32), [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32), [[UV12:%[0-9]+]]:_(s32), [[UV13:%[0-9]+]]:_(s32), [[UV14:%[0-9]+]]:_(s32), [[UV15:%[0-9]+]]:_(s32), [[UV16:%[0-9]+]]:_(s32), [[UV17:%[0-9]+]]:_(s32), [[UV18:%[0-9]+]]:_(s32), [[UV19:%[0-9]+]]:_(s32), [[UV20:%[0-9]+]]:_(s32), [[UV21:%[0-9]+]]:_(s32), [[UV22:%[0-9]+]]:_(s32), [[UV23:%[0-9]+]]:_(s32), [[UV24:%[0-9]+]]:_(s32), [[UV25:%[0-9]+]]:_(s32), [[UV26:%[0-9]+]]:_(s32), [[UV27:%[0-9]+]]:_(s32), [[UV28:%[0-9]+]]:_(s32), [[UV29:%[0-9]+]]:_(s32), [[UV30:%[0-9]+]]:_(s32), [[UV31:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[ZEXT]](s1024)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   $vgpr2 = COPY [[UV2]](s32)
-  ; CHECK:   $vgpr3 = COPY [[UV3]](s32)
-  ; CHECK:   $vgpr4 = COPY [[UV4]](s32)
-  ; CHECK:   $vgpr5 = COPY [[UV5]](s32)
-  ; CHECK:   $vgpr6 = COPY [[UV6]](s32)
-  ; CHECK:   $vgpr7 = COPY [[UV7]](s32)
-  ; CHECK:   $vgpr8 = COPY [[UV8]](s32)
-  ; CHECK:   $vgpr9 = COPY [[UV9]](s32)
-  ; CHECK:   $vgpr10 = COPY [[UV10]](s32)
-  ; CHECK:   $vgpr11 = COPY [[UV11]](s32)
-  ; CHECK:   $vgpr12 = COPY [[UV12]](s32)
-  ; CHECK:   $vgpr13 = COPY [[UV13]](s32)
-  ; CHECK:   $vgpr14 = COPY [[UV14]](s32)
-  ; CHECK:   $vgpr15 = COPY [[UV15]](s32)
-  ; CHECK:   $vgpr16 = COPY [[UV16]](s32)
-  ; CHECK:   $vgpr17 = COPY [[UV17]](s32)
-  ; CHECK:   $vgpr18 = COPY [[UV18]](s32)
-  ; CHECK:   $vgpr19 = COPY [[UV19]](s32)
-  ; CHECK:   $vgpr20 = COPY [[UV20]](s32)
-  ; CHECK:   $vgpr21 = COPY [[UV21]](s32)
-  ; CHECK:   $vgpr22 = COPY [[UV22]](s32)
-  ; CHECK:   $vgpr23 = COPY [[UV23]](s32)
-  ; CHECK:   $vgpr24 = COPY [[UV24]](s32)
-  ; CHECK:   $vgpr25 = COPY [[UV25]](s32)
-  ; CHECK:   $vgpr26 = COPY [[UV26]](s32)
-  ; CHECK:   $vgpr27 = COPY [[UV27]](s32)
-  ; CHECK:   $vgpr28 = COPY [[UV28]](s32)
-  ; CHECK:   $vgpr29 = COPY [[UV29]](s32)
-  ; CHECK:   $vgpr30 = COPY [[UV30]](s32)
-  ; CHECK:   $vgpr31 = COPY [[UV31]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15, implicit $vgpr16, implicit $vgpr17, implicit $vgpr18, implicit $vgpr19, implicit $vgpr20, implicit $vgpr21, implicit $vgpr22, implicit $vgpr23, implicit $vgpr24, implicit $vgpr25, implicit $vgpr26, implicit $vgpr27, implicit $vgpr28, implicit $vgpr29, implicit $vgpr30, implicit $vgpr31
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s1022) = G_LOAD [[DEF]](p1) :: (load (s1022) from `i1022 addrspace(1)* undef`, align 8, addrspace 1)
+  ; CHECK-NEXT:   [[ZEXT:%[0-9]+]]:_(s1024) = G_ZEXT [[LOAD]](s1022)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32), [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32), [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32), [[UV12:%[0-9]+]]:_(s32), [[UV13:%[0-9]+]]:_(s32), [[UV14:%[0-9]+]]:_(s32), [[UV15:%[0-9]+]]:_(s32), [[UV16:%[0-9]+]]:_(s32), [[UV17:%[0-9]+]]:_(s32), [[UV18:%[0-9]+]]:_(s32), [[UV19:%[0-9]+]]:_(s32), [[UV20:%[0-9]+]]:_(s32), [[UV21:%[0-9]+]]:_(s32), [[UV22:%[0-9]+]]:_(s32), [[UV23:%[0-9]+]]:_(s32), [[UV24:%[0-9]+]]:_(s32), [[UV25:%[0-9]+]]:_(s32), [[UV26:%[0-9]+]]:_(s32), [[UV27:%[0-9]+]]:_(s32), [[UV28:%[0-9]+]]:_(s32), [[UV29:%[0-9]+]]:_(s32), [[UV30:%[0-9]+]]:_(s32), [[UV31:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[ZEXT]](s1024)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
+  ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
+  ; CHECK-NEXT:   $vgpr4 = COPY [[UV4]](s32)
+  ; CHECK-NEXT:   $vgpr5 = COPY [[UV5]](s32)
+  ; CHECK-NEXT:   $vgpr6 = COPY [[UV6]](s32)
+  ; CHECK-NEXT:   $vgpr7 = COPY [[UV7]](s32)
+  ; CHECK-NEXT:   $vgpr8 = COPY [[UV8]](s32)
+  ; CHECK-NEXT:   $vgpr9 = COPY [[UV9]](s32)
+  ; CHECK-NEXT:   $vgpr10 = COPY [[UV10]](s32)
+  ; CHECK-NEXT:   $vgpr11 = COPY [[UV11]](s32)
+  ; CHECK-NEXT:   $vgpr12 = COPY [[UV12]](s32)
+  ; CHECK-NEXT:   $vgpr13 = COPY [[UV13]](s32)
+  ; CHECK-NEXT:   $vgpr14 = COPY [[UV14]](s32)
+  ; CHECK-NEXT:   $vgpr15 = COPY [[UV15]](s32)
+  ; CHECK-NEXT:   $vgpr16 = COPY [[UV16]](s32)
+  ; CHECK-NEXT:   $vgpr17 = COPY [[UV17]](s32)
+  ; CHECK-NEXT:   $vgpr18 = COPY [[UV18]](s32)
+  ; CHECK-NEXT:   $vgpr19 = COPY [[UV19]](s32)
+  ; CHECK-NEXT:   $vgpr20 = COPY [[UV20]](s32)
+  ; CHECK-NEXT:   $vgpr21 = COPY [[UV21]](s32)
+  ; CHECK-NEXT:   $vgpr22 = COPY [[UV22]](s32)
+  ; CHECK-NEXT:   $vgpr23 = COPY [[UV23]](s32)
+  ; CHECK-NEXT:   $vgpr24 = COPY [[UV24]](s32)
+  ; CHECK-NEXT:   $vgpr25 = COPY [[UV25]](s32)
+  ; CHECK-NEXT:   $vgpr26 = COPY [[UV26]](s32)
+  ; CHECK-NEXT:   $vgpr27 = COPY [[UV27]](s32)
+  ; CHECK-NEXT:   $vgpr28 = COPY [[UV28]](s32)
+  ; CHECK-NEXT:   $vgpr29 = COPY [[UV29]](s32)
+  ; CHECK-NEXT:   $vgpr30 = COPY [[UV30]](s32)
+  ; CHECK-NEXT:   $vgpr31 = COPY [[UV31]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15, implicit $vgpr16, implicit $vgpr17, implicit $vgpr18, implicit $vgpr19, implicit $vgpr20, implicit $vgpr21, implicit $vgpr22, implicit $vgpr23, implicit $vgpr24, implicit $vgpr25, implicit $vgpr26, implicit $vgpr27, implicit $vgpr28, implicit $vgpr29, implicit $vgpr30, implicit $vgpr31
   %val = load i1022, i1022 addrspace(1)* undef
   ret i1022 %val
 }
 
 %struct.with.ptrs = type { <32 x i32>, i32 addrspace(3)*, i32 addrspace(1)*, <2 x i8 addrspace(1)*> }
 
 define %struct.with.ptrs @ptr_in_struct_func_void() #0 {
   ; CHECK-LABEL: name: ptr_in_struct_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(p5) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   [[LOAD:%[0-9]+]]:_(<32 x s32>) = G_LOAD [[DEF]](p1) :: (volatile load (<32 x s32>) from `%struct.with.ptrs addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 128
-  ; CHECK:   [[PTR_ADD:%[0-9]+]]:_(p1) = G_PTR_ADD [[DEF]], [[C]](s64)
-  ; CHECK:   [[LOAD1:%[0-9]+]]:_(p3) = G_LOAD [[PTR_ADD]](p1) :: (volatile load (p3) from `%struct.with.ptrs addrspace(1)* undef` + 128, align 128, addrspace 1)
-  ; CHECK:   [[C1:%[0-9]+]]:_(s64) = G_CONSTANT i64 136
-  ; CHECK:   [[PTR_ADD1:%[0-9]+]]:_(p1) = G_PTR_ADD [[DEF]], [[C1]](s64)
-  ; CHECK:   [[LOAD2:%[0-9]+]]:_(p1) = G_LOAD [[PTR_ADD1]](p1) :: (volatile load (p1) from `%struct.with.ptrs addrspace(1)* undef` + 136, addrspace 1)
-  ; CHECK:   [[C2:%[0-9]+]]:_(s64) = G_CONSTANT i64 144
-  ; CHECK:   [[PTR_ADD2:%[0-9]+]]:_(p1) = G_PTR_ADD [[DEF]], [[C2]](s64)
-  ; CHECK:   [[LOAD3:%[0-9]+]]:_(<2 x p1>) = G_LOAD [[PTR_ADD2]](p1) :: (volatile load (<2 x p1>) from `%struct.with.ptrs addrspace(1)* undef` + 144, addrspace 1)
-  ; CHECK:   G_STORE [[LOAD]](<32 x s32>), [[COPY]](p5) :: (store (<32 x s32>), addrspace 5)
-  ; CHECK:   [[C3:%[0-9]+]]:_(s32) = G_CONSTANT i32 128
-  ; CHECK:   [[PTR_ADD3:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY]], [[C3]](s32)
-  ; CHECK:   G_STORE [[LOAD1]](p3), [[PTR_ADD3]](p5) :: (store (p3), align 128, addrspace 5)
-  ; CHECK:   [[C4:%[0-9]+]]:_(s32) = G_CONSTANT i32 136
-  ; CHECK:   [[PTR_ADD4:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY]], [[C4]](s32)
-  ; CHECK:   G_STORE [[LOAD2]](p1), [[PTR_ADD4]](p5) :: (store (p1), addrspace 5)
-  ; CHECK:   [[C5:%[0-9]+]]:_(s32) = G_CONSTANT i32 144
-  ; CHECK:   [[PTR_ADD5:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY]], [[C5]](s32)
-  ; CHECK:   G_STORE [[LOAD3]](<2 x p1>), [[PTR_ADD5]](p5) :: (store (<2 x p1>), addrspace 5)
-  ; CHECK:   [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-  ; CHECK:   S_SETPC_B64_return [[COPY2]]
+  ; CHECK-NEXT:   liveins: $vgpr0
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(p5) = COPY $vgpr0
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(<32 x s32>) = G_LOAD [[DEF]](p1) :: (volatile load (<32 x s32>) from `%struct.with.ptrs addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 128
+  ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p1) = G_PTR_ADD [[DEF]], [[C]](s64)
+  ; CHECK-NEXT:   [[LOAD1:%[0-9]+]]:_(p3) = G_LOAD [[PTR_ADD]](p1) :: (volatile load (p3) from `%struct.with.ptrs addrspace(1)* undef` + 128, align 128, addrspace 1)
+  ; CHECK-NEXT:   [[C1:%[0-9]+]]:_(s64) = G_CONSTANT i64 136
+  ; CHECK-NEXT:   [[PTR_ADD1:%[0-9]+]]:_(p1) = G_PTR_ADD [[DEF]], [[C1]](s64)
+  ; CHECK-NEXT:   [[LOAD2:%[0-9]+]]:_(p1) = G_LOAD [[PTR_ADD1]](p1) :: (volatile load (p1) from `%struct.with.ptrs addrspace(1)* undef` + 136, addrspace 1)
+  ; CHECK-NEXT:   [[C2:%[0-9]+]]:_(s64) = G_CONSTANT i64 144
+  ; CHECK-NEXT:   [[PTR_ADD2:%[0-9]+]]:_(p1) = G_PTR_ADD [[DEF]], [[C2]](s64)
+  ; CHECK-NEXT:   [[LOAD3:%[0-9]+]]:_(<2 x p1>) = G_LOAD [[PTR_ADD2]](p1) :: (volatile load (<2 x p1>) from `%struct.with.ptrs addrspace(1)* undef` + 144, addrspace 1)
+  ; CHECK-NEXT:   G_STORE [[LOAD]](<32 x s32>), [[COPY]](p5) :: (store (<32 x s32>), addrspace 5)
+  ; CHECK-NEXT:   [[C3:%[0-9]+]]:_(s32) = G_CONSTANT i32 128
+  ; CHECK-NEXT:   [[PTR_ADD3:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY]], [[C3]](s32)
+  ; CHECK-NEXT:   G_STORE [[LOAD1]](p3), [[PTR_ADD3]](p5) :: (store (p3), align 128, addrspace 5)
+  ; CHECK-NEXT:   [[C4:%[0-9]+]]:_(s32) = G_CONSTANT i32 136
+  ; CHECK-NEXT:   [[PTR_ADD4:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY]], [[C4]](s32)
+  ; CHECK-NEXT:   G_STORE [[LOAD2]](p1), [[PTR_ADD4]](p5) :: (store (p1), addrspace 5)
+  ; CHECK-NEXT:   [[C5:%[0-9]+]]:_(s32) = G_CONSTANT i32 144
+  ; CHECK-NEXT:   [[PTR_ADD5:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY]], [[C5]](s32)
+  ; CHECK-NEXT:   G_STORE [[LOAD3]](<2 x p1>), [[PTR_ADD5]](p5) :: (store (<2 x p1>), addrspace 5)
+  ; CHECK-NEXT:   SI_RETURN
   %val = load volatile %struct.with.ptrs, %struct.with.ptrs addrspace(1)* undef
   ret %struct.with.ptrs %val
 }
 
 attributes #0 = { nounwind }

llvm/test/CodeGen/AMDGPU/GlobalISel/global-value.ll

 @internal_constant32 = internal addrspace(6) constant i32 9, align 4
 @internal_global = internal addrspace(1) global i32 9, align 4
 @internal_other = internal addrspace(999) global i32 9, align 4
 
 
 define i32 addrspace(4)* @external_constant_got() {
   ; GCN-LABEL: name: external_constant_got
   ; GCN: bb.1 (%ir-block.0):
-  ; GCN:   liveins: $sgpr30_sgpr31
-  ; GCN:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN:   [[SI_PC_ADD_REL_OFFSET:%[0-9]+]]:sreg_64(p4) = SI_PC_ADD_REL_OFFSET target-flags(amdgpu-gotprel32-lo) @external_constant + 4, target-flags(amdgpu-gotprel32-hi) @external_constant + 12, implicit-def $scc
   ; GCN:   [[LOAD:%[0-9]+]]:_(p4) = G_LOAD [[SI_PC_ADD_REL_OFFSET]](p4) :: (dereferenceable invariant load (p4) from got, addrspace 4)
   ; GCN:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD]](p4)
   ; GCN:   $vgpr0 = COPY [[UV]](s32)
   ; GCN:   $vgpr1 = COPY [[UV1]](s32)
-  ; GCN:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; GCN:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1
+  ; GCN:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   ret i32 addrspace(4)* @external_constant
 }
 
 define i32 addrspace(1)* @external_global_got() {
   ; GCN-LABEL: name: external_global_got
   ; GCN: bb.1 (%ir-block.0):
-  ; GCN:   liveins: $sgpr30_sgpr31
-  ; GCN:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN:   [[SI_PC_ADD_REL_OFFSET:%[0-9]+]]:sreg_64(p4) = SI_PC_ADD_REL_OFFSET target-flags(amdgpu-gotprel32-lo) @external_global + 4, target-flags(amdgpu-gotprel32-hi) @external_global + 12, implicit-def $scc
   ; GCN:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[SI_PC_ADD_REL_OFFSET]](p4) :: (dereferenceable invariant load (p1) from got, addrspace 4)
   ; GCN:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD]](p1)
   ; GCN:   $vgpr0 = COPY [[UV]](s32)
   ; GCN:   $vgpr1 = COPY [[UV1]](s32)
-  ; GCN:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; GCN:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1
+  ; GCN:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   ret i32 addrspace(1)* @external_global
 }
 
 define i32 addrspace(999)* @external_other_got() {
   ; GCN-LABEL: name: external_other_got
   ; GCN: bb.1 (%ir-block.0):
-  ; GCN:   liveins: $sgpr30_sgpr31
-  ; GCN:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN:   [[SI_PC_ADD_REL_OFFSET:%[0-9]+]]:sreg_64(p4) = SI_PC_ADD_REL_OFFSET target-flags(amdgpu-gotprel32-lo) @external_other + 4, target-flags(amdgpu-gotprel32-hi) @external_other + 12, implicit-def $scc
   ; GCN:   [[LOAD:%[0-9]+]]:_(p999) = G_LOAD [[SI_PC_ADD_REL_OFFSET]](p4) :: (dereferenceable invariant load (p999) from got, addrspace 4)
   ; GCN:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[LOAD]](p999)
   ; GCN:   $vgpr0 = COPY [[UV]](s32)
   ; GCN:   $vgpr1 = COPY [[UV1]](s32)
-  ; GCN:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; GCN:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1
+  ; GCN:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   ret i32 addrspace(999)* @external_other
 }
 
 define i32 addrspace(4)* @internal_constant_pcrel() {
   ; GCN-LABEL: name: internal_constant_pcrel
   ; GCN: bb.1 (%ir-block.0):
-  ; GCN:   liveins: $sgpr30_sgpr31
-  ; GCN:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN:   [[SI_PC_ADD_REL_OFFSET:%[0-9]+]]:sreg_64(p4) = SI_PC_ADD_REL_OFFSET target-flags(amdgpu-rel32-lo) @internal_constant + 4, target-flags(amdgpu-rel32-hi) @internal_constant + 12, implicit-def $scc
   ; GCN:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[SI_PC_ADD_REL_OFFSET]](p4)
   ; GCN:   $vgpr0 = COPY [[UV]](s32)
   ; GCN:   $vgpr1 = COPY [[UV1]](s32)
-  ; GCN:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; GCN:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1
+  ; GCN:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   ret i32 addrspace(4)* @internal_constant
 }
 
 define i32 addrspace(1)* @internal_global_pcrel() {
   ; GCN-LABEL: name: internal_global_pcrel
   ; GCN: bb.1 (%ir-block.0):
-  ; GCN:   liveins: $sgpr30_sgpr31
-  ; GCN:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN:   [[SI_PC_ADD_REL_OFFSET:%[0-9]+]]:sreg_64(p1) = SI_PC_ADD_REL_OFFSET target-flags(amdgpu-rel32-lo) @internal_global + 4, target-flags(amdgpu-rel32-hi) @internal_global + 12, implicit-def $scc
   ; GCN:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[SI_PC_ADD_REL_OFFSET]](p1)
   ; GCN:   $vgpr0 = COPY [[UV]](s32)
   ; GCN:   $vgpr1 = COPY [[UV1]](s32)
-  ; GCN:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; GCN:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1
+  ; GCN:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   ret i32 addrspace(1)* @internal_global
 }
 
 define i32 addrspace(999)* @internal_other_pcrel() {
   ; GCN-LABEL: name: internal_other_pcrel
   ; GCN: bb.1 (%ir-block.0):
-  ; GCN:   liveins: $sgpr30_sgpr31
-  ; GCN:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN:   [[SI_PC_ADD_REL_OFFSET:%[0-9]+]]:sreg_64(p999) = SI_PC_ADD_REL_OFFSET target-flags(amdgpu-rel32-lo) @internal_other + 4, target-flags(amdgpu-rel32-hi) @internal_other + 12, implicit-def $scc
   ; GCN:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[SI_PC_ADD_REL_OFFSET]](p999)
   ; GCN:   $vgpr0 = COPY [[UV]](s32)
   ; GCN:   $vgpr1 = COPY [[UV1]](s32)
-  ; GCN:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; GCN:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0, implicit $vgpr1
+  ; GCN:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   ret i32 addrspace(999)* @internal_other
 }
 
 define i32 addrspace(6)* @external_constant32_got() {
   ; GCN-LABEL: name: external_constant32_got
   ; GCN: bb.1 (%ir-block.0):
-  ; GCN:   liveins: $sgpr30_sgpr31
-  ; GCN:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN:   [[SI_PC_ADD_REL_OFFSET:%[0-9]+]]:sreg_64(p4) = SI_PC_ADD_REL_OFFSET target-flags(amdgpu-gotprel32-lo) @external_constant32 + 4, target-flags(amdgpu-gotprel32-hi) @external_constant32 + 12, implicit-def $scc
   ; GCN:   [[LOAD:%[0-9]+]]:_(p4) = G_LOAD [[SI_PC_ADD_REL_OFFSET]](p4) :: (dereferenceable invariant load (p4) from got, addrspace 4)
   ; GCN:   [[EXTRACT:%[0-9]+]]:_(p6) = G_EXTRACT [[LOAD]](p4), 0
   ; GCN:   $vgpr0 = COPY [[EXTRACT]](p6)
-  ; GCN:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; GCN:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0
+  ; GCN:   SI_RETURN implicit $vgpr0
   ret i32 addrspace(6)* @external_constant32
 }
 
 define i32 addrspace(6)* @internal_constant32_pcrel() {
   ; GCN-LABEL: name: internal_constant32_pcrel
   ; GCN: bb.1 (%ir-block.0):
-  ; GCN:   liveins: $sgpr30_sgpr31
-  ; GCN:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN:   [[SI_PC_ADD_REL_OFFSET:%[0-9]+]]:sreg_64(p4) = SI_PC_ADD_REL_OFFSET target-flags(amdgpu-rel32-lo) @internal_constant32 + 4, target-flags(amdgpu-rel32-hi) @internal_constant32 + 12, implicit-def $scc
   ; GCN:   [[EXTRACT:%[0-9]+]]:_(p6) = G_EXTRACT [[SI_PC_ADD_REL_OFFSET]](p4), 0
   ; GCN:   $vgpr0 = COPY [[EXTRACT]](p6)
-  ; GCN:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; GCN:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0
+  ; GCN:   SI_RETURN implicit $vgpr0
   ret i32 addrspace(6)* @internal_constant32
 }

llvm/test/CodeGen/AMDGPU/GlobalISel/irtranslator-atomicrmw.ll

 ; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
 ; RUN: llc -global-isel -march=amdgcn -mcpu=fiji -O0 -stop-after=irtranslator -o - %s | FileCheck %s
 
 define float @test_atomicrmw_fadd(float addrspace(3)* %addr) {
   ; CHECK-LABEL: name: test_atomicrmw_fadd
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $vgpr0, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(p3) = COPY $vgpr0
-  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00
   ; CHECK-NEXT:   [[ATOMICRMW_FADD:%[0-9]+]]:_(s32) = G_ATOMICRMW_FADD [[COPY]](p3), [[C]] :: (load store seq_cst (s32) on %ir.addr, addrspace 3)
   ; CHECK-NEXT:   $vgpr0 = COPY [[ATOMICRMW_FADD]](s32)
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY2]], implicit $vgpr0
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %oldval = atomicrmw fadd float addrspace(3)* %addr, float 1.0 seq_cst
   ret float %oldval
 }
 
 define float @test_atomicrmw_fsub(float addrspace(3)* %addr) {
   ; CHECK-LABEL: name: test_atomicrmw_fsub
   ; CHECK: bb.1 (%ir-block.0):
   ; CHECK-NEXT:   successors: %bb.2(0x80000000)
-  ; CHECK-NEXT:   liveins: $vgpr0, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(p3) = COPY $vgpr0
-  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:_(s64) = G_CONSTANT i64 0
   ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(s32) = G_LOAD [[COPY]](p3) :: (load (s32) from %ir.addr, addrspace 3)
   ; CHECK-NEXT:   G_BR %bb.2
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.2.atomicrmw.start:
   ; CHECK-NEXT:   successors: %bb.3(0x40000000), %bb.2(0x40000000)
   ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   [[PHI:%[0-9]+]]:_(s64) = G_PHI %9(s64), %bb.2, [[C1]](s64), %bb.1
-  ; CHECK-NEXT:   [[PHI1:%[0-9]+]]:_(s32) = G_PHI [[LOAD]](s32), %bb.1, %7(s32), %bb.2
+  ; CHECK-NEXT:   [[PHI:%[0-9]+]]:_(s64) = G_PHI %8(s64), %bb.2, [[C1]](s64), %bb.1
+  ; CHECK-NEXT:   [[PHI1:%[0-9]+]]:_(s32) = G_PHI [[LOAD]](s32), %bb.1, %6(s32), %bb.2
   ; CHECK-NEXT:   [[FSUB:%[0-9]+]]:_(s32) = G_FSUB [[PHI1]], [[C]]
   ; CHECK-NEXT:   [[ATOMIC_CMPXCHG_WITH_SUCCESS:%[0-9]+]]:_(s32), [[ATOMIC_CMPXCHG_WITH_SUCCESS1:%[0-9]+]]:_(s1) = G_ATOMIC_CMPXCHG_WITH_SUCCESS [[COPY]](p3), [[PHI1]], [[FSUB]] :: (load store seq_cst seq_cst (s32) on %ir.2, addrspace 3)
   ; CHECK-NEXT:   [[INT:%[0-9]+]]:_(s64) = G_INTRINSIC intrinsic(@llvm.amdgcn.if.break), [[ATOMIC_CMPXCHG_WITH_SUCCESS1]](s1), [[PHI]](s64)
   ; CHECK-NEXT:   [[INT1:%[0-9]+]]:_(s1) = G_INTRINSIC_W_SIDE_EFFECTS intrinsic(@llvm.amdgcn.loop), [[INT]](s64)
   ; CHECK-NEXT:   G_BRCOND [[INT1]](s1), %bb.3
   ; CHECK-NEXT:   G_BR %bb.2
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.3.atomicrmw.end:
   ; CHECK-NEXT:   [[PHI2:%[0-9]+]]:_(s32) = G_PHI [[ATOMIC_CMPXCHG_WITH_SUCCESS]](s32), %bb.2
   ; CHECK-NEXT:   [[PHI3:%[0-9]+]]:_(s64) = G_PHI [[INT]](s64), %bb.2
   ; CHECK-NEXT:   G_INTRINSIC_W_SIDE_EFFECTS intrinsic(@llvm.amdgcn.end.cf), [[PHI3]](s64)
   ; CHECK-NEXT:   $vgpr0 = COPY [[PHI2]](s32)
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY2]], implicit $vgpr0
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %oldval = atomicrmw fsub float addrspace(3)* %addr, float 1.0 seq_cst
   ret float %oldval
 }

llvm/test/CodeGen/AMDGPU/GlobalISel/irtranslator-call-abi-attribute-hints.ll

define amdgpu_kernel void @kernel_call_no_other_sgprs() {
   ; CHECK-NEXT:   S_ENDPGM 0
   call void @extern() "amdgpu-no-dispatch-id" "amdgpu-no-dispatch-ptr" "amdgpu-no-queue-ptr" "amdgpu-no-workgroup-id-x" "amdgpu-no-workgroup-id-y" "amdgpu-no-workgroup-id-z"
   ret void
 }
 
 define void @func_call_no_workitem_ids() {
   ; CHECK-LABEL: name: func_call_no_workitem_ids
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr_32 = COPY $sgpr14
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr_32 = COPY $sgpr13
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr_32 = COPY $sgpr12
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr10_sgpr11
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr8_sgpr9
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:sgpr_64 = COPY $sgpr6_sgpr7
   ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr4_sgpr5
-  ; CHECK-NEXT:   [[COPY7:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; CHECK-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @extern
-  ; CHECK-NEXT:   [[COPY8:%[0-9]+]]:_(p4) = COPY [[COPY6]]
-  ; CHECK-NEXT:   [[COPY9:%[0-9]+]]:_(p4) = COPY [[COPY5]]
-  ; CHECK-NEXT:   [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY4]]
-  ; CHECK-NEXT:   [[COPY11:%[0-9]+]]:_(s64) = COPY [[COPY3]]
-  ; CHECK-NEXT:   [[COPY12:%[0-9]+]]:_(s32) = COPY [[COPY2]]
-  ; CHECK-NEXT:   [[COPY13:%[0-9]+]]:_(s32) = COPY [[COPY1]]
-  ; CHECK-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY]]
-  ; CHECK-NEXT:   [[COPY15:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY15]](<4 x s32>)
-  ; CHECK-NEXT:   $sgpr4_sgpr5 = COPY [[COPY8]](p4)
-  ; CHECK-NEXT:   $sgpr6_sgpr7 = COPY [[COPY9]](p4)
-  ; CHECK-NEXT:   $sgpr8_sgpr9 = COPY [[COPY10]](p4)
-  ; CHECK-NEXT:   $sgpr10_sgpr11 = COPY [[COPY11]](s64)
-  ; CHECK-NEXT:   $sgpr12 = COPY [[COPY12]](s32)
-  ; CHECK-NEXT:   $sgpr13 = COPY [[COPY13]](s32)
-  ; CHECK-NEXT:   $sgpr14 = COPY [[COPY14]](s32)
+  ; CHECK-NEXT:   [[COPY7:%[0-9]+]]:_(p4) = COPY [[COPY6]]
+  ; CHECK-NEXT:   [[COPY8:%[0-9]+]]:_(p4) = COPY [[COPY5]]
+  ; CHECK-NEXT:   [[COPY9:%[0-9]+]]:_(p4) = COPY [[COPY4]]
+  ; CHECK-NEXT:   [[COPY10:%[0-9]+]]:_(s64) = COPY [[COPY3]]
+  ; CHECK-NEXT:   [[COPY11:%[0-9]+]]:_(s32) = COPY [[COPY2]]
+  ; CHECK-NEXT:   [[COPY12:%[0-9]+]]:_(s32) = COPY [[COPY1]]
+  ; CHECK-NEXT:   [[COPY13:%[0-9]+]]:_(s32) = COPY [[COPY]]
+  ; CHECK-NEXT:   [[COPY14:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY14]](<4 x s32>)
+  ; CHECK-NEXT:   $sgpr4_sgpr5 = COPY [[COPY7]](p4)
+  ; CHECK-NEXT:   $sgpr6_sgpr7 = COPY [[COPY8]](p4)
+  ; CHECK-NEXT:   $sgpr8_sgpr9 = COPY [[COPY9]](p4)
+  ; CHECK-NEXT:   $sgpr10_sgpr11 = COPY [[COPY10]](s64)
+  ; CHECK-NEXT:   $sgpr12 = COPY [[COPY11]](s32)
+  ; CHECK-NEXT:   $sgpr13 = COPY [[COPY12]](s32)
+  ; CHECK-NEXT:   $sgpr14 = COPY [[COPY13]](s32)
   ; CHECK-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @extern, csr_amdgpu_highregs, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr4_sgpr5, implicit $sgpr6_sgpr7, implicit $sgpr8_sgpr9, implicit $sgpr10_sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14
   ; CHECK-NEXT:   ADJCALLSTACKDOWN 0, 0, implicit-def $scc
-  ; CHECK-NEXT:   [[COPY16:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY7]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY16]]
+  ; CHECK-NEXT:   SI_RETURN
   call void @extern() "amdgpu-no-workitem-id-x" "amdgpu-no-workitem-id-y" "amdgpu-no-workitem-id-z"
   ret void
 }
 
 define void @func_call_no_workgroup_ids() {
   ; CHECK-LABEL: name: func_call_no_workgroup_ids
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:vgpr_32(s32) = COPY $vgpr31
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr10_sgpr11
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr8_sgpr9
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr6_sgpr7
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr4_sgpr5
-  ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; CHECK-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @extern
-  ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:_(p4) = COPY [[COPY4]]
-  ; CHECK-NEXT:   [[COPY7:%[0-9]+]]:_(p4) = COPY [[COPY3]]
-  ; CHECK-NEXT:   [[COPY8:%[0-9]+]]:_(p4) = COPY [[COPY2]]
-  ; CHECK-NEXT:   [[COPY9:%[0-9]+]]:_(s64) = COPY [[COPY1]]
-  ; CHECK-NEXT:   [[COPY10:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
-  ; CHECK-NEXT:   [[COPY11:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY11]](<4 x s32>)
-  ; CHECK-NEXT:   $sgpr4_sgpr5 = COPY [[COPY6]](p4)
-  ; CHECK-NEXT:   $sgpr6_sgpr7 = COPY [[COPY7]](p4)
-  ; CHECK-NEXT:   $sgpr8_sgpr9 = COPY [[COPY8]](p4)
-  ; CHECK-NEXT:   $sgpr10_sgpr11 = COPY [[COPY9]](s64)
-  ; CHECK-NEXT:   $vgpr31 = COPY [[COPY10]](s32)
+  ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:_(p4) = COPY [[COPY4]]
+  ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:_(p4) = COPY [[COPY3]]
+  ; CHECK-NEXT:   [[COPY7:%[0-9]+]]:_(p4) = COPY [[COPY2]]
+  ; CHECK-NEXT:   [[COPY8:%[0-9]+]]:_(s64) = COPY [[COPY1]]
+  ; CHECK-NEXT:   [[COPY9:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
+  ; CHECK-NEXT:   [[COPY10:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY10]](<4 x s32>)
+  ; CHECK-NEXT:   $sgpr4_sgpr5 = COPY [[COPY5]](p4)
+  ; CHECK-NEXT:   $sgpr6_sgpr7 = COPY [[COPY6]](p4)
+  ; CHECK-NEXT:   $sgpr8_sgpr9 = COPY [[COPY7]](p4)
+  ; CHECK-NEXT:   $sgpr10_sgpr11 = COPY [[COPY8]](s64)
+  ; CHECK-NEXT:   $vgpr31 = COPY [[COPY9]](s32)
   ; CHECK-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @extern, csr_amdgpu_highregs, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr4_sgpr5, implicit $sgpr6_sgpr7, implicit $sgpr8_sgpr9, implicit $sgpr10_sgpr11, implicit $vgpr31
   ; CHECK-NEXT:   ADJCALLSTACKDOWN 0, 0, implicit-def $scc
-  ; CHECK-NEXT:   [[COPY12:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY5]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY12]]
+  ; CHECK-NEXT:   SI_RETURN
   call void @extern() "amdgpu-no-workgroup-id-x" "amdgpu-no-workgroup-id-y" "amdgpu-no-workgroup-id-z"
   ret void
 }
 
 define void @func_call_no_other_sgprs() {
   ; CHECK-LABEL: name: func_call_no_other_sgprs
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $vgpr31, $sgpr8_sgpr9, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr31, $sgpr8_sgpr9
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:vgpr_32(s32) = COPY $vgpr31
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr8_sgpr9
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; CHECK-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @extern
-  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(p4) = COPY [[COPY1]]
-  ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
-  ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY5]](<4 x s32>)
-  ; CHECK-NEXT:   $sgpr8_sgpr9 = COPY [[COPY3]](p4)
-  ; CHECK-NEXT:   $vgpr31 = COPY [[COPY4]](s32)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(p4) = COPY [[COPY1]]
+  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
+  ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY4]](<4 x s32>)
+  ; CHECK-NEXT:   $sgpr8_sgpr9 = COPY [[COPY2]](p4)
+  ; CHECK-NEXT:   $vgpr31 = COPY [[COPY3]](s32)
   ; CHECK-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @extern, csr_amdgpu_highregs, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr8_sgpr9, implicit $vgpr31
   ; CHECK-NEXT:   ADJCALLSTACKDOWN 0, 0, implicit-def $scc
-  ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY6]]
+  ; CHECK-NEXT:   SI_RETURN
   call void @extern() "amdgpu-no-dispatch-id" "amdgpu-no-dispatch-ptr" "amdgpu-no-queue-ptr" "amdgpu-no-workgroup-id-x" "amdgpu-no-workgroup-id-y" "amdgpu-no-workgroup-id-z"
   ret void
 }

llvm/test/CodeGen/AMDGPU/GlobalISel/irtranslator-call-implicit-args.ll

define amdgpu_kernel void @test_call_external_void_func_i32([17 x i8]) #0 {
   ; GFX908-NEXT:   S_ENDPGM 0
   call void @external_void_func_i32(i32 42)
   ret void
 }
 
 define void @test_func_call_external_void_func_i32() #0 {
   ; GFX900-LABEL: name: test_func_call_external_void_func_i32
   ; GFX900: bb.1 (%ir-block.0):
-  ; GFX900-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11, $sgpr30_sgpr31
+  ; GFX900-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11
   ; GFX900-NEXT: {{  $}}
   ; GFX900-NEXT:   [[COPY:%[0-9]+]]:vgpr_32(s32) = COPY $vgpr31
   ; GFX900-NEXT:   [[COPY1:%[0-9]+]]:sgpr_32 = COPY $sgpr14
   ; GFX900-NEXT:   [[COPY2:%[0-9]+]]:sgpr_32 = COPY $sgpr13
   ; GFX900-NEXT:   [[COPY3:%[0-9]+]]:sgpr_32 = COPY $sgpr12
   ; GFX900-NEXT:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr10_sgpr11
   ; GFX900-NEXT:   [[COPY5:%[0-9]+]]:sgpr_64 = COPY $sgpr8_sgpr9
   ; GFX900-NEXT:   [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr6_sgpr7
   ; GFX900-NEXT:   [[COPY7:%[0-9]+]]:sgpr_64 = COPY $sgpr4_sgpr5
-  ; GFX900-NEXT:   [[COPY8:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GFX900-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 99
   ; GFX900-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; GFX900-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @external_void_func_i32
-  ; GFX900-NEXT:   [[COPY9:%[0-9]+]]:_(p4) = COPY [[COPY7]]
-  ; GFX900-NEXT:   [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY6]]
-  ; GFX900-NEXT:   [[COPY11:%[0-9]+]]:_(p4) = COPY [[COPY5]]
-  ; GFX900-NEXT:   [[COPY12:%[0-9]+]]:_(s64) = COPY [[COPY4]]
-  ; GFX900-NEXT:   [[COPY13:%[0-9]+]]:_(s32) = COPY [[COPY3]]
-  ; GFX900-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY2]]
-  ; GFX900-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY1]]
-  ; GFX900-NEXT:   [[COPY16:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
+  ; GFX900-NEXT:   [[COPY8:%[0-9]+]]:_(p4) = COPY [[COPY7]]
+  ; GFX900-NEXT:   [[COPY9:%[0-9]+]]:_(p4) = COPY [[COPY6]]
+  ; GFX900-NEXT:   [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY5]]
+  ; GFX900-NEXT:   [[COPY11:%[0-9]+]]:_(s64) = COPY [[COPY4]]
+  ; GFX900-NEXT:   [[COPY12:%[0-9]+]]:_(s32) = COPY [[COPY3]]
+  ; GFX900-NEXT:   [[COPY13:%[0-9]+]]:_(s32) = COPY [[COPY2]]
+  ; GFX900-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY1]]
+  ; GFX900-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
   ; GFX900-NEXT:   $vgpr0 = COPY [[C]](s32)
-  ; GFX900-NEXT:   [[COPY17:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GFX900-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY17]](<4 x s32>)
-  ; GFX900-NEXT:   $sgpr4_sgpr5 = COPY [[COPY9]](p4)
-  ; GFX900-NEXT:   $sgpr6_sgpr7 = COPY [[COPY10]](p4)
-  ; GFX900-NEXT:   $sgpr8_sgpr9 = COPY [[COPY11]](p4)
-  ; GFX900-NEXT:   $sgpr10_sgpr11 = COPY [[COPY12]](s64)
-  ; GFX900-NEXT:   $sgpr12 = COPY [[COPY13]](s32)
-  ; GFX900-NEXT:   $sgpr13 = COPY [[COPY14]](s32)
-  ; GFX900-NEXT:   $sgpr14 = COPY [[COPY15]](s32)
-  ; GFX900-NEXT:   $vgpr31 = COPY [[COPY16]](s32)
+  ; GFX900-NEXT:   [[COPY16:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GFX900-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY16]](<4 x s32>)
+  ; GFX900-NEXT:   $sgpr4_sgpr5 = COPY [[COPY8]](p4)
+  ; GFX900-NEXT:   $sgpr6_sgpr7 = COPY [[COPY9]](p4)
+  ; GFX900-NEXT:   $sgpr8_sgpr9 = COPY [[COPY10]](p4)
+  ; GFX900-NEXT:   $sgpr10_sgpr11 = COPY [[COPY11]](s64)
+  ; GFX900-NEXT:   $sgpr12 = COPY [[COPY12]](s32)
+  ; GFX900-NEXT:   $sgpr13 = COPY [[COPY13]](s32)
+  ; GFX900-NEXT:   $sgpr14 = COPY [[COPY14]](s32)
+  ; GFX900-NEXT:   $vgpr31 = COPY [[COPY15]](s32)
   ; GFX900-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @external_void_func_i32, csr_amdgpu_highregs, implicit $vgpr0, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr4_sgpr5, implicit $sgpr6_sgpr7, implicit $sgpr8_sgpr9, implicit $sgpr10_sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $vgpr31
   ; GFX900-NEXT:   ADJCALLSTACKDOWN 0, 0, implicit-def $scc
-  ; GFX900-NEXT:   [[COPY18:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY8]]
-  ; GFX900-NEXT:   S_SETPC_B64_return [[COPY18]]
+  ; GFX900-NEXT:   SI_RETURN
   ; GFX908-LABEL: name: test_func_call_external_void_func_i32
   ; GFX908: bb.1 (%ir-block.0):
-  ; GFX908-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11, $sgpr30_sgpr31
+  ; GFX908-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11
   ; GFX908-NEXT: {{  $}}
   ; GFX908-NEXT:   [[COPY:%[0-9]+]]:vgpr_32(s32) = COPY $vgpr31
   ; GFX908-NEXT:   [[COPY1:%[0-9]+]]:sgpr_32 = COPY $sgpr14
   ; GFX908-NEXT:   [[COPY2:%[0-9]+]]:sgpr_32 = COPY $sgpr13
   ; GFX908-NEXT:   [[COPY3:%[0-9]+]]:sgpr_32 = COPY $sgpr12
   ; GFX908-NEXT:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr10_sgpr11
   ; GFX908-NEXT:   [[COPY5:%[0-9]+]]:sgpr_64 = COPY $sgpr8_sgpr9
   ; GFX908-NEXT:   [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr6_sgpr7
   ; GFX908-NEXT:   [[COPY7:%[0-9]+]]:sgpr_64 = COPY $sgpr4_sgpr5
-  ; GFX908-NEXT:   [[COPY8:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GFX908-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 99
   ; GFX908-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; GFX908-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @external_void_func_i32
-  ; GFX908-NEXT:   [[COPY9:%[0-9]+]]:_(p4) = COPY [[COPY7]]
-  ; GFX908-NEXT:   [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY6]]
-  ; GFX908-NEXT:   [[COPY11:%[0-9]+]]:_(p4) = COPY [[COPY5]]
-  ; GFX908-NEXT:   [[COPY12:%[0-9]+]]:_(s64) = COPY [[COPY4]]
-  ; GFX908-NEXT:   [[COPY13:%[0-9]+]]:_(s32) = COPY [[COPY3]]
-  ; GFX908-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY2]]
-  ; GFX908-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY1]]
-  ; GFX908-NEXT:   [[COPY16:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
+  ; GFX908-NEXT:   [[COPY8:%[0-9]+]]:_(p4) = COPY [[COPY7]]
+  ; GFX908-NEXT:   [[COPY9:%[0-9]+]]:_(p4) = COPY [[COPY6]]
+  ; GFX908-NEXT:   [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY5]]
+  ; GFX908-NEXT:   [[COPY11:%[0-9]+]]:_(s64) = COPY [[COPY4]]
+  ; GFX908-NEXT:   [[COPY12:%[0-9]+]]:_(s32) = COPY [[COPY3]]
+  ; GFX908-NEXT:   [[COPY13:%[0-9]+]]:_(s32) = COPY [[COPY2]]
+  ; GFX908-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY1]]
+  ; GFX908-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
   ; GFX908-NEXT:   $vgpr0 = COPY [[C]](s32)
-  ; GFX908-NEXT:   [[COPY17:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GFX908-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY17]](<4 x s32>)
-  ; GFX908-NEXT:   $sgpr4_sgpr5 = COPY [[COPY9]](p4)
-  ; GFX908-NEXT:   $sgpr6_sgpr7 = COPY [[COPY10]](p4)
-  ; GFX908-NEXT:   $sgpr8_sgpr9 = COPY [[COPY11]](p4)
-  ; GFX908-NEXT:   $sgpr10_sgpr11 = COPY [[COPY12]](s64)
-  ; GFX908-NEXT:   $sgpr12 = COPY [[COPY13]](s32)
-  ; GFX908-NEXT:   $sgpr13 = COPY [[COPY14]](s32)
-  ; GFX908-NEXT:   $sgpr14 = COPY [[COPY15]](s32)
-  ; GFX908-NEXT:   $vgpr31 = COPY [[COPY16]](s32)
+  ; GFX908-NEXT:   [[COPY16:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GFX908-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY16]](<4 x s32>)
+  ; GFX908-NEXT:   $sgpr4_sgpr5 = COPY [[COPY8]](p4)
+  ; GFX908-NEXT:   $sgpr6_sgpr7 = COPY [[COPY9]](p4)
+  ; GFX908-NEXT:   $sgpr8_sgpr9 = COPY [[COPY10]](p4)
+  ; GFX908-NEXT:   $sgpr10_sgpr11 = COPY [[COPY11]](s64)
+  ; GFX908-NEXT:   $sgpr12 = COPY [[COPY12]](s32)
+  ; GFX908-NEXT:   $sgpr13 = COPY [[COPY13]](s32)
+  ; GFX908-NEXT:   $sgpr14 = COPY [[COPY14]](s32)
+  ; GFX908-NEXT:   $vgpr31 = COPY [[COPY15]](s32)
   ; GFX908-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @external_void_func_i32, csr_amdgpu_highregs, implicit $vgpr0, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr4_sgpr5, implicit $sgpr6_sgpr7, implicit $sgpr8_sgpr9, implicit $sgpr10_sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $vgpr31
   ; GFX908-NEXT:   ADJCALLSTACKDOWN 0, 0, implicit-def $scc
-  ; GFX908-NEXT:   [[COPY18:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY8]]
-  ; GFX908-NEXT:   S_SETPC_B64_return [[COPY18]]
+  ; GFX908-NEXT:   SI_RETURN
   call void @external_void_func_i32(i32 99)
   ret void
 }
 
 ; Explicit argument is split between registers ad the stack due to v31
 ; being used for workitem IDs.
 define amdgpu_kernel void @test_call_external_void_func_v32i32([17 x i8]) #0 {
   ; GFX900-LABEL: name: test_call_external_void_func_v32i32
define amdgpu_kernel void @test_call_external_void_func_v32i32([17 x i8]) #0 {
   ; GFX908-NEXT:   S_ENDPGM 0
   call void @external_void_func_v32i32(<32 x i32> zeroinitializer)
   ret void
 }
 
 define void @test_func_call_external_void_func_v32i32([17 x i8]) #0 {
   ; GFX900-LABEL: name: test_func_call_external_void_func_v32i32
   ; GFX900: bb.1 (%ir-block.1):
-  ; GFX900-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7, $vgpr8, $vgpr9, $vgpr10, $vgpr11, $vgpr12, $vgpr13, $vgpr14, $vgpr15, $vgpr16, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11, $sgpr30_sgpr31
+  ; GFX900-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7, $vgpr8, $vgpr9, $vgpr10, $vgpr11, $vgpr12, $vgpr13, $vgpr14, $vgpr15, $vgpr16, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11
   ; GFX900-NEXT: {{  $}}
   ; GFX900-NEXT:   [[COPY:%[0-9]+]]:vgpr_32(s32) = COPY $vgpr31
   ; GFX900-NEXT:   [[COPY1:%[0-9]+]]:sgpr_32 = COPY $sgpr14
   ; GFX900-NEXT:   [[COPY2:%[0-9]+]]:sgpr_32 = COPY $sgpr13
   ; GFX900-NEXT:   [[COPY3:%[0-9]+]]:sgpr_32 = COPY $sgpr12
   ; GFX900-NEXT:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr10_sgpr11
   ; GFX900-NEXT:   [[COPY5:%[0-9]+]]:sgpr_64 = COPY $sgpr8_sgpr9
   ; GFX900-NEXT:   [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr6_sgpr7
define void @test_func_call_external_void_func_v32i32([17 x i8]) #0 {
   ; GFX900-NEXT:   [[TRUNC28:%[0-9]+]]:_(s16) = G_TRUNC [[COPY22]](s32)
   ; GFX900-NEXT:   [[TRUNC29:%[0-9]+]]:_(s8) = G_TRUNC [[TRUNC28]](s16)
   ; GFX900-NEXT:   [[COPY23:%[0-9]+]]:_(s32) = COPY $vgpr15
   ; GFX900-NEXT:   [[TRUNC30:%[0-9]+]]:_(s16) = G_TRUNC [[COPY23]](s32)
   ; GFX900-NEXT:   [[TRUNC31:%[0-9]+]]:_(s8) = G_TRUNC [[TRUNC30]](s16)
   ; GFX900-NEXT:   [[COPY24:%[0-9]+]]:_(s32) = COPY $vgpr16
   ; GFX900-NEXT:   [[TRUNC32:%[0-9]+]]:_(s16) = G_TRUNC [[COPY24]](s32)
   ; GFX900-NEXT:   [[TRUNC33:%[0-9]+]]:_(s8) = G_TRUNC [[TRUNC32]](s16)
-  ; GFX900-NEXT:   [[COPY25:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GFX900-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
   ; GFX900-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:_(<32 x s32>) = G_BUILD_VECTOR [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32)
   ; GFX900-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; GFX900-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @external_void_func_v32i32
-  ; GFX900-NEXT:   [[COPY26:%[0-9]+]]:_(p4) = COPY [[COPY7]]
-  ; GFX900-NEXT:   [[COPY27:%[0-9]+]]:_(p4) = COPY [[COPY6]]
-  ; GFX900-NEXT:   [[COPY28:%[0-9]+]]:_(p4) = COPY [[COPY5]]
-  ; GFX900-NEXT:   [[COPY29:%[0-9]+]]:_(s64) = COPY [[COPY4]]
-  ; GFX900-NEXT:   [[COPY30:%[0-9]+]]:_(s32) = COPY [[COPY3]]
-  ; GFX900-NEXT:   [[COPY31:%[0-9]+]]:_(s32) = COPY [[COPY2]]
-  ; GFX900-NEXT:   [[COPY32:%[0-9]+]]:_(s32) = COPY [[COPY1]]
-  ; GFX900-NEXT:   [[COPY33:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
+  ; GFX900-NEXT:   [[COPY25:%[0-9]+]]:_(p4) = COPY [[COPY7]]
+  ; GFX900-NEXT:   [[COPY26:%[0-9]+]]:_(p4) = COPY [[COPY6]]
+  ; GFX900-NEXT:   [[COPY27:%[0-9]+]]:_(p4) = COPY [[COPY5]]
+  ; GFX900-NEXT:   [[COPY28:%[0-9]+]]:_(s64) = COPY [[COPY4]]
+  ; GFX900-NEXT:   [[COPY29:%[0-9]+]]:_(s32) = COPY [[COPY3]]
+  ; GFX900-NEXT:   [[COPY30:%[0-9]+]]:_(s32) = COPY [[COPY2]]
+  ; GFX900-NEXT:   [[COPY31:%[0-9]+]]:_(s32) = COPY [[COPY1]]
+  ; GFX900-NEXT:   [[COPY32:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
   ; GFX900-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32), [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32), [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32), [[UV12:%[0-9]+]]:_(s32), [[UV13:%[0-9]+]]:_(s32), [[UV14:%[0-9]+]]:_(s32), [[UV15:%[0-9]+]]:_(s32), [[UV16:%[0-9]+]]:_(s32), [[UV17:%[0-9]+]]:_(s32), [[UV18:%[0-9]+]]:_(s32), [[UV19:%[0-9]+]]:_(s32), [[UV20:%[0-9]+]]:_(s32), [[UV21:%[0-9]+]]:_(s32), [[UV22:%[0-9]+]]:_(s32), [[UV23:%[0-9]+]]:_(s32), [[UV24:%[0-9]+]]:_(s32), [[UV25:%[0-9]+]]:_(s32), [[UV26:%[0-9]+]]:_(s32), [[UV27:%[0-9]+]]:_(s32), [[UV28:%[0-9]+]]:_(s32), [[UV29:%[0-9]+]]:_(s32), [[UV30:%[0-9]+]]:_(s32), [[UV31:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR]](<32 x s32>)
-  ; GFX900-NEXT:   [[COPY34:%[0-9]+]]:_(p5) = COPY $sgpr32
+  ; GFX900-NEXT:   [[COPY33:%[0-9]+]]:_(p5) = COPY $sgpr32
   ; GFX900-NEXT:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
-  ; GFX900-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY34]], [[C1]](s32)
+  ; GFX900-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY33]], [[C1]](s32)
   ; GFX900-NEXT:   G_STORE [[UV31]](s32), [[PTR_ADD]](p5) :: (store (s32) into stack, align 16, addrspace 5)
   ; GFX900-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; GFX900-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; GFX900-NEXT:   $vgpr2 = COPY [[UV2]](s32)
   ; GFX900-NEXT:   $vgpr3 = COPY [[UV3]](s32)
   ; GFX900-NEXT:   $vgpr4 = COPY [[UV4]](s32)
   ; GFX900-NEXT:   $vgpr5 = COPY [[UV5]](s32)
   ; GFX900-NEXT:   $vgpr6 = COPY [[UV6]](s32)
define void @test_func_call_external_void_func_v32i32([17 x i8]) #0 {
   ; GFX900-NEXT:   $vgpr23 = COPY [[UV23]](s32)
   ; GFX900-NEXT:   $vgpr24 = COPY [[UV24]](s32)
   ; GFX900-NEXT:   $vgpr25 = COPY [[UV25]](s32)
   ; GFX900-NEXT:   $vgpr26 = COPY [[UV26]](s32)
   ; GFX900-NEXT:   $vgpr27 = COPY [[UV27]](s32)
   ; GFX900-NEXT:   $vgpr28 = COPY [[UV28]](s32)
   ; GFX900-NEXT:   $vgpr29 = COPY [[UV29]](s32)
   ; GFX900-NEXT:   $vgpr30 = COPY [[UV30]](s32)
-  ; GFX900-NEXT:   [[COPY35:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GFX900-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY35]](<4 x s32>)
-  ; GFX900-NEXT:   $sgpr4_sgpr5 = COPY [[COPY26]](p4)
-  ; GFX900-NEXT:   $sgpr6_sgpr7 = COPY [[COPY27]](p4)
-  ; GFX900-NEXT:   $sgpr8_sgpr9 = COPY [[COPY28]](p4)
-  ; GFX900-NEXT:   $sgpr10_sgpr11 = COPY [[COPY29]](s64)
-  ; GFX900-NEXT:   $sgpr12 = COPY [[COPY30]](s32)
-  ; GFX900-NEXT:   $sgpr13 = COPY [[COPY31]](s32)
-  ; GFX900-NEXT:   $sgpr14 = COPY [[COPY32]](s32)
-  ; GFX900-NEXT:   $vgpr31 = COPY [[COPY33]](s32)
+  ; GFX900-NEXT:   [[COPY34:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GFX900-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY34]](<4 x s32>)
+  ; GFX900-NEXT:   $sgpr4_sgpr5 = COPY [[COPY25]](p4)
+  ; GFX900-NEXT:   $sgpr6_sgpr7 = COPY [[COPY26]](p4)
+  ; GFX900-NEXT:   $sgpr8_sgpr9 = COPY [[COPY27]](p4)
+  ; GFX900-NEXT:   $sgpr10_sgpr11 = COPY [[COPY28]](s64)
+  ; GFX900-NEXT:   $sgpr12 = COPY [[COPY29]](s32)
+  ; GFX900-NEXT:   $sgpr13 = COPY [[COPY30]](s32)
+  ; GFX900-NEXT:   $sgpr14 = COPY [[COPY31]](s32)
+  ; GFX900-NEXT:   $vgpr31 = COPY [[COPY32]](s32)
   ; GFX900-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @external_void_func_v32i32, csr_amdgpu_highregs, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15, implicit $vgpr16, implicit $vgpr17, implicit $vgpr18, implicit $vgpr19, implicit $vgpr20, implicit $vgpr21, implicit $vgpr22, implicit $vgpr23, implicit $vgpr24, implicit $vgpr25, implicit $vgpr26, implicit $vgpr27, implicit $vgpr28, implicit $vgpr29, implicit $vgpr30, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr4_sgpr5, implicit $sgpr6_sgpr7, implicit $sgpr8_sgpr9, implicit $sgpr10_sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $vgpr31
   ; GFX900-NEXT:   ADJCALLSTACKDOWN 0, 4, implicit-def $scc
-  ; GFX900-NEXT:   [[COPY36:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY25]]
-  ; GFX900-NEXT:   S_SETPC_B64_return [[COPY36]]
+  ; GFX900-NEXT:   SI_RETURN
   ; GFX908-LABEL: name: test_func_call_external_void_func_v32i32
   ; GFX908: bb.1 (%ir-block.1):
-  ; GFX908-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7, $vgpr8, $vgpr9, $vgpr10, $vgpr11, $vgpr12, $vgpr13, $vgpr14, $vgpr15, $vgpr16, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11, $sgpr30_sgpr31
+  ; GFX908-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7, $vgpr8, $vgpr9, $vgpr10, $vgpr11, $vgpr12, $vgpr13, $vgpr14, $vgpr15, $vgpr16, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11
   ; GFX908-NEXT: {{  $}}
   ; GFX908-NEXT:   [[COPY:%[0-9]+]]:vgpr_32(s32) = COPY $vgpr31
   ; GFX908-NEXT:   [[COPY1:%[0-9]+]]:sgpr_32 = COPY $sgpr14
   ; GFX908-NEXT:   [[COPY2:%[0-9]+]]:sgpr_32 = COPY $sgpr13
   ; GFX908-NEXT:   [[COPY3:%[0-9]+]]:sgpr_32 = COPY $sgpr12
   ; GFX908-NEXT:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr10_sgpr11
   ; GFX908-NEXT:   [[COPY5:%[0-9]+]]:sgpr_64 = COPY $sgpr8_sgpr9
   ; GFX908-NEXT:   [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr6_sgpr7
define void @test_func_call_external_void_func_v32i32([17 x i8]) #0 {
   ; GFX908-NEXT:   [[TRUNC28:%[0-9]+]]:_(s16) = G_TRUNC [[COPY22]](s32)
   ; GFX908-NEXT:   [[TRUNC29:%[0-9]+]]:_(s8) = G_TRUNC [[TRUNC28]](s16)
   ; GFX908-NEXT:   [[COPY23:%[0-9]+]]:_(s32) = COPY $vgpr15
   ; GFX908-NEXT:   [[TRUNC30:%[0-9]+]]:_(s16) = G_TRUNC [[COPY23]](s32)
   ; GFX908-NEXT:   [[TRUNC31:%[0-9]+]]:_(s8) = G_TRUNC [[TRUNC30]](s16)
   ; GFX908-NEXT:   [[COPY24:%[0-9]+]]:_(s32) = COPY $vgpr16
   ; GFX908-NEXT:   [[TRUNC32:%[0-9]+]]:_(s16) = G_TRUNC [[COPY24]](s32)
   ; GFX908-NEXT:   [[TRUNC33:%[0-9]+]]:_(s8) = G_TRUNC [[TRUNC32]](s16)
-  ; GFX908-NEXT:   [[COPY25:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GFX908-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
   ; GFX908-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:_(<32 x s32>) = G_BUILD_VECTOR [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32)
   ; GFX908-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; GFX908-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @external_void_func_v32i32
-  ; GFX908-NEXT:   [[COPY26:%[0-9]+]]:_(p4) = COPY [[COPY7]]
-  ; GFX908-NEXT:   [[COPY27:%[0-9]+]]:_(p4) = COPY [[COPY6]]
-  ; GFX908-NEXT:   [[COPY28:%[0-9]+]]:_(p4) = COPY [[COPY5]]
-  ; GFX908-NEXT:   [[COPY29:%[0-9]+]]:_(s64) = COPY [[COPY4]]
-  ; GFX908-NEXT:   [[COPY30:%[0-9]+]]:_(s32) = COPY [[COPY3]]
-  ; GFX908-NEXT:   [[COPY31:%[0-9]+]]:_(s32) = COPY [[COPY2]]
-  ; GFX908-NEXT:   [[COPY32:%[0-9]+]]:_(s32) = COPY [[COPY1]]
-  ; GFX908-NEXT:   [[COPY33:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
+  ; GFX908-NEXT:   [[COPY25:%[0-9]+]]:_(p4) = COPY [[COPY7]]
+  ; GFX908-NEXT:   [[COPY26:%[0-9]+]]:_(p4) = COPY [[COPY6]]
+  ; GFX908-NEXT:   [[COPY27:%[0-9]+]]:_(p4) = COPY [[COPY5]]
+  ; GFX908-NEXT:   [[COPY28:%[0-9]+]]:_(s64) = COPY [[COPY4]]
+  ; GFX908-NEXT:   [[COPY29:%[0-9]+]]:_(s32) = COPY [[COPY3]]
+  ; GFX908-NEXT:   [[COPY30:%[0-9]+]]:_(s32) = COPY [[COPY2]]
+  ; GFX908-NEXT:   [[COPY31:%[0-9]+]]:_(s32) = COPY [[COPY1]]
+  ; GFX908-NEXT:   [[COPY32:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
   ; GFX908-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32), [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32), [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32), [[UV12:%[0-9]+]]:_(s32), [[UV13:%[0-9]+]]:_(s32), [[UV14:%[0-9]+]]:_(s32), [[UV15:%[0-9]+]]:_(s32), [[UV16:%[0-9]+]]:_(s32), [[UV17:%[0-9]+]]:_(s32), [[UV18:%[0-9]+]]:_(s32), [[UV19:%[0-9]+]]:_(s32), [[UV20:%[0-9]+]]:_(s32), [[UV21:%[0-9]+]]:_(s32), [[UV22:%[0-9]+]]:_(s32), [[UV23:%[0-9]+]]:_(s32), [[UV24:%[0-9]+]]:_(s32), [[UV25:%[0-9]+]]:_(s32), [[UV26:%[0-9]+]]:_(s32), [[UV27:%[0-9]+]]:_(s32), [[UV28:%[0-9]+]]:_(s32), [[UV29:%[0-9]+]]:_(s32), [[UV30:%[0-9]+]]:_(s32), [[UV31:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR]](<32 x s32>)
-  ; GFX908-NEXT:   [[COPY34:%[0-9]+]]:_(p5) = COPY $sgpr32
+  ; GFX908-NEXT:   [[COPY33:%[0-9]+]]:_(p5) = COPY $sgpr32
   ; GFX908-NEXT:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
-  ; GFX908-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY34]], [[C1]](s32)
+  ; GFX908-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY33]], [[C1]](s32)
   ; GFX908-NEXT:   G_STORE [[UV31]](s32), [[PTR_ADD]](p5) :: (store (s32) into stack, align 16, addrspace 5)
   ; GFX908-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; GFX908-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; GFX908-NEXT:   $vgpr2 = COPY [[UV2]](s32)
   ; GFX908-NEXT:   $vgpr3 = COPY [[UV3]](s32)
   ; GFX908-NEXT:   $vgpr4 = COPY [[UV4]](s32)
   ; GFX908-NEXT:   $vgpr5 = COPY [[UV5]](s32)
   ; GFX908-NEXT:   $vgpr6 = COPY [[UV6]](s32)
define void @test_func_call_external_void_func_v32i32([17 x i8]) #0 {
   ; GFX908-NEXT:   $vgpr23 = COPY [[UV23]](s32)
   ; GFX908-NEXT:   $vgpr24 = COPY [[UV24]](s32)
   ; GFX908-NEXT:   $vgpr25 = COPY [[UV25]](s32)
   ; GFX908-NEXT:   $vgpr26 = COPY [[UV26]](s32)
   ; GFX908-NEXT:   $vgpr27 = COPY [[UV27]](s32)
   ; GFX908-NEXT:   $vgpr28 = COPY [[UV28]](s32)
   ; GFX908-NEXT:   $vgpr29 = COPY [[UV29]](s32)
   ; GFX908-NEXT:   $vgpr30 = COPY [[UV30]](s32)
-  ; GFX908-NEXT:   [[COPY35:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GFX908-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY35]](<4 x s32>)
-  ; GFX908-NEXT:   $sgpr4_sgpr5 = COPY [[COPY26]](p4)
-  ; GFX908-NEXT:   $sgpr6_sgpr7 = COPY [[COPY27]](p4)
-  ; GFX908-NEXT:   $sgpr8_sgpr9 = COPY [[COPY28]](p4)
-  ; GFX908-NEXT:   $sgpr10_sgpr11 = COPY [[COPY29]](s64)
-  ; GFX908-NEXT:   $sgpr12 = COPY [[COPY30]](s32)
-  ; GFX908-NEXT:   $sgpr13 = COPY [[COPY31]](s32)
-  ; GFX908-NEXT:   $sgpr14 = COPY [[COPY32]](s32)
-  ; GFX908-NEXT:   $vgpr31 = COPY [[COPY33]](s32)
+  ; GFX908-NEXT:   [[COPY34:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GFX908-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY34]](<4 x s32>)
+  ; GFX908-NEXT:   $sgpr4_sgpr5 = COPY [[COPY25]](p4)
+  ; GFX908-NEXT:   $sgpr6_sgpr7 = COPY [[COPY26]](p4)
+  ; GFX908-NEXT:   $sgpr8_sgpr9 = COPY [[COPY27]](p4)
+  ; GFX908-NEXT:   $sgpr10_sgpr11 = COPY [[COPY28]](s64)
+  ; GFX908-NEXT:   $sgpr12 = COPY [[COPY29]](s32)
+  ; GFX908-NEXT:   $sgpr13 = COPY [[COPY30]](s32)
+  ; GFX908-NEXT:   $sgpr14 = COPY [[COPY31]](s32)
+  ; GFX908-NEXT:   $vgpr31 = COPY [[COPY32]](s32)
   ; GFX908-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @external_void_func_v32i32, csr_amdgpu_highregs, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15, implicit $vgpr16, implicit $vgpr17, implicit $vgpr18, implicit $vgpr19, implicit $vgpr20, implicit $vgpr21, implicit $vgpr22, implicit $vgpr23, implicit $vgpr24, implicit $vgpr25, implicit $vgpr26, implicit $vgpr27, implicit $vgpr28, implicit $vgpr29, implicit $vgpr30, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr4_sgpr5, implicit $sgpr6_sgpr7, implicit $sgpr8_sgpr9, implicit $sgpr10_sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $vgpr31
   ; GFX908-NEXT:   ADJCALLSTACKDOWN 0, 4, implicit-def $scc
-  ; GFX908-NEXT:   [[COPY36:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY25]]
-  ; GFX908-NEXT:   S_SETPC_B64_return [[COPY36]]
+  ; GFX908-NEXT:   SI_RETURN
   call void @external_void_func_v32i32(<32 x i32> zeroinitializer)
   ret void
 }
 
 ; FIXME: Should fold out parts with known 0 id.
 
 define amdgpu_kernel void @test_only_workitem_id_x() #0 !reqd_work_group_size !0 {
   ; GFX900-LABEL: name: test_only_workitem_id_x

llvm/test/CodeGen/AMDGPU/GlobalISel/irtranslator-call-non-fixed.ll

 ; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
 ; RUN: llc -global-isel -stop-after=irtranslator -mtriple=amdgcn-mesa-mesa3d -mcpu=gfx900 -verify-machineinstrs -o - %s | FileCheck -enable-var-scope %s
 
 ; amdgpu_gfx calling convention
 declare hidden amdgpu_gfx void @external_gfx_void_func_void() #0
 declare hidden amdgpu_gfx void @external_gfx_void_func_i32(i32) #0
 declare hidden amdgpu_gfx void @external_gfx_void_func_i32_inreg(i32 inreg) #0
 declare hidden amdgpu_gfx void @external_gfx_void_func_struct_i8_i32({ i8, i32 }) #0
 declare hidden amdgpu_gfx void @external_gfx_void_func_struct_i8_i32_inreg({ i8, i32 } inreg) #0
 
 define amdgpu_gfx void @test_gfx_call_external_void_func_void() #0 {
   ; CHECK-LABEL: name: test_gfx_call_external_void_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $sgpr30_sgpr31
-  ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; CHECK-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @external_gfx_void_func_void
-  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY1]](<4 x s32>)
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY]](<4 x s32>)
   ; CHECK-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @external_gfx_void_func_void, csr_amdgpu_si_gfx, implicit $sgpr0_sgpr1_sgpr2_sgpr3
   ; CHECK-NEXT:   ADJCALLSTACKDOWN 0, 0, implicit-def $scc
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:gfx_ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK-NEXT:   S_SETPC_B64_return_gfx [[COPY2]]
+  ; CHECK-NEXT:   SI_RETURN
   call amdgpu_gfx void @external_gfx_void_func_void()
   ret void
 }
 
 define amdgpu_gfx void @test_gfx_call_external_void_func_i32_imm(i32) #0 {
   ; CHECK-LABEL: name: test_gfx_call_external_void_func_i32_imm
   ; CHECK: bb.1 (%ir-block.1):
-  ; CHECK-NEXT:   liveins: $vgpr0, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 42
   ; CHECK-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; CHECK-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @external_gfx_void_func_i32
   ; CHECK-NEXT:   $vgpr0 = COPY [[C]](s32)
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY2]](<4 x s32>)
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY1]](<4 x s32>)
   ; CHECK-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @external_gfx_void_func_i32, csr_amdgpu_si_gfx, implicit $vgpr0, implicit $sgpr0_sgpr1_sgpr2_sgpr3
   ; CHECK-NEXT:   ADJCALLSTACKDOWN 0, 0, implicit-def $scc
-  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:gfx_ccr_sgpr_64 = COPY [[COPY1]]
-  ; CHECK-NEXT:   S_SETPC_B64_return_gfx [[COPY3]]
+  ; CHECK-NEXT:   SI_RETURN
   call amdgpu_gfx void @external_gfx_void_func_i32(i32 42)
   ret void
 }
 
 define amdgpu_gfx void @test_gfx_call_external_void_func_i32_imm_inreg(i32 inreg) #0 {
   ; CHECK-LABEL: name: test_gfx_call_external_void_func_i32_imm_inreg
   ; CHECK: bb.1 (%ir-block.1):
-  ; CHECK-NEXT:   liveins: $sgpr4, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $sgpr4
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $sgpr4
-  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 42
   ; CHECK-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; CHECK-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @external_gfx_void_func_i32_inreg
   ; CHECK-NEXT:   $sgpr4 = COPY [[C]](s32)
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY2]](<4 x s32>)
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY1]](<4 x s32>)
   ; CHECK-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @external_gfx_void_func_i32_inreg, csr_amdgpu_si_gfx, implicit $sgpr4, implicit $sgpr0_sgpr1_sgpr2_sgpr3
   ; CHECK-NEXT:   ADJCALLSTACKDOWN 0, 0, implicit-def $scc
-  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:gfx_ccr_sgpr_64 = COPY [[COPY1]]
-  ; CHECK-NEXT:   S_SETPC_B64_return_gfx [[COPY3]]
+  ; CHECK-NEXT:   SI_RETURN
   call amdgpu_gfx void @external_gfx_void_func_i32_inreg(i32 inreg 42)
   ret void
 }
 
 define amdgpu_gfx void @test_gfx_call_external_void_func_struct_i8_i32() #0 {
   ; CHECK-LABEL: name: test_gfx_call_external_void_func_struct_i8_i32
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $sgpr30_sgpr31
-  ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
   ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (load (p1) from `{ i8, i32 } addrspace(1)* addrspace(4)* undef`, addrspace 4)
   ; CHECK-NEXT:   [[LOAD1:%[0-9]+]]:_(s8) = G_LOAD [[LOAD]](p1) :: (load (s8) from %ir.ptr0, align 4, addrspace 1)
   ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 4
   ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p1) = G_PTR_ADD [[LOAD]], [[C]](s64)
   ; CHECK-NEXT:   [[LOAD2:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD]](p1) :: (load (s32) from %ir.ptr0 + 4, addrspace 1)
   ; CHECK-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; CHECK-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @external_gfx_void_func_struct_i8_i32
   ; CHECK-NEXT:   [[ANYEXT:%[0-9]+]]:_(s16) = G_ANYEXT [[LOAD1]](s8)
   ; CHECK-NEXT:   [[ANYEXT1:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT]](s16)
   ; CHECK-NEXT:   $vgpr0 = COPY [[ANYEXT1]](s32)
   ; CHECK-NEXT:   $vgpr1 = COPY [[LOAD2]](s32)
-  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY1]](<4 x s32>)
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY]](<4 x s32>)
   ; CHECK-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @external_gfx_void_func_struct_i8_i32, csr_amdgpu_si_gfx, implicit $vgpr0, implicit $vgpr1, implicit $sgpr0_sgpr1_sgpr2_sgpr3
   ; CHECK-NEXT:   ADJCALLSTACKDOWN 0, 0, implicit-def $scc
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:gfx_ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK-NEXT:   S_SETPC_B64_return_gfx [[COPY2]]
+  ; CHECK-NEXT:   SI_RETURN
   %ptr0 = load { i8, i32 } addrspace(1)*, { i8, i32 } addrspace(1)* addrspace(4)* undef
   %val = load { i8, i32 }, { i8, i32 } addrspace(1)* %ptr0
   call amdgpu_gfx void @external_gfx_void_func_struct_i8_i32({ i8, i32 } %val)
   ret void
 }
 
 define amdgpu_gfx void @test_gfx_call_external_void_func_struct_i8_i32_inreg() #0 {
   ; CHECK-LABEL: name: test_gfx_call_external_void_func_struct_i8_i32_inreg
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $sgpr30_sgpr31
-  ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
   ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (load (p1) from `{ i8, i32 } addrspace(1)* addrspace(4)* undef`, addrspace 4)
   ; CHECK-NEXT:   [[LOAD1:%[0-9]+]]:_(s8) = G_LOAD [[LOAD]](p1) :: (load (s8) from %ir.ptr0, align 4, addrspace 1)
   ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 4
   ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p1) = G_PTR_ADD [[LOAD]], [[C]](s64)
   ; CHECK-NEXT:   [[LOAD2:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD]](p1) :: (load (s32) from %ir.ptr0 + 4, addrspace 1)
   ; CHECK-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; CHECK-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @external_gfx_void_func_struct_i8_i32_inreg
   ; CHECK-NEXT:   [[ANYEXT:%[0-9]+]]:_(s16) = G_ANYEXT [[LOAD1]](s8)
   ; CHECK-NEXT:   [[ANYEXT1:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT]](s16)
   ; CHECK-NEXT:   $sgpr4 = COPY [[ANYEXT1]](s32)
   ; CHECK-NEXT:   $sgpr5 = COPY [[LOAD2]](s32)
-  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY1]](<4 x s32>)
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY]](<4 x s32>)
   ; CHECK-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @external_gfx_void_func_struct_i8_i32_inreg, csr_amdgpu_si_gfx, implicit $sgpr4, implicit $sgpr5, implicit $sgpr0_sgpr1_sgpr2_sgpr3
   ; CHECK-NEXT:   ADJCALLSTACKDOWN 0, 0, implicit-def $scc
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:gfx_ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK-NEXT:   S_SETPC_B64_return_gfx [[COPY2]]
+  ; CHECK-NEXT:   SI_RETURN
   %ptr0 = load { i8, i32 } addrspace(1)*, { i8, i32 } addrspace(1)* addrspace(4)* undef
   %val = load { i8, i32 }, { i8, i32 } addrspace(1)* %ptr0
   call amdgpu_gfx void @external_gfx_void_func_struct_i8_i32_inreg({ i8, i32 } inreg %val)
   ret void
 }
 
 attributes #0 = { nounwind }
 attributes #1 = { nounwind readnone }
 attributes #2 = { nounwind noinline }

llvm/test/CodeGen/AMDGPU/GlobalISel/irtranslator-call-return-values.ll

define amdgpu_kernel void @test_call_external_i32_func_i32_imm(i32 addrspace(1)* %out) #0 {
   %val = call i32 @external_i32_func_i32(i32 42)
   store volatile i32 %val, i32 addrspace(1)* %out
   ret void
 }
 
 define amdgpu_gfx void @test_gfx_call_external_i32_func_i32_imm(i32 addrspace(1)* %out) #0 {
   ; GCN-LABEL: name: test_gfx_call_external_i32_func_i32_imm
   ; GCN: bb.1 (%ir-block.0):
-  ; GCN-NEXT:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+  ; GCN-NEXT:   liveins: $vgpr0, $vgpr1
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; GCN-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
   ; GCN-NEXT:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
-  ; GCN-NEXT:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 42
   ; GCN-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; GCN-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @external_gfx_i32_func_i32
   ; GCN-NEXT:   $vgpr0 = COPY [[C]](s32)
-  ; GCN-NEXT:   [[COPY3:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY3]](<4 x s32>)
+  ; GCN-NEXT:   [[COPY2:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY2]](<4 x s32>)
   ; GCN-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @external_gfx_i32_func_i32, csr_amdgpu_si_gfx, implicit $vgpr0, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit-def $vgpr0
-  ; GCN-NEXT:   [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; GCN-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; GCN-NEXT:   ADJCALLSTACKDOWN 0, 0, implicit-def $scc
-  ; GCN-NEXT:   G_STORE [[COPY4]](s32), [[MV]](p1) :: (volatile store (s32) into %ir.out, addrspace 1)
-  ; GCN-NEXT:   [[COPY5:%[0-9]+]]:gfx_ccr_sgpr_64 = COPY [[COPY2]]
-  ; GCN-NEXT:   S_SETPC_B64_return_gfx [[COPY5]]
+  ; GCN-NEXT:   G_STORE [[COPY3]](s32), [[MV]](p1) :: (volatile store (s32) into %ir.out, addrspace 1)
+  ; GCN-NEXT:   SI_RETURN
   %val = call amdgpu_gfx i32 @external_gfx_i32_func_i32(i32 42)
   store volatile i32 %val, i32 addrspace(1)* %out
   ret void
 }
 
 define amdgpu_kernel void @test_call_external_i1_func_void() #0 {
   ; GCN-LABEL: name: test_call_external_i1_func_void
   ; GCN: bb.1 (%ir-block.0):
define amdgpu_kernel void @test_call_external_i1_func_void() #0 {
   %val = call i1 @external_i1_func_void()
   store volatile i1 %val, i1 addrspace(1)* undef
   ret void
 }
 
 define amdgpu_gfx void @test_gfx_call_external_i1_func_void() #0 {
   ; GCN-LABEL: name: test_gfx_call_external_i1_func_void
   ; GCN: bb.1 (%ir-block.0):
-  ; GCN-NEXT:   liveins: $sgpr30_sgpr31
-  ; GCN-NEXT: {{  $}}
-  ; GCN-NEXT:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
   ; GCN-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; GCN-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @external_gfx_i1_func_void
-  ; GCN-NEXT:   [[COPY1:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY1]](<4 x s32>)
+  ; GCN-NEXT:   [[COPY:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY]](<4 x s32>)
   ; GCN-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @external_gfx_i1_func_void, csr_amdgpu_si_gfx, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit-def $vgpr0
-  ; GCN-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; GCN-NEXT:   [[TRUNC:%[0-9]+]]:_(s1) = G_TRUNC [[COPY2]](s32)
+  ; GCN-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; GCN-NEXT:   [[TRUNC:%[0-9]+]]:_(s1) = G_TRUNC [[COPY1]](s32)
   ; GCN-NEXT:   ADJCALLSTACKDOWN 0, 0, implicit-def $scc
   ; GCN-NEXT:   G_STORE [[TRUNC]](s1), [[DEF]](p1) :: (volatile store (s1) into `i1 addrspace(1)* undef`, addrspace 1)
-  ; GCN-NEXT:   [[COPY3:%[0-9]+]]:gfx_ccr_sgpr_64 = COPY [[COPY]]
-  ; GCN-NEXT:   S_SETPC_B64_return_gfx [[COPY3]]
+  ; GCN-NEXT:   SI_RETURN
   %val = call amdgpu_gfx i1 @external_gfx_i1_func_void()
   store volatile i1 %val, i1 addrspace(1)* undef
   ret void
 }
 
 define amdgpu_kernel void @test_call_external_i1_zeroext_func_void() #0 {
   ; GCN-LABEL: name: test_call_external_i1_zeroext_func_void
   ; GCN: bb.1 (%ir-block.0):
define amdgpu_kernel void @test_call_external_i8_func_void() #0 {
   %val = call i8 @external_i8_func_void()
   store volatile i8 %val, i8 addrspace(1)* undef
   ret void
 }
 
 define amdgpu_gfx void @test_gfx_call_external_i8_func_void() #0 {
   ; GCN-LABEL: name: test_gfx_call_external_i8_func_void
   ; GCN: bb.1 (%ir-block.0):
-  ; GCN-NEXT:   liveins: $sgpr30_sgpr31
-  ; GCN-NEXT: {{  $}}
-  ; GCN-NEXT:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
   ; GCN-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; GCN-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @external_gfx_i8_func_void
-  ; GCN-NEXT:   [[COPY1:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY1]](<4 x s32>)
+  ; GCN-NEXT:   [[COPY:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY]](<4 x s32>)
   ; GCN-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @external_gfx_i8_func_void, csr_amdgpu_si_gfx, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit-def $vgpr0
-  ; GCN-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; GCN-NEXT:   [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
+  ; GCN-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; GCN-NEXT:   [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
   ; GCN-NEXT:   [[TRUNC1:%[0-9]+]]:_(s8) = G_TRUNC [[TRUNC]](s16)
   ; GCN-NEXT:   ADJCALLSTACKDOWN 0, 0, implicit-def $scc
   ; GCN-NEXT:   G_STORE [[TRUNC1]](s8), [[DEF]](p1) :: (volatile store (s8) into `i8 addrspace(1)* undef`, addrspace 1)
-  ; GCN-NEXT:   [[COPY3:%[0-9]+]]:gfx_ccr_sgpr_64 = COPY [[COPY]]
-  ; GCN-NEXT:   S_SETPC_B64_return_gfx [[COPY3]]
+  ; GCN-NEXT:   SI_RETURN
   %val = call amdgpu_gfx i8 @external_gfx_i8_func_void()
   store volatile i8 %val, i8 addrspace(1)* undef
   ret void
 }
 
 define amdgpu_kernel void @test_call_external_i8_zeroext_func_void() #0 {
   ; GCN-LABEL: name: test_call_external_i8_zeroext_func_void
   ; GCN: bb.1 (%ir-block.0):
define amdgpu_kernel void @test_call_external_i32_func_void() #0 {
   %val = call i32 @external_i32_func_void()
   store volatile i32 %val, i32 addrspace(1)* undef
   ret void
 }
 
 define amdgpu_gfx void @test_gfx_call_external_i32_func_void() #0 {
   ; GCN-LABEL: name: test_gfx_call_external_i32_func_void
   ; GCN: bb.1 (%ir-block.0):
-  ; GCN-NEXT:   liveins: $sgpr30_sgpr31
-  ; GCN-NEXT: {{  $}}
-  ; GCN-NEXT:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
   ; GCN-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; GCN-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @external_gfx_i32_func_void
-  ; GCN-NEXT:   [[COPY1:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY1]](<4 x s32>)
+  ; GCN-NEXT:   [[COPY:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY]](<4 x s32>)
   ; GCN-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @external_gfx_i32_func_void, csr_amdgpu_si_gfx, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit-def $vgpr0
-  ; GCN-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; GCN-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; GCN-NEXT:   ADJCALLSTACKDOWN 0, 0, implicit-def $scc
-  ; GCN-NEXT:   G_STORE [[COPY2]](s32), [[DEF]](p1) :: (volatile store (s32) into `i32 addrspace(1)* undef`, addrspace 1)
-  ; GCN-NEXT:   [[COPY3:%[0-9]+]]:gfx_ccr_sgpr_64 = COPY [[COPY]]
-  ; GCN-NEXT:   S_SETPC_B64_return_gfx [[COPY3]]
+  ; GCN-NEXT:   G_STORE [[COPY1]](s32), [[DEF]](p1) :: (volatile store (s32) into `i32 addrspace(1)* undef`, addrspace 1)
+  ; GCN-NEXT:   SI_RETURN
   %val = call amdgpu_gfx i32 @external_gfx_i32_func_void()
   store volatile i32 %val, i32 addrspace(1)* undef
   ret void
 }
 
 define amdgpu_kernel void @test_call_external_i48_func_void() #0 {
   ; GCN-LABEL: name: test_call_external_i48_func_void
   ; GCN: bb.1 (%ir-block.0):
define amdgpu_kernel void @test_call_external_i32_i64_func_void() #0 {
   store volatile i32 %val.0, i32 addrspace(1)* undef
   store volatile i64 %val.1, i64 addrspace(1)* undef
   ret void
 }
 
 define amdgpu_gfx void @test_gfx_call_external_i32_i64_func_void() #0 {
   ; GCN-LABEL: name: test_gfx_call_external_i32_i64_func_void
   ; GCN: bb.1 (%ir-block.0):
-  ; GCN-NEXT:   liveins: $sgpr30_sgpr31
-  ; GCN-NEXT: {{  $}}
-  ; GCN-NEXT:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; GCN-NEXT:   [[COPY1:%[0-9]+]]:_(p1) = COPY [[DEF]](p1)
+  ; GCN-NEXT:   [[COPY:%[0-9]+]]:_(p1) = COPY [[DEF]](p1)
   ; GCN-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; GCN-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @external_gfx_i32_i64_func_void
-  ; GCN-NEXT:   [[COPY2:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY2]](<4 x s32>)
+  ; GCN-NEXT:   [[COPY1:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY1]](<4 x s32>)
   ; GCN-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @external_gfx_i32_i64_func_void, csr_amdgpu_si_gfx, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit-def $vgpr0, implicit-def $vgpr1, implicit-def $vgpr2
-  ; GCN-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; GCN-NEXT:   [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; GCN-NEXT:   [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; GCN-NEXT:   [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-  ; GCN-NEXT:   ADJCALLSTACKDOWN 0, 0, implicit-def $scc
-  ; GCN-NEXT:   G_STORE [[COPY3]](s32), [[DEF]](p1) :: (volatile store (s32) into `i32 addrspace(1)* undef`, addrspace 1)
-  ; GCN-NEXT:   G_STORE [[MV]](s64), [[COPY1]](p1) :: (volatile store (s64) into `i64 addrspace(1)* undef`, addrspace 1)
-  ; GCN-NEXT:   [[COPY6:%[0-9]+]]:gfx_ccr_sgpr_64 = COPY [[COPY]]
-  ; GCN-NEXT:   S_SETPC_B64_return_gfx [[COPY6]]
+  ; GCN-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; GCN-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; GCN-NEXT:   [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr2
+  ; GCN-NEXT:   [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY3]](s32), [[COPY4]](s32)
+  ; GCN-NEXT:   ADJCALLSTACKDOWN 0, 0, implicit-def $scc
+  ; GCN-NEXT:   G_STORE [[COPY2]](s32), [[DEF]](p1) :: (volatile store (s32) into `i32 addrspace(1)* undef`, addrspace 1)
+  ; GCN-NEXT:   G_STORE [[MV]](s64), [[COPY]](p1) :: (volatile store (s64) into `i64 addrspace(1)* undef`, addrspace 1)
+  ; GCN-NEXT:   SI_RETURN
   %val = call amdgpu_gfx { i32, i64 } @external_gfx_i32_i64_func_void()
   %val.0 = extractvalue { i32, i64 } %val, 0
   %val.1 = extractvalue { i32, i64 } %val, 1
   store volatile i32 %val.0, i32 addrspace(1)* undef
   store volatile i64 %val.1, i64 addrspace(1)* undef
   ret void
 }
 

llvm/test/CodeGen/AMDGPU/GlobalISel/irtranslator-call.ll

define amdgpu_kernel void @test_call_external_void_func_void() #0 {
   ; CHECK-NEXT:   S_ENDPGM 0
   call void @external_void_func_void()
   ret void
 }
 
 define amdgpu_gfx void @test_gfx_call_external_void_func_void() #0 {
   ; CHECK-LABEL: name: test_gfx_call_external_void_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $sgpr30_sgpr31
-  ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; CHECK-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @external_gfx_void_func_void
-  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY1]](<4 x s32>)
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY]](<4 x s32>)
   ; CHECK-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @external_gfx_void_func_void, csr_amdgpu_si_gfx, implicit $sgpr0_sgpr1_sgpr2_sgpr3
   ; CHECK-NEXT:   ADJCALLSTACKDOWN 0, 0, implicit-def $scc
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:gfx_ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK-NEXT:   S_SETPC_B64_return_gfx [[COPY2]]
+  ; CHECK-NEXT:   SI_RETURN
   call amdgpu_gfx void @external_gfx_void_func_void()
   ret void
 }
 
 define void @test_func_call_external_void_func_void() #0 {
   ; CHECK-LABEL: name: test_func_call_external_void_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:vgpr_32(s32) = COPY $vgpr31
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr_32 = COPY $sgpr14
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr_32 = COPY $sgpr13
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr_32 = COPY $sgpr12
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr10_sgpr11
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:sgpr_64 = COPY $sgpr8_sgpr9
   ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr6_sgpr7
   ; CHECK-NEXT:   [[COPY7:%[0-9]+]]:sgpr_64 = COPY $sgpr4_sgpr5
-  ; CHECK-NEXT:   [[COPY8:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; CHECK-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @external_void_func_void
-  ; CHECK-NEXT:   [[COPY9:%[0-9]+]]:_(p4) = COPY [[COPY7]]
-  ; CHECK-NEXT:   [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY6]]
-  ; CHECK-NEXT:   [[COPY11:%[0-9]+]]:_(p4) = COPY [[COPY5]]
-  ; CHECK-NEXT:   [[COPY12:%[0-9]+]]:_(s64) = COPY [[COPY4]]
-  ; CHECK-NEXT:   [[COPY13:%[0-9]+]]:_(s32) = COPY [[COPY3]]
-  ; CHECK-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY2]]
-  ; CHECK-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY1]]
-  ; CHECK-NEXT:   [[COPY16:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
-  ; CHECK-NEXT:   [[COPY17:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY17]](<4 x s32>)
-  ; CHECK-NEXT:   $sgpr4_sgpr5 = COPY [[COPY9]](p4)
-  ; CHECK-NEXT:   $sgpr6_sgpr7 = COPY [[COPY10]](p4)
-  ; CHECK-NEXT:   $sgpr8_sgpr9 = COPY [[COPY11]](p4)
-  ; CHECK-NEXT:   $sgpr10_sgpr11 = COPY [[COPY12]](s64)
-  ; CHECK-NEXT:   $sgpr12 = COPY [[COPY13]](s32)
-  ; CHECK-NEXT:   $sgpr13 = COPY [[COPY14]](s32)
-  ; CHECK-NEXT:   $sgpr14 = COPY [[COPY15]](s32)
-  ; CHECK-NEXT:   $vgpr31 = COPY [[COPY16]](s32)
+  ; CHECK-NEXT:   [[COPY8:%[0-9]+]]:_(p4) = COPY [[COPY7]]
+  ; CHECK-NEXT:   [[COPY9:%[0-9]+]]:_(p4) = COPY [[COPY6]]
+  ; CHECK-NEXT:   [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY5]]
+  ; CHECK-NEXT:   [[COPY11:%[0-9]+]]:_(s64) = COPY [[COPY4]]
+  ; CHECK-NEXT:   [[COPY12:%[0-9]+]]:_(s32) = COPY [[COPY3]]
+  ; CHECK-NEXT:   [[COPY13:%[0-9]+]]:_(s32) = COPY [[COPY2]]
+  ; CHECK-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY1]]
+  ; CHECK-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
+  ; CHECK-NEXT:   [[COPY16:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY16]](<4 x s32>)
+  ; CHECK-NEXT:   $sgpr4_sgpr5 = COPY [[COPY8]](p4)
+  ; CHECK-NEXT:   $sgpr6_sgpr7 = COPY [[COPY9]](p4)
+  ; CHECK-NEXT:   $sgpr8_sgpr9 = COPY [[COPY10]](p4)
+  ; CHECK-NEXT:   $sgpr10_sgpr11 = COPY [[COPY11]](s64)
+  ; CHECK-NEXT:   $sgpr12 = COPY [[COPY12]](s32)
+  ; CHECK-NEXT:   $sgpr13 = COPY [[COPY13]](s32)
+  ; CHECK-NEXT:   $sgpr14 = COPY [[COPY14]](s32)
+  ; CHECK-NEXT:   $vgpr31 = COPY [[COPY15]](s32)
   ; CHECK-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @external_void_func_void, csr_amdgpu_highregs, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr4_sgpr5, implicit $sgpr6_sgpr7, implicit $sgpr8_sgpr9, implicit $sgpr10_sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $vgpr31
   ; CHECK-NEXT:   ADJCALLSTACKDOWN 0, 0, implicit-def $scc
-  ; CHECK-NEXT:   [[COPY18:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY8]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY18]]
+  ; CHECK-NEXT:   SI_RETURN
   call void @external_void_func_void()
   ret void
 }
 
 define amdgpu_kernel void @test_call_external_void_func_empty_struct() #0 {
   ; CHECK-LABEL: name: test_call_external_void_func_empty_struct
   ; CHECK: bb.1 (%ir-block.0):
   ; CHECK-NEXT:   liveins: $sgpr14, $sgpr15, $sgpr16, $vgpr0, $vgpr1, $vgpr2, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11
define amdgpu_kernel void @test_call_external_void_func_i32_imm(i32) #0 {
   ; CHECK-NEXT:   S_ENDPGM 0
   call void @external_void_func_i32(i32 42)
   ret void
 }
 
 define amdgpu_gfx void @test_gfx_call_external_void_func_i32_imm(i32) #0 {
   ; CHECK-LABEL: name: test_gfx_call_external_void_func_i32_imm
   ; CHECK: bb.1 (%ir-block.1):
-  ; CHECK-NEXT:   liveins: $vgpr0, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 42
   ; CHECK-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; CHECK-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @external_gfx_void_func_i32
   ; CHECK-NEXT:   $vgpr0 = COPY [[C]](s32)
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY2]](<4 x s32>)
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY1]](<4 x s32>)
   ; CHECK-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @external_gfx_void_func_i32, csr_amdgpu_si_gfx, implicit $vgpr0, implicit $sgpr0_sgpr1_sgpr2_sgpr3
   ; CHECK-NEXT:   ADJCALLSTACKDOWN 0, 0, implicit-def $scc
-  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:gfx_ccr_sgpr_64 = COPY [[COPY1]]
-  ; CHECK-NEXT:   S_SETPC_B64_return_gfx [[COPY3]]
+  ; CHECK-NEXT:   SI_RETURN
   call amdgpu_gfx void @external_gfx_void_func_i32(i32 42)
   ret void
 }
 
 define amdgpu_gfx void @test_gfx_call_external_void_func_i32_imm_inreg(i32 inreg) #0 {
   ; CHECK-LABEL: name: test_gfx_call_external_void_func_i32_imm_inreg
   ; CHECK: bb.1 (%ir-block.1):
-  ; CHECK-NEXT:   liveins: $sgpr4, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $sgpr4
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $sgpr4
-  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 42
   ; CHECK-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; CHECK-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @external_gfx_void_func_i32_inreg
   ; CHECK-NEXT:   $sgpr4 = COPY [[C]](s32)
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY2]](<4 x s32>)
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY1]](<4 x s32>)
   ; CHECK-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @external_gfx_void_func_i32_inreg, csr_amdgpu_si_gfx, implicit $sgpr4, implicit $sgpr0_sgpr1_sgpr2_sgpr3
   ; CHECK-NEXT:   ADJCALLSTACKDOWN 0, 0, implicit-def $scc
-  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:gfx_ccr_sgpr_64 = COPY [[COPY1]]
-  ; CHECK-NEXT:   S_SETPC_B64_return_gfx [[COPY3]]
+  ; CHECK-NEXT:   SI_RETURN
   call amdgpu_gfx void @external_gfx_void_func_i32_inreg(i32 inreg 42)
   ret void
 }
 
 define amdgpu_kernel void @test_call_external_void_func_i64_imm() #0 {
   ; CHECK-LABEL: name: test_call_external_void_func_i64_imm
   ; CHECK: bb.1 (%ir-block.0):
   ; CHECK-NEXT:   liveins: $sgpr14, $sgpr15, $sgpr16, $vgpr0, $vgpr1, $vgpr2, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11
define amdgpu_kernel void @test_call_external_void_func_struct_i8_i32() #0 {
   %val = load { i8, i32 }, { i8, i32 } addrspace(1)* %ptr0
   call void @external_void_func_struct_i8_i32({ i8, i32 } %val)
   ret void
 }
 
 define amdgpu_gfx void @test_gfx_call_external_void_func_struct_i8_i32() #0 {
   ; CHECK-LABEL: name: test_gfx_call_external_void_func_struct_i8_i32
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $sgpr30_sgpr31
-  ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
   ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (load (p1) from `{ i8, i32 } addrspace(1)* addrspace(4)* undef`, addrspace 4)
   ; CHECK-NEXT:   [[LOAD1:%[0-9]+]]:_(s8) = G_LOAD [[LOAD]](p1) :: (load (s8) from %ir.ptr0, align 4, addrspace 1)
   ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 4
   ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p1) = G_PTR_ADD [[LOAD]], [[C]](s64)
   ; CHECK-NEXT:   [[LOAD2:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD]](p1) :: (load (s32) from %ir.ptr0 + 4, addrspace 1)
   ; CHECK-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; CHECK-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @external_gfx_void_func_struct_i8_i32
   ; CHECK-NEXT:   [[ANYEXT:%[0-9]+]]:_(s16) = G_ANYEXT [[LOAD1]](s8)
   ; CHECK-NEXT:   [[ANYEXT1:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT]](s16)
   ; CHECK-NEXT:   $vgpr0 = COPY [[ANYEXT1]](s32)
   ; CHECK-NEXT:   $vgpr1 = COPY [[LOAD2]](s32)
-  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY1]](<4 x s32>)
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY]](<4 x s32>)
   ; CHECK-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @external_gfx_void_func_struct_i8_i32, csr_amdgpu_si_gfx, implicit $vgpr0, implicit $vgpr1, implicit $sgpr0_sgpr1_sgpr2_sgpr3
   ; CHECK-NEXT:   ADJCALLSTACKDOWN 0, 0, implicit-def $scc
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:gfx_ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK-NEXT:   S_SETPC_B64_return_gfx [[COPY2]]
+  ; CHECK-NEXT:   SI_RETURN
   %ptr0 = load { i8, i32 } addrspace(1)*, { i8, i32 } addrspace(1)* addrspace(4)* undef
   %val = load { i8, i32 }, { i8, i32 } addrspace(1)* %ptr0
   call amdgpu_gfx void @external_gfx_void_func_struct_i8_i32({ i8, i32 } %val)
   ret void
 }
 
 define amdgpu_gfx void @test_gfx_call_external_void_func_struct_i8_i32_inreg() #0 {
   ; CHECK-LABEL: name: test_gfx_call_external_void_func_struct_i8_i32_inreg
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $sgpr30_sgpr31
-  ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p4) = G_IMPLICIT_DEF
   ; CHECK-NEXT:   [[LOAD:%[0-9]+]]:_(p1) = G_LOAD [[DEF]](p4) :: (load (p1) from `{ i8, i32 } addrspace(1)* addrspace(4)* undef`, addrspace 4)
   ; CHECK-NEXT:   [[LOAD1:%[0-9]+]]:_(s8) = G_LOAD [[LOAD]](p1) :: (load (s8) from %ir.ptr0, align 4, addrspace 1)
   ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 4
   ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p1) = G_PTR_ADD [[LOAD]], [[C]](s64)
   ; CHECK-NEXT:   [[LOAD2:%[0-9]+]]:_(s32) = G_LOAD [[PTR_ADD]](p1) :: (load (s32) from %ir.ptr0 + 4, addrspace 1)
   ; CHECK-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; CHECK-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @external_gfx_void_func_struct_i8_i32_inreg
   ; CHECK-NEXT:   [[ANYEXT:%[0-9]+]]:_(s16) = G_ANYEXT [[LOAD1]](s8)
   ; CHECK-NEXT:   [[ANYEXT1:%[0-9]+]]:_(s32) = G_ANYEXT [[ANYEXT]](s16)
   ; CHECK-NEXT:   $sgpr4 = COPY [[ANYEXT1]](s32)
   ; CHECK-NEXT:   $sgpr5 = COPY [[LOAD2]](s32)
-  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY1]](<4 x s32>)
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY]](<4 x s32>)
   ; CHECK-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @external_gfx_void_func_struct_i8_i32_inreg, csr_amdgpu_si_gfx, implicit $sgpr4, implicit $sgpr5, implicit $sgpr0_sgpr1_sgpr2_sgpr3
   ; CHECK-NEXT:   ADJCALLSTACKDOWN 0, 0, implicit-def $scc
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:gfx_ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK-NEXT:   S_SETPC_B64_return_gfx [[COPY2]]
+  ; CHECK-NEXT:   SI_RETURN
   %ptr0 = load { i8, i32 } addrspace(1)*, { i8, i32 } addrspace(1)* addrspace(4)* undef
   %val = load { i8, i32 }, { i8, i32 } addrspace(1)* %ptr0
   call amdgpu_gfx void @external_gfx_void_func_struct_i8_i32_inreg({ i8, i32 } inreg %val)
   ret void
 }
 
 define amdgpu_kernel void @test_call_external_void_func_byval_struct_i8_i32() #0 {
   ; CHECK-LABEL: name: test_call_external_void_func_byval_struct_i8_i32
define amdgpu_kernel void @test_call_external_void_func_byval_struct_i8_i32() #0 {
   ret void
 }
 
 declare void @void_func_byval_a3i32_byval_i8_align32([3 x i32] addrspace(5)* byval([3 x i32]) %arg0, i8 addrspace(5)* byval(i8) align 32 %arg1, i32 %arg2) #0
 
 define void @call_byval_3ai32_byval_i8_align32([3 x i32] addrspace(5)* %incoming0, i8 addrspace(5)* align 32 %incoming1) #0 {
   ; CHECK-LABEL: name: call_byval_3ai32_byval_i8_align32
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr0, $vgpr1, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr0, $vgpr1, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:vgpr_32(s32) = COPY $vgpr31
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr_32 = COPY $sgpr14
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr_32 = COPY $sgpr13
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr_32 = COPY $sgpr12
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr10_sgpr11
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:sgpr_64 = COPY $sgpr8_sgpr9
   ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr6_sgpr7
   ; CHECK-NEXT:   [[COPY7:%[0-9]+]]:sgpr_64 = COPY $sgpr4_sgpr5
   ; CHECK-NEXT:   [[COPY8:%[0-9]+]]:_(p5) = COPY $vgpr0
   ; CHECK-NEXT:   [[COPY9:%[0-9]+]]:_(p5) = COPY $vgpr1
-  ; CHECK-NEXT:   [[COPY10:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 999
   ; CHECK-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; CHECK-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @void_func_byval_a3i32_byval_i8_align32
-  ; CHECK-NEXT:   [[COPY11:%[0-9]+]]:_(p4) = COPY [[COPY7]]
-  ; CHECK-NEXT:   [[COPY12:%[0-9]+]]:_(p4) = COPY [[COPY6]]
-  ; CHECK-NEXT:   [[COPY13:%[0-9]+]]:_(p4) = COPY [[COPY5]]
-  ; CHECK-NEXT:   [[COPY14:%[0-9]+]]:_(s64) = COPY [[COPY4]]
-  ; CHECK-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY3]]
-  ; CHECK-NEXT:   [[COPY16:%[0-9]+]]:_(s32) = COPY [[COPY2]]
-  ; CHECK-NEXT:   [[COPY17:%[0-9]+]]:_(s32) = COPY [[COPY1]]
-  ; CHECK-NEXT:   [[COPY18:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
-  ; CHECK-NEXT:   [[COPY19:%[0-9]+]]:_(p5) = COPY $sgpr32
+  ; CHECK-NEXT:   [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY7]]
+  ; CHECK-NEXT:   [[COPY11:%[0-9]+]]:_(p4) = COPY [[COPY6]]
+  ; CHECK-NEXT:   [[COPY12:%[0-9]+]]:_(p4) = COPY [[COPY5]]
+  ; CHECK-NEXT:   [[COPY13:%[0-9]+]]:_(s64) = COPY [[COPY4]]
+  ; CHECK-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY3]]
+  ; CHECK-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY2]]
+  ; CHECK-NEXT:   [[COPY16:%[0-9]+]]:_(s32) = COPY [[COPY1]]
+  ; CHECK-NEXT:   [[COPY17:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
+  ; CHECK-NEXT:   [[COPY18:%[0-9]+]]:_(p5) = COPY $sgpr32
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY19]], [[C1]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY18]], [[C1]](s32)
   ; CHECK-NEXT:   [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 12
   ; CHECK-NEXT:   G_MEMCPY [[PTR_ADD]](p5), [[COPY8]](p5), [[C2]](s32), 0 :: (dereferenceable store (s96) into stack, align 4, addrspace 5), (dereferenceable load (s96) from %ir.incoming0, align 4, addrspace 5)
   ; CHECK-NEXT:   [[C3:%[0-9]+]]:_(s32) = G_CONSTANT i32 32
-  ; CHECK-NEXT:   [[PTR_ADD1:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY19]], [[C3]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD1:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY18]], [[C3]](s32)
   ; CHECK-NEXT:   [[C4:%[0-9]+]]:_(s32) = G_CONSTANT i32 1
   ; CHECK-NEXT:   G_MEMCPY [[PTR_ADD1]](p5), [[COPY9]](p5), [[C4]](s32), 0 :: (dereferenceable store (s8) into stack + 32, align 32, addrspace 5), (dereferenceable load (s8) from %ir.incoming1, align 32, addrspace 5)
   ; CHECK-NEXT:   $vgpr0 = COPY [[C]](s32)
-  ; CHECK-NEXT:   [[COPY20:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY20]](<4 x s32>)
-  ; CHECK-NEXT:   $sgpr4_sgpr5 = COPY [[COPY11]](p4)
-  ; CHECK-NEXT:   $sgpr6_sgpr7 = COPY [[COPY12]](p4)
-  ; CHECK-NEXT:   $sgpr8_sgpr9 = COPY [[COPY13]](p4)
-  ; CHECK-NEXT:   $sgpr10_sgpr11 = COPY [[COPY14]](s64)
-  ; CHECK-NEXT:   $sgpr12 = COPY [[COPY15]](s32)
-  ; CHECK-NEXT:   $sgpr13 = COPY [[COPY16]](s32)
-  ; CHECK-NEXT:   $sgpr14 = COPY [[COPY17]](s32)
-  ; CHECK-NEXT:   $vgpr31 = COPY [[COPY18]](s32)
+  ; CHECK-NEXT:   [[COPY19:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY19]](<4 x s32>)
+  ; CHECK-NEXT:   $sgpr4_sgpr5 = COPY [[COPY10]](p4)
+  ; CHECK-NEXT:   $sgpr6_sgpr7 = COPY [[COPY11]](p4)
+  ; CHECK-NEXT:   $sgpr8_sgpr9 = COPY [[COPY12]](p4)
+  ; CHECK-NEXT:   $sgpr10_sgpr11 = COPY [[COPY13]](s64)
+  ; CHECK-NEXT:   $sgpr12 = COPY [[COPY14]](s32)
+  ; CHECK-NEXT:   $sgpr13 = COPY [[COPY15]](s32)
+  ; CHECK-NEXT:   $sgpr14 = COPY [[COPY16]](s32)
+  ; CHECK-NEXT:   $vgpr31 = COPY [[COPY17]](s32)
   ; CHECK-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @void_func_byval_a3i32_byval_i8_align32, csr_amdgpu_highregs, implicit $vgpr0, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr4_sgpr5, implicit $sgpr6_sgpr7, implicit $sgpr8_sgpr9, implicit $sgpr10_sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $vgpr31
   ; CHECK-NEXT:   ADJCALLSTACKDOWN 0, 36, implicit-def $scc
-  ; CHECK-NEXT:   [[COPY21:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY10]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY21]]
+  ; CHECK-NEXT:   SI_RETURN
   call void @void_func_byval_a3i32_byval_i8_align32([3 x i32] addrspace(5)* byval([3 x i32]) %incoming0, i8 addrspace(5)* align 32 %incoming1, i32 999)
   ret void
 }
 
 declare void @void_func_byval_a4i64_align4([4 x i64] addrspace(5)* byval([4 x i64]) align 4 %arg0) #0
 
 ; Make sure we are aware of the higher alignment of the incoming value
 ; than implied by the outgoing byval alignment in the memory operand.
 define void @call_byval_a4i64_align4_higher_source_align([4 x i64] addrspace(5)* align 256 %incoming_high_align) #0 {
   ; CHECK-LABEL: name: call_byval_a4i64_align4_higher_source_align
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr0, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr0, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:vgpr_32(s32) = COPY $vgpr31
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr_32 = COPY $sgpr14
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr_32 = COPY $sgpr13
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr_32 = COPY $sgpr12
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr10_sgpr11
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:sgpr_64 = COPY $sgpr8_sgpr9
   ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr6_sgpr7
   ; CHECK-NEXT:   [[COPY7:%[0-9]+]]:sgpr_64 = COPY $sgpr4_sgpr5
   ; CHECK-NEXT:   [[COPY8:%[0-9]+]]:_(p5) = COPY $vgpr0
-  ; CHECK-NEXT:   [[COPY9:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; CHECK-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @void_func_byval_a4i64_align4
-  ; CHECK-NEXT:   [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY7]]
-  ; CHECK-NEXT:   [[COPY11:%[0-9]+]]:_(p4) = COPY [[COPY6]]
-  ; CHECK-NEXT:   [[COPY12:%[0-9]+]]:_(p4) = COPY [[COPY5]]
-  ; CHECK-NEXT:   [[COPY13:%[0-9]+]]:_(s64) = COPY [[COPY4]]
-  ; CHECK-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY3]]
-  ; CHECK-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY2]]
-  ; CHECK-NEXT:   [[COPY16:%[0-9]+]]:_(s32) = COPY [[COPY1]]
-  ; CHECK-NEXT:   [[COPY17:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
-  ; CHECK-NEXT:   [[COPY18:%[0-9]+]]:_(p5) = COPY $sgpr32
+  ; CHECK-NEXT:   [[COPY9:%[0-9]+]]:_(p4) = COPY [[COPY7]]
+  ; CHECK-NEXT:   [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY6]]
+  ; CHECK-NEXT:   [[COPY11:%[0-9]+]]:_(p4) = COPY [[COPY5]]
+  ; CHECK-NEXT:   [[COPY12:%[0-9]+]]:_(s64) = COPY [[COPY4]]
+  ; CHECK-NEXT:   [[COPY13:%[0-9]+]]:_(s32) = COPY [[COPY3]]
+  ; CHECK-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY2]]
+  ; CHECK-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY1]]
+  ; CHECK-NEXT:   [[COPY16:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
+  ; CHECK-NEXT:   [[COPY17:%[0-9]+]]:_(p5) = COPY $sgpr32
   ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY18]], [[C]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C]](s32)
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 32
   ; CHECK-NEXT:   G_MEMCPY [[PTR_ADD]](p5), [[COPY8]](p5), [[C1]](s32), 0 :: (dereferenceable store (s256) into stack, align 4, addrspace 5), (dereferenceable load (s256) from %ir.incoming_high_align, align 256, addrspace 5)
-  ; CHECK-NEXT:   [[COPY19:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY19]](<4 x s32>)
-  ; CHECK-NEXT:   $sgpr4_sgpr5 = COPY [[COPY10]](p4)
-  ; CHECK-NEXT:   $sgpr6_sgpr7 = COPY [[COPY11]](p4)
-  ; CHECK-NEXT:   $sgpr8_sgpr9 = COPY [[COPY12]](p4)
-  ; CHECK-NEXT:   $sgpr10_sgpr11 = COPY [[COPY13]](s64)
-  ; CHECK-NEXT:   $sgpr12 = COPY [[COPY14]](s32)
-  ; CHECK-NEXT:   $sgpr13 = COPY [[COPY15]](s32)
-  ; CHECK-NEXT:   $sgpr14 = COPY [[COPY16]](s32)
-  ; CHECK-NEXT:   $vgpr31 = COPY [[COPY17]](s32)
+  ; CHECK-NEXT:   [[COPY18:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY18]](<4 x s32>)
+  ; CHECK-NEXT:   $sgpr4_sgpr5 = COPY [[COPY9]](p4)
+  ; CHECK-NEXT:   $sgpr6_sgpr7 = COPY [[COPY10]](p4)
+  ; CHECK-NEXT:   $sgpr8_sgpr9 = COPY [[COPY11]](p4)
+  ; CHECK-NEXT:   $sgpr10_sgpr11 = COPY [[COPY12]](s64)
+  ; CHECK-NEXT:   $sgpr12 = COPY [[COPY13]](s32)
+  ; CHECK-NEXT:   $sgpr13 = COPY [[COPY14]](s32)
+  ; CHECK-NEXT:   $sgpr14 = COPY [[COPY15]](s32)
+  ; CHECK-NEXT:   $vgpr31 = COPY [[COPY16]](s32)
   ; CHECK-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @void_func_byval_a4i64_align4, csr_amdgpu_highregs, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr4_sgpr5, implicit $sgpr6_sgpr7, implicit $sgpr8_sgpr9, implicit $sgpr10_sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $vgpr31
   ; CHECK-NEXT:   ADJCALLSTACKDOWN 0, 32, implicit-def $scc
-  ; CHECK-NEXT:   [[COPY20:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY9]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY20]]
+  ; CHECK-NEXT:   SI_RETURN
   call void @void_func_byval_a4i64_align4([4 x i64] addrspace(5)* byval([4 x i64]) align 4 %incoming_high_align)
   ret void
 }
 
 define amdgpu_kernel void @test_call_external_void_func_v2i8() #0 {
   ; CHECK-LABEL: name: test_call_external_void_func_v2i8
   ; CHECK: bb.1 (%ir-block.0):
   ; CHECK-NEXT:   liveins: $sgpr14, $sgpr15, $sgpr16, $vgpr0, $vgpr1, $vgpr2, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11

 entry:
   call void @stack_passed_f64_arg(<32 x i32> %val, double %tmp)
   ret void
 }
 
 define void @stack_12xv3i32() #0 {
   ; CHECK-LABEL: name: stack_12xv3i32
   ; CHECK: bb.1.entry:
-  ; CHECK-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:vgpr_32(s32) = COPY $vgpr31
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr_32 = COPY $sgpr14
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr_32 = COPY $sgpr13
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr_32 = COPY $sgpr12
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr10_sgpr11
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:sgpr_64 = COPY $sgpr8_sgpr9
   ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr6_sgpr7
   ; CHECK-NEXT:   [[COPY7:%[0-9]+]]:sgpr_64 = COPY $sgpr4_sgpr5
-  ; CHECK-NEXT:   [[COPY8:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[C]](s32), [[C]](s32), [[C]](s32)
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 1
   ; CHECK-NEXT:   [[BUILD_VECTOR1:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[C1]](s32), [[C1]](s32), [[C1]](s32)
   ; CHECK-NEXT:   [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 2
   ; CHECK-NEXT:   [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[C2]](s32), [[C2]](s32), [[C2]](s32)
   ; CHECK-NEXT:   [[C3:%[0-9]+]]:_(s32) = G_CONSTANT i32 3
   ; CHECK-NEXT:   [[BUILD_VECTOR3:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[C3]](s32), [[C3]](s32), [[C3]](s32)
define void @stack_12xv3i32() #0 {
   ; CHECK-NEXT:   [[C12:%[0-9]+]]:_(s32) = G_CONSTANT i32 12
   ; CHECK-NEXT:   [[BUILD_VECTOR10:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[C10]](s32), [[C11]](s32), [[C12]](s32)
   ; CHECK-NEXT:   [[C13:%[0-9]+]]:_(s32) = G_CONSTANT i32 13
   ; CHECK-NEXT:   [[C14:%[0-9]+]]:_(s32) = G_CONSTANT i32 14
   ; CHECK-NEXT:   [[C15:%[0-9]+]]:_(s32) = G_CONSTANT i32 15
   ; CHECK-NEXT:   [[BUILD_VECTOR11:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[C13]](s32), [[C14]](s32), [[C15]](s32)
   ; CHECK-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; CHECK-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @external_void_func_12xv3i32
-  ; CHECK-NEXT:   [[COPY9:%[0-9]+]]:_(p4) = COPY [[COPY7]]
-  ; CHECK-NEXT:   [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY6]]
-  ; CHECK-NEXT:   [[COPY11:%[0-9]+]]:_(p4) = COPY [[COPY5]]
-  ; CHECK-NEXT:   [[COPY12:%[0-9]+]]:_(s64) = COPY [[COPY4]]
-  ; CHECK-NEXT:   [[COPY13:%[0-9]+]]:_(s32) = COPY [[COPY3]]
-  ; CHECK-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY2]]
-  ; CHECK-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY1]]
-  ; CHECK-NEXT:   [[COPY16:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
+  ; CHECK-NEXT:   [[COPY8:%[0-9]+]]:_(p4) = COPY [[COPY7]]
+  ; CHECK-NEXT:   [[COPY9:%[0-9]+]]:_(p4) = COPY [[COPY6]]
+  ; CHECK-NEXT:   [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY5]]
+  ; CHECK-NEXT:   [[COPY11:%[0-9]+]]:_(s64) = COPY [[COPY4]]
+  ; CHECK-NEXT:   [[COPY12:%[0-9]+]]:_(s32) = COPY [[COPY3]]
+  ; CHECK-NEXT:   [[COPY13:%[0-9]+]]:_(s32) = COPY [[COPY2]]
+  ; CHECK-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY1]]
+  ; CHECK-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
   ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR]](<3 x s32>)
   ; CHECK-NEXT:   [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR1]](<3 x s32>)
   ; CHECK-NEXT:   [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32), [[UV8:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR2]](<3 x s32>)
   ; CHECK-NEXT:   [[UV9:%[0-9]+]]:_(s32), [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR3]](<3 x s32>)
   ; CHECK-NEXT:   [[UV12:%[0-9]+]]:_(s32), [[UV13:%[0-9]+]]:_(s32), [[UV14:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR4]](<3 x s32>)
   ; CHECK-NEXT:   [[UV15:%[0-9]+]]:_(s32), [[UV16:%[0-9]+]]:_(s32), [[UV17:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR5]](<3 x s32>)
   ; CHECK-NEXT:   [[UV18:%[0-9]+]]:_(s32), [[UV19:%[0-9]+]]:_(s32), [[UV20:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR6]](<3 x s32>)
   ; CHECK-NEXT:   [[UV21:%[0-9]+]]:_(s32), [[UV22:%[0-9]+]]:_(s32), [[UV23:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR7]](<3 x s32>)
   ; CHECK-NEXT:   [[UV24:%[0-9]+]]:_(s32), [[UV25:%[0-9]+]]:_(s32), [[UV26:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR8]](<3 x s32>)
   ; CHECK-NEXT:   [[UV27:%[0-9]+]]:_(s32), [[UV28:%[0-9]+]]:_(s32), [[UV29:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR9]](<3 x s32>)
   ; CHECK-NEXT:   [[UV30:%[0-9]+]]:_(s32), [[UV31:%[0-9]+]]:_(s32), [[UV32:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR10]](<3 x s32>)
-  ; CHECK-NEXT:   [[COPY17:%[0-9]+]]:_(p5) = COPY $sgpr32
+  ; CHECK-NEXT:   [[COPY16:%[0-9]+]]:_(p5) = COPY $sgpr32
   ; CHECK-NEXT:   [[C16:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C16]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY16]], [[C16]](s32)
   ; CHECK-NEXT:   G_STORE [[UV31]](s32), [[PTR_ADD]](p5) :: (store (s32) into stack, align 16, addrspace 5)
   ; CHECK-NEXT:   [[C17:%[0-9]+]]:_(s32) = G_CONSTANT i32 4
-  ; CHECK-NEXT:   [[PTR_ADD1:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C17]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD1:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY16]], [[C17]](s32)
   ; CHECK-NEXT:   G_STORE [[UV32]](s32), [[PTR_ADD1]](p5) :: (store (s32) into stack + 4, addrspace 5)
   ; CHECK-NEXT:   [[UV33:%[0-9]+]]:_(s32), [[UV34:%[0-9]+]]:_(s32), [[UV35:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR11]](<3 x s32>)
   ; CHECK-NEXT:   [[C18:%[0-9]+]]:_(s32) = G_CONSTANT i32 8
-  ; CHECK-NEXT:   [[PTR_ADD2:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C18]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD2:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY16]], [[C18]](s32)
   ; CHECK-NEXT:   G_STORE [[UV33]](s32), [[PTR_ADD2]](p5) :: (store (s32) into stack + 8, align 8, addrspace 5)
   ; CHECK-NEXT:   [[C19:%[0-9]+]]:_(s32) = G_CONSTANT i32 12
-  ; CHECK-NEXT:   [[PTR_ADD3:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C19]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD3:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY16]], [[C19]](s32)
   ; CHECK-NEXT:   G_STORE [[UV34]](s32), [[PTR_ADD3]](p5) :: (store (s32) into stack + 12, addrspace 5)
   ; CHECK-NEXT:   [[C20:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
-  ; CHECK-NEXT:   [[PTR_ADD4:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C20]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD4:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY16]], [[C20]](s32)
   ; CHECK-NEXT:   G_STORE [[UV35]](s32), [[PTR_ADD4]](p5) :: (store (s32) into stack + 16, align 16, addrspace 5)
   ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
   ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
   ; CHECK-NEXT:   $vgpr4 = COPY [[UV4]](s32)
   ; CHECK-NEXT:   $vgpr5 = COPY [[UV5]](s32)
   ; CHECK-NEXT:   $vgpr6 = COPY [[UV6]](s32)
define void @stack_12xv3i32() #0 {
   ; CHECK-NEXT:   $vgpr23 = COPY [[UV23]](s32)
   ; CHECK-NEXT:   $vgpr24 = COPY [[UV24]](s32)
   ; CHECK-NEXT:   $vgpr25 = COPY [[UV25]](s32)
   ; CHECK-NEXT:   $vgpr26 = COPY [[UV26]](s32)
   ; CHECK-NEXT:   $vgpr27 = COPY [[UV27]](s32)
   ; CHECK-NEXT:   $vgpr28 = COPY [[UV28]](s32)
   ; CHECK-NEXT:   $vgpr29 = COPY [[UV29]](s32)
   ; CHECK-NEXT:   $vgpr30 = COPY [[UV30]](s32)
-  ; CHECK-NEXT:   [[COPY18:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY18]](<4 x s32>)
-  ; CHECK-NEXT:   $sgpr4_sgpr5 = COPY [[COPY9]](p4)
-  ; CHECK-NEXT:   $sgpr6_sgpr7 = COPY [[COPY10]](p4)
-  ; CHECK-NEXT:   $sgpr8_sgpr9 = COPY [[COPY11]](p4)
-  ; CHECK-NEXT:   $sgpr10_sgpr11 = COPY [[COPY12]](s64)
-  ; CHECK-NEXT:   $sgpr12 = COPY [[COPY13]](s32)
-  ; CHECK-NEXT:   $sgpr13 = COPY [[COPY14]](s32)
-  ; CHECK-NEXT:   $sgpr14 = COPY [[COPY15]](s32)
-  ; CHECK-NEXT:   $vgpr31 = COPY [[COPY16]](s32)
+  ; CHECK-NEXT:   [[COPY17:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY17]](<4 x s32>)
+  ; CHECK-NEXT:   $sgpr4_sgpr5 = COPY [[COPY8]](p4)
+  ; CHECK-NEXT:   $sgpr6_sgpr7 = COPY [[COPY9]](p4)
+  ; CHECK-NEXT:   $sgpr8_sgpr9 = COPY [[COPY10]](p4)
+  ; CHECK-NEXT:   $sgpr10_sgpr11 = COPY [[COPY11]](s64)
+  ; CHECK-NEXT:   $sgpr12 = COPY [[COPY12]](s32)
+  ; CHECK-NEXT:   $sgpr13 = COPY [[COPY13]](s32)
+  ; CHECK-NEXT:   $sgpr14 = COPY [[COPY14]](s32)
+  ; CHECK-NEXT:   $vgpr31 = COPY [[COPY15]](s32)
   ; CHECK-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @external_void_func_12xv3i32, csr_amdgpu_highregs, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15, implicit $vgpr16, implicit $vgpr17, implicit $vgpr18, implicit $vgpr19, implicit $vgpr20, implicit $vgpr21, implicit $vgpr22, implicit $vgpr23, implicit $vgpr24, implicit $vgpr25, implicit $vgpr26, implicit $vgpr27, implicit $vgpr28, implicit $vgpr29, implicit $vgpr30, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr4_sgpr5, implicit $sgpr6_sgpr7, implicit $sgpr8_sgpr9, implicit $sgpr10_sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $vgpr31
   ; CHECK-NEXT:   ADJCALLSTACKDOWN 0, 20, implicit-def $scc
-  ; CHECK-NEXT:   [[COPY19:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY8]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY19]]
+  ; CHECK-NEXT:   SI_RETURN
 entry:
   call void @external_void_func_12xv3i32(
       <3 x i32> <i32 0, i32 0, i32 0>,
       <3 x i32> <i32 1, i32 1, i32 1>,
       <3 x i32> <i32 2, i32 2, i32 2>,
       <3 x i32> <i32 3, i32 3, i32 3>,
       <3 x i32> <i32 4, i32 4, i32 4>,
       <3 x i32> <i32 5, i32 5, i32 5>,
       <3 x i32> <i32 6, i32 6, i32 6>,
       <3 x i32> <i32 7, i32 7, i32 7>,
       <3 x i32> <i32 8, i32 8, i32 8>,
       <3 x i32> <i32 9, i32 9, i32 9>,
       <3 x i32> <i32 10, i32 11, i32 12>,
       <3 x i32> <i32 13, i32 14, i32 15>)
   ret void
 }
 
 define void @stack_12xv3f32() #0 {
   ; CHECK-LABEL: name: stack_12xv3f32
   ; CHECK: bb.1.entry:
-  ; CHECK-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:vgpr_32(s32) = COPY $vgpr31
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr_32 = COPY $sgpr14
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr_32 = COPY $sgpr13
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr_32 = COPY $sgpr12
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr10_sgpr11
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:sgpr_64 = COPY $sgpr8_sgpr9
   ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr6_sgpr7
   ; CHECK-NEXT:   [[COPY7:%[0-9]+]]:sgpr_64 = COPY $sgpr4_sgpr5
-  ; CHECK-NEXT:   [[COPY8:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 0.000000e+00
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[C]](s32), [[C]](s32), [[C]](s32)
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00
   ; CHECK-NEXT:   [[BUILD_VECTOR1:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[C1]](s32), [[C1]](s32), [[C1]](s32)
   ; CHECK-NEXT:   [[C2:%[0-9]+]]:_(s32) = G_FCONSTANT float 2.000000e+00
   ; CHECK-NEXT:   [[BUILD_VECTOR2:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[C2]](s32), [[C2]](s32), [[C2]](s32)
   ; CHECK-NEXT:   [[C3:%[0-9]+]]:_(s32) = G_FCONSTANT float 3.000000e+00
   ; CHECK-NEXT:   [[BUILD_VECTOR3:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[C3]](s32), [[C3]](s32), [[C3]](s32)
define void @stack_12xv3f32() #0 {
   ; CHECK-NEXT:   [[C12:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.200000e+01
   ; CHECK-NEXT:   [[BUILD_VECTOR10:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[C10]](s32), [[C11]](s32), [[C12]](s32)
   ; CHECK-NEXT:   [[C13:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.300000e+01
   ; CHECK-NEXT:   [[C14:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.400000e+01
   ; CHECK-NEXT:   [[C15:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.500000e+01
   ; CHECK-NEXT:   [[BUILD_VECTOR11:%[0-9]+]]:_(<3 x s32>) = G_BUILD_VECTOR [[C13]](s32), [[C14]](s32), [[C15]](s32)
   ; CHECK-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; CHECK-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @external_void_func_12xv3f32
-  ; CHECK-NEXT:   [[COPY9:%[0-9]+]]:_(p4) = COPY [[COPY7]]
-  ; CHECK-NEXT:   [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY6]]
-  ; CHECK-NEXT:   [[COPY11:%[0-9]+]]:_(p4) = COPY [[COPY5]]
-  ; CHECK-NEXT:   [[COPY12:%[0-9]+]]:_(s64) = COPY [[COPY4]]
-  ; CHECK-NEXT:   [[COPY13:%[0-9]+]]:_(s32) = COPY [[COPY3]]
-  ; CHECK-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY2]]
-  ; CHECK-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY1]]
-  ; CHECK-NEXT:   [[COPY16:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
+  ; CHECK-NEXT:   [[COPY8:%[0-9]+]]:_(p4) = COPY [[COPY7]]
+  ; CHECK-NEXT:   [[COPY9:%[0-9]+]]:_(p4) = COPY [[COPY6]]
+  ; CHECK-NEXT:   [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY5]]
+  ; CHECK-NEXT:   [[COPY11:%[0-9]+]]:_(s64) = COPY [[COPY4]]
+  ; CHECK-NEXT:   [[COPY12:%[0-9]+]]:_(s32) = COPY [[COPY3]]
+  ; CHECK-NEXT:   [[COPY13:%[0-9]+]]:_(s32) = COPY [[COPY2]]
+  ; CHECK-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY1]]
+  ; CHECK-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
   ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR]](<3 x s32>)
   ; CHECK-NEXT:   [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR1]](<3 x s32>)
   ; CHECK-NEXT:   [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32), [[UV8:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR2]](<3 x s32>)
   ; CHECK-NEXT:   [[UV9:%[0-9]+]]:_(s32), [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR3]](<3 x s32>)
   ; CHECK-NEXT:   [[UV12:%[0-9]+]]:_(s32), [[UV13:%[0-9]+]]:_(s32), [[UV14:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR4]](<3 x s32>)
   ; CHECK-NEXT:   [[UV15:%[0-9]+]]:_(s32), [[UV16:%[0-9]+]]:_(s32), [[UV17:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR5]](<3 x s32>)
   ; CHECK-NEXT:   [[UV18:%[0-9]+]]:_(s32), [[UV19:%[0-9]+]]:_(s32), [[UV20:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR6]](<3 x s32>)
   ; CHECK-NEXT:   [[UV21:%[0-9]+]]:_(s32), [[UV22:%[0-9]+]]:_(s32), [[UV23:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR7]](<3 x s32>)
   ; CHECK-NEXT:   [[UV24:%[0-9]+]]:_(s32), [[UV25:%[0-9]+]]:_(s32), [[UV26:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR8]](<3 x s32>)
   ; CHECK-NEXT:   [[UV27:%[0-9]+]]:_(s32), [[UV28:%[0-9]+]]:_(s32), [[UV29:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR9]](<3 x s32>)
   ; CHECK-NEXT:   [[UV30:%[0-9]+]]:_(s32), [[UV31:%[0-9]+]]:_(s32), [[UV32:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR10]](<3 x s32>)
-  ; CHECK-NEXT:   [[COPY17:%[0-9]+]]:_(p5) = COPY $sgpr32
+  ; CHECK-NEXT:   [[COPY16:%[0-9]+]]:_(p5) = COPY $sgpr32
   ; CHECK-NEXT:   [[C16:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C16]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY16]], [[C16]](s32)
   ; CHECK-NEXT:   G_STORE [[UV31]](s32), [[PTR_ADD]](p5) :: (store (s32) into stack, align 16, addrspace 5)
   ; CHECK-NEXT:   [[C17:%[0-9]+]]:_(s32) = G_CONSTANT i32 4
-  ; CHECK-NEXT:   [[PTR_ADD1:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C17]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD1:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY16]], [[C17]](s32)
   ; CHECK-NEXT:   G_STORE [[UV32]](s32), [[PTR_ADD1]](p5) :: (store (s32) into stack + 4, addrspace 5)
   ; CHECK-NEXT:   [[UV33:%[0-9]+]]:_(s32), [[UV34:%[0-9]+]]:_(s32), [[UV35:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR11]](<3 x s32>)
   ; CHECK-NEXT:   [[C18:%[0-9]+]]:_(s32) = G_CONSTANT i32 8
-  ; CHECK-NEXT:   [[PTR_ADD2:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C18]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD2:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY16]], [[C18]](s32)
   ; CHECK-NEXT:   G_STORE [[UV33]](s32), [[PTR_ADD2]](p5) :: (store (s32) into stack + 8, align 8, addrspace 5)
   ; CHECK-NEXT:   [[C19:%[0-9]+]]:_(s32) = G_CONSTANT i32 12
-  ; CHECK-NEXT:   [[PTR_ADD3:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C19]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD3:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY16]], [[C19]](s32)
   ; CHECK-NEXT:   G_STORE [[UV34]](s32), [[PTR_ADD3]](p5) :: (store (s32) into stack + 12, addrspace 5)
   ; CHECK-NEXT:   [[C20:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
-  ; CHECK-NEXT:   [[PTR_ADD4:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C20]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD4:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY16]], [[C20]](s32)
   ; CHECK-NEXT:   G_STORE [[UV35]](s32), [[PTR_ADD4]](p5) :: (store (s32) into stack + 16, align 16, addrspace 5)
   ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
   ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
   ; CHECK-NEXT:   $vgpr4 = COPY [[UV4]](s32)
   ; CHECK-NEXT:   $vgpr5 = COPY [[UV5]](s32)
   ; CHECK-NEXT:   $vgpr6 = COPY [[UV6]](s32)
define void @stack_12xv3f32() #0 {
   ; CHECK-NEXT:   $vgpr23 = COPY [[UV23]](s32)
   ; CHECK-NEXT:   $vgpr24 = COPY [[UV24]](s32)
   ; CHECK-NEXT:   $vgpr25 = COPY [[UV25]](s32)
   ; CHECK-NEXT:   $vgpr26 = COPY [[UV26]](s32)
   ; CHECK-NEXT:   $vgpr27 = COPY [[UV27]](s32)
   ; CHECK-NEXT:   $vgpr28 = COPY [[UV28]](s32)
   ; CHECK-NEXT:   $vgpr29 = COPY [[UV29]](s32)
   ; CHECK-NEXT:   $vgpr30 = COPY [[UV30]](s32)
-  ; CHECK-NEXT:   [[COPY18:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY18]](<4 x s32>)
-  ; CHECK-NEXT:   $sgpr4_sgpr5 = COPY [[COPY9]](p4)
-  ; CHECK-NEXT:   $sgpr6_sgpr7 = COPY [[COPY10]](p4)
-  ; CHECK-NEXT:   $sgpr8_sgpr9 = COPY [[COPY11]](p4)
-  ; CHECK-NEXT:   $sgpr10_sgpr11 = COPY [[COPY12]](s64)
-  ; CHECK-NEXT:   $sgpr12 = COPY [[COPY13]](s32)
-  ; CHECK-NEXT:   $sgpr13 = COPY [[COPY14]](s32)
-  ; CHECK-NEXT:   $sgpr14 = COPY [[COPY15]](s32)
-  ; CHECK-NEXT:   $vgpr31 = COPY [[COPY16]](s32)
+  ; CHECK-NEXT:   [[COPY17:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY17]](<4 x s32>)
+  ; CHECK-NEXT:   $sgpr4_sgpr5 = COPY [[COPY8]](p4)
+  ; CHECK-NEXT:   $sgpr6_sgpr7 = COPY [[COPY9]](p4)
+  ; CHECK-NEXT:   $sgpr8_sgpr9 = COPY [[COPY10]](p4)
+  ; CHECK-NEXT:   $sgpr10_sgpr11 = COPY [[COPY11]](s64)
+  ; CHECK-NEXT:   $sgpr12 = COPY [[COPY12]](s32)
+  ; CHECK-NEXT:   $sgpr13 = COPY [[COPY13]](s32)
+  ; CHECK-NEXT:   $sgpr14 = COPY [[COPY14]](s32)
+  ; CHECK-NEXT:   $vgpr31 = COPY [[COPY15]](s32)
   ; CHECK-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @external_void_func_12xv3f32, csr_amdgpu_highregs, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15, implicit $vgpr16, implicit $vgpr17, implicit $vgpr18, implicit $vgpr19, implicit $vgpr20, implicit $vgpr21, implicit $vgpr22, implicit $vgpr23, implicit $vgpr24, implicit $vgpr25, implicit $vgpr26, implicit $vgpr27, implicit $vgpr28, implicit $vgpr29, implicit $vgpr30, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr4_sgpr5, implicit $sgpr6_sgpr7, implicit $sgpr8_sgpr9, implicit $sgpr10_sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $vgpr31
   ; CHECK-NEXT:   ADJCALLSTACKDOWN 0, 20, implicit-def $scc
-  ; CHECK-NEXT:   [[COPY19:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY8]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY19]]
+  ; CHECK-NEXT:   SI_RETURN
 entry:
   call void @external_void_func_12xv3f32(
       <3 x float> <float 0.0, float 0.0, float 0.0>,
       <3 x float> <float 1.0, float 1.0, float 1.0>,
       <3 x float> <float 2.0, float 2.0, float 2.0>,
       <3 x float> <float 3.0, float 3.0, float 3.0>,
       <3 x float> <float 4.0, float 4.0, float 4.0>,
       <3 x float> <float 5.0, float 5.0, float 5.0>,
       <3 x float> <float 6.0, float 6.0, float 6.0>,
       <3 x float> <float 7.0, float 7.0, float 7.0>,
       <3 x float> <float 8.0, float 8.0, float 8.0>,
       <3 x float> <float 9.0, float 9.0, float 9.0>,
       <3 x float> <float 10.0, float 11.0, float 12.0>,
       <3 x float> <float 13.0, float 14.0, float 15.0>)
   ret void
 }
 
 define void @stack_8xv5i32() #0 {
   ; CHECK-LABEL: name: stack_8xv5i32
   ; CHECK: bb.1.entry:
-  ; CHECK-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:vgpr_32(s32) = COPY $vgpr31
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr_32 = COPY $sgpr14
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr_32 = COPY $sgpr13
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr_32 = COPY $sgpr12
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr10_sgpr11
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:sgpr_64 = COPY $sgpr8_sgpr9
   ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr6_sgpr7
   ; CHECK-NEXT:   [[COPY7:%[0-9]+]]:sgpr_64 = COPY $sgpr4_sgpr5
-  ; CHECK-NEXT:   [[COPY8:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:_(<5 x s32>) = G_BUILD_VECTOR [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32)
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 1
   ; CHECK-NEXT:   [[BUILD_VECTOR1:%[0-9]+]]:_(<5 x s32>) = G_BUILD_VECTOR [[C1]](s32), [[C1]](s32), [[C1]](s32), [[C1]](s32), [[C1]](s32)
   ; CHECK-NEXT:   [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 2
   ; CHECK-NEXT:   [[BUILD_VECTOR2:%[0-9]+]]:_(<5 x s32>) = G_BUILD_VECTOR [[C2]](s32), [[C2]](s32), [[C2]](s32), [[C2]](s32), [[C2]](s32)
   ; CHECK-NEXT:   [[C3:%[0-9]+]]:_(s32) = G_CONSTANT i32 3
   ; CHECK-NEXT:   [[BUILD_VECTOR3:%[0-9]+]]:_(<5 x s32>) = G_BUILD_VECTOR [[C3]](s32), [[C3]](s32), [[C3]](s32), [[C3]](s32), [[C3]](s32)
define void @stack_8xv5i32() #0 {
   ; CHECK-NEXT:   [[C11:%[0-9]+]]:_(s32) = G_CONSTANT i32 11
   ; CHECK-NEXT:   [[C12:%[0-9]+]]:_(s32) = G_CONSTANT i32 12
   ; CHECK-NEXT:   [[C13:%[0-9]+]]:_(s32) = G_CONSTANT i32 13
   ; CHECK-NEXT:   [[C14:%[0-9]+]]:_(s32) = G_CONSTANT i32 14
   ; CHECK-NEXT:   [[C15:%[0-9]+]]:_(s32) = G_CONSTANT i32 15
   ; CHECK-NEXT:   [[BUILD_VECTOR7:%[0-9]+]]:_(<5 x s32>) = G_BUILD_VECTOR [[C11]](s32), [[C12]](s32), [[C13]](s32), [[C14]](s32), [[C15]](s32)
   ; CHECK-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; CHECK-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @external_void_func_8xv5i32
-  ; CHECK-NEXT:   [[COPY9:%[0-9]+]]:_(p4) = COPY [[COPY7]]
-  ; CHECK-NEXT:   [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY6]]
-  ; CHECK-NEXT:   [[COPY11:%[0-9]+]]:_(p4) = COPY [[COPY5]]
-  ; CHECK-NEXT:   [[COPY12:%[0-9]+]]:_(s64) = COPY [[COPY4]]
-  ; CHECK-NEXT:   [[COPY13:%[0-9]+]]:_(s32) = COPY [[COPY3]]
-  ; CHECK-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY2]]
-  ; CHECK-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY1]]
-  ; CHECK-NEXT:   [[COPY16:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
+  ; CHECK-NEXT:   [[COPY8:%[0-9]+]]:_(p4) = COPY [[COPY7]]
+  ; CHECK-NEXT:   [[COPY9:%[0-9]+]]:_(p4) = COPY [[COPY6]]
+  ; CHECK-NEXT:   [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY5]]
+  ; CHECK-NEXT:   [[COPY11:%[0-9]+]]:_(s64) = COPY [[COPY4]]
+  ; CHECK-NEXT:   [[COPY12:%[0-9]+]]:_(s32) = COPY [[COPY3]]
+  ; CHECK-NEXT:   [[COPY13:%[0-9]+]]:_(s32) = COPY [[COPY2]]
+  ; CHECK-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY1]]
+  ; CHECK-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
   ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR]](<5 x s32>)
   ; CHECK-NEXT:   [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32), [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR1]](<5 x s32>)
   ; CHECK-NEXT:   [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32), [[UV12:%[0-9]+]]:_(s32), [[UV13:%[0-9]+]]:_(s32), [[UV14:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR2]](<5 x s32>)
   ; CHECK-NEXT:   [[UV15:%[0-9]+]]:_(s32), [[UV16:%[0-9]+]]:_(s32), [[UV17:%[0-9]+]]:_(s32), [[UV18:%[0-9]+]]:_(s32), [[UV19:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR3]](<5 x s32>)
   ; CHECK-NEXT:   [[UV20:%[0-9]+]]:_(s32), [[UV21:%[0-9]+]]:_(s32), [[UV22:%[0-9]+]]:_(s32), [[UV23:%[0-9]+]]:_(s32), [[UV24:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR4]](<5 x s32>)
   ; CHECK-NEXT:   [[UV25:%[0-9]+]]:_(s32), [[UV26:%[0-9]+]]:_(s32), [[UV27:%[0-9]+]]:_(s32), [[UV28:%[0-9]+]]:_(s32), [[UV29:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR5]](<5 x s32>)
   ; CHECK-NEXT:   [[UV30:%[0-9]+]]:_(s32), [[UV31:%[0-9]+]]:_(s32), [[UV32:%[0-9]+]]:_(s32), [[UV33:%[0-9]+]]:_(s32), [[UV34:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR6]](<5 x s32>)
-  ; CHECK-NEXT:   [[COPY17:%[0-9]+]]:_(p5) = COPY $sgpr32
+  ; CHECK-NEXT:   [[COPY16:%[0-9]+]]:_(p5) = COPY $sgpr32
   ; CHECK-NEXT:   [[C16:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C16]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY16]], [[C16]](s32)
   ; CHECK-NEXT:   G_STORE [[UV31]](s32), [[PTR_ADD]](p5) :: (store (s32) into stack, align 16, addrspace 5)
   ; CHECK-NEXT:   [[C17:%[0-9]+]]:_(s32) = G_CONSTANT i32 4
-  ; CHECK-NEXT:   [[PTR_ADD1:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C17]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD1:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY16]], [[C17]](s32)
   ; CHECK-NEXT:   G_STORE [[UV32]](s32), [[PTR_ADD1]](p5) :: (store (s32) into stack + 4, addrspace 5)
   ; CHECK-NEXT:   [[C18:%[0-9]+]]:_(s32) = G_CONSTANT i32 8
-  ; CHECK-NEXT:   [[PTR_ADD2:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C18]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD2:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY16]], [[C18]](s32)
   ; CHECK-NEXT:   G_STORE [[UV33]](s32), [[PTR_ADD2]](p5) :: (store (s32) into stack + 8, align 8, addrspace 5)
   ; CHECK-NEXT:   [[C19:%[0-9]+]]:_(s32) = G_CONSTANT i32 12
-  ; CHECK-NEXT:   [[PTR_ADD3:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C19]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD3:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY16]], [[C19]](s32)
   ; CHECK-NEXT:   G_STORE [[UV34]](s32), [[PTR_ADD3]](p5) :: (store (s32) into stack + 12, addrspace 5)
   ; CHECK-NEXT:   [[UV35:%[0-9]+]]:_(s32), [[UV36:%[0-9]+]]:_(s32), [[UV37:%[0-9]+]]:_(s32), [[UV38:%[0-9]+]]:_(s32), [[UV39:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR7]](<5 x s32>)
   ; CHECK-NEXT:   [[C20:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
-  ; CHECK-NEXT:   [[PTR_ADD4:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C20]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD4:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY16]], [[C20]](s32)
   ; CHECK-NEXT:   G_STORE [[UV35]](s32), [[PTR_ADD4]](p5) :: (store (s32) into stack + 16, align 16, addrspace 5)
   ; CHECK-NEXT:   [[C21:%[0-9]+]]:_(s32) = G_CONSTANT i32 20
-  ; CHECK-NEXT:   [[PTR_ADD5:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C21]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD5:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY16]], [[C21]](s32)
   ; CHECK-NEXT:   G_STORE [[UV36]](s32), [[PTR_ADD5]](p5) :: (store (s32) into stack + 20, addrspace 5)
   ; CHECK-NEXT:   [[C22:%[0-9]+]]:_(s32) = G_CONSTANT i32 24
-  ; CHECK-NEXT:   [[PTR_ADD6:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C22]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD6:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY16]], [[C22]](s32)
   ; CHECK-NEXT:   G_STORE [[UV37]](s32), [[PTR_ADD6]](p5) :: (store (s32) into stack + 24, align 8, addrspace 5)
   ; CHECK-NEXT:   [[C23:%[0-9]+]]:_(s32) = G_CONSTANT i32 28
-  ; CHECK-NEXT:   [[PTR_ADD7:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C23]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD7:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY16]], [[C23]](s32)
   ; CHECK-NEXT:   G_STORE [[UV38]](s32), [[PTR_ADD7]](p5) :: (store (s32) into stack + 28, addrspace 5)
   ; CHECK-NEXT:   [[C24:%[0-9]+]]:_(s32) = G_CONSTANT i32 32
-  ; CHECK-NEXT:   [[PTR_ADD8:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C24]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD8:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY16]], [[C24]](s32)
   ; CHECK-NEXT:   G_STORE [[UV39]](s32), [[PTR_ADD8]](p5) :: (store (s32) into stack + 32, align 16, addrspace 5)
   ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
   ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
   ; CHECK-NEXT:   $vgpr4 = COPY [[UV4]](s32)
   ; CHECK-NEXT:   $vgpr5 = COPY [[UV5]](s32)
   ; CHECK-NEXT:   $vgpr6 = COPY [[UV6]](s32)
define void @stack_8xv5i32() #0 {
   ; CHECK-NEXT:   $vgpr23 = COPY [[UV23]](s32)
   ; CHECK-NEXT:   $vgpr24 = COPY [[UV24]](s32)
   ; CHECK-NEXT:   $vgpr25 = COPY [[UV25]](s32)
   ; CHECK-NEXT:   $vgpr26 = COPY [[UV26]](s32)
   ; CHECK-NEXT:   $vgpr27 = COPY [[UV27]](s32)
   ; CHECK-NEXT:   $vgpr28 = COPY [[UV28]](s32)
   ; CHECK-NEXT:   $vgpr29 = COPY [[UV29]](s32)
   ; CHECK-NEXT:   $vgpr30 = COPY [[UV30]](s32)
-  ; CHECK-NEXT:   [[COPY18:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY18]](<4 x s32>)
-  ; CHECK-NEXT:   $sgpr4_sgpr5 = COPY [[COPY9]](p4)
-  ; CHECK-NEXT:   $sgpr6_sgpr7 = COPY [[COPY10]](p4)
-  ; CHECK-NEXT:   $sgpr8_sgpr9 = COPY [[COPY11]](p4)
-  ; CHECK-NEXT:   $sgpr10_sgpr11 = COPY [[COPY12]](s64)
-  ; CHECK-NEXT:   $sgpr12 = COPY [[COPY13]](s32)
-  ; CHECK-NEXT:   $sgpr13 = COPY [[COPY14]](s32)
-  ; CHECK-NEXT:   $sgpr14 = COPY [[COPY15]](s32)
-  ; CHECK-NEXT:   $vgpr31 = COPY [[COPY16]](s32)
+  ; CHECK-NEXT:   [[COPY17:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY17]](<4 x s32>)
+  ; CHECK-NEXT:   $sgpr4_sgpr5 = COPY [[COPY8]](p4)
+  ; CHECK-NEXT:   $sgpr6_sgpr7 = COPY [[COPY9]](p4)
+  ; CHECK-NEXT:   $sgpr8_sgpr9 = COPY [[COPY10]](p4)
+  ; CHECK-NEXT:   $sgpr10_sgpr11 = COPY [[COPY11]](s64)
+  ; CHECK-NEXT:   $sgpr12 = COPY [[COPY12]](s32)
+  ; CHECK-NEXT:   $sgpr13 = COPY [[COPY13]](s32)
+  ; CHECK-NEXT:   $sgpr14 = COPY [[COPY14]](s32)
+  ; CHECK-NEXT:   $vgpr31 = COPY [[COPY15]](s32)
   ; CHECK-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @external_void_func_8xv5i32, csr_amdgpu_highregs, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15, implicit $vgpr16, implicit $vgpr17, implicit $vgpr18, implicit $vgpr19, implicit $vgpr20, implicit $vgpr21, implicit $vgpr22, implicit $vgpr23, implicit $vgpr24, implicit $vgpr25, implicit $vgpr26, implicit $vgpr27, implicit $vgpr28, implicit $vgpr29, implicit $vgpr30, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr4_sgpr5, implicit $sgpr6_sgpr7, implicit $sgpr8_sgpr9, implicit $sgpr10_sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $vgpr31
   ; CHECK-NEXT:   ADJCALLSTACKDOWN 0, 36, implicit-def $scc
-  ; CHECK-NEXT:   [[COPY19:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY8]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY19]]
+  ; CHECK-NEXT:   SI_RETURN
 entry:
   call void @external_void_func_8xv5i32(
       <5 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0>,
       <5 x i32> <i32 1, i32 1, i32 1, i32 1, i32 1>,
       <5 x i32> <i32 2, i32 2, i32 2, i32 2, i32 2>,
       <5 x i32> <i32 3, i32 3, i32 3, i32 3, i32 3>,
       <5 x i32> <i32 4, i32 4, i32 4, i32 4, i32 4>,
       <5 x i32> <i32 5, i32 5, i32 5, i32 5, i32 5>,
       <5 x i32> <i32 6, i32 7, i32 8, i32 9, i32 10>,
       <5 x i32> <i32 11, i32 12, i32 13, i32 14, i32 15>)
   ret void
 }
 
 define void @stack_8xv5f32() #0 {
   ; CHECK-LABEL: name: stack_8xv5f32
   ; CHECK: bb.1.entry:
-  ; CHECK-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:vgpr_32(s32) = COPY $vgpr31
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr_32 = COPY $sgpr14
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr_32 = COPY $sgpr13
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr_32 = COPY $sgpr12
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr10_sgpr11
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:sgpr_64 = COPY $sgpr8_sgpr9
   ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr6_sgpr7
   ; CHECK-NEXT:   [[COPY7:%[0-9]+]]:sgpr_64 = COPY $sgpr4_sgpr5
-  ; CHECK-NEXT:   [[COPY8:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_FCONSTANT float 0.000000e+00
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:_(<5 x s32>) = G_BUILD_VECTOR [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32), [[C]](s32)
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.000000e+00
   ; CHECK-NEXT:   [[BUILD_VECTOR1:%[0-9]+]]:_(<5 x s32>) = G_BUILD_VECTOR [[C1]](s32), [[C1]](s32), [[C1]](s32), [[C1]](s32), [[C1]](s32)
   ; CHECK-NEXT:   [[C2:%[0-9]+]]:_(s32) = G_FCONSTANT float 2.000000e+00
   ; CHECK-NEXT:   [[BUILD_VECTOR2:%[0-9]+]]:_(<5 x s32>) = G_BUILD_VECTOR [[C2]](s32), [[C2]](s32), [[C2]](s32), [[C2]](s32), [[C2]](s32)
   ; CHECK-NEXT:   [[C3:%[0-9]+]]:_(s32) = G_FCONSTANT float 3.000000e+00
   ; CHECK-NEXT:   [[BUILD_VECTOR3:%[0-9]+]]:_(<5 x s32>) = G_BUILD_VECTOR [[C3]](s32), [[C3]](s32), [[C3]](s32), [[C3]](s32), [[C3]](s32)
define void @stack_8xv5f32() #0 {
   ; CHECK-NEXT:   [[C11:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.100000e+01
   ; CHECK-NEXT:   [[C12:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.200000e+01
   ; CHECK-NEXT:   [[C13:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.300000e+01
   ; CHECK-NEXT:   [[C14:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.400000e+01
   ; CHECK-NEXT:   [[C15:%[0-9]+]]:_(s32) = G_FCONSTANT float 1.500000e+01
   ; CHECK-NEXT:   [[BUILD_VECTOR7:%[0-9]+]]:_(<5 x s32>) = G_BUILD_VECTOR [[C11]](s32), [[C12]](s32), [[C13]](s32), [[C14]](s32), [[C15]](s32)
   ; CHECK-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; CHECK-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @external_void_func_8xv5f32
-  ; CHECK-NEXT:   [[COPY9:%[0-9]+]]:_(p4) = COPY [[COPY7]]
-  ; CHECK-NEXT:   [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY6]]
-  ; CHECK-NEXT:   [[COPY11:%[0-9]+]]:_(p4) = COPY [[COPY5]]
-  ; CHECK-NEXT:   [[COPY12:%[0-9]+]]:_(s64) = COPY [[COPY4]]
-  ; CHECK-NEXT:   [[COPY13:%[0-9]+]]:_(s32) = COPY [[COPY3]]
-  ; CHECK-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY2]]
-  ; CHECK-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY1]]
-  ; CHECK-NEXT:   [[COPY16:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
+  ; CHECK-NEXT:   [[COPY8:%[0-9]+]]:_(p4) = COPY [[COPY7]]
+  ; CHECK-NEXT:   [[COPY9:%[0-9]+]]:_(p4) = COPY [[COPY6]]
+  ; CHECK-NEXT:   [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY5]]
+  ; CHECK-NEXT:   [[COPY11:%[0-9]+]]:_(s64) = COPY [[COPY4]]
+  ; CHECK-NEXT:   [[COPY12:%[0-9]+]]:_(s32) = COPY [[COPY3]]
+  ; CHECK-NEXT:   [[COPY13:%[0-9]+]]:_(s32) = COPY [[COPY2]]
+  ; CHECK-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY1]]
+  ; CHECK-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
   ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32), [[UV4:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR]](<5 x s32>)
   ; CHECK-NEXT:   [[UV5:%[0-9]+]]:_(s32), [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32), [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR1]](<5 x s32>)
   ; CHECK-NEXT:   [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32), [[UV12:%[0-9]+]]:_(s32), [[UV13:%[0-9]+]]:_(s32), [[UV14:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR2]](<5 x s32>)
   ; CHECK-NEXT:   [[UV15:%[0-9]+]]:_(s32), [[UV16:%[0-9]+]]:_(s32), [[UV17:%[0-9]+]]:_(s32), [[UV18:%[0-9]+]]:_(s32), [[UV19:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR3]](<5 x s32>)
   ; CHECK-NEXT:   [[UV20:%[0-9]+]]:_(s32), [[UV21:%[0-9]+]]:_(s32), [[UV22:%[0-9]+]]:_(s32), [[UV23:%[0-9]+]]:_(s32), [[UV24:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR4]](<5 x s32>)
   ; CHECK-NEXT:   [[UV25:%[0-9]+]]:_(s32), [[UV26:%[0-9]+]]:_(s32), [[UV27:%[0-9]+]]:_(s32), [[UV28:%[0-9]+]]:_(s32), [[UV29:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR5]](<5 x s32>)
   ; CHECK-NEXT:   [[UV30:%[0-9]+]]:_(s32), [[UV31:%[0-9]+]]:_(s32), [[UV32:%[0-9]+]]:_(s32), [[UV33:%[0-9]+]]:_(s32), [[UV34:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR6]](<5 x s32>)
-  ; CHECK-NEXT:   [[COPY17:%[0-9]+]]:_(p5) = COPY $sgpr32
+  ; CHECK-NEXT:   [[COPY16:%[0-9]+]]:_(p5) = COPY $sgpr32
   ; CHECK-NEXT:   [[C16:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
-  ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C16]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY16]], [[C16]](s32)
   ; CHECK-NEXT:   G_STORE [[UV31]](s32), [[PTR_ADD]](p5) :: (store (s32) into stack, align 16, addrspace 5)
   ; CHECK-NEXT:   [[C17:%[0-9]+]]:_(s32) = G_CONSTANT i32 4
-  ; CHECK-NEXT:   [[PTR_ADD1:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C17]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD1:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY16]], [[C17]](s32)
   ; CHECK-NEXT:   G_STORE [[UV32]](s32), [[PTR_ADD1]](p5) :: (store (s32) into stack + 4, addrspace 5)
   ; CHECK-NEXT:   [[C18:%[0-9]+]]:_(s32) = G_CONSTANT i32 8
-  ; CHECK-NEXT:   [[PTR_ADD2:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C18]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD2:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY16]], [[C18]](s32)
   ; CHECK-NEXT:   G_STORE [[UV33]](s32), [[PTR_ADD2]](p5) :: (store (s32) into stack + 8, align 8, addrspace 5)
   ; CHECK-NEXT:   [[C19:%[0-9]+]]:_(s32) = G_CONSTANT i32 12
-  ; CHECK-NEXT:   [[PTR_ADD3:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C19]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD3:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY16]], [[C19]](s32)
   ; CHECK-NEXT:   G_STORE [[UV34]](s32), [[PTR_ADD3]](p5) :: (store (s32) into stack + 12, addrspace 5)
   ; CHECK-NEXT:   [[UV35:%[0-9]+]]:_(s32), [[UV36:%[0-9]+]]:_(s32), [[UV37:%[0-9]+]]:_(s32), [[UV38:%[0-9]+]]:_(s32), [[UV39:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR7]](<5 x s32>)
   ; CHECK-NEXT:   [[C20:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
-  ; CHECK-NEXT:   [[PTR_ADD4:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C20]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD4:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY16]], [[C20]](s32)
   ; CHECK-NEXT:   G_STORE [[UV35]](s32), [[PTR_ADD4]](p5) :: (store (s32) into stack + 16, align 16, addrspace 5)
   ; CHECK-NEXT:   [[C21:%[0-9]+]]:_(s32) = G_CONSTANT i32 20
-  ; CHECK-NEXT:   [[PTR_ADD5:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C21]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD5:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY16]], [[C21]](s32)
   ; CHECK-NEXT:   G_STORE [[UV36]](s32), [[PTR_ADD5]](p5) :: (store (s32) into stack + 20, addrspace 5)
   ; CHECK-NEXT:   [[C22:%[0-9]+]]:_(s32) = G_CONSTANT i32 24
-  ; CHECK-NEXT:   [[PTR_ADD6:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C22]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD6:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY16]], [[C22]](s32)
   ; CHECK-NEXT:   G_STORE [[UV37]](s32), [[PTR_ADD6]](p5) :: (store (s32) into stack + 24, align 8, addrspace 5)
   ; CHECK-NEXT:   [[C23:%[0-9]+]]:_(s32) = G_CONSTANT i32 28
-  ; CHECK-NEXT:   [[PTR_ADD7:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C23]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD7:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY16]], [[C23]](s32)
   ; CHECK-NEXT:   G_STORE [[UV38]](s32), [[PTR_ADD7]](p5) :: (store (s32) into stack + 28, addrspace 5)
   ; CHECK-NEXT:   [[C24:%[0-9]+]]:_(s32) = G_CONSTANT i32 32
-  ; CHECK-NEXT:   [[PTR_ADD8:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY17]], [[C24]](s32)
+  ; CHECK-NEXT:   [[PTR_ADD8:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY16]], [[C24]](s32)
   ; CHECK-NEXT:   G_STORE [[UV39]](s32), [[PTR_ADD8]](p5) :: (store (s32) into stack + 32, align 16, addrspace 5)
   ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
   ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
   ; CHECK-NEXT:   $vgpr4 = COPY [[UV4]](s32)
   ; CHECK-NEXT:   $vgpr5 = COPY [[UV5]](s32)
   ; CHECK-NEXT:   $vgpr6 = COPY [[UV6]](s32)
define void @stack_8xv5f32() #0 {
   ; CHECK-NEXT:   $vgpr23 = COPY [[UV23]](s32)
   ; CHECK-NEXT:   $vgpr24 = COPY [[UV24]](s32)
   ; CHECK-NEXT:   $vgpr25 = COPY [[UV25]](s32)
   ; CHECK-NEXT:   $vgpr26 = COPY [[UV26]](s32)
   ; CHECK-NEXT:   $vgpr27 = COPY [[UV27]](s32)
   ; CHECK-NEXT:   $vgpr28 = COPY [[UV28]](s32)
   ; CHECK-NEXT:   $vgpr29 = COPY [[UV29]](s32)
   ; CHECK-NEXT:   $vgpr30 = COPY [[UV30]](s32)
-  ; CHECK-NEXT:   [[COPY18:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY18]](<4 x s32>)
-  ; CHECK-NEXT:   $sgpr4_sgpr5 = COPY [[COPY9]](p4)
-  ; CHECK-NEXT:   $sgpr6_sgpr7 = COPY [[COPY10]](p4)
-  ; CHECK-NEXT:   $sgpr8_sgpr9 = COPY [[COPY11]](p4)
-  ; CHECK-NEXT:   $sgpr10_sgpr11 = COPY [[COPY12]](s64)
-  ; CHECK-NEXT:   $sgpr12 = COPY [[COPY13]](s32)
-  ; CHECK-NEXT:   $sgpr13 = COPY [[COPY14]](s32)
-  ; CHECK-NEXT:   $sgpr14 = COPY [[COPY15]](s32)
-  ; CHECK-NEXT:   $vgpr31 = COPY [[COPY16]](s32)
+  ; CHECK-NEXT:   [[COPY17:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY17]](<4 x s32>)
+  ; CHECK-NEXT:   $sgpr4_sgpr5 = COPY [[COPY8]](p4)
+  ; CHECK-NEXT:   $sgpr6_sgpr7 = COPY [[COPY9]](p4)
+  ; CHECK-NEXT:   $sgpr8_sgpr9 = COPY [[COPY10]](p4)
+  ; CHECK-NEXT:   $sgpr10_sgpr11 = COPY [[COPY11]](s64)
+  ; CHECK-NEXT:   $sgpr12 = COPY [[COPY12]](s32)
+  ; CHECK-NEXT:   $sgpr13 = COPY [[COPY13]](s32)
+  ; CHECK-NEXT:   $sgpr14 = COPY [[COPY14]](s32)
+  ; CHECK-NEXT:   $vgpr31 = COPY [[COPY15]](s32)
   ; CHECK-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @external_void_func_8xv5f32, csr_amdgpu_highregs, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15, implicit $vgpr16, implicit $vgpr17, implicit $vgpr18, implicit $vgpr19, implicit $vgpr20, implicit $vgpr21, implicit $vgpr22, implicit $vgpr23, implicit $vgpr24, implicit $vgpr25, implicit $vgpr26, implicit $vgpr27, implicit $vgpr28, implicit $vgpr29, implicit $vgpr30, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr4_sgpr5, implicit $sgpr6_sgpr7, implicit $sgpr8_sgpr9, implicit $sgpr10_sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $vgpr31
   ; CHECK-NEXT:   ADJCALLSTACKDOWN 0, 36, implicit-def $scc
-  ; CHECK-NEXT:   [[COPY19:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY8]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY19]]
+  ; CHECK-NEXT:   SI_RETURN
 entry:
   call void @external_void_func_8xv5f32(
       <5 x float> <float 0.0, float 0.0, float 0.0, float 0.0, float 0.0>,
       <5 x float> <float 1.0, float 1.0, float 1.0, float 1.0, float 1.0>,
       <5 x float> <float 2.0, float 2.0, float 2.0, float 2.0, float 2.0>,
       <5 x float> <float 3.0, float 3.0, float 3.0, float 3.0, float 3.0>,
       <5 x float> <float 4.0, float 4.0, float 4.0, float 4.0, float 4.0>,
       <5 x float> <float 5.0, float 5.0, float 5.0, float 5.0, float 5.0>,
       <5 x float> <float 6.0, float 7.0, float 8.0, float 9.0, float 10.0>,
       <5 x float> <float 11.0, float 12.0, float 13.0, float 14.0, float 15.0>)
   ret void
 }
 
 attributes #0 = { nounwind }
 attributes #1 = { nounwind readnone }
 attributes #2 = { nounwind noinline }

llvm/test/CodeGen/AMDGPU/GlobalISel/irtranslator-constantexpr.ll

 ; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
 ; RUN: llc -global-isel -mtriple=amdgcn-amd-amdhsa -verify-machineinstrs -stop-after=irtranslator -o - %s | FileCheck %s
 
 @var = global i32 undef
 
 define i32 @test() {
   ; CHECK-LABEL: name: test
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 -1
-  ; CHECK:   [[INTTOPTR:%[0-9]+]]:_(p0) = G_INTTOPTR [[C]](s32)
-  ; CHECK:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @var
-  ; CHECK:   [[ICMP:%[0-9]+]]:_(s1) = G_ICMP intpred(eq), [[INTTOPTR]](p0), [[GV]]
-  ; CHECK:   [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[ICMP]](s1)
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY [[ZEXT]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(s32) = COPY [[COPY1]](s32)
-  ; CHECK:   [[COPY3:%[0-9]+]]:_(s32) = COPY [[COPY2]](s32)
-  ; CHECK:   [[COPY4:%[0-9]+]]:_(s32) = COPY [[COPY3]](s32)
-  ; CHECK:   $vgpr0 = COPY [[COPY4]](s32)
-  ; CHECK:   [[COPY5:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY5]], implicit $vgpr0
+  ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 -1
+  ; CHECK-NEXT:   [[INTTOPTR:%[0-9]+]]:_(p0) = G_INTTOPTR [[C]](s32)
+  ; CHECK-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @var
+  ; CHECK-NEXT:   [[ICMP:%[0-9]+]]:_(s1) = G_ICMP intpred(eq), [[INTTOPTR]](p0), [[GV]]
+  ; CHECK-NEXT:   [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[ICMP]](s1)
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY [[ZEXT]](s32)
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY [[COPY1]](s32)
+  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY [[COPY2]](s32)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[COPY3]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   ret i32 bitcast (<1 x i32> <i32 extractelement (<1 x i32> bitcast (i32 zext (i1 icmp eq (i32* @var, i32* inttoptr (i32 -1 to i32*)) to i32) to <1 x i32>), i64 0)> to i32)
 }
 
 @gint = external addrspace(1) global i8, align 4
 
 ; Technically we should be able to fold away the compare to true, but
 ; currently constexpr doesn't understand null in non-0 address spaces.
 define amdgpu_kernel void @constantexpr_select_0() {
   ; CHECK-LABEL: name: constantexpr_select_0
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   [[GV:%[0-9]+]]:_(p1) = G_GLOBAL_VALUE @gint
-  ; CHECK:   [[C:%[0-9]+]]:_(p1) = G_CONSTANT i64 0
-  ; CHECK:   [[ICMP:%[0-9]+]]:_(s1) = G_ICMP intpred(eq), [[GV]](p1), [[C]]
-  ; CHECK:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 1
-  ; CHECK:   [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
-  ; CHECK:   [[SELECT:%[0-9]+]]:_(s32) = G_SELECT [[ICMP]](s1), [[C1]], [[C2]]
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   G_STORE [[SELECT]](s32), [[DEF]](p1) :: (store (s32) into `i32 addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   S_ENDPGM 0
+  ; CHECK-NEXT:   [[GV:%[0-9]+]]:_(p1) = G_GLOBAL_VALUE @gint
+  ; CHECK-NEXT:   [[C:%[0-9]+]]:_(p1) = G_CONSTANT i64 0
+  ; CHECK-NEXT:   [[ICMP:%[0-9]+]]:_(s1) = G_ICMP intpred(eq), [[GV]](p1), [[C]]
+  ; CHECK-NEXT:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 1
+  ; CHECK-NEXT:   [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
+  ; CHECK-NEXT:   [[SELECT:%[0-9]+]]:_(s32) = G_SELECT [[ICMP]](s1), [[C1]], [[C2]]
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   G_STORE [[SELECT]](s32), [[DEF]](p1) :: (store (s32) into `i32 addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   S_ENDPGM 0
   store i32 select (i1 icmp eq (i8 addrspace(1)* @gint, i8 addrspace(1)* null), i32 1, i32 0), i32 addrspace(1)* undef, align 4
   ret void
 }
 
 define amdgpu_kernel void @constantexpr_select_1() {
   ; CHECK-LABEL: name: constantexpr_select_1
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 1024
-  ; CHECK:   [[INTTOPTR:%[0-9]+]]:_(p1) = G_INTTOPTR [[C]](s64)
-  ; CHECK:   [[GV:%[0-9]+]]:_(p1) = G_GLOBAL_VALUE @gint
-  ; CHECK:   [[ICMP:%[0-9]+]]:_(s1) = G_ICMP intpred(eq), [[INTTOPTR]](p1), [[GV]]
-  ; CHECK:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 1
-  ; CHECK:   [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
-  ; CHECK:   [[SELECT:%[0-9]+]]:_(s32) = G_SELECT [[ICMP]](s1), [[C1]], [[C2]]
-  ; CHECK:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK:   G_STORE [[SELECT]](s32), [[DEF]](p1) :: (store (s32) into `i32 addrspace(1)* undef`, addrspace 1)
-  ; CHECK:   S_ENDPGM 0
+  ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 1024
+  ; CHECK-NEXT:   [[INTTOPTR:%[0-9]+]]:_(p1) = G_INTTOPTR [[C]](s64)
+  ; CHECK-NEXT:   [[GV:%[0-9]+]]:_(p1) = G_GLOBAL_VALUE @gint
+  ; CHECK-NEXT:   [[ICMP:%[0-9]+]]:_(s1) = G_ICMP intpred(eq), [[INTTOPTR]](p1), [[GV]]
+  ; CHECK-NEXT:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 1
+  ; CHECK-NEXT:   [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
+  ; CHECK-NEXT:   [[SELECT:%[0-9]+]]:_(s32) = G_SELECT [[ICMP]](s1), [[C1]], [[C2]]
+  ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; CHECK-NEXT:   G_STORE [[SELECT]](s32), [[DEF]](p1) :: (store (s32) into `i32 addrspace(1)* undef`, addrspace 1)
+  ; CHECK-NEXT:   S_ENDPGM 0
   store i32 select (i1 icmp eq (i8 addrspace(1)* @gint, i8 addrspace(1)* inttoptr (i64 1024 to i8 addrspace(1)*)), i32 1, i32 0), i32 addrspace(1)* undef, align 4
   ret void
 }

llvm/test/CodeGen/AMDGPU/GlobalISel/irtranslator-constrained-fp.ll

 ; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
 ; RUN: llc -global-isel -march=amdgcn -stop-after=irtranslator -verify-machineinstrs %s -o - | FileCheck %s
 
 define float @v_constained_fadd_f32_fpexcept_strict(float %x, float %y) #0 {
   ; CHECK-LABEL: name: v_constained_fadd_f32_fpexcept_strict
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[STRICT_FADD:%[0-9]+]]:_(s32) = G_STRICT_FADD [[COPY]], [[COPY1]]
   ; CHECK-NEXT:   $vgpr0 = COPY [[STRICT_FADD]](s32)
-  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = call float @llvm.experimental.constrained.fadd.f32(float %x, float %y, metadata !"round.tonearest", metadata !"fpexcept.strict")
   ret float %val
 }
 
 define float @v_constained_fadd_f32_fpexcept_strict_flags(float %x, float %y) #0 {
   ; CHECK-LABEL: name: v_constained_fadd_f32_fpexcept_strict_flags
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[STRICT_FADD:%[0-9]+]]:_(s32) = nsz G_STRICT_FADD [[COPY]], [[COPY1]]
   ; CHECK-NEXT:   $vgpr0 = COPY [[STRICT_FADD]](s32)
-  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = call nsz float @llvm.experimental.constrained.fadd.f32(float %x, float %y, metadata !"round.tonearest", metadata !"fpexcept.strict")
   ret float %val
 }
 
 define float @v_constained_fadd_f32_fpexcept_ignore(float %x, float %y) #0 {
   ; CHECK-LABEL: name: v_constained_fadd_f32_fpexcept_ignore
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK-NEXT:   %3:_(s32) = nofpexcept G_STRICT_FADD [[COPY]], [[COPY1]]
-  ; CHECK-NEXT:   $vgpr0 = COPY %3(s32)
-  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; CHECK-NEXT:   %2:_(s32) = nofpexcept G_STRICT_FADD [[COPY]], [[COPY1]]
+  ; CHECK-NEXT:   $vgpr0 = COPY %2(s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = call float @llvm.experimental.constrained.fadd.f32(float %x, float %y, metadata !"round.tonearest", metadata !"fpexcept.ignore")
   ret float %val
 }
 
 define float @v_constained_fadd_f32_fpexcept_ignore_flags(float %x, float %y) #0 {
   ; CHECK-LABEL: name: v_constained_fadd_f32_fpexcept_ignore_flags
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK-NEXT:   %3:_(s32) = nsz nofpexcept G_STRICT_FADD [[COPY]], [[COPY1]]
-  ; CHECK-NEXT:   $vgpr0 = COPY %3(s32)
-  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; CHECK-NEXT:   %2:_(s32) = nsz nofpexcept G_STRICT_FADD [[COPY]], [[COPY1]]
+  ; CHECK-NEXT:   $vgpr0 = COPY %2(s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = call nsz float @llvm.experimental.constrained.fadd.f32(float %x, float %y, metadata !"round.tonearest", metadata !"fpexcept.ignore")
   ret float %val
 }
 
 define float @v_constained_fadd_f32_fpexcept_maytrap(float %x, float %y) #0 {
   ; CHECK-LABEL: name: v_constained_fadd_f32_fpexcept_maytrap
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[STRICT_FADD:%[0-9]+]]:_(s32) = G_STRICT_FADD [[COPY]], [[COPY1]]
   ; CHECK-NEXT:   $vgpr0 = COPY [[STRICT_FADD]](s32)
-  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = call float @llvm.experimental.constrained.fadd.f32(float %x, float %y, metadata !"round.tonearest", metadata !"fpexcept.maytrap")
   ret float %val
 }
 
 define <2 x float> @v_constained_fadd_v2f32_fpexcept_strict(<2 x float> %x, <2 x float> %y) #0 {
   ; CHECK-LABEL: name: v_constained_fadd_v2f32_fpexcept_strict
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32)
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
   ; CHECK-NEXT:   [[BUILD_VECTOR1:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY2]](s32), [[COPY3]](s32)
-  ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[STRICT_FADD:%[0-9]+]]:_(<2 x s32>) = G_STRICT_FADD [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
   ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[STRICT_FADD]](<2 x s32>)
   ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY4]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY5]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %val = call <2 x float> @llvm.experimental.constrained.fadd.v2f32(<2 x float> %x, <2 x float> %y, metadata !"round.tonearest", metadata !"fpexcept.strict")
   ret <2 x float> %val
 }
 
 define <2 x float> @v_constained_fadd_v2f32_fpexcept_ignore(<2 x float> %x, <2 x float> %y) #0 {
   ; CHECK-LABEL: name: v_constained_fadd_v2f32_fpexcept_ignore
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32)
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
   ; CHECK-NEXT:   [[BUILD_VECTOR1:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY2]](s32), [[COPY3]](s32)
-  ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK-NEXT:   %7:_(<2 x s32>) = nofpexcept G_STRICT_FADD [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
-  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES %7(<2 x s32>)
+  ; CHECK-NEXT:   %6:_(<2 x s32>) = nofpexcept G_STRICT_FADD [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES %6(<2 x s32>)
   ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY4]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY5]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %val = call <2 x float> @llvm.experimental.constrained.fadd.v2f32(<2 x float> %x, <2 x float> %y, metadata !"round.tonearest", metadata !"fpexcept.ignore")
   ret <2 x float> %val
 }
 
 define <2 x float> @v_constained_fadd_v2f32_fpexcept_maytrap(<2 x float> %x, <2 x float> %y) #0 {
   ; CHECK-LABEL: name: v_constained_fadd_v2f32_fpexcept_maytrap
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32)
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
   ; CHECK-NEXT:   [[BUILD_VECTOR1:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY2]](s32), [[COPY3]](s32)
-  ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[STRICT_FADD:%[0-9]+]]:_(<2 x s32>) = G_STRICT_FADD [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
   ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[STRICT_FADD]](<2 x s32>)
   ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY4]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY5]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %val = call <2 x float> @llvm.experimental.constrained.fadd.v2f32(<2 x float> %x, <2 x float> %y, metadata !"round.tonearest", metadata !"fpexcept.maytrap")
   ret <2 x float> %val
 }
 
 define float @v_constained_fsub_f32_fpexcept_ignore_flags(float %x, float %y) #0 {
   ; CHECK-LABEL: name: v_constained_fsub_f32_fpexcept_ignore_flags
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK-NEXT:   %3:_(s32) = nsz nofpexcept G_STRICT_FSUB [[COPY]], [[COPY1]]
-  ; CHECK-NEXT:   $vgpr0 = COPY %3(s32)
-  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; CHECK-NEXT:   %2:_(s32) = nsz nofpexcept G_STRICT_FSUB [[COPY]], [[COPY1]]
+  ; CHECK-NEXT:   $vgpr0 = COPY %2(s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = call nsz float @llvm.experimental.constrained.fsub.f32(float %x, float %y, metadata !"round.tonearest", metadata !"fpexcept.ignore")
   ret float %val
 }
 
 define float @v_constained_fmul_f32_fpexcept_ignore_flags(float %x, float %y) #0 {
   ; CHECK-LABEL: name: v_constained_fmul_f32_fpexcept_ignore_flags
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK-NEXT:   %3:_(s32) = nsz nofpexcept G_STRICT_FMUL [[COPY]], [[COPY1]]
-  ; CHECK-NEXT:   $vgpr0 = COPY %3(s32)
-  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; CHECK-NEXT:   %2:_(s32) = nsz nofpexcept G_STRICT_FMUL [[COPY]], [[COPY1]]
+  ; CHECK-NEXT:   $vgpr0 = COPY %2(s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = call nsz float @llvm.experimental.constrained.fmul.f32(float %x, float %y, metadata !"round.tonearest", metadata !"fpexcept.ignore")
   ret float %val
 }
 
 define float @v_constained_fdiv_f32_fpexcept_ignore_flags(float %x, float %y) #0 {
   ; CHECK-LABEL: name: v_constained_fdiv_f32_fpexcept_ignore_flags
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK-NEXT:   %3:_(s32) = nsz nofpexcept G_STRICT_FDIV [[COPY]], [[COPY1]]
-  ; CHECK-NEXT:   $vgpr0 = COPY %3(s32)
-  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; CHECK-NEXT:   %2:_(s32) = nsz nofpexcept G_STRICT_FDIV [[COPY]], [[COPY1]]
+  ; CHECK-NEXT:   $vgpr0 = COPY %2(s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = call nsz float @llvm.experimental.constrained.fdiv.f32(float %x, float %y, metadata !"round.tonearest", metadata !"fpexcept.ignore")
   ret float %val
 }
 
 define float @v_constained_frem_f32_fpexcept_ignore_flags(float %x, float %y) #0 {
   ; CHECK-LABEL: name: v_constained_frem_f32_fpexcept_ignore_flags
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK-NEXT:   %3:_(s32) = nsz nofpexcept G_STRICT_FREM [[COPY]], [[COPY1]]
-  ; CHECK-NEXT:   $vgpr0 = COPY %3(s32)
-  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; CHECK-NEXT:   %2:_(s32) = nsz nofpexcept G_STRICT_FREM [[COPY]], [[COPY1]]
+  ; CHECK-NEXT:   $vgpr0 = COPY %2(s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = call nsz float @llvm.experimental.constrained.frem.f32(float %x, float %y, metadata !"round.tonearest", metadata !"fpexcept.ignore")
   ret float %val
 }
 
 define float @v_constained_fma_f32_fpexcept_ignore_flags(float %x, float %y, float %z) #0 {
   ; CHECK-LABEL: name: v_constained_fma_f32_fpexcept_ignore_flags
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK-NEXT:   %4:_(s32) = nsz nofpexcept G_STRICT_FMA [[COPY]], [[COPY1]], [[COPY2]]
-  ; CHECK-NEXT:   $vgpr0 = COPY %4(s32)
-  ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+  ; CHECK-NEXT:   %3:_(s32) = nsz nofpexcept G_STRICT_FMA [[COPY]], [[COPY1]], [[COPY2]]
+  ; CHECK-NEXT:   $vgpr0 = COPY %3(s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = call nsz float @llvm.experimental.constrained.fma.f32(float %x, float %y, float %z, metadata !"round.tonearest", metadata !"fpexcept.ignore")
   ret float %val
 }
 
 define float @v_constained_sqrt_f32_fpexcept_strict(float %x) #0 {
   ; CHECK-LABEL: name: v_constained_sqrt_f32_fpexcept_strict
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $vgpr0, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[STRICT_FSQRT:%[0-9]+]]:_(s32) = G_STRICT_FSQRT [[COPY]]
   ; CHECK-NEXT:   $vgpr0 = COPY [[STRICT_FSQRT]](s32)
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY2]], implicit $vgpr0
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %val = call float @llvm.experimental.constrained.sqrt.f32(float %x, metadata !"round.tonearest", metadata !"fpexcept.strict")
   ret float %val
 }
 
 declare float @llvm.experimental.constrained.fadd.f32(float, float, metadata, metadata) #1
 declare <2 x float> @llvm.experimental.constrained.fadd.v2f32(<2 x float>, <2 x float>, metadata, metadata) #1
 declare <3 x float> @llvm.experimental.constrained.fadd.v3f32(<3 x float>, <3 x float>, metadata, metadata) #1
 declare float @llvm.experimental.constrained.fsub.f32(float, float, metadata, metadata) #1
 declare float @llvm.experimental.constrained.fmul.f32(float, float, metadata, metadata) #1
 declare float @llvm.experimental.constrained.fdiv.f32(float, float, metadata, metadata) #1
 declare float @llvm.experimental.constrained.frem.f32(float, float, metadata, metadata) #1
 declare float @llvm.experimental.constrained.fma.f32(float, float, float, metadata, metadata) #1
 declare float @llvm.experimental.constrained.sqrt.f32(float, metadata, metadata) #1
 
 attributes #0 = { strictfp }
 attributes #1 = { inaccessiblememonly nounwind willreturn }

llvm/test/CodeGen/AMDGPU/GlobalISel/irtranslator-fixed-function-abi-vgpr-args.ll

 ; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
 ; RUN: llc -global-isel -mtriple=amdgcn-amd-amdhsa -amdgpu-fixed-function-abi=1 -stop-after=irtranslator -o - %s | FileCheck --check-prefix=FIXED %s
 ; RUN: llc -global-isel -mtriple=amdgcn-amd-amdhsa -amdgpu-fixed-function-abi=0 -stop-after=irtranslator -o - %s | FileCheck --check-prefix=VARABI %s
 
 ; Make sure arg1 is not allocated in v31, which is reserved for
 ; workitem IDs with -amdgpu-fixed-function-abi.
 
 define void @void_a31i32_i32([31 x i32] %arg0, i32 %arg1) {
   ; FIXED-LABEL: name: void_a31i32_i32
   ; FIXED: bb.1 (%ir-block.0):
-  ; FIXED:   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, $sgpr30_sgpr31
-  ; FIXED:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; FIXED:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; FIXED:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; FIXED:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
-  ; FIXED:   [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
-  ; FIXED:   [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
-  ; FIXED:   [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
-  ; FIXED:   [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
-  ; FIXED:   [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
-  ; FIXED:   [[COPY9:%[0-9]+]]:_(s32) = COPY $vgpr9
-  ; FIXED:   [[COPY10:%[0-9]+]]:_(s32) = COPY $vgpr10
-  ; FIXED:   [[COPY11:%[0-9]+]]:_(s32) = COPY $vgpr11
-  ; FIXED:   [[COPY12:%[0-9]+]]:_(s32) = COPY $vgpr12
-  ; FIXED:   [[COPY13:%[0-9]+]]:_(s32) = COPY $vgpr13
-  ; FIXED:   [[COPY14:%[0-9]+]]:_(s32) = COPY $vgpr14
-  ; FIXED:   [[COPY15:%[0-9]+]]:_(s32) = COPY $vgpr15
-  ; FIXED:   [[COPY16:%[0-9]+]]:_(s32) = COPY $vgpr16
-  ; FIXED:   [[COPY17:%[0-9]+]]:_(s32) = COPY $vgpr17
-  ; FIXED:   [[COPY18:%[0-9]+]]:_(s32) = COPY $vgpr18
-  ; FIXED:   [[COPY19:%[0-9]+]]:_(s32) = COPY $vgpr19
-  ; FIXED:   [[COPY20:%[0-9]+]]:_(s32) = COPY $vgpr20
-  ; FIXED:   [[COPY21:%[0-9]+]]:_(s32) = COPY $vgpr21
-  ; FIXED:   [[COPY22:%[0-9]+]]:_(s32) = COPY $vgpr22
-  ; FIXED:   [[COPY23:%[0-9]+]]:_(s32) = COPY $vgpr23
-  ; FIXED:   [[COPY24:%[0-9]+]]:_(s32) = COPY $vgpr24
-  ; FIXED:   [[COPY25:%[0-9]+]]:_(s32) = COPY $vgpr25
-  ; FIXED:   [[COPY26:%[0-9]+]]:_(s32) = COPY $vgpr26
-  ; FIXED:   [[COPY27:%[0-9]+]]:_(s32) = COPY $vgpr27
-  ; FIXED:   [[COPY28:%[0-9]+]]:_(s32) = COPY $vgpr28
-  ; FIXED:   [[COPY29:%[0-9]+]]:_(s32) = COPY $vgpr29
-  ; FIXED:   [[COPY30:%[0-9]+]]:_(s32) = COPY $vgpr30
-  ; FIXED:   [[FRAME_INDEX:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.0
-  ; FIXED:   [[LOAD:%[0-9]+]]:_(s32) = G_LOAD [[FRAME_INDEX]](p5) :: (invariant load (s32) from %fixed-stack.0, align 16, addrspace 5)
-  ; FIXED:   [[COPY31:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; FIXED:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; FIXED:   G_STORE [[LOAD]](s32), [[DEF]](p1) :: (store (s32) into `i32 addrspace(1)* undef`, addrspace 1)
-  ; FIXED:   [[COPY32:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY31]]
-  ; FIXED:   S_SETPC_B64_return [[COPY32]]
+  ; FIXED-NEXT:   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
+  ; FIXED-NEXT: {{  $}}
+  ; FIXED-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; FIXED-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; FIXED-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
+  ; FIXED-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
+  ; FIXED-NEXT:   [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
+  ; FIXED-NEXT:   [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
+  ; FIXED-NEXT:   [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
+  ; FIXED-NEXT:   [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
+  ; FIXED-NEXT:   [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
+  ; FIXED-NEXT:   [[COPY9:%[0-9]+]]:_(s32) = COPY $vgpr9
+  ; FIXED-NEXT:   [[COPY10:%[0-9]+]]:_(s32) = COPY $vgpr10
+  ; FIXED-NEXT:   [[COPY11:%[0-9]+]]:_(s32) = COPY $vgpr11
+  ; FIXED-NEXT:   [[COPY12:%[0-9]+]]:_(s32) = COPY $vgpr12
+  ; FIXED-NEXT:   [[COPY13:%[0-9]+]]:_(s32) = COPY $vgpr13
+  ; FIXED-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY $vgpr14
+  ; FIXED-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY $vgpr15
+  ; FIXED-NEXT:   [[COPY16:%[0-9]+]]:_(s32) = COPY $vgpr16
+  ; FIXED-NEXT:   [[COPY17:%[0-9]+]]:_(s32) = COPY $vgpr17
+  ; FIXED-NEXT:   [[COPY18:%[0-9]+]]:_(s32) = COPY $vgpr18
+  ; FIXED-NEXT:   [[COPY19:%[0-9]+]]:_(s32) = COPY $vgpr19
+  ; FIXED-NEXT:   [[COPY20:%[0-9]+]]:_(s32) = COPY $vgpr20
+  ; FIXED-NEXT:   [[COPY21:%[0-9]+]]:_(s32) = COPY $vgpr21
+  ; FIXED-NEXT:   [[COPY22:%[0-9]+]]:_(s32) = COPY $vgpr22
+  ; FIXED-NEXT:   [[COPY23:%[0-9]+]]:_(s32) = COPY $vgpr23
+  ; FIXED-NEXT:   [[COPY24:%[0-9]+]]:_(s32) = COPY $vgpr24
+  ; FIXED-NEXT:   [[COPY25:%[0-9]+]]:_(s32) = COPY $vgpr25
+  ; FIXED-NEXT:   [[COPY26:%[0-9]+]]:_(s32) = COPY $vgpr26
+  ; FIXED-NEXT:   [[COPY27:%[0-9]+]]:_(s32) = COPY $vgpr27
+  ; FIXED-NEXT:   [[COPY28:%[0-9]+]]:_(s32) = COPY $vgpr28
+  ; FIXED-NEXT:   [[COPY29:%[0-9]+]]:_(s32) = COPY $vgpr29
+  ; FIXED-NEXT:   [[COPY30:%[0-9]+]]:_(s32) = COPY $vgpr30
+  ; FIXED-NEXT:   [[FRAME_INDEX:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.0
+  ; FIXED-NEXT:   [[LOAD:%[0-9]+]]:_(s32) = G_LOAD [[FRAME_INDEX]](p5) :: (invariant load (s32) from %fixed-stack.0, align 16, addrspace 5)
+  ; FIXED-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; FIXED-NEXT:   G_STORE [[LOAD]](s32), [[DEF]](p1) :: (store (s32) into `i32 addrspace(1)* undef`, addrspace 1)
+  ; FIXED-NEXT:   SI_RETURN
   ; VARABI-LABEL: name: void_a31i32_i32
   ; VARABI: bb.1 (%ir-block.0):
-  ; VARABI:   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, $sgpr30_sgpr31
-  ; VARABI:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; VARABI:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; VARABI:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; VARABI:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
-  ; VARABI:   [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
-  ; VARABI:   [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
-  ; VARABI:   [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
-  ; VARABI:   [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
-  ; VARABI:   [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
-  ; VARABI:   [[COPY9:%[0-9]+]]:_(s32) = COPY $vgpr9
-  ; VARABI:   [[COPY10:%[0-9]+]]:_(s32) = COPY $vgpr10
-  ; VARABI:   [[COPY11:%[0-9]+]]:_(s32) = COPY $vgpr11
-  ; VARABI:   [[COPY12:%[0-9]+]]:_(s32) = COPY $vgpr12
-  ; VARABI:   [[COPY13:%[0-9]+]]:_(s32) = COPY $vgpr13
-  ; VARABI:   [[COPY14:%[0-9]+]]:_(s32) = COPY $vgpr14
-  ; VARABI:   [[COPY15:%[0-9]+]]:_(s32) = COPY $vgpr15
-  ; VARABI:   [[COPY16:%[0-9]+]]:_(s32) = COPY $vgpr16
-  ; VARABI:   [[COPY17:%[0-9]+]]:_(s32) = COPY $vgpr17
-  ; VARABI:   [[COPY18:%[0-9]+]]:_(s32) = COPY $vgpr18
-  ; VARABI:   [[COPY19:%[0-9]+]]:_(s32) = COPY $vgpr19
-  ; VARABI:   [[COPY20:%[0-9]+]]:_(s32) = COPY $vgpr20
-  ; VARABI:   [[COPY21:%[0-9]+]]:_(s32) = COPY $vgpr21
-  ; VARABI:   [[COPY22:%[0-9]+]]:_(s32) = COPY $vgpr22
-  ; VARABI:   [[COPY23:%[0-9]+]]:_(s32) = COPY $vgpr23
-  ; VARABI:   [[COPY24:%[0-9]+]]:_(s32) = COPY $vgpr24
-  ; VARABI:   [[COPY25:%[0-9]+]]:_(s32) = COPY $vgpr25
-  ; VARABI:   [[COPY26:%[0-9]+]]:_(s32) = COPY $vgpr26
-  ; VARABI:   [[COPY27:%[0-9]+]]:_(s32) = COPY $vgpr27
-  ; VARABI:   [[COPY28:%[0-9]+]]:_(s32) = COPY $vgpr28
-  ; VARABI:   [[COPY29:%[0-9]+]]:_(s32) = COPY $vgpr29
-  ; VARABI:   [[COPY30:%[0-9]+]]:_(s32) = COPY $vgpr30
-  ; VARABI:   [[COPY31:%[0-9]+]]:_(s32) = COPY $vgpr31
-  ; VARABI:   [[COPY32:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; VARABI:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; VARABI:   G_STORE [[COPY31]](s32), [[DEF]](p1) :: (store (s32) into `i32 addrspace(1)* undef`, addrspace 1)
-  ; VARABI:   [[COPY33:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY32]]
-  ; VARABI:   S_SETPC_B64_return [[COPY33]]
+  ; VARABI-NEXT:   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
+  ; VARABI-NEXT: {{  $}}
+  ; VARABI-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; VARABI-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; VARABI-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
+  ; VARABI-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
+  ; VARABI-NEXT:   [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
+  ; VARABI-NEXT:   [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
+  ; VARABI-NEXT:   [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
+  ; VARABI-NEXT:   [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
+  ; VARABI-NEXT:   [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr8
+  ; VARABI-NEXT:   [[COPY9:%[0-9]+]]:_(s32) = COPY $vgpr9
+  ; VARABI-NEXT:   [[COPY10:%[0-9]+]]:_(s32) = COPY $vgpr10
+  ; VARABI-NEXT:   [[COPY11:%[0-9]+]]:_(s32) = COPY $vgpr11
+  ; VARABI-NEXT:   [[COPY12:%[0-9]+]]:_(s32) = COPY $vgpr12
+  ; VARABI-NEXT:   [[COPY13:%[0-9]+]]:_(s32) = COPY $vgpr13
+  ; VARABI-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY $vgpr14
+  ; VARABI-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY $vgpr15
+  ; VARABI-NEXT:   [[COPY16:%[0-9]+]]:_(s32) = COPY $vgpr16
+  ; VARABI-NEXT:   [[COPY17:%[0-9]+]]:_(s32) = COPY $vgpr17
+  ; VARABI-NEXT:   [[COPY18:%[0-9]+]]:_(s32) = COPY $vgpr18
+  ; VARABI-NEXT:   [[COPY19:%[0-9]+]]:_(s32) = COPY $vgpr19
+  ; VARABI-NEXT:   [[COPY20:%[0-9]+]]:_(s32) = COPY $vgpr20
+  ; VARABI-NEXT:   [[COPY21:%[0-9]+]]:_(s32) = COPY $vgpr21
+  ; VARABI-NEXT:   [[COPY22:%[0-9]+]]:_(s32) = COPY $vgpr22
+  ; VARABI-NEXT:   [[COPY23:%[0-9]+]]:_(s32) = COPY $vgpr23
+  ; VARABI-NEXT:   [[COPY24:%[0-9]+]]:_(s32) = COPY $vgpr24
+  ; VARABI-NEXT:   [[COPY25:%[0-9]+]]:_(s32) = COPY $vgpr25
+  ; VARABI-NEXT:   [[COPY26:%[0-9]+]]:_(s32) = COPY $vgpr26
+  ; VARABI-NEXT:   [[COPY27:%[0-9]+]]:_(s32) = COPY $vgpr27
+  ; VARABI-NEXT:   [[COPY28:%[0-9]+]]:_(s32) = COPY $vgpr28
+  ; VARABI-NEXT:   [[COPY29:%[0-9]+]]:_(s32) = COPY $vgpr29
+  ; VARABI-NEXT:   [[COPY30:%[0-9]+]]:_(s32) = COPY $vgpr30
+  ; VARABI-NEXT:   [[COPY31:%[0-9]+]]:_(s32) = COPY $vgpr31
+  ; VARABI-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
+  ; VARABI-NEXT:   G_STORE [[COPY31]](s32), [[DEF]](p1) :: (store (s32) into `i32 addrspace(1)* undef`, addrspace 1)
+  ; VARABI-NEXT:   SI_RETURN
   store i32 %arg1, i32 addrspace(1)* undef
   ret void
 }

llvm/test/CodeGen/AMDGPU/GlobalISel/irtranslator-getelementptr.ll

 ; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
 ; RUN: llc -global-isel -mtriple=amdgcn-amd-amdhsa -stop-after=irtranslator -o - %s | FileCheck %s
 
 ; Test 64-bit pointer with 64-bit index
 define <2 x i32 addrspace(1)*> @vector_gep_v2p1_index_v2i64(<2 x i32 addrspace(1)*> %ptr, <2 x i64> %idx) {
   ; CHECK-LABEL: name: vector_gep_v2p1_index_v2i64
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
   ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
   ; CHECK-NEXT:   [[MV1:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:_(<2 x p1>) = G_BUILD_VECTOR [[MV]](p1), [[MV1]](p1)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
   ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
   ; CHECK-NEXT:   [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
   ; CHECK-NEXT:   [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
   ; CHECK-NEXT:   [[MV3:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY6]](s32), [[COPY7]](s32)
   ; CHECK-NEXT:   [[BUILD_VECTOR1:%[0-9]+]]:_(<2 x s64>) = G_BUILD_VECTOR [[MV2]](s64), [[MV3]](s64)
-  ; CHECK-NEXT:   [[COPY8:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 4
   ; CHECK-NEXT:   [[BUILD_VECTOR2:%[0-9]+]]:_(<2 x s64>) = G_BUILD_VECTOR [[C]](s64), [[C]](s64)
   ; CHECK-NEXT:   [[MUL:%[0-9]+]]:_(<2 x s64>) = G_MUL [[BUILD_VECTOR1]], [[BUILD_VECTOR2]]
   ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:_(<2 x p1>) = G_PTR_ADD [[BUILD_VECTOR]], [[MUL]](<2 x s64>)
-  ; CHECK-NEXT:   [[COPY9:%[0-9]+]]:_(<2 x p1>) = COPY [[PTR_ADD]](<2 x p1>)
-  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY9]](<2 x p1>)
+  ; CHECK-NEXT:   [[COPY8:%[0-9]+]]:_(<2 x p1>) = COPY [[PTR_ADD]](<2 x p1>)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY8]](<2 x p1>)
   ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
   ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
-  ; CHECK-NEXT:   [[COPY10:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY8]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY10]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
   %gep = getelementptr i32, <2 x i32 addrspace(1)*> %ptr, <2 x i64> %idx
   ret <2 x i32 addrspace(1)*> %gep
 }
 
 ; Test 32-bit pointer with 32-bit index
 define <2 x i32 addrspace(3)*> @vector_gep_v2p3_index_v2i32(<2 x i32 addrspace(3)*> %ptr, <2 x i32> %idx) {
   ; CHECK-LABEL: name: vector_gep_v2p3_index_v2i32
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(p3) = COPY $vgpr0
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(p3) = COPY $vgpr1
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:_(<2 x p3>) = G_BUILD_VECTOR [[COPY]](p3), [[COPY1]](p3)
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
   ; CHECK-NEXT:   [[BUILD_VECTOR1:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY2]](s32), [[COPY3]](s32)
-  ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 4
   ; CHECK-NEXT:   [[BUILD_VECTOR2:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[C]](s32), [[C]](s32)
   ; CHECK-NEXT:   [[MUL:%[0-9]+]]:_(<2 x s32>) = G_MUL [[BUILD_VECTOR1]], [[BUILD_VECTOR2]]
   ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:_(<2 x p3>) = G_PTR_ADD [[BUILD_VECTOR]], [[MUL]](<2 x s32>)
-  ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:_(<2 x p3>) = COPY [[PTR_ADD]](<2 x p3>)
-  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY5]](<2 x p3>)
+  ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:_(<2 x p3>) = COPY [[PTR_ADD]](<2 x p3>)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY4]](<2 x p3>)
   ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY4]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY6]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %gep = getelementptr i32, <2 x i32 addrspace(3)*> %ptr, <2 x i32> %idx
   ret <2 x i32 addrspace(3)*> %gep
 }
 
 ; Test 64-bit pointer with 32-bit index
 define <2 x i32 addrspace(1)*> @vector_gep_v2p1_index_v2i32(<2 x i32 addrspace(1)*> %ptr, <2 x i32> %idx) {
   ; CHECK-LABEL: name: vector_gep_v2p1_index_v2i32
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
   ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
   ; CHECK-NEXT:   [[MV1:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:_(<2 x p1>) = G_BUILD_VECTOR [[MV]](p1), [[MV1]](p1)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
   ; CHECK-NEXT:   [[BUILD_VECTOR1:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY4]](s32), [[COPY5]](s32)
-  ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[SEXT:%[0-9]+]]:_(<2 x s64>) = G_SEXT [[BUILD_VECTOR1]](<2 x s32>)
   ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 4
   ; CHECK-NEXT:   [[BUILD_VECTOR2:%[0-9]+]]:_(<2 x s64>) = G_BUILD_VECTOR [[C]](s64), [[C]](s64)
   ; CHECK-NEXT:   [[MUL:%[0-9]+]]:_(<2 x s64>) = G_MUL [[SEXT]], [[BUILD_VECTOR2]]
   ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:_(<2 x p1>) = G_PTR_ADD [[BUILD_VECTOR]], [[MUL]](<2 x s64>)
-  ; CHECK-NEXT:   [[COPY7:%[0-9]+]]:_(<2 x p1>) = COPY [[PTR_ADD]](<2 x p1>)
-  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY7]](<2 x p1>)
+  ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:_(<2 x p1>) = COPY [[PTR_ADD]](<2 x p1>)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY6]](<2 x p1>)
   ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
   ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
-  ; CHECK-NEXT:   [[COPY8:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY8]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
   %gep = getelementptr i32, <2 x i32 addrspace(1)*> %ptr, <2 x i32> %idx
   ret <2 x i32 addrspace(1)*> %gep
 }
 
 ; Test 64-bit pointer with 64-bit scalar index
 define <2 x i32 addrspace(1)*> @vector_gep_v2p1_index_i64(<2 x i32 addrspace(1)*> %ptr, i64 %idx) {
   ; CHECK-LABEL: name: vector_gep_v2p1_index_i64
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
   ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
   ; CHECK-NEXT:   [[MV1:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:_(<2 x p1>) = G_BUILD_VECTOR [[MV]](p1), [[MV1]](p1)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
   ; CHECK-NEXT:   [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
-  ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[BUILD_VECTOR1:%[0-9]+]]:_(<2 x s64>) = G_BUILD_VECTOR [[MV2]](s64), [[MV2]](s64)
-  ; CHECK-NEXT:   [[COPY7:%[0-9]+]]:_(<2 x s64>) = COPY [[BUILD_VECTOR1]](<2 x s64>)
+  ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:_(<2 x s64>) = COPY [[BUILD_VECTOR1]](<2 x s64>)
   ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 4
   ; CHECK-NEXT:   [[BUILD_VECTOR2:%[0-9]+]]:_(<2 x s64>) = G_BUILD_VECTOR [[C]](s64), [[C]](s64)
-  ; CHECK-NEXT:   [[MUL:%[0-9]+]]:_(<2 x s64>) = G_MUL [[COPY7]], [[BUILD_VECTOR2]]
+  ; CHECK-NEXT:   [[MUL:%[0-9]+]]:_(<2 x s64>) = G_MUL [[COPY6]], [[BUILD_VECTOR2]]
   ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:_(<2 x p1>) = G_PTR_ADD [[BUILD_VECTOR]], [[MUL]](<2 x s64>)
-  ; CHECK-NEXT:   [[COPY8:%[0-9]+]]:_(<2 x p1>) = COPY [[PTR_ADD]](<2 x p1>)
-  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY8]](<2 x p1>)
+  ; CHECK-NEXT:   [[COPY7:%[0-9]+]]:_(<2 x p1>) = COPY [[PTR_ADD]](<2 x p1>)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY7]](<2 x p1>)
   ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
   ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
-  ; CHECK-NEXT:   [[COPY9:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY6]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY9]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
   %gep = getelementptr i32, <2 x i32 addrspace(1)*> %ptr, i64 %idx
   ret <2 x i32 addrspace(1)*> %gep
 }
 
 ; Test 64-bit pointer with 32-bit scalar index
 define <2 x i32 addrspace(1)*> @vector_gep_v2p1_index_i32(<2 x i32 addrspace(1)*> %ptr, i32 %idx) {
   ; CHECK-LABEL: name: vector_gep_v2p1_index_i32
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
   ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
   ; CHECK-NEXT:   [[MV1:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:_(<2 x p1>) = G_BUILD_VECTOR [[MV]](p1), [[MV1]](p1)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
-  ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[BUILD_VECTOR1:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY4]](s32), [[COPY4]](s32)
   ; CHECK-NEXT:   [[SEXT:%[0-9]+]]:_(<2 x s64>) = G_SEXT [[BUILD_VECTOR1]](<2 x s32>)
   ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 4
   ; CHECK-NEXT:   [[BUILD_VECTOR2:%[0-9]+]]:_(<2 x s64>) = G_BUILD_VECTOR [[C]](s64), [[C]](s64)
   ; CHECK-NEXT:   [[MUL:%[0-9]+]]:_(<2 x s64>) = G_MUL [[SEXT]], [[BUILD_VECTOR2]]
   ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:_(<2 x p1>) = G_PTR_ADD [[BUILD_VECTOR]], [[MUL]](<2 x s64>)
-  ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:_(<2 x p1>) = COPY [[PTR_ADD]](<2 x p1>)
-  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY6]](<2 x p1>)
+  ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:_(<2 x p1>) = COPY [[PTR_ADD]](<2 x p1>)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY5]](<2 x p1>)
   ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
   ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
-  ; CHECK-NEXT:   [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY5]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY7]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
   %gep = getelementptr i32, <2 x i32 addrspace(1)*> %ptr, i32 %idx
   ret <2 x i32 addrspace(1)*> %gep
 }
 
 ; Test 64-bit pointer with 64-bit constant, non-splat
 define <2 x i32 addrspace(1)*> @vector_gep_v2p1_index_v2i64_constant(<2 x i32 addrspace(1)*> %ptr, <2 x i64> %idx) {
   ; CHECK-LABEL: name: vector_gep_v2p1_index_v2i64_constant
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
   ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
   ; CHECK-NEXT:   [[MV1:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
   ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:_(<2 x p1>) = G_BUILD_VECTOR [[MV]](p1), [[MV1]](p1)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
   ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
   ; CHECK-NEXT:   [[COPY7:%[0-9]+]]:_(s32) = COPY $vgpr7
   ; CHECK-NEXT:   [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY4]](s32), [[COPY5]](s32)
   ; CHECK-NEXT:   [[MV3:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY6]](s32), [[COPY7]](s32)
   ; CHECK-NEXT:   [[BUILD_VECTOR1:%[0-9]+]]:_(<2 x s64>) = G_BUILD_VECTOR [[MV2]](s64), [[MV3]](s64)
-  ; CHECK-NEXT:   [[COPY8:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 1
   ; CHECK-NEXT:   [[C1:%[0-9]+]]:_(s64) = G_CONSTANT i64 2
   ; CHECK-NEXT:   [[BUILD_VECTOR2:%[0-9]+]]:_(<2 x s64>) = G_BUILD_VECTOR [[C]](s64), [[C1]](s64)
   ; CHECK-NEXT:   [[C2:%[0-9]+]]:_(s64) = G_CONSTANT i64 4
   ; CHECK-NEXT:   [[BUILD_VECTOR3:%[0-9]+]]:_(<2 x s64>) = G_BUILD_VECTOR [[C2]](s64), [[C2]](s64)
   ; CHECK-NEXT:   [[C3:%[0-9]+]]:_(s64) = G_CONSTANT i64 8
   ; CHECK-NEXT:   [[BUILD_VECTOR4:%[0-9]+]]:_(<2 x s64>) = G_BUILD_VECTOR [[C2]](s64), [[C3]](s64)
   ; CHECK-NEXT:   [[PTR_ADD:%[0-9]+]]:_(<2 x p1>) = G_PTR_ADD [[BUILD_VECTOR]], [[BUILD_VECTOR4]](<2 x s64>)
-  ; CHECK-NEXT:   [[COPY9:%[0-9]+]]:_(<2 x p1>) = COPY [[PTR_ADD]](<2 x p1>)
-  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY9]](<2 x p1>)
+  ; CHECK-NEXT:   [[COPY8:%[0-9]+]]:_(<2 x p1>) = COPY [[PTR_ADD]](<2 x p1>)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY8]](<2 x p1>)
   ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; CHECK-NEXT:   $vgpr2 = COPY [[UV2]](s32)
   ; CHECK-NEXT:   $vgpr3 = COPY [[UV3]](s32)
-  ; CHECK-NEXT:   [[COPY10:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY8]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY10]], implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
   %gep = getelementptr i32, <2 x i32 addrspace(1)*> %ptr, <2 x i64> <i64 1, i64 2>
   ret <2 x i32 addrspace(1)*> %gep
 }

llvm/test/CodeGen/AMDGPU/GlobalISel/irtranslator-indirect-call.ll

define amdgpu_kernel void @test_indirect_call_sgpr_ptr(void()* %fptr) {
   ; CHECK-NEXT:   S_ENDPGM 0
   call void %fptr()
   ret void
 }
 
 define amdgpu_gfx void @test_gfx_indirect_call_sgpr_ptr(void()* %fptr) {
   ; CHECK-LABEL: name: test_gfx_indirect_call_sgpr_ptr
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
   ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(p0) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
-  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY3]](<4 x s32>)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY2]](<4 x s32>)
   ; CHECK-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[MV]](p0), 0, csr_amdgpu_si_gfx, implicit $sgpr0_sgpr1_sgpr2_sgpr3
   ; CHECK-NEXT:   ADJCALLSTACKDOWN 0, 0, implicit-def $scc
-  ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:gfx_ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK-NEXT:   S_SETPC_B64_return_gfx [[COPY4]]
+  ; CHECK-NEXT:   SI_RETURN
   call amdgpu_gfx void %fptr()
   ret void
 }

llvm/test/CodeGen/AMDGPU/GlobalISel/irtranslator-inline-asm.ll

define amdgpu_kernel void @asm_simple_agpr_clobber() {
   ; CHECK-NEXT:   S_ENDPGM 0
   call void asm sideeffect "; def a0", "~{a0}"(), !srcloc !0
   ret void
 }
 
 define i32 @asm_vgpr_early_clobber() {
   ; CHECK-LABEL: name: asm_vgpr_early_clobber
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $sgpr30_sgpr31
-  ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK-NEXT:   INLINEASM &"v_mov_b32 $0, 7; v_mov_b32 $1, 7", 1 /* sideeffect attdialect */, 1835019 /* regdef-ec:VGPR_32 */, def early-clobber %1, 1835019 /* regdef-ec:VGPR_32 */, def early-clobber %2, !0
+  ; CHECK-NEXT:   INLINEASM &"v_mov_b32 $0, 7; v_mov_b32 $1, 7", 1 /* sideeffect attdialect */, 1835019 /* regdef-ec:VGPR_32 */, def early-clobber %0, 1835019 /* regdef-ec:VGPR_32 */, def early-clobber %1, !0
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY %0
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY %1
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY %2
-  ; CHECK-NEXT:   [[ADD:%[0-9]+]]:_(s32) = G_ADD [[COPY1]], [[COPY2]]
+  ; CHECK-NEXT:   [[ADD:%[0-9]+]]:_(s32) = G_ADD [[COPY]], [[COPY1]]
   ; CHECK-NEXT:   $vgpr0 = COPY [[ADD]](s32)
-  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   call { i32, i32 } asm sideeffect "v_mov_b32 $0, 7; v_mov_b32 $1, 7", "=&v,=&v"(), !srcloc !0
   %asmresult = extractvalue { i32, i32 } %1, 0
   %asmresult1 = extractvalue { i32, i32 } %1, 1
   %add = add i32 %asmresult, %asmresult1
   ret i32 %add
 }
 
 define i32 @test_specific_vgpr_output() nounwind {
   ; CHECK-LABEL: name: test_specific_vgpr_output
   ; CHECK: bb.1.entry:
-  ; CHECK-NEXT:   liveins: $sgpr30_sgpr31
-  ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   INLINEASM &"v_mov_b32 v1, 7", 0 /* attdialect */, 10 /* regdef */, implicit-def $vgpr1
-  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK-NEXT:   $vgpr0 = COPY [[COPY1]](s32)
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY2]], implicit $vgpr0
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   $vgpr0 = COPY [[COPY]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
 entry:
   %0 = tail call i32 asm "v_mov_b32 v1, 7", "={v1}"() nounwind
   ret i32 %0
 }
 
 define i32 @test_single_vgpr_output() nounwind {
   ; CHECK-LABEL: name: test_single_vgpr_output
   ; CHECK: bb.1.entry:
-  ; CHECK-NEXT:   liveins: $sgpr30_sgpr31
-  ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK-NEXT:   INLINEASM &"v_mov_b32 $0, 7", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def %1
-  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY %1
-  ; CHECK-NEXT:   $vgpr0 = COPY [[COPY1]](s32)
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY2]], implicit $vgpr0
+  ; CHECK-NEXT:   INLINEASM &"v_mov_b32 $0, 7", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def %0
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY %0
+  ; CHECK-NEXT:   $vgpr0 = COPY [[COPY]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
 entry:
   %0 = tail call i32 asm "v_mov_b32 $0, 7", "=v"() nounwind
   ret i32 %0
 }
 
 define i32 @test_single_sgpr_output_s32() nounwind {
   ; CHECK-LABEL: name: test_single_sgpr_output_s32
   ; CHECK: bb.1.entry:
-  ; CHECK-NEXT:   liveins: $sgpr30_sgpr31
-  ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK-NEXT:   INLINEASM &"s_mov_b32 $0, 7", 0 /* attdialect */, 1966090 /* regdef:SReg_32 */, def %1
-  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY %1
-  ; CHECK-NEXT:   $vgpr0 = COPY [[COPY1]](s32)
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY2]], implicit $vgpr0
+  ; CHECK-NEXT:   INLINEASM &"s_mov_b32 $0, 7", 0 /* attdialect */, 1966090 /* regdef:SReg_32 */, def %0
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY %0
+  ; CHECK-NEXT:   $vgpr0 = COPY [[COPY]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
 entry:
   %0 = tail call i32 asm "s_mov_b32 $0, 7", "=s"() nounwind
   ret i32 %0
 }
 
 ; Check support for returning several floats
 define float @test_multiple_register_outputs_same() #0 {
   ; CHECK-LABEL: name: test_multiple_register_outputs_same
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $sgpr30_sgpr31
-  ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK-NEXT:   INLINEASM &"v_mov_b32 $0, 0; v_mov_b32 $1, 1", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def %1, 1835018 /* regdef:VGPR_32 */, def %2
+  ; CHECK-NEXT:   INLINEASM &"v_mov_b32 $0, 0; v_mov_b32 $1, 1", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def %0, 1835018 /* regdef:VGPR_32 */, def %1
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY %0
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY %1
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY %2
-  ; CHECK-NEXT:   [[FADD:%[0-9]+]]:_(s32) = G_FADD [[COPY1]], [[COPY2]]
+  ; CHECK-NEXT:   [[FADD:%[0-9]+]]:_(s32) = G_FADD [[COPY]], [[COPY1]]
   ; CHECK-NEXT:   $vgpr0 = COPY [[FADD]](s32)
-  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %1 = call { float, float } asm "v_mov_b32 $0, 0; v_mov_b32 $1, 1", "=v,=v"()
   %asmresult = extractvalue { float, float } %1, 0
   %asmresult1 = extractvalue { float, float } %1, 1
   %add = fadd float %asmresult, %asmresult1
   ret float %add
 }
 
 ; Check support for returning several floats
 define double @test_multiple_register_outputs_mixed() #0 {
   ; CHECK-LABEL: name: test_multiple_register_outputs_mixed
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $sgpr30_sgpr31
-  ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK-NEXT:   INLINEASM &"v_mov_b32 $0, 0; v_add_f64 $1, 0, 0", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def %1, 2949130 /* regdef:VReg_64 */, def %2
-  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY %1
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s64) = COPY %2
-  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY2]](s64)
+  ; CHECK-NEXT:   INLINEASM &"v_mov_b32 $0, 0; v_add_f64 $1, 0, 0", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def %0, 2949130 /* regdef:VReg_64 */, def %1
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY %0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s64) = COPY %1
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](s64)
   ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %1 = call { float, double } asm "v_mov_b32 $0, 0; v_add_f64 $1, 0, 0", "=v,=v"()
   %asmresult = extractvalue { float, double } %1, 1
   ret double %asmresult
 }
 
 
 define float @test_vector_output() nounwind {
   ; CHECK-LABEL: name: test_vector_output
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $sgpr30_sgpr31
-  ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
   ; CHECK-NEXT:   INLINEASM &"v_add_f64 $0, 0, 0", 1 /* sideeffect attdialect */, 10 /* regdef */, implicit-def $vgpr14_vgpr15
-  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(<2 x s32>) = COPY $vgpr14_vgpr15
-  ; CHECK-NEXT:   [[EVEC:%[0-9]+]]:_(s32) = G_EXTRACT_VECTOR_ELT [[COPY1]](<2 x s32>), [[C]](s32)
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(<2 x s32>) = COPY $vgpr14_vgpr15
+  ; CHECK-NEXT:   [[EVEC:%[0-9]+]]:_(s32) = G_EXTRACT_VECTOR_ELT [[COPY]](<2 x s32>), [[C]](s32)
   ; CHECK-NEXT:   $vgpr0 = COPY [[EVEC]](s32)
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY2]], implicit $vgpr0
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %1 = tail call <2 x float> asm sideeffect "v_add_f64 $0, 0, 0", "={v[14:15]}"() nounwind
   %2 = extractelement <2 x float> %1, i32 0
   ret float %2
 }
 
 define amdgpu_kernel void @test_input_vgpr_imm() {
   ; CHECK-LABEL: name: test_input_vgpr_imm
   ; CHECK: bb.1 (%ir-block.0):
define amdgpu_kernel void @test_input_imm() {
   call void asm sideeffect "s_mov_b32 s0, $0", "i"(i32 42)
   call void asm sideeffect "s_mov_b64 s[0:1], $0", "i"(i64 42)
   ret void
 }
 
 define float @test_input_vgpr(i32 %src) nounwind {
   ; CHECK-LABEL: name: test_input_vgpr
   ; CHECK: bb.1.entry:
-  ; CHECK-NEXT:   liveins: $vgpr0, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:vgpr_32 = COPY [[COPY]](s32)
-  ; CHECK-NEXT:   INLINEASM &"v_add_f32 $0, 1.0, $1", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def %2, 1835017 /* reguse:VGPR_32 */, [[COPY2]]
-  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY %2
-  ; CHECK-NEXT:   $vgpr0 = COPY [[COPY3]](s32)
-  ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:vgpr_32 = COPY [[COPY]](s32)
+  ; CHECK-NEXT:   INLINEASM &"v_add_f32 $0, 1.0, $1", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def %1, 1835017 /* reguse:VGPR_32 */, [[COPY1]]
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY %1
+  ; CHECK-NEXT:   $vgpr0 = COPY [[COPY2]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
 entry:
   %0 = tail call float asm "v_add_f32 $0, 1.0, $1", "=v,v"(i32 %src) nounwind
   ret float %0
 }
 
 define i32 @test_memory_constraint(i32 addrspace(3)* %a) nounwind {
   ; CHECK-LABEL: name: test_memory_constraint
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $vgpr0, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(p3) = COPY $vgpr0
-  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK-NEXT:   INLINEASM &"ds_read_b32 $0, $1", 8 /* mayload attdialect */, 1835018 /* regdef:VGPR_32 */, def %2, 196622 /* mem:m */, [[COPY]](p3)
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY %2
-  ; CHECK-NEXT:   $vgpr0 = COPY [[COPY2]](s32)
-  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; CHECK-NEXT:   INLINEASM &"ds_read_b32 $0, $1", 8 /* mayload attdialect */, 1835018 /* regdef:VGPR_32 */, def %1, 196622 /* mem:m */, [[COPY]](p3)
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY %1
+  ; CHECK-NEXT:   $vgpr0 = COPY [[COPY1]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %1 = tail call i32 asm "ds_read_b32 $0, $1", "=v,*m"(i32 addrspace(3)* %a)
   ret i32 %1
 }
 
 define i32 @test_vgpr_matching_constraint(i32 %a) nounwind {
   ; CHECK-LABEL: name: test_vgpr_matching_constraint
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $vgpr0, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 1
   ; CHECK-NEXT:   [[AND:%[0-9]+]]:_(s32) = G_AND [[COPY]], [[C]]
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:vgpr_32 = COPY [[AND]](s32)
-  ; CHECK-NEXT:   INLINEASM &";", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def %4, 2147483657 /* reguse tiedto:$0 */, [[COPY2]](tied-def 3)
-  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY %4
-  ; CHECK-NEXT:   $vgpr0 = COPY [[COPY3]](s32)
-  ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:vgpr_32 = COPY [[AND]](s32)
+  ; CHECK-NEXT:   INLINEASM &";", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def %3, 2147483657 /* reguse tiedto:$0 */, [[COPY1]](tied-def 3)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY %3
+  ; CHECK-NEXT:   $vgpr0 = COPY [[COPY2]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %and = and i32 %a, 1
   %asm = call i32 asm sideeffect ";", "=v,0"(i32 %and)
   ret i32 %asm
 }
 
 define i32 @test_sgpr_matching_constraint() nounwind {
   ; CHECK-LABEL: name: test_sgpr_matching_constraint
   ; CHECK: bb.1.entry:
-  ; CHECK-NEXT:   liveins: $sgpr30_sgpr31
-  ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK-NEXT:   INLINEASM &"s_mov_b32 $0, 7", 0 /* attdialect */, 1966090 /* regdef:SReg_32 */, def %1
-  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY %1
-  ; CHECK-NEXT:   INLINEASM &"s_mov_b32 $0, 8", 0 /* attdialect */, 1966090 /* regdef:SReg_32 */, def %3
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY %3
+  ; CHECK-NEXT:   INLINEASM &"s_mov_b32 $0, 7", 0 /* attdialect */, 1966090 /* regdef:SReg_32 */, def %0
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY %0
+  ; CHECK-NEXT:   INLINEASM &"s_mov_b32 $0, 8", 0 /* attdialect */, 1966090 /* regdef:SReg_32 */, def %2
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY %2
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY [[COPY]](s32)
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY [[COPY1]](s32)
-  ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:sreg_32 = COPY [[COPY2]](s32)
-  ; CHECK-NEXT:   INLINEASM &"s_add_u32 $0, $1, $2", 0 /* attdialect */, 1966090 /* regdef:SReg_32 */, def %5, 1966089 /* reguse:SReg_32 */, [[COPY3]], 2147483657 /* reguse tiedto:$0 */, [[COPY4]](tied-def 3)
-  ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:_(s32) = COPY %5
-  ; CHECK-NEXT:   $vgpr0 = COPY [[COPY5]](s32)
-  ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY6]], implicit $vgpr0
+  ; CHECK-NEXT:   INLINEASM &"s_add_u32 $0, $1, $2", 0 /* attdialect */, 1966090 /* regdef:SReg_32 */, def %4, 1966089 /* reguse:SReg_32 */, [[COPY2]], 2147483657 /* reguse tiedto:$0 */, [[COPY3]](tied-def 3)
+  ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:_(s32) = COPY %4
+  ; CHECK-NEXT:   $vgpr0 = COPY [[COPY4]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
 entry:
   %asm0 = tail call i32 asm "s_mov_b32 $0, 7", "=s"() nounwind
   %asm1 = tail call i32 asm "s_mov_b32 $0, 8", "=s"() nounwind
   %asm2 = tail call i32 asm "s_add_u32 $0, $1, $2", "=s,s,0"(i32 %asm0, i32 %asm1) nounwind
   ret i32 %asm2
 }
 
 define void @test_many_matching_constraints(i32 %a, i32 %b, i32 %c) nounwind {
   ; CHECK-LABEL: name: test_many_matching_constraints
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
-  ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[COPY2]](s32)
-  ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[COPY]](s32)
-  ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[COPY1]](s32)
-  ; CHECK-NEXT:   INLINEASM &"; ", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def %4, 1835018 /* regdef:VGPR_32 */, def %5, 1835018 /* regdef:VGPR_32 */, def %6, 2147483657 /* reguse tiedto:$0 */, [[COPY4]](tied-def 3), 2147614729 /* reguse tiedto:$2 */, [[COPY5]](tied-def 7), 2147549193 /* reguse tiedto:$1 */, [[COPY6]](tied-def 5)
+  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:vgpr_32 = COPY [[COPY2]](s32)
+  ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[COPY]](s32)
+  ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[COPY1]](s32)
+  ; CHECK-NEXT:   INLINEASM &"; ", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def %3, 1835018 /* regdef:VGPR_32 */, def %4, 1835018 /* regdef:VGPR_32 */, def %5, 2147483657 /* reguse tiedto:$0 */, [[COPY3]](tied-def 3), 2147614729 /* reguse tiedto:$2 */, [[COPY4]](tied-def 7), 2147549193 /* reguse tiedto:$1 */, [[COPY5]](tied-def 5)
+  ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:_(s32) = COPY %3
   ; CHECK-NEXT:   [[COPY7:%[0-9]+]]:_(s32) = COPY %4
   ; CHECK-NEXT:   [[COPY8:%[0-9]+]]:_(s32) = COPY %5
-  ; CHECK-NEXT:   [[COPY9:%[0-9]+]]:_(s32) = COPY %6
+  ; CHECK-NEXT:   G_STORE [[COPY6]](s32), [[DEF]](p1) :: (store (s32) into `i32 addrspace(1)* undef`, addrspace 1)
   ; CHECK-NEXT:   G_STORE [[COPY7]](s32), [[DEF]](p1) :: (store (s32) into `i32 addrspace(1)* undef`, addrspace 1)
   ; CHECK-NEXT:   G_STORE [[COPY8]](s32), [[DEF]](p1) :: (store (s32) into `i32 addrspace(1)* undef`, addrspace 1)
-  ; CHECK-NEXT:   G_STORE [[COPY9]](s32), [[DEF]](p1) :: (store (s32) into `i32 addrspace(1)* undef`, addrspace 1)
-  ; CHECK-NEXT:   [[COPY10:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY10]]
+  ; CHECK-NEXT:   SI_RETURN
   %asm = call {i32, i32, i32} asm sideeffect "; ", "=v,=v,=v,0,2,1"(i32 %c, i32 %a, i32 %b)
   %asmresult0 = extractvalue  {i32, i32, i32} %asm, 0
   store i32 %asmresult0, i32 addrspace(1)* undef
   %asmresult1 = extractvalue  {i32, i32, i32} %asm, 1
   store i32 %asmresult1, i32 addrspace(1)* undef
   %asmresult2 = extractvalue  {i32, i32, i32} %asm, 2
   store i32 %asmresult2, i32 addrspace(1)* undef
   ret void
 }
 
 define i32 @test_sgpr_to_vgpr_move_matching_constraint() nounwind {
   ; CHECK-LABEL: name: test_sgpr_to_vgpr_move_matching_constraint
   ; CHECK: bb.1.entry:
-  ; CHECK-NEXT:   liveins: $sgpr30_sgpr31
-  ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK-NEXT:   INLINEASM &"s_mov_b32 $0, 7", 0 /* attdialect */, 1966090 /* regdef:SReg_32 */, def %1
-  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY %1
-  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:vgpr_32 = COPY [[COPY1]](s32)
-  ; CHECK-NEXT:   INLINEASM &"v_mov_b32 $0, $1", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def %3, 2147483657 /* reguse tiedto:$0 */, [[COPY2]](tied-def 3)
-  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY %3
-  ; CHECK-NEXT:   $vgpr0 = COPY [[COPY3]](s32)
-  ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK-NEXT:   S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+  ; CHECK-NEXT:   INLINEASM &"s_mov_b32 $0, 7", 0 /* attdialect */, 1966090 /* regdef:SReg_32 */, def %0
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY %0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:vgpr_32 = COPY [[COPY]](s32)
+  ; CHECK-NEXT:   INLINEASM &"v_mov_b32 $0, $1", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def %2, 2147483657 /* reguse tiedto:$0 */, [[COPY1]](tied-def 3)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY %2
+  ; CHECK-NEXT:   $vgpr0 = COPY [[COPY2]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
 entry:
   %asm0 = tail call i32 asm "s_mov_b32 $0, 7", "=s"() nounwind
   %asm1 = tail call i32 asm "v_mov_b32 $0, $1", "=v,0"(i32 %asm0) nounwind
   ret i32 %asm1
 }
 
 define amdgpu_kernel void @asm_constraint_n_n()  {
   ; CHECK-LABEL: name: asm_constraint_n_n
   ; CHECK: bb.1 (%ir-block.0):
   ; CHECK-NEXT:   INLINEASM &"s_trap ${0:n}", 1 /* sideeffect attdialect */, 13 /* imm */, 10
   ; CHECK-NEXT:   S_ENDPGM 0
   tail call void asm sideeffect "s_trap ${0:n}", "n"(i32 10) #1
   ret void
 }
 
 !0 = !{i32 70}

llvm/test/CodeGen/AMDGPU/GlobalISel/irtranslator-memory-intrinsics.ll

 ; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
 ; RUN: llc -global-isel -march=amdgcn -O0 -stop-after=irtranslator -verify-machineinstrs %s -o - | FileCheck %s
 
 ; Size operand should be the minimum of the two pointer sizes.
 
 define void @test_memcpy_p1_p3_i64(i8 addrspace(1)* %dst, i8 addrspace(3)* %src) {
   ; CHECK-LABEL: name: test_memcpy_p1_p3_i64
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(p3) = COPY $vgpr2
-  ; CHECK:   [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 256
-  ; CHECK:   [[TRUNC:%[0-9]+]]:_(s32) = G_TRUNC [[C]](s64)
-  ; CHECK:   G_MEMCPY [[MV]](p1), [[COPY2]](p3), [[TRUNC]](s32), 0 :: (store (s8) into %ir.dst, addrspace 1), (load (s8) from %ir.src, addrspace 3)
-  ; CHECK:   [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-  ; CHECK:   S_SETPC_B64_return [[COPY4]]
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(p3) = COPY $vgpr2
+  ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 256
+  ; CHECK-NEXT:   [[TRUNC:%[0-9]+]]:_(s32) = G_TRUNC [[C]](s64)
+  ; CHECK-NEXT:   G_MEMCPY [[MV]](p1), [[COPY2]](p3), [[TRUNC]](s32), 0 :: (store (s8) into %ir.dst, addrspace 1), (load (s8) from %ir.src, addrspace 3)
+  ; CHECK-NEXT:   SI_RETURN
   call void @llvm.memcpy.p1i8.p3i8.i64(i8 addrspace(1)* %dst, i8 addrspace(3)* %src, i64 256, i1 false)
   ret void
 }
 
 define void @test_memcpy_p1_p3_i32(i8 addrspace(1)* %dst, i8 addrspace(3)* %src) {
   ; CHECK-LABEL: name: test_memcpy_p1_p3_i32
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(p3) = COPY $vgpr2
-  ; CHECK:   [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 256
-  ; CHECK:   G_MEMCPY [[MV]](p1), [[COPY2]](p3), [[C]](s32), 0 :: (store (s8) into %ir.dst, addrspace 1), (load (s8) from %ir.src, addrspace 3)
-  ; CHECK:   [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-  ; CHECK:   S_SETPC_B64_return [[COPY4]]
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(p3) = COPY $vgpr2
+  ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 256
+  ; CHECK-NEXT:   G_MEMCPY [[MV]](p1), [[COPY2]](p3), [[C]](s32), 0 :: (store (s8) into %ir.dst, addrspace 1), (load (s8) from %ir.src, addrspace 3)
+  ; CHECK-NEXT:   SI_RETURN
   call void @llvm.memcpy.p1i8.p3i8.i32(i8 addrspace(1)* %dst, i8 addrspace(3)* %src, i32 256, i1 false)
   ret void
 }
 
 define void @test_memcpy_p1_p3_i16(i8 addrspace(1)* %dst, i8 addrspace(3)* %src) {
   ; CHECK-LABEL: name: test_memcpy_p1_p3_i16
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(p3) = COPY $vgpr2
-  ; CHECK:   [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[C:%[0-9]+]]:_(s16) = G_CONSTANT i16 256
-  ; CHECK:   [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[C]](s16)
-  ; CHECK:   G_MEMCPY [[MV]](p1), [[COPY2]](p3), [[ZEXT]](s32), 0 :: (store (s8) into %ir.dst, addrspace 1), (load (s8) from %ir.src, addrspace 3)
-  ; CHECK:   [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-  ; CHECK:   S_SETPC_B64_return [[COPY4]]
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(p3) = COPY $vgpr2
+  ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s16) = G_CONSTANT i16 256
+  ; CHECK-NEXT:   [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[C]](s16)
+  ; CHECK-NEXT:   G_MEMCPY [[MV]](p1), [[COPY2]](p3), [[ZEXT]](s32), 0 :: (store (s8) into %ir.dst, addrspace 1), (load (s8) from %ir.src, addrspace 3)
+  ; CHECK-NEXT:   SI_RETURN
   call void @llvm.memcpy.p1i8.p3i8.i16(i8 addrspace(1)* %dst, i8 addrspace(3)* %src, i16 256, i1 false)
   ret void
 }
 
 define void @test_memcpy_p3_p1_i64(i8 addrspace(3)* %dst, i8 addrspace(1)* %src) {
   ; CHECK-LABEL: name: test_memcpy_p3_p1_i64
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(p3) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; CHECK:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY1]](s32), [[COPY2]](s32)
-  ; CHECK:   [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 256
-  ; CHECK:   [[TRUNC:%[0-9]+]]:_(s32) = G_TRUNC [[C]](s64)
-  ; CHECK:   G_MEMCPY [[COPY]](p3), [[MV]](p1), [[TRUNC]](s32), 0 :: (store (s8) into %ir.dst, addrspace 3), (load (s8) from %ir.src, addrspace 1)
-  ; CHECK:   [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-  ; CHECK:   S_SETPC_B64_return [[COPY4]]
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(p3) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
+  ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY1]](s32), [[COPY2]](s32)
+  ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 256
+  ; CHECK-NEXT:   [[TRUNC:%[0-9]+]]:_(s32) = G_TRUNC [[C]](s64)
+  ; CHECK-NEXT:   G_MEMCPY [[COPY]](p3), [[MV]](p1), [[TRUNC]](s32), 0 :: (store (s8) into %ir.dst, addrspace 3), (load (s8) from %ir.src, addrspace 1)
+  ; CHECK-NEXT:   SI_RETURN
   call void @llvm.memcpy.p3i8.p1i8.i64(i8 addrspace(3)* %dst, i8 addrspace(1)* %src, i64 256, i1 false)
   ret void
 }
 
 define void @test_memcpy_p3_p1_i32(i8 addrspace(3)* %dst, i8 addrspace(1)* %src) {
   ; CHECK-LABEL: name: test_memcpy_p3_p1_i32
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(p3) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; CHECK:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY1]](s32), [[COPY2]](s32)
-  ; CHECK:   [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 256
-  ; CHECK:   G_MEMCPY [[COPY]](p3), [[MV]](p1), [[C]](s32), 0 :: (store (s8) into %ir.dst, addrspace 3), (load (s8) from %ir.src, addrspace 1)
-  ; CHECK:   [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-  ; CHECK:   S_SETPC_B64_return [[COPY4]]
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(p3) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
+  ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY1]](s32), [[COPY2]](s32)
+  ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 256
+  ; CHECK-NEXT:   G_MEMCPY [[COPY]](p3), [[MV]](p1), [[C]](s32), 0 :: (store (s8) into %ir.dst, addrspace 3), (load (s8) from %ir.src, addrspace 1)
+  ; CHECK-NEXT:   SI_RETURN
   call void @llvm.memcpy.p3i8.p1i8.i32(i8 addrspace(3)* %dst, i8 addrspace(1)* %src, i32 256, i1 false)
   ret void
 }
 
 define void @test_memcpy_p3_p1_i16(i8 addrspace(3)* %dst, i8 addrspace(1)* %src) {
   ; CHECK-LABEL: name: test_memcpy_p3_p1_i16
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(p3) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; CHECK:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY1]](s32), [[COPY2]](s32)
-  ; CHECK:   [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[C:%[0-9]+]]:_(s16) = G_CONSTANT i16 256
-  ; CHECK:   [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[C]](s16)
-  ; CHECK:   G_MEMCPY [[COPY]](p3), [[MV]](p1), [[ZEXT]](s32), 0 :: (store (s8) into %ir.dst, addrspace 3), (load (s8) from %ir.src, addrspace 1)
-  ; CHECK:   [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-  ; CHECK:   S_SETPC_B64_return [[COPY4]]
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(p3) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
+  ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY1]](s32), [[COPY2]](s32)
+  ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s16) = G_CONSTANT i16 256
+  ; CHECK-NEXT:   [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[C]](s16)
+  ; CHECK-NEXT:   G_MEMCPY [[COPY]](p3), [[MV]](p1), [[ZEXT]](s32), 0 :: (store (s8) into %ir.dst, addrspace 3), (load (s8) from %ir.src, addrspace 1)
+  ; CHECK-NEXT:   SI_RETURN
   call void @llvm.memcpy.p3i8.p1i8.i16(i8 addrspace(3)* %dst, i8 addrspace(1)* %src, i16 256, i1 false)
   ret void
 }
 
 define void @test_memmove_p1_p3_i64(i8 addrspace(1)* %dst, i8 addrspace(3)* %src) {
   ; CHECK-LABEL: name: test_memmove_p1_p3_i64
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(p3) = COPY $vgpr2
-  ; CHECK:   [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 256
-  ; CHECK:   [[TRUNC:%[0-9]+]]:_(s32) = G_TRUNC [[C]](s64)
-  ; CHECK:   G_MEMMOVE [[MV]](p1), [[COPY2]](p3), [[TRUNC]](s32), 0 :: (store (s8) into %ir.dst, addrspace 1), (load (s8) from %ir.src, addrspace 3)
-  ; CHECK:   [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-  ; CHECK:   S_SETPC_B64_return [[COPY4]]
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(p3) = COPY $vgpr2
+  ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 256
+  ; CHECK-NEXT:   [[TRUNC:%[0-9]+]]:_(s32) = G_TRUNC [[C]](s64)
+  ; CHECK-NEXT:   G_MEMMOVE [[MV]](p1), [[COPY2]](p3), [[TRUNC]](s32), 0 :: (store (s8) into %ir.dst, addrspace 1), (load (s8) from %ir.src, addrspace 3)
+  ; CHECK-NEXT:   SI_RETURN
   call void @llvm.memmove.p1i8.p3i8.i64(i8 addrspace(1)* %dst, i8 addrspace(3)* %src, i64 256, i1 false)
   ret void
 }
 
 define void @test_memmove_p1_p3_i32(i8 addrspace(1)* %dst, i8 addrspace(3)* %src) {
   ; CHECK-LABEL: name: test_memmove_p1_p3_i32
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(p3) = COPY $vgpr2
-  ; CHECK:   [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 256
-  ; CHECK:   G_MEMMOVE [[MV]](p1), [[COPY2]](p3), [[C]](s32), 0 :: (store (s8) into %ir.dst, addrspace 1), (load (s8) from %ir.src, addrspace 3)
-  ; CHECK:   [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-  ; CHECK:   S_SETPC_B64_return [[COPY4]]
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(p3) = COPY $vgpr2
+  ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 256
+  ; CHECK-NEXT:   G_MEMMOVE [[MV]](p1), [[COPY2]](p3), [[C]](s32), 0 :: (store (s8) into %ir.dst, addrspace 1), (load (s8) from %ir.src, addrspace 3)
+  ; CHECK-NEXT:   SI_RETURN
   call void @llvm.memmove.p1i8.p3i8.i32(i8 addrspace(1)* %dst, i8 addrspace(3)* %src, i32 256, i1 false)
   ret void
 }
 
 define void @test_memmove_p1_p3_i16(i8 addrspace(1)* %dst, i8 addrspace(3)* %src) {
   ; CHECK-LABEL: name: test_memmove_p1_p3_i16
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(p3) = COPY $vgpr2
-  ; CHECK:   [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[C:%[0-9]+]]:_(s16) = G_CONSTANT i16 256
-  ; CHECK:   [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[C]](s16)
-  ; CHECK:   G_MEMMOVE [[MV]](p1), [[COPY2]](p3), [[ZEXT]](s32), 0 :: (store (s8) into %ir.dst, addrspace 1), (load (s8) from %ir.src, addrspace 3)
-  ; CHECK:   [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-  ; CHECK:   S_SETPC_B64_return [[COPY4]]
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(p3) = COPY $vgpr2
+  ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s16) = G_CONSTANT i16 256
+  ; CHECK-NEXT:   [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[C]](s16)
+  ; CHECK-NEXT:   G_MEMMOVE [[MV]](p1), [[COPY2]](p3), [[ZEXT]](s32), 0 :: (store (s8) into %ir.dst, addrspace 1), (load (s8) from %ir.src, addrspace 3)
+  ; CHECK-NEXT:   SI_RETURN
   call void @llvm.memmove.p1i8.p3i8.i16(i8 addrspace(1)* %dst, i8 addrspace(3)* %src, i16 256, i1 false)
   ret void
 }
 
 define void @test_memset_p1_i64(i8 addrspace(1)* %dst, i8 %val) {
   ; CHECK-LABEL: name: test_memset_p1_i64
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; CHECK:   [[TRUNC:%[0-9]+]]:_(s8) = G_TRUNC [[COPY2]](s32)
-  ; CHECK:   [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 256
-  ; CHECK:   G_MEMSET [[MV]](p1), [[TRUNC]](s8), [[C]](s64), 0 :: (store (s8) into %ir.dst, addrspace 1)
-  ; CHECK:   [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-  ; CHECK:   S_SETPC_B64_return [[COPY4]]
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
+  ; CHECK-NEXT:   [[TRUNC:%[0-9]+]]:_(s8) = G_TRUNC [[COPY2]](s32)
+  ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 256
+  ; CHECK-NEXT:   G_MEMSET [[MV]](p1), [[TRUNC]](s8), [[C]](s64), 0 :: (store (s8) into %ir.dst, addrspace 1)
+  ; CHECK-NEXT:   SI_RETURN
   call void @llvm.memset.p1i8.i64(i8 addrspace(1)* %dst, i8 %val, i64 256, i1 false)
   ret void
 }
 
 define void @test_memset_p1_i32(i8 addrspace(1)* %dst, i8 %val) {
   ; CHECK-LABEL: name: test_memset_p1_i32
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; CHECK:   [[TRUNC:%[0-9]+]]:_(s8) = G_TRUNC [[COPY2]](s32)
-  ; CHECK:   [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 256
-  ; CHECK:   [[ZEXT:%[0-9]+]]:_(s64) = G_ZEXT [[C]](s32)
-  ; CHECK:   G_MEMSET [[MV]](p1), [[TRUNC]](s8), [[ZEXT]](s64), 0 :: (store (s8) into %ir.dst, addrspace 1)
-  ; CHECK:   [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-  ; CHECK:   S_SETPC_B64_return [[COPY4]]
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
+  ; CHECK-NEXT:   [[TRUNC:%[0-9]+]]:_(s8) = G_TRUNC [[COPY2]](s32)
+  ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 256
+  ; CHECK-NEXT:   [[ZEXT:%[0-9]+]]:_(s64) = G_ZEXT [[C]](s32)
+  ; CHECK-NEXT:   G_MEMSET [[MV]](p1), [[TRUNC]](s8), [[ZEXT]](s64), 0 :: (store (s8) into %ir.dst, addrspace 1)
+  ; CHECK-NEXT:   SI_RETURN
   call void @llvm.memset.p1i8.i32(i8 addrspace(1)* %dst, i8 %val, i32 256, i1 false)
   ret void
 }
 
 define void @test_memset_p1_i16(i8 addrspace(1)* %dst, i8 %val) {
   ; CHECK-LABEL: name: test_memset_p1_i16
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; CHECK:   [[TRUNC:%[0-9]+]]:_(s8) = G_TRUNC [[COPY2]](s32)
-  ; CHECK:   [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[C:%[0-9]+]]:_(s16) = G_CONSTANT i16 256
-  ; CHECK:   [[ZEXT:%[0-9]+]]:_(s64) = G_ZEXT [[C]](s16)
-  ; CHECK:   G_MEMSET [[MV]](p1), [[TRUNC]](s8), [[ZEXT]](s64), 0 :: (store (s8) into %ir.dst, addrspace 1)
-  ; CHECK:   [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-  ; CHECK:   S_SETPC_B64_return [[COPY4]]
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
+  ; CHECK-NEXT:   [[TRUNC:%[0-9]+]]:_(s8) = G_TRUNC [[COPY2]](s32)
+  ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s16) = G_CONSTANT i16 256
+  ; CHECK-NEXT:   [[ZEXT:%[0-9]+]]:_(s64) = G_ZEXT [[C]](s16)
+  ; CHECK-NEXT:   G_MEMSET [[MV]](p1), [[TRUNC]](s8), [[ZEXT]](s64), 0 :: (store (s8) into %ir.dst, addrspace 1)
+  ; CHECK-NEXT:   SI_RETURN
   call void @llvm.memset.p1i8.i16(i8 addrspace(1)* %dst, i8 %val, i16 256, i1 false)
   ret void
 }
 
 define void @test_memset_p3_i64(i8 addrspace(3)* %dst, i8 %val) {
   ; CHECK-LABEL: name: test_memset_p3_i64
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(p3) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[TRUNC:%[0-9]+]]:_(s8) = G_TRUNC [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 256
-  ; CHECK:   [[TRUNC1:%[0-9]+]]:_(s32) = G_TRUNC [[C]](s64)
-  ; CHECK:   G_MEMSET [[COPY]](p3), [[TRUNC]](s8), [[TRUNC1]](s32), 0 :: (store (s8) into %ir.dst, addrspace 3)
-  ; CHECK:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK:   S_SETPC_B64_return [[COPY3]]
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(p3) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[TRUNC:%[0-9]+]]:_(s8) = G_TRUNC [[COPY1]](s32)
+  ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 256
+  ; CHECK-NEXT:   [[TRUNC1:%[0-9]+]]:_(s32) = G_TRUNC [[C]](s64)
+  ; CHECK-NEXT:   G_MEMSET [[COPY]](p3), [[TRUNC]](s8), [[TRUNC1]](s32), 0 :: (store (s8) into %ir.dst, addrspace 3)
+  ; CHECK-NEXT:   SI_RETURN
   call void @llvm.memset.p3i8.i64(i8 addrspace(3)* %dst, i8 %val, i64 256, i1 false)
   ret void
 }
 
 define void @test_memset_p3_i32(i8 addrspace(3)* %dst, i8 %val) {
   ; CHECK-LABEL: name: test_memset_p3_i32
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(p3) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[TRUNC:%[0-9]+]]:_(s8) = G_TRUNC [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 256
-  ; CHECK:   G_MEMSET [[COPY]](p3), [[TRUNC]](s8), [[C]](s32), 0 :: (store (s8) into %ir.dst, addrspace 3)
-  ; CHECK:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK:   S_SETPC_B64_return [[COPY3]]
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(p3) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[TRUNC:%[0-9]+]]:_(s8) = G_TRUNC [[COPY1]](s32)
+  ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 256
+  ; CHECK-NEXT:   G_MEMSET [[COPY]](p3), [[TRUNC]](s8), [[C]](s32), 0 :: (store (s8) into %ir.dst, addrspace 3)
+  ; CHECK-NEXT:   SI_RETURN
   call void @llvm.memset.p3i8.i32(i8 addrspace(3)* %dst, i8 %val, i32 256, i1 false)
   ret void
 }
 
 define void @test_memset_p3_i16(i8 addrspace(3)* %dst, i8 %val) {
   ; CHECK-LABEL: name: test_memset_p3_i16
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(p3) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[TRUNC:%[0-9]+]]:_(s8) = G_TRUNC [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[C:%[0-9]+]]:_(s16) = G_CONSTANT i16 256
-  ; CHECK:   [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[C]](s16)
-  ; CHECK:   G_MEMSET [[COPY]](p3), [[TRUNC]](s8), [[ZEXT]](s32), 0 :: (store (s8) into %ir.dst, addrspace 3)
-  ; CHECK:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK:   S_SETPC_B64_return [[COPY3]]
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(p3) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[TRUNC:%[0-9]+]]:_(s8) = G_TRUNC [[COPY1]](s32)
+  ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s16) = G_CONSTANT i16 256
+  ; CHECK-NEXT:   [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[C]](s16)
+  ; CHECK-NEXT:   G_MEMSET [[COPY]](p3), [[TRUNC]](s8), [[ZEXT]](s32), 0 :: (store (s8) into %ir.dst, addrspace 3)
+  ; CHECK-NEXT:   SI_RETURN
   call void @llvm.memset.p3i8.i16(i8 addrspace(3)* %dst, i8 %val, i16 256, i1 false)
   ret void
 }
 
 declare void @llvm.memcpy.p1i8.p3i8.i64(i8 addrspace(1)* noalias nocapture writeonly, i8 addrspace(3)* noalias nocapture readonly, i64, i1 immarg) #0
 declare void @llvm.memcpy.p1i8.p3i8.i32(i8 addrspace(1)* noalias nocapture writeonly, i8 addrspace(3)* noalias nocapture readonly, i32, i1 immarg) #0
 declare void @llvm.memcpy.p1i8.p3i8.i16(i8 addrspace(1)* noalias nocapture writeonly, i8 addrspace(3)* noalias nocapture readonly, i16, i1 immarg) #0
 declare void @llvm.memcpy.p3i8.p1i8.i64(i8 addrspace(3)* noalias nocapture writeonly, i8 addrspace(1)* noalias nocapture readonly, i64, i1 immarg) #0

llvm/test/CodeGen/AMDGPU/GlobalISel/irtranslator-metadata.ll

 ; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
 ; RUN: llc -global-isel -march=amdgcn -mcpu=gfx900 -stop-after=irtranslator -o - %s | FileCheck %s
 ; Make sure intrinsics with metadata arguments are translated
 
 define i32 @reloc_constant() {
   ; CHECK-LABEL: name: reloc_constant
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; CHECK:   [[INT0:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.reloc.constant), !0
   ; We cannot have any specific metadata check here as ConstantAsMetadata is printed as <raw_ptr_val>
   ; CHECK:   [[INT1:%[0-9]+]]:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.reloc.constant), <0x{{[0-9a-f]+}}>
   ; CHECK:   [[SUM:%[0-9]+]]:_(s32) = G_ADD [[INT0]], [[INT1]]
   ; CHECK:   $vgpr0 = COPY [[SUM]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; CHECK:   S_SETPC_B64_return [[COPY1]], implicit $vgpr0
+  ; CHECK:   SI_RETURN implicit $vgpr0
   %val0 = call i32 @llvm.amdgcn.reloc.constant(metadata !0)
   %val1 = call i32 @llvm.amdgcn.reloc.constant(metadata i32 4)
   %res = add i32 %val0, %val1
   ret i32 %res
 }
 
 declare i32 @llvm.amdgcn.reloc.constant(metadata) #0
 
 attributes #0 = { nounwind readnone speculatable willreturn }
 
 !0 = !{!"arst"}

llvm/test/CodeGen/AMDGPU/GlobalISel/irtranslator-ptrmask.ll

 ; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
 ; RUN: llc -global-isel -mtriple=amdgcn-amd-amdhsa -verify-machineinstrs -stop-after=irtranslator < %s | FileCheck %s
 
 define i8* @ptrmask_flat_i64(i8* %ptr, i64 %mask) {
   ; CHECK-LABEL: name: ptrmask_flat_i64
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[MV:%[0-9]+]]:_(p0) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; CHECK:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
-  ; CHECK:   [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
-  ; CHECK:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[PTRMASK:%[0-9]+]]:_(p0) = G_PTRMASK [[MV]], [[MV1]](s64)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[PTRMASK]](p0)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   [[COPY5:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY4]]
-  ; CHECK:   S_SETPC_B64_return [[COPY5]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(p0) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
+  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
+  ; CHECK-NEXT:   [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
+  ; CHECK-NEXT:   [[PTRMASK:%[0-9]+]]:_(p0) = G_PTRMASK [[MV]], [[MV1]](s64)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[PTRMASK]](p0)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %masked = call i8* @llvm.ptrmask.p0i8.i64(i8* %ptr, i64 %mask)
   ret i8* %masked
 }
 
 define i8* @ptrmask_flat_i32(i8* %ptr, i32 %mask) {
   ; CHECK-LABEL: name: ptrmask_flat_i32
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[MV:%[0-9]+]]:_(p0) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; CHECK:   [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[PTRMASK:%[0-9]+]]:_(p0) = G_PTRMASK [[MV]], [[COPY2]](s32)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[PTRMASK]](p0)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-  ; CHECK:   S_SETPC_B64_return [[COPY4]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(p0) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
+  ; CHECK-NEXT:   [[PTRMASK:%[0-9]+]]:_(p0) = G_PTRMASK [[MV]], [[COPY2]](s32)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[PTRMASK]](p0)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %masked = call i8* @llvm.ptrmask.p0i8.i32(i8* %ptr, i32 %mask)
   ret i8* %masked
 }
 
 define i8* @ptrmask_flat_i16(i8* %ptr, i16 %mask) {
   ; CHECK-LABEL: name: ptrmask_flat_i16
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[MV:%[0-9]+]]:_(p0) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; CHECK:   [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
-  ; CHECK:   [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[PTRMASK:%[0-9]+]]:_(p0) = G_PTRMASK [[MV]], [[TRUNC]](s16)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[PTRMASK]](p0)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-  ; CHECK:   S_SETPC_B64_return [[COPY4]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(p0) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
+  ; CHECK-NEXT:   [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY2]](s32)
+  ; CHECK-NEXT:   [[PTRMASK:%[0-9]+]]:_(p0) = G_PTRMASK [[MV]], [[TRUNC]](s16)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[PTRMASK]](p0)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %masked = call i8* @llvm.ptrmask.p0i8.i16(i8* %ptr, i16 %mask)
   ret i8* %masked
 }
 
 define i8* @ptrmask_flat_i1(i8* %ptr, i1 %mask) {
   ; CHECK-LABEL: name: ptrmask_flat_i1
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[MV:%[0-9]+]]:_(p0) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; CHECK:   [[TRUNC:%[0-9]+]]:_(s1) = G_TRUNC [[COPY2]](s32)
-  ; CHECK:   [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[PTRMASK:%[0-9]+]]:_(p0) = G_PTRMASK [[MV]], [[TRUNC]](s1)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[PTRMASK]](p0)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-  ; CHECK:   S_SETPC_B64_return [[COPY4]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(p0) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
+  ; CHECK-NEXT:   [[TRUNC:%[0-9]+]]:_(s1) = G_TRUNC [[COPY2]](s32)
+  ; CHECK-NEXT:   [[PTRMASK:%[0-9]+]]:_(p0) = G_PTRMASK [[MV]], [[TRUNC]](s1)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[PTRMASK]](p0)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %masked = call i8* @llvm.ptrmask.p0i8.i1(i8* %ptr, i1 %mask)
   ret i8* %masked
 }
 
 define i8 addrspace(3)* @ptrmask_local_i64(i8 addrspace(3)* %ptr, i64 %mask) {
   ; CHECK-LABEL: name: ptrmask_local_i64
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(p3) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; CHECK:   [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY1]](s32), [[COPY2]](s32)
-  ; CHECK:   [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[PTRMASK:%[0-9]+]]:_(p3) = G_PTRMASK [[COPY]], [[MV]](s64)
-  ; CHECK:   $vgpr0 = COPY [[PTRMASK]](p3)
-  ; CHECK:   [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-  ; CHECK:   S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(p3) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
+  ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY1]](s32), [[COPY2]](s32)
+  ; CHECK-NEXT:   [[PTRMASK:%[0-9]+]]:_(p3) = G_PTRMASK [[COPY]], [[MV]](s64)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[PTRMASK]](p3)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %masked = call i8 addrspace(3)* @llvm.ptrmask.p3i8.i64(i8 addrspace(3)* %ptr, i64 %mask)
   ret i8 addrspace(3)* %masked
 }
 
 define i8 addrspace(3)* @ptrmask_local_i32(i8 addrspace(3)* %ptr, i32 %mask) {
   ; CHECK-LABEL: name: ptrmask_local_i32
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(p3) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[PTRMASK:%[0-9]+]]:_(p3) = G_PTRMASK [[COPY]], [[COPY1]](s32)
-  ; CHECK:   $vgpr0 = COPY [[PTRMASK]](p3)
-  ; CHECK:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(p3) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[PTRMASK:%[0-9]+]]:_(p3) = G_PTRMASK [[COPY]], [[COPY1]](s32)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[PTRMASK]](p3)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %masked = call i8 addrspace(3)* @llvm.ptrmask.p3i8.i32(i8 addrspace(3)* %ptr, i32 %mask)
   ret i8 addrspace(3)* %masked
 }
 
 define i8 addrspace(3)* @ptrmask_local_i16(i8 addrspace(3)* %ptr, i16 %mask) {
   ; CHECK-LABEL: name: ptrmask_local_i16
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(p3) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[PTRMASK:%[0-9]+]]:_(p3) = G_PTRMASK [[COPY]], [[TRUNC]](s16)
-  ; CHECK:   $vgpr0 = COPY [[PTRMASK]](p3)
-  ; CHECK:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(p3) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
+  ; CHECK-NEXT:   [[PTRMASK:%[0-9]+]]:_(p3) = G_PTRMASK [[COPY]], [[TRUNC]](s16)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[PTRMASK]](p3)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %masked = call i8 addrspace(3)* @llvm.ptrmask.p3i8.i16(i8 addrspace(3)* %ptr, i16 %mask)
   ret i8 addrspace(3)* %masked
 }
 
 define i8 addrspace(3)* @ptrmask_local_i1(i8 addrspace(3)* %ptr, i1 %mask) {
   ; CHECK-LABEL: name: ptrmask_local_i1
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(p3) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[TRUNC:%[0-9]+]]:_(s1) = G_TRUNC [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[PTRMASK:%[0-9]+]]:_(p3) = G_PTRMASK [[COPY]], [[TRUNC]](s1)
-  ; CHECK:   $vgpr0 = COPY [[PTRMASK]](p3)
-  ; CHECK:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(p3) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[TRUNC:%[0-9]+]]:_(s1) = G_TRUNC [[COPY1]](s32)
+  ; CHECK-NEXT:   [[PTRMASK:%[0-9]+]]:_(p3) = G_PTRMASK [[COPY]], [[TRUNC]](s1)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[PTRMASK]](p3)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %masked = call i8 addrspace(3)* @llvm.ptrmask.p3i8.i1(i8 addrspace(3)* %ptr, i1 %mask)
   ret i8 addrspace(3)* %masked
 }
 
 ; Seems to not work
 ; define <2 x i8*> @ptrmask_flat_i64_v2(<2 x i8*> %ptr, <2 x i64> %mask) {
 ;   %masked = call <2 x i8*> @llvm.ptrmask.v2p0i8.v2i64(<2 x i8*> %ptr, <2 x i64> %mask)
 ;   ret <2 x i8*> %masked

llvm/test/CodeGen/AMDGPU/GlobalISel/irtranslator-sat.ll

 ; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
 ; RUN: llc -march=amdgcn -global-isel -stop-after=irtranslator %s -o - | FileCheck %s
 
 define i16 @uaddsat_i16(i16 %lhs, i16 %rhs) {
   ; CHECK-LABEL: name: uaddsat_i16
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[UADDSAT:%[0-9]+]]:_(s16) = G_UADDSAT [[TRUNC]], [[TRUNC1]]
-  ; CHECK:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[UADDSAT]](s16)
-  ; CHECK:   $vgpr0 = COPY [[ANYEXT]](s32)
-  ; CHECK:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
+  ; CHECK-NEXT:   [[UADDSAT:%[0-9]+]]:_(s16) = G_UADDSAT [[TRUNC]], [[TRUNC1]]
+  ; CHECK-NEXT:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[UADDSAT]](s16)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[ANYEXT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %res = call i16 @llvm.uadd.sat.i16(i16 %lhs, i16 %rhs)
   ret i16 %res
 }
 declare i16 @llvm.uadd.sat.i16(i16, i16)
 
 define i32 @uaddsat_i32(i32 %lhs, i32 %rhs) {
   ; CHECK-LABEL: name: uaddsat_i32
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[UADDSAT:%[0-9]+]]:_(s32) = G_UADDSAT [[COPY]], [[COPY1]]
-  ; CHECK:   $vgpr0 = COPY [[UADDSAT]](s32)
-  ; CHECK:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[UADDSAT:%[0-9]+]]:_(s32) = G_UADDSAT [[COPY]], [[COPY1]]
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UADDSAT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %res = call i32 @llvm.uadd.sat.i32(i32 %lhs, i32 %rhs)
   ret i32 %res
 }
 declare i32 @llvm.uadd.sat.i32(i32, i32)
 
 define i64 @uaddsat_i64(i64 %lhs, i64 %rhs) {
   ; CHECK-LABEL: name: uaddsat_i64
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; CHECK:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
-  ; CHECK:   [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
-  ; CHECK:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[UADDSAT:%[0-9]+]]:_(s64) = G_UADDSAT [[MV]], [[MV1]]
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UADDSAT]](s64)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   [[COPY5:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY4]]
-  ; CHECK:   S_SETPC_B64_return [[COPY5]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
+  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
+  ; CHECK-NEXT:   [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
+  ; CHECK-NEXT:   [[UADDSAT:%[0-9]+]]:_(s64) = G_UADDSAT [[MV]], [[MV1]]
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UADDSAT]](s64)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %res = call i64 @llvm.uadd.sat.i64(i64 %lhs, i64 %rhs)
   ret i64 %res
 }
 declare i64 @llvm.uadd.sat.i64(i64, i64)
 
 define <2 x i32> @uaddsat_v2i32(<2 x i32> %lhs, <2 x i32> %rhs) {
   ; CHECK-LABEL: name: uaddsat_v2i32
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; CHECK:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
-  ; CHECK:   [[BUILD_VECTOR1:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY2]](s32), [[COPY3]](s32)
-  ; CHECK:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[UADDSAT:%[0-9]+]]:_(<2 x s32>) = G_UADDSAT [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UADDSAT]](<2 x s32>)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   [[COPY5:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY4]]
-  ; CHECK:   S_SETPC_B64_return [[COPY5]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
+  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
+  ; CHECK-NEXT:   [[BUILD_VECTOR1:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY2]](s32), [[COPY3]](s32)
+  ; CHECK-NEXT:   [[UADDSAT:%[0-9]+]]:_(<2 x s32>) = G_UADDSAT [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UADDSAT]](<2 x s32>)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %res = call <2 x i32> @llvm.uadd.sat.v2i32(<2 x i32> %lhs, <2 x i32> %rhs)
   ret <2 x i32> %res
 }
 declare <2 x i32> @llvm.uadd.sat.v2i32(<2 x i32>, <2 x i32>)
 
 define i16 @saddsat_i16(i16 %lhs, i16 %rhs) {
   ; CHECK-LABEL: name: saddsat_i16
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[SADDSAT:%[0-9]+]]:_(s16) = G_SADDSAT [[TRUNC]], [[TRUNC1]]
-  ; CHECK:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[SADDSAT]](s16)
-  ; CHECK:   $vgpr0 = COPY [[ANYEXT]](s32)
-  ; CHECK:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
+  ; CHECK-NEXT:   [[SADDSAT:%[0-9]+]]:_(s16) = G_SADDSAT [[TRUNC]], [[TRUNC1]]
+  ; CHECK-NEXT:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[SADDSAT]](s16)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[ANYEXT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %res = call i16 @llvm.sadd.sat.i16(i16 %lhs, i16 %rhs)
   ret i16 %res
 }
 declare i16 @llvm.sadd.sat.i16(i16, i16)
 
 define i32 @saddsat_i32(i32 %lhs, i32 %rhs) {
   ; CHECK-LABEL: name: saddsat_i32
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[SADDSAT:%[0-9]+]]:_(s32) = G_SADDSAT [[COPY]], [[COPY1]]
-  ; CHECK:   $vgpr0 = COPY [[SADDSAT]](s32)
-  ; CHECK:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[SADDSAT:%[0-9]+]]:_(s32) = G_SADDSAT [[COPY]], [[COPY1]]
+  ; CHECK-NEXT:   $vgpr0 = COPY [[SADDSAT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %res = call i32 @llvm.sadd.sat.i32(i32 %lhs, i32 %rhs)
   ret i32 %res
 }
 declare i32 @llvm.sadd.sat.i32(i32, i32)
 
 define i64 @saddsat_i64(i64 %lhs, i64 %rhs) {
   ; CHECK-LABEL: name: saddsat_i64
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; CHECK:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
-  ; CHECK:   [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
-  ; CHECK:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[SADDSAT:%[0-9]+]]:_(s64) = G_SADDSAT [[MV]], [[MV1]]
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[SADDSAT]](s64)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   [[COPY5:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY4]]
-  ; CHECK:   S_SETPC_B64_return [[COPY5]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
+  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
+  ; CHECK-NEXT:   [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
+  ; CHECK-NEXT:   [[SADDSAT:%[0-9]+]]:_(s64) = G_SADDSAT [[MV]], [[MV1]]
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[SADDSAT]](s64)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %res = call i64 @llvm.sadd.sat.i64(i64 %lhs, i64 %rhs)
   ret i64 %res
 }
 declare i64 @llvm.sadd.sat.i64(i64, i64)
 
 define <2 x i32> @saddsat_v2i32(<2 x i32> %lhs, <2 x i32> %rhs) {
   ; CHECK-LABEL: name: saddsat_v2i32
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; CHECK:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
-  ; CHECK:   [[BUILD_VECTOR1:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY2]](s32), [[COPY3]](s32)
-  ; CHECK:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[SADDSAT:%[0-9]+]]:_(<2 x s32>) = G_SADDSAT [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[SADDSAT]](<2 x s32>)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   [[COPY5:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY4]]
-  ; CHECK:   S_SETPC_B64_return [[COPY5]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
+  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
+  ; CHECK-NEXT:   [[BUILD_VECTOR1:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY2]](s32), [[COPY3]](s32)
+  ; CHECK-NEXT:   [[SADDSAT:%[0-9]+]]:_(<2 x s32>) = G_SADDSAT [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[SADDSAT]](<2 x s32>)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %res = call <2 x i32> @llvm.sadd.sat.v2i32(<2 x i32> %lhs, <2 x i32> %rhs)
   ret <2 x i32> %res
 }
 declare <2 x i32> @llvm.sadd.sat.v2i32(<2 x i32>, <2 x i32>)
 
 define i16 @usubsat_i16(i16 %lhs, i16 %rhs) {
   ; CHECK-LABEL: name: usubsat_i16
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[USUBSAT:%[0-9]+]]:_(s16) = G_USUBSAT [[TRUNC]], [[TRUNC1]]
-  ; CHECK:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[USUBSAT]](s16)
-  ; CHECK:   $vgpr0 = COPY [[ANYEXT]](s32)
-  ; CHECK:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
+  ; CHECK-NEXT:   [[USUBSAT:%[0-9]+]]:_(s16) = G_USUBSAT [[TRUNC]], [[TRUNC1]]
+  ; CHECK-NEXT:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[USUBSAT]](s16)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[ANYEXT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %res = call i16 @llvm.usub.sat.i16(i16 %lhs, i16 %rhs)
   ret i16 %res
 }
 declare i16 @llvm.usub.sat.i16(i16, i16)
 
 define i32 @usubsat_i32(i32 %lhs, i32 %rhs) {
   ; CHECK-LABEL: name: usubsat_i32
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[USUBSAT:%[0-9]+]]:_(s32) = G_USUBSAT [[COPY]], [[COPY1]]
-  ; CHECK:   $vgpr0 = COPY [[USUBSAT]](s32)
-  ; CHECK:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[USUBSAT:%[0-9]+]]:_(s32) = G_USUBSAT [[COPY]], [[COPY1]]
+  ; CHECK-NEXT:   $vgpr0 = COPY [[USUBSAT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %res = call i32 @llvm.usub.sat.i32(i32 %lhs, i32 %rhs)
   ret i32 %res
 }
 declare i32 @llvm.usub.sat.i32(i32, i32)
 
 define i64 @usubsat_i64(i64 %lhs, i64 %rhs) {
   ; CHECK-LABEL: name: usubsat_i64
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; CHECK:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
-  ; CHECK:   [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
-  ; CHECK:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[USUBSAT:%[0-9]+]]:_(s64) = G_USUBSAT [[MV]], [[MV1]]
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[USUBSAT]](s64)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   [[COPY5:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY4]]
-  ; CHECK:   S_SETPC_B64_return [[COPY5]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
+  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
+  ; CHECK-NEXT:   [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
+  ; CHECK-NEXT:   [[USUBSAT:%[0-9]+]]:_(s64) = G_USUBSAT [[MV]], [[MV1]]
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[USUBSAT]](s64)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %res = call i64 @llvm.usub.sat.i64(i64 %lhs, i64 %rhs)
   ret i64 %res
 }
 declare i64 @llvm.usub.sat.i64(i64, i64)
 
 define <2 x i32> @usubsat_v2i32(<2 x i32> %lhs, <2 x i32> %rhs) {
   ; CHECK-LABEL: name: usubsat_v2i32
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; CHECK:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
-  ; CHECK:   [[BUILD_VECTOR1:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY2]](s32), [[COPY3]](s32)
-  ; CHECK:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[USUBSAT:%[0-9]+]]:_(<2 x s32>) = G_USUBSAT [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[USUBSAT]](<2 x s32>)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   [[COPY5:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY4]]
-  ; CHECK:   S_SETPC_B64_return [[COPY5]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
+  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
+  ; CHECK-NEXT:   [[BUILD_VECTOR1:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY2]](s32), [[COPY3]](s32)
+  ; CHECK-NEXT:   [[USUBSAT:%[0-9]+]]:_(<2 x s32>) = G_USUBSAT [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[USUBSAT]](<2 x s32>)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %res = call <2 x i32> @llvm.usub.sat.v2i32(<2 x i32> %lhs, <2 x i32> %rhs)
   ret <2 x i32> %res
 }
 declare <2 x i32> @llvm.usub.sat.v2i32(<2 x i32>, <2 x i32>)
 
 define i16 @ssubsat_i16(i16 %lhs, i16 %rhs) {
   ; CHECK-LABEL: name: ssubsat_i16
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[SSUBSAT:%[0-9]+]]:_(s16) = G_SSUBSAT [[TRUNC]], [[TRUNC1]]
-  ; CHECK:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[SSUBSAT]](s16)
-  ; CHECK:   $vgpr0 = COPY [[ANYEXT]](s32)
-  ; CHECK:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
+  ; CHECK-NEXT:   [[SSUBSAT:%[0-9]+]]:_(s16) = G_SSUBSAT [[TRUNC]], [[TRUNC1]]
+  ; CHECK-NEXT:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[SSUBSAT]](s16)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[ANYEXT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %res = call i16 @llvm.ssub.sat.i16(i16 %lhs, i16 %rhs)
   ret i16 %res
 }
 declare i16 @llvm.ssub.sat.i16(i16, i16)
 
 define i32 @ssubsat_i32(i32 %lhs, i32 %rhs) {
   ; CHECK-LABEL: name: ssubsat_i32
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[SSUBSAT:%[0-9]+]]:_(s32) = G_SSUBSAT [[COPY]], [[COPY1]]
-  ; CHECK:   $vgpr0 = COPY [[SSUBSAT]](s32)
-  ; CHECK:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[SSUBSAT:%[0-9]+]]:_(s32) = G_SSUBSAT [[COPY]], [[COPY1]]
+  ; CHECK-NEXT:   $vgpr0 = COPY [[SSUBSAT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %res = call i32 @llvm.ssub.sat.i32(i32 %lhs, i32 %rhs)
   ret i32 %res
 }
 declare i32 @llvm.ssub.sat.i32(i32, i32)
 
 define i64 @ssubsat_i64(i64 %lhs, i64 %rhs) {
   ; CHECK-LABEL: name: ssubsat_i64
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; CHECK:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
-  ; CHECK:   [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
-  ; CHECK:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[SSUBSAT:%[0-9]+]]:_(s64) = G_SSUBSAT [[MV]], [[MV1]]
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[SSUBSAT]](s64)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   [[COPY5:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY4]]
-  ; CHECK:   S_SETPC_B64_return [[COPY5]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
+  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
+  ; CHECK-NEXT:   [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
+  ; CHECK-NEXT:   [[SSUBSAT:%[0-9]+]]:_(s64) = G_SSUBSAT [[MV]], [[MV1]]
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[SSUBSAT]](s64)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %res = call i64 @llvm.ssub.sat.i64(i64 %lhs, i64 %rhs)
   ret i64 %res
 }
 declare i64 @llvm.ssub.sat.i64(i64, i64)
 
 define <2 x i32> @ssubsat_v2i32(<2 x i32> %lhs, <2 x i32> %rhs) {
   ; CHECK-LABEL: name: ssubsat_v2i32
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; CHECK:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
-  ; CHECK:   [[BUILD_VECTOR1:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY2]](s32), [[COPY3]](s32)
-  ; CHECK:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[SSUBSAT:%[0-9]+]]:_(<2 x s32>) = G_SSUBSAT [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[SSUBSAT]](<2 x s32>)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   [[COPY5:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY4]]
-  ; CHECK:   S_SETPC_B64_return [[COPY5]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
+  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
+  ; CHECK-NEXT:   [[BUILD_VECTOR1:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY2]](s32), [[COPY3]](s32)
+  ; CHECK-NEXT:   [[SSUBSAT:%[0-9]+]]:_(<2 x s32>) = G_SSUBSAT [[BUILD_VECTOR]], [[BUILD_VECTOR1]]
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[SSUBSAT]](<2 x s32>)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %res = call <2 x i32> @llvm.ssub.sat.v2i32(<2 x i32> %lhs, <2 x i32> %rhs)
   ret <2 x i32> %res
 }
 declare <2 x i32> @llvm.ssub.sat.v2i32(<2 x i32>, <2 x i32>)
 
 define i16 @ushlsat_i16(i16 %lhs, i16 %rhs) {
   ; CHECK-LABEL: name: ushlsat_i16
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[USHLSAT:%[0-9]+]]:_(s16) = G_USHLSAT [[TRUNC]], [[TRUNC1]](s16)
-  ; CHECK:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[USHLSAT]](s16)
-  ; CHECK:   $vgpr0 = COPY [[ANYEXT]](s32)
-  ; CHECK:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
+  ; CHECK-NEXT:   [[USHLSAT:%[0-9]+]]:_(s16) = G_USHLSAT [[TRUNC]], [[TRUNC1]](s16)
+  ; CHECK-NEXT:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[USHLSAT]](s16)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[ANYEXT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %res = call i16 @llvm.ushl.sat.i16(i16 %lhs, i16 %rhs)
   ret i16 %res
 }
 declare i16 @llvm.ushl.sat.i16(i16, i16)
 
 define i32 @ushlsat_i32(i32 %lhs, i32 %rhs) {
   ; CHECK-LABEL: name: ushlsat_i32
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[USHLSAT:%[0-9]+]]:_(s32) = G_USHLSAT [[COPY]], [[COPY1]](s32)
-  ; CHECK:   $vgpr0 = COPY [[USHLSAT]](s32)
-  ; CHECK:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[USHLSAT:%[0-9]+]]:_(s32) = G_USHLSAT [[COPY]], [[COPY1]](s32)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[USHLSAT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %res = call i32 @llvm.ushl.sat.i32(i32 %lhs, i32 %rhs)
   ret i32 %res
 }
 declare i32 @llvm.ushl.sat.i32(i32, i32)
 
 define i64 @ushlsat_i64(i64 %lhs, i64 %rhs) {
   ; CHECK-LABEL: name: ushlsat_i64
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; CHECK:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
-  ; CHECK:   [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
-  ; CHECK:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[USHLSAT:%[0-9]+]]:_(s64) = G_USHLSAT [[MV]], [[MV1]](s64)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[USHLSAT]](s64)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   [[COPY5:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY4]]
-  ; CHECK:   S_SETPC_B64_return [[COPY5]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
+  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
+  ; CHECK-NEXT:   [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
+  ; CHECK-NEXT:   [[USHLSAT:%[0-9]+]]:_(s64) = G_USHLSAT [[MV]], [[MV1]](s64)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[USHLSAT]](s64)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %res = call i64 @llvm.ushl.sat.i64(i64 %lhs, i64 %rhs)
   ret i64 %res
 }
 declare i64 @llvm.ushl.sat.i64(i64, i64)
 
 define <2 x i32> @ushlsat_v2i32(<2 x i32> %lhs, <2 x i32> %rhs) {
   ; CHECK-LABEL: name: ushlsat_v2i32
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; CHECK:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
-  ; CHECK:   [[BUILD_VECTOR1:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY2]](s32), [[COPY3]](s32)
-  ; CHECK:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[USHLSAT:%[0-9]+]]:_(<2 x s32>) = G_USHLSAT [[BUILD_VECTOR]], [[BUILD_VECTOR1]](<2 x s32>)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[USHLSAT]](<2 x s32>)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   [[COPY5:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY4]]
-  ; CHECK:   S_SETPC_B64_return [[COPY5]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
+  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
+  ; CHECK-NEXT:   [[BUILD_VECTOR1:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY2]](s32), [[COPY3]](s32)
+  ; CHECK-NEXT:   [[USHLSAT:%[0-9]+]]:_(<2 x s32>) = G_USHLSAT [[BUILD_VECTOR]], [[BUILD_VECTOR1]](<2 x s32>)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[USHLSAT]](<2 x s32>)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %res = call <2 x i32> @llvm.ushl.sat.v2i32(<2 x i32> %lhs, <2 x i32> %rhs)
   ret <2 x i32> %res
 }
 declare <2 x i32> @llvm.ushl.sat.v2i32(<2 x i32>, <2 x i32>)
 
 define i16 @sshlsat_i16(i16 %lhs, i16 %rhs) {
   ; CHECK-LABEL: name: sshlsat_i16
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[SSHLSAT:%[0-9]+]]:_(s16) = G_SSHLSAT [[TRUNC]], [[TRUNC1]](s16)
-  ; CHECK:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[SSHLSAT]](s16)
-  ; CHECK:   $vgpr0 = COPY [[ANYEXT]](s32)
-  ; CHECK:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY]](s32)
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[COPY1]](s32)
+  ; CHECK-NEXT:   [[SSHLSAT:%[0-9]+]]:_(s16) = G_SSHLSAT [[TRUNC]], [[TRUNC1]](s16)
+  ; CHECK-NEXT:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[SSHLSAT]](s16)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[ANYEXT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %res = call i16 @llvm.sshl.sat.i16(i16 %lhs, i16 %rhs)
   ret i16 %res
 }
 declare i16 @llvm.sshl.sat.i16(i16, i16)
 
 define i32 @sshlsat_i32(i32 %lhs, i32 %rhs) {
   ; CHECK-LABEL: name: sshlsat_i32
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[SSHLSAT:%[0-9]+]]:_(s32) = G_SSHLSAT [[COPY]], [[COPY1]](s32)
-  ; CHECK:   $vgpr0 = COPY [[SSHLSAT]](s32)
-  ; CHECK:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; CHECK:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[SSHLSAT:%[0-9]+]]:_(s32) = G_SSHLSAT [[COPY]], [[COPY1]](s32)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[SSHLSAT]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %res = call i32 @llvm.sshl.sat.i32(i32 %lhs, i32 %rhs)
   ret i32 %res
 }
 declare i32 @llvm.sshl.sat.i32(i32, i32)
 
 define i64 @sshlsat_i64(i64 %lhs, i64 %rhs) {
   ; CHECK-LABEL: name: sshlsat_i64
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; CHECK:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
-  ; CHECK:   [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
-  ; CHECK:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[SSHLSAT:%[0-9]+]]:_(s64) = G_SSHLSAT [[MV]], [[MV1]](s64)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[SSHLSAT]](s64)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   [[COPY5:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY4]]
-  ; CHECK:   S_SETPC_B64_return [[COPY5]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY]](s32), [[COPY1]](s32)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
+  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
+  ; CHECK-NEXT:   [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY2]](s32), [[COPY3]](s32)
+  ; CHECK-NEXT:   [[SSHLSAT:%[0-9]+]]:_(s64) = G_SSHLSAT [[MV]], [[MV1]](s64)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[SSHLSAT]](s64)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %res = call i64 @llvm.sshl.sat.i64(i64 %lhs, i64 %rhs)
   ret i64 %res
 }
 declare i64 @llvm.sshl.sat.i64(i64, i64)
 
 define <2 x i32> @sshlsat_v2i32(<2 x i32> %lhs, <2 x i32> %rhs) {
   ; CHECK-LABEL: name: sshlsat_v2i32
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
-  ; CHECK:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; CHECK:   [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32)
-  ; CHECK:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; CHECK:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
-  ; CHECK:   [[BUILD_VECTOR1:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY2]](s32), [[COPY3]](s32)
-  ; CHECK:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[SSHLSAT:%[0-9]+]]:_(<2 x s32>) = G_SSHLSAT [[BUILD_VECTOR]], [[BUILD_VECTOR1]](<2 x s32>)
-  ; CHECK:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[SSHLSAT]](<2 x s32>)
-  ; CHECK:   $vgpr0 = COPY [[UV]](s32)
-  ; CHECK:   $vgpr1 = COPY [[UV1]](s32)
-  ; CHECK:   [[COPY5:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY4]]
-  ; CHECK:   S_SETPC_B64_return [[COPY5]], implicit $vgpr0, implicit $vgpr1
+  ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+  ; CHECK-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY]](s32), [[COPY1]](s32)
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
+  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
+  ; CHECK-NEXT:   [[BUILD_VECTOR1:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY2]](s32), [[COPY3]](s32)
+  ; CHECK-NEXT:   [[SSHLSAT:%[0-9]+]]:_(<2 x s32>) = G_SSHLSAT [[BUILD_VECTOR]], [[BUILD_VECTOR1]](<2 x s32>)
+  ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[SSHLSAT]](<2 x s32>)
+  ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
+  ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
+  ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %res = call <2 x i32> @llvm.sshl.sat.v2i32(<2 x i32> %lhs, <2 x i32> %rhs)
   ret <2 x i32> %res
 }
 declare <2 x i32> @llvm.sshl.sat.v2i32(<2 x i32>, <2 x i32>)

llvm/test/CodeGen/AMDGPU/GlobalISel/irtranslator-sibling-call.ll

 ; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
 ; RUN: llc -global-isel -amdgpu-fixed-function-abi -stop-after=irtranslator -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 -amdgpu-sroa=0 -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefix=GCN %s
 ; This is a copy of sibling-call.ll, but stops after the IRTranslator.
 
 define fastcc i32 @i32_fastcc_i32_i32(i32 %arg0, i32 %arg1) #1 {
   ; GCN-LABEL: name: i32_fastcc_i32_i32
   ; GCN: bb.1 (%ir-block.0):
-  ; GCN-NEXT:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+  ; GCN-NEXT:   liveins: $vgpr0, $vgpr1
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; GCN-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; GCN-NEXT:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   [[ADD:%[0-9]+]]:_(s32) = G_ADD [[COPY]], [[COPY1]]
   ; GCN-NEXT:   $vgpr0 = COPY [[ADD]](s32)
-  ; GCN-NEXT:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; GCN-NEXT:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; GCN-NEXT:   SI_RETURN implicit $vgpr0
   %add0 = add i32 %arg0, %arg1
   ret i32 %add0
 }
 
 define fastcc i32 @i32_fastcc_i32_i32_stack_object(i32 %arg0, i32 %arg1) #1 {
   ; GCN-LABEL: name: i32_fastcc_i32_i32_stack_object
   ; GCN: bb.1 (%ir-block.0):
-  ; GCN-NEXT:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+  ; GCN-NEXT:   liveins: $vgpr0, $vgpr1
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; GCN-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; GCN-NEXT:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 9
   ; GCN-NEXT:   [[FRAME_INDEX:%[0-9]+]]:_(p5) = G_FRAME_INDEX %stack.0.alloca
   ; GCN-NEXT:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 20
   ; GCN-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[FRAME_INDEX]], [[C1]](s32)
   ; GCN-NEXT:   G_STORE [[C]](s32), [[PTR_ADD]](p5) :: (volatile store (s32) into %ir.gep, addrspace 5)
   ; GCN-NEXT:   [[ADD:%[0-9]+]]:_(s32) = G_ADD [[COPY]], [[COPY1]]
   ; GCN-NEXT:   $vgpr0 = COPY [[ADD]](s32)
-  ; GCN-NEXT:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; GCN-NEXT:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; GCN-NEXT:   SI_RETURN implicit $vgpr0
   %alloca = alloca [16 x i32], align 4, addrspace(5)
   %gep = getelementptr inbounds [16 x i32], [16 x i32] addrspace(5)* %alloca, i32 0, i32 5
   store volatile i32 9, i32 addrspace(5)* %gep
   %add0 = add i32 %arg0, %arg1
   ret i32 %add0
 }
 
 define hidden fastcc i32 @sibling_call_i32_fastcc_i32_i32(i32 %a, i32 %b, i32 %c) #1 {
   ; GCN-LABEL: name: sibling_call_i32_fastcc_i32_i32
   ; GCN: bb.1.entry:
-  ; GCN-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
+  ; GCN-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; GCN-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
   ; GCN-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; GCN-NEXT:   [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   [[GV:%[0-9]+]]:sreg_64(p0) = G_GLOBAL_VALUE @i32_fastcc_i32_i32
   ; GCN-NEXT:   $vgpr0 = COPY [[COPY]](s32)
   ; GCN-NEXT:   $vgpr1 = COPY [[COPY1]](s32)
-  ; GCN-NEXT:   [[COPY4:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY4]](<4 x s32>)
+  ; GCN-NEXT:   [[COPY3:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY3]](<4 x s32>)
   ; GCN-NEXT:   SI_TCRETURN [[GV]](p0), @i32_fastcc_i32_i32, 0, csr_amdgpu_highregs, implicit $vgpr0, implicit $vgpr1, implicit $sgpr0_sgpr1_sgpr2_sgpr3
 entry:
   %ret = tail call fastcc i32 @i32_fastcc_i32_i32(i32 %a, i32 %b)
   ret i32 %ret
 }
 
 define fastcc i32 @sibling_call_i32_fastcc_i32_i32_stack_object(i32 %a, i32 %b, i32 %c) #1 {
   ; GCN-LABEL: name: sibling_call_i32_fastcc_i32_i32_stack_object
   ; GCN: bb.1.entry:
-  ; GCN-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
+  ; GCN-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; GCN-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
   ; GCN-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; GCN-NEXT:   [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 9
   ; GCN-NEXT:   [[FRAME_INDEX:%[0-9]+]]:_(p5) = G_FRAME_INDEX %stack.0.alloca
   ; GCN-NEXT:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 20
   ; GCN-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[FRAME_INDEX]], [[C1]](s32)
   ; GCN-NEXT:   G_STORE [[C]](s32), [[PTR_ADD]](p5) :: (volatile store (s32) into %ir.gep, addrspace 5)
   ; GCN-NEXT:   [[GV:%[0-9]+]]:sreg_64(p0) = G_GLOBAL_VALUE @i32_fastcc_i32_i32
   ; GCN-NEXT:   $vgpr0 = COPY [[COPY]](s32)
   ; GCN-NEXT:   $vgpr1 = COPY [[COPY1]](s32)
-  ; GCN-NEXT:   [[COPY4:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY4]](<4 x s32>)
+  ; GCN-NEXT:   [[COPY3:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY3]](<4 x s32>)
   ; GCN-NEXT:   SI_TCRETURN [[GV]](p0), @i32_fastcc_i32_i32, 0, csr_amdgpu_highregs, implicit $vgpr0, implicit $vgpr1, implicit $sgpr0_sgpr1_sgpr2_sgpr3
 entry:
   %alloca = alloca [16 x i32], align 4, addrspace(5)
   %gep = getelementptr inbounds [16 x i32], [16 x i32] addrspace(5)* %alloca, i32 0, i32 5
   store volatile i32 9, i32 addrspace(5)* %gep
   %ret = tail call fastcc i32 @i32_fastcc_i32_i32(i32 %a, i32 %b)
   ret i32 %ret
 }
 
 define fastcc i32 @sibling_call_i32_fastcc_i32_i32_callee_stack_object(i32 %a, i32 %b, i32 %c) #1 {
   ; GCN-LABEL: name: sibling_call_i32_fastcc_i32_i32_callee_stack_object
   ; GCN: bb.1.entry:
-  ; GCN-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
+  ; GCN-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; GCN-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
   ; GCN-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; GCN-NEXT:   [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 9
   ; GCN-NEXT:   [[FRAME_INDEX:%[0-9]+]]:_(p5) = G_FRAME_INDEX %stack.0.alloca
   ; GCN-NEXT:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 20
   ; GCN-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[FRAME_INDEX]], [[C1]](s32)
   ; GCN-NEXT:   G_STORE [[C]](s32), [[PTR_ADD]](p5) :: (volatile store (s32) into %ir.gep, addrspace 5)
   ; GCN-NEXT:   [[GV:%[0-9]+]]:sreg_64(p0) = G_GLOBAL_VALUE @i32_fastcc_i32_i32_stack_object
   ; GCN-NEXT:   $vgpr0 = COPY [[COPY]](s32)
   ; GCN-NEXT:   $vgpr1 = COPY [[COPY1]](s32)
-  ; GCN-NEXT:   [[COPY4:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY4]](<4 x s32>)
+  ; GCN-NEXT:   [[COPY3:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY3]](<4 x s32>)
   ; GCN-NEXT:   SI_TCRETURN [[GV]](p0), @i32_fastcc_i32_i32_stack_object, 0, csr_amdgpu_highregs, implicit $vgpr0, implicit $vgpr1, implicit $sgpr0_sgpr1_sgpr2_sgpr3
 entry:
   %alloca = alloca [16 x i32], align 4, addrspace(5)
   %gep = getelementptr inbounds [16 x i32], [16 x i32] addrspace(5)* %alloca, i32 0, i32 5
   store volatile i32 9, i32 addrspace(5)* %gep
   %ret = tail call fastcc i32 @i32_fastcc_i32_i32_stack_object(i32 %a, i32 %b)
   ret i32 %ret
 }
 
 define fastcc void @sibling_call_i32_fastcc_i32_i32_unused_result(i32 %a, i32 %b, i32 %c) #1 {
   ; GCN-LABEL: name: sibling_call_i32_fastcc_i32_i32_unused_result
   ; GCN: bb.1.entry:
-  ; GCN-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
+  ; GCN-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; GCN-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
   ; GCN-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; GCN-NEXT:   [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   [[GV:%[0-9]+]]:sreg_64(p0) = G_GLOBAL_VALUE @i32_fastcc_i32_i32
   ; GCN-NEXT:   $vgpr0 = COPY [[COPY]](s32)
   ; GCN-NEXT:   $vgpr1 = COPY [[COPY1]](s32)
-  ; GCN-NEXT:   [[COPY4:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY4]](<4 x s32>)
+  ; GCN-NEXT:   [[COPY3:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY3]](<4 x s32>)
   ; GCN-NEXT:   SI_TCRETURN [[GV]](p0), @i32_fastcc_i32_i32, 0, csr_amdgpu_highregs, implicit $vgpr0, implicit $vgpr1, implicit $sgpr0_sgpr1_sgpr2_sgpr3
 entry:
   %ret = tail call fastcc i32 @i32_fastcc_i32_i32(i32 %a, i32 %b)
   ret void
 }
 
 ; It doesn't make sense to do a tail from a kernel
 define amdgpu_kernel void @kernel_call_i32_fastcc_i32_i32_unused_result(i32 %a, i32 %b, i32 %c) #1 {

 entry:
   %ret = tail call fastcc i32 @i32_fastcc_i32_i32(i32 %a, i32 %b)
   ret void
 }
 
 define hidden fastcc i32 @i32_fastcc_i32_byval_i32(i32 %arg0, i32 addrspace(5)* byval(i32) align 4 %arg1) #1 {
   ; GCN-LABEL: name: i32_fastcc_i32_byval_i32
   ; GCN: bb.1 (%ir-block.0):
-  ; GCN-NEXT:   liveins: $vgpr0, $sgpr30_sgpr31
+  ; GCN-NEXT:   liveins: $vgpr0
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; GCN-NEXT:   [[FRAME_INDEX:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.0
   ; GCN-NEXT:   [[COPY1:%[0-9]+]]:_(p5) = COPY [[FRAME_INDEX]](p5)
-  ; GCN-NEXT:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   [[LOAD:%[0-9]+]]:_(s32) = G_LOAD [[COPY1]](p5) :: (dereferenceable load (s32) from %ir.arg1, addrspace 5)
   ; GCN-NEXT:   [[ADD:%[0-9]+]]:_(s32) = G_ADD [[COPY]], [[LOAD]]
   ; GCN-NEXT:   $vgpr0 = COPY [[ADD]](s32)
-  ; GCN-NEXT:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; GCN-NEXT:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; GCN-NEXT:   SI_RETURN implicit $vgpr0
   %arg1.load = load i32, i32 addrspace(5)* %arg1, align 4
   %add0 = add i32 %arg0, %arg1.load
   ret i32 %add0
 }
 
 ; Tail call disallowed with byval in parent.
 define fastcc i32 @sibling_call_i32_fastcc_i32_byval_i32_byval_parent(i32 %a, i32 addrspace(5)* byval(i32) %b.byval, i32 %c) #1 {
   ; GCN-LABEL: name: sibling_call_i32_fastcc_i32_byval_i32_byval_parent
   ; GCN: bb.1.entry:
-  ; GCN-NEXT:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+  ; GCN-NEXT:   liveins: $vgpr0, $vgpr1
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; GCN-NEXT:   [[FRAME_INDEX:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.0
   ; GCN-NEXT:   [[COPY1:%[0-9]+]]:_(p5) = COPY [[FRAME_INDEX]](p5)
   ; GCN-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; GCN-NEXT:   [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; GCN-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @i32_fastcc_i32_byval_i32
-  ; GCN-NEXT:   [[COPY4:%[0-9]+]]:_(p5) = COPY $sgpr32
+  ; GCN-NEXT:   [[COPY3:%[0-9]+]]:_(p5) = COPY $sgpr32
   ; GCN-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
-  ; GCN-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY4]], [[C]](s32)
+  ; GCN-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY3]], [[C]](s32)
   ; GCN-NEXT:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 4
   ; GCN-NEXT:   G_MEMCPY [[PTR_ADD]](p5), [[COPY1]](p5), [[C1]](s32), 0 :: (dereferenceable store (s32) into stack, addrspace 5), (dereferenceable load (s32) from %ir.b.byval, addrspace 5)
   ; GCN-NEXT:   $vgpr0 = COPY [[COPY]](s32)
-  ; GCN-NEXT:   [[COPY5:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY5]](<4 x s32>)
+  ; GCN-NEXT:   [[COPY4:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY4]](<4 x s32>)
   ; GCN-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @i32_fastcc_i32_byval_i32, csr_amdgpu_highregs, implicit $vgpr0, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit-def $vgpr0
-  ; GCN-NEXT:   [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; GCN-NEXT:   [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; GCN-NEXT:   ADJCALLSTACKDOWN 0, 4, implicit-def $scc
-  ; GCN-NEXT:   $vgpr0 = COPY [[COPY6]](s32)
-  ; GCN-NEXT:   [[COPY7:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-  ; GCN-NEXT:   S_SETPC_B64_return [[COPY7]], implicit $vgpr0
+  ; GCN-NEXT:   $vgpr0 = COPY [[COPY5]](s32)
+  ; GCN-NEXT:   SI_RETURN implicit $vgpr0
 entry:
   %ret = tail call fastcc i32 @i32_fastcc_i32_byval_i32(i32 %a, i32 addrspace(5)* byval(i32) %b.byval)
   ret i32 %ret
 }
 
 ; Tail call disallowed with byval in parent, not callee. The stack
 ; usage of incoming arguments must be <= the outgoing stack
 ; arguments.
 define fastcc i32 @sibling_call_i32_fastcc_i32_byval_i32(i32 %a, [32 x i32] %large) #1 {
   ; GCN-LABEL: name: sibling_call_i32_fastcc_i32_byval_i32
   ; GCN: bb.1.entry:
-  ; GCN-NEXT:   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, $sgpr30_sgpr31
+  ; GCN-NEXT:   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
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; GCN-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
   ; GCN-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
   ; GCN-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
   ; GCN-NEXT:   [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
   ; GCN-NEXT:   [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
   ; GCN-NEXT:   [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
define fastcc i32 @sibling_call_i32_fastcc_i32_byval_i32(i32 %a, [32 x i32] %large) #1 {
   ; GCN-NEXT:   [[COPY27:%[0-9]+]]:_(s32) = COPY $vgpr27
   ; GCN-NEXT:   [[COPY28:%[0-9]+]]:_(s32) = COPY $vgpr28
   ; GCN-NEXT:   [[COPY29:%[0-9]+]]:_(s32) = COPY $vgpr29
   ; GCN-NEXT:   [[COPY30:%[0-9]+]]:_(s32) = COPY $vgpr30
   ; GCN-NEXT:   [[FRAME_INDEX:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.2
   ; GCN-NEXT:   [[LOAD:%[0-9]+]]:_(s32) = G_LOAD [[FRAME_INDEX]](p5) :: (invariant load (s32) from %fixed-stack.2, align 16, addrspace 5)
   ; GCN-NEXT:   [[FRAME_INDEX1:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.1
   ; GCN-NEXT:   [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[FRAME_INDEX1]](p5) :: (invariant load (s32) from %fixed-stack.1, addrspace 5)
-  ; GCN-NEXT:   [[COPY31:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
   ; GCN-NEXT:   [[INTTOPTR:%[0-9]+]]:_(p5) = G_INTTOPTR [[C]](s32)
   ; GCN-NEXT:   [[GV:%[0-9]+]]:sreg_64(p0) = G_GLOBAL_VALUE @i32_fastcc_i32_byval_i32
   ; GCN-NEXT:   [[FRAME_INDEX2:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.0
   ; GCN-NEXT:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 4
   ; GCN-NEXT:   G_MEMCPY [[FRAME_INDEX2]](p5), [[INTTOPTR]](p5), [[C1]](s32), 0 :: (dereferenceable store (s32) into %fixed-stack.0, align 16, addrspace 5), (dereferenceable load (s32) from `i32 addrspace(5)* inttoptr (i32 16 to i32 addrspace(5)*)`, align 16, addrspace 5)
   ; GCN-NEXT:   $vgpr0 = COPY [[COPY]](s32)
-  ; GCN-NEXT:   [[COPY32:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY32]](<4 x s32>)
+  ; GCN-NEXT:   [[COPY31:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY31]](<4 x s32>)
   ; GCN-NEXT:   SI_TCRETURN [[GV]](p0), @i32_fastcc_i32_byval_i32, 0, csr_amdgpu_highregs, implicit $vgpr0, implicit $sgpr0_sgpr1_sgpr2_sgpr3
 entry:
   %ret = tail call fastcc i32 @i32_fastcc_i32_byval_i32(i32 %a, i32 addrspace(5)* byval(i32) inttoptr (i32 16 to i32 addrspace(5)*))
   ret i32 %ret
 }
 
 define fastcc i32 @i32_fastcc_i32_i32_a32i32(i32 %arg0, i32 %arg1, [32 x i32] %large) #1 {
   ; GCN-LABEL: name: i32_fastcc_i32_i32_a32i32
   ; GCN: bb.1 (%ir-block.0):
-  ; GCN-NEXT:   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, $sgpr30_sgpr31
+  ; GCN-NEXT:   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
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; GCN-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
   ; GCN-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
   ; GCN-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
   ; GCN-NEXT:   [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
   ; GCN-NEXT:   [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
   ; GCN-NEXT:   [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
define fastcc i32 @i32_fastcc_i32_i32_a32i32(i32 %arg0, i32 %arg1, [32 x i32] %large) #1 {
   ; GCN-NEXT:   [[COPY29:%[0-9]+]]:_(s32) = COPY $vgpr29
   ; GCN-NEXT:   [[COPY30:%[0-9]+]]:_(s32) = COPY $vgpr30
   ; GCN-NEXT:   [[FRAME_INDEX:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.2
   ; GCN-NEXT:   [[LOAD:%[0-9]+]]:_(s32) = G_LOAD [[FRAME_INDEX]](p5) :: (invariant load (s32) from %fixed-stack.2, align 16, addrspace 5)
   ; GCN-NEXT:   [[FRAME_INDEX1:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.1
   ; GCN-NEXT:   [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[FRAME_INDEX1]](p5) :: (invariant load (s32) from %fixed-stack.1, addrspace 5)
   ; GCN-NEXT:   [[FRAME_INDEX2:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.0
   ; GCN-NEXT:   [[LOAD2:%[0-9]+]]:_(s32) = G_LOAD [[FRAME_INDEX2]](p5) :: (invariant load (s32) from %fixed-stack.0, align 8, addrspace 5)
-  ; GCN-NEXT:   [[COPY31:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   [[ADD:%[0-9]+]]:_(s32) = G_ADD [[COPY]], [[COPY1]]
   ; GCN-NEXT:   [[ADD1:%[0-9]+]]:_(s32) = G_ADD [[ADD]], [[LOAD1]]
   ; GCN-NEXT:   [[ADD2:%[0-9]+]]:_(s32) = G_ADD [[ADD1]], [[LOAD2]]
   ; GCN-NEXT:   $vgpr0 = COPY [[ADD2]](s32)
-  ; GCN-NEXT:   [[COPY32:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY31]]
-  ; GCN-NEXT:   S_SETPC_B64_return [[COPY32]], implicit $vgpr0
+  ; GCN-NEXT:   SI_RETURN implicit $vgpr0
   %val_firststack = extractvalue [32 x i32] %large, 30
   %val_laststack = extractvalue [32 x i32] %large, 31
   %add0 = add i32 %arg0, %arg1
   %add1 = add i32 %add0, %val_firststack
   %add2 = add i32 %add1, %val_laststack
   ret i32 %add2
 }
 
 define fastcc i32 @sibling_call_i32_fastcc_i32_i32_a32i32(i32 %a, i32 %b, [32 x i32] %c) #1 {
   ; GCN-LABEL: name: sibling_call_i32_fastcc_i32_i32_a32i32
   ; GCN: bb.1.entry:
-  ; GCN-NEXT:   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, $sgpr30_sgpr31
+  ; GCN-NEXT:   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
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; GCN-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
   ; GCN-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
   ; GCN-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
   ; GCN-NEXT:   [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
   ; GCN-NEXT:   [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
   ; GCN-NEXT:   [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
define fastcc i32 @sibling_call_i32_fastcc_i32_i32_a32i32(i32 %a, i32 %b, [32 x i32] %c) #1 {
   ; GCN-NEXT:   [[COPY29:%[0-9]+]]:_(s32) = COPY $vgpr29
   ; GCN-NEXT:   [[COPY30:%[0-9]+]]:_(s32) = COPY $vgpr30
   ; GCN-NEXT:   [[FRAME_INDEX:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.5
   ; GCN-NEXT:   [[LOAD:%[0-9]+]]:_(s32) = G_LOAD [[FRAME_INDEX]](p5) :: (invariant load (s32) from %fixed-stack.5, align 16, addrspace 5)
   ; GCN-NEXT:   [[FRAME_INDEX1:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.4
   ; GCN-NEXT:   [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[FRAME_INDEX1]](p5) :: (invariant load (s32) from %fixed-stack.4, addrspace 5)
   ; GCN-NEXT:   [[FRAME_INDEX2:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.3
   ; GCN-NEXT:   [[LOAD2:%[0-9]+]]:_(s32) = G_LOAD [[FRAME_INDEX2]](p5) :: (invariant load (s32) from %fixed-stack.3, align 8, addrspace 5)
-  ; GCN-NEXT:   [[COPY31:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   [[GV:%[0-9]+]]:sreg_64(p0) = G_GLOBAL_VALUE @i32_fastcc_i32_i32_a32i32
   ; GCN-NEXT:   [[FRAME_INDEX3:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.2
   ; GCN-NEXT:   G_STORE [[LOAD]](s32), [[FRAME_INDEX3]](p5) :: (store (s32) into %fixed-stack.2, align 16, addrspace 5)
   ; GCN-NEXT:   [[FRAME_INDEX4:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.1
   ; GCN-NEXT:   G_STORE [[LOAD1]](s32), [[FRAME_INDEX4]](p5) :: (store (s32) into %fixed-stack.1, addrspace 5)
   ; GCN-NEXT:   [[FRAME_INDEX5:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.0
   ; GCN-NEXT:   G_STORE [[LOAD2]](s32), [[FRAME_INDEX5]](p5) :: (store (s32) into %fixed-stack.0, align 8, addrspace 5)
   ; GCN-NEXT:   $vgpr0 = COPY [[COPY]](s32)
define fastcc i32 @sibling_call_i32_fastcc_i32_i32_a32i32(i32 %a, i32 %b, [32 x i32] %c) #1 {
   ; GCN-NEXT:   $vgpr23 = COPY [[COPY23]](s32)
   ; GCN-NEXT:   $vgpr24 = COPY [[COPY24]](s32)
   ; GCN-NEXT:   $vgpr25 = COPY [[COPY25]](s32)
   ; GCN-NEXT:   $vgpr26 = COPY [[COPY26]](s32)
   ; GCN-NEXT:   $vgpr27 = COPY [[COPY27]](s32)
   ; GCN-NEXT:   $vgpr28 = COPY [[COPY28]](s32)
   ; GCN-NEXT:   $vgpr29 = COPY [[COPY29]](s32)
   ; GCN-NEXT:   $vgpr30 = COPY [[COPY30]](s32)
-  ; GCN-NEXT:   [[COPY32:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY32]](<4 x s32>)
+  ; GCN-NEXT:   [[COPY31:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY31]](<4 x s32>)
   ; GCN-NEXT:   SI_TCRETURN [[GV]](p0), @i32_fastcc_i32_i32_a32i32, 0, csr_amdgpu_highregs, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15, implicit $vgpr16, implicit $vgpr17, implicit $vgpr18, implicit $vgpr19, implicit $vgpr20, implicit $vgpr21, implicit $vgpr22, implicit $vgpr23, implicit $vgpr24, implicit $vgpr25, implicit $vgpr26, implicit $vgpr27, implicit $vgpr28, implicit $vgpr29, implicit $vgpr30, implicit $sgpr0_sgpr1_sgpr2_sgpr3
 entry:
   %ret = tail call fastcc i32 @i32_fastcc_i32_i32_a32i32(i32 %a, i32 %b, [32 x i32] %c)
   ret i32 %ret
 }
 
 define fastcc i32 @sibling_call_i32_fastcc_i32_i32_a32i32_stack_object(i32 %a, i32 %b, [32 x i32] %c) #1 {
   ; GCN-LABEL: name: sibling_call_i32_fastcc_i32_i32_a32i32_stack_object
   ; GCN: bb.1.entry:
-  ; GCN-NEXT:   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, $sgpr30_sgpr31
+  ; GCN-NEXT:   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
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; GCN-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
   ; GCN-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
   ; GCN-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
   ; GCN-NEXT:   [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
   ; GCN-NEXT:   [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
   ; GCN-NEXT:   [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
define fastcc i32 @sibling_call_i32_fastcc_i32_i32_a32i32_stack_object(i32 %a, i32 %b, [32 x i32] %c) #1 {
   ; GCN-NEXT:   [[COPY29:%[0-9]+]]:_(s32) = COPY $vgpr29
   ; GCN-NEXT:   [[COPY30:%[0-9]+]]:_(s32) = COPY $vgpr30
   ; GCN-NEXT:   [[FRAME_INDEX:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.5
   ; GCN-NEXT:   [[LOAD:%[0-9]+]]:_(s32) = G_LOAD [[FRAME_INDEX]](p5) :: (invariant load (s32) from %fixed-stack.5, align 16, addrspace 5)
   ; GCN-NEXT:   [[FRAME_INDEX1:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.4
   ; GCN-NEXT:   [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[FRAME_INDEX1]](p5) :: (invariant load (s32) from %fixed-stack.4, addrspace 5)
   ; GCN-NEXT:   [[FRAME_INDEX2:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.3
   ; GCN-NEXT:   [[LOAD2:%[0-9]+]]:_(s32) = G_LOAD [[FRAME_INDEX2]](p5) :: (invariant load (s32) from %fixed-stack.3, align 8, addrspace 5)
-  ; GCN-NEXT:   [[COPY31:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 9
   ; GCN-NEXT:   [[FRAME_INDEX3:%[0-9]+]]:_(p5) = G_FRAME_INDEX %stack.0.alloca
   ; GCN-NEXT:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 20
   ; GCN-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[FRAME_INDEX3]], [[C1]](s32)
   ; GCN-NEXT:   G_STORE [[C]](s32), [[PTR_ADD]](p5) :: (volatile store (s32) into %ir.gep, addrspace 5)
   ; GCN-NEXT:   [[GV:%[0-9]+]]:sreg_64(p0) = G_GLOBAL_VALUE @i32_fastcc_i32_i32_a32i32
   ; GCN-NEXT:   [[FRAME_INDEX4:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.2
   ; GCN-NEXT:   G_STORE [[LOAD]](s32), [[FRAME_INDEX4]](p5) :: (store (s32) into %fixed-stack.2, align 16, addrspace 5)
define fastcc i32 @sibling_call_i32_fastcc_i32_i32_a32i32_stack_object(i32 %a, i32 %b, [32 x i32] %c) #1 {
   ; GCN-NEXT:   $vgpr23 = COPY [[COPY23]](s32)
   ; GCN-NEXT:   $vgpr24 = COPY [[COPY24]](s32)
   ; GCN-NEXT:   $vgpr25 = COPY [[COPY25]](s32)
   ; GCN-NEXT:   $vgpr26 = COPY [[COPY26]](s32)
   ; GCN-NEXT:   $vgpr27 = COPY [[COPY27]](s32)
   ; GCN-NEXT:   $vgpr28 = COPY [[COPY28]](s32)
   ; GCN-NEXT:   $vgpr29 = COPY [[COPY29]](s32)
   ; GCN-NEXT:   $vgpr30 = COPY [[COPY30]](s32)
-  ; GCN-NEXT:   [[COPY32:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY32]](<4 x s32>)
+  ; GCN-NEXT:   [[COPY31:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY31]](<4 x s32>)
   ; GCN-NEXT:   SI_TCRETURN [[GV]](p0), @i32_fastcc_i32_i32_a32i32, 0, csr_amdgpu_highregs, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15, implicit $vgpr16, implicit $vgpr17, implicit $vgpr18, implicit $vgpr19, implicit $vgpr20, implicit $vgpr21, implicit $vgpr22, implicit $vgpr23, implicit $vgpr24, implicit $vgpr25, implicit $vgpr26, implicit $vgpr27, implicit $vgpr28, implicit $vgpr29, implicit $vgpr30, implicit $sgpr0_sgpr1_sgpr2_sgpr3
 entry:
   %alloca = alloca [16 x i32], align 4, addrspace(5)
   %gep = getelementptr inbounds [16 x i32], [16 x i32] addrspace(5)* %alloca, i32 0, i32 5
   store volatile i32 9, i32 addrspace(5)* %gep
   %ret = tail call fastcc i32 @i32_fastcc_i32_i32_a32i32(i32 %a, i32 %b, [32 x i32] %c)
   ret i32 %ret
 }
 
 ; If the callee requires more stack argument space than the caller,
 ; don't do a tail call.
 ; TODO: Do we really need this restriction?
 define fastcc i32 @no_sibling_call_callee_more_stack_space(i32 %a, i32 %b) #1 {
   ; GCN-LABEL: name: no_sibling_call_callee_more_stack_space
   ; GCN: bb.1.entry:
-  ; GCN-NEXT:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+  ; GCN-NEXT:   liveins: $vgpr0, $vgpr1
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; GCN-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
-  ; GCN-NEXT:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
   ; GCN-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; GCN-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @i32_fastcc_i32_i32_a32i32
-  ; GCN-NEXT:   [[COPY3:%[0-9]+]]:_(p5) = COPY $sgpr32
+  ; GCN-NEXT:   [[COPY2:%[0-9]+]]:_(p5) = COPY $sgpr32
   ; GCN-NEXT:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
-  ; GCN-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY3]], [[C1]](s32)
+  ; GCN-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY2]], [[C1]](s32)
   ; GCN-NEXT:   G_STORE [[C]](s32), [[PTR_ADD]](p5) :: (store (s32) into stack, align 16, addrspace 5)
   ; GCN-NEXT:   [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 4
-  ; GCN-NEXT:   [[PTR_ADD1:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY3]], [[C2]](s32)
+  ; GCN-NEXT:   [[PTR_ADD1:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY2]], [[C2]](s32)
   ; GCN-NEXT:   G_STORE [[C]](s32), [[PTR_ADD1]](p5) :: (store (s32) into stack + 4, addrspace 5)
   ; GCN-NEXT:   [[C3:%[0-9]+]]:_(s32) = G_CONSTANT i32 8
-  ; GCN-NEXT:   [[PTR_ADD2:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY3]], [[C3]](s32)
+  ; GCN-NEXT:   [[PTR_ADD2:%[0-9]+]]:_(p5) = G_PTR_ADD [[COPY2]], [[C3]](s32)
   ; GCN-NEXT:   G_STORE [[C]](s32), [[PTR_ADD2]](p5) :: (store (s32) into stack + 8, align 8, addrspace 5)
   ; GCN-NEXT:   $vgpr0 = COPY [[COPY]](s32)
   ; GCN-NEXT:   $vgpr1 = COPY [[COPY1]](s32)
   ; GCN-NEXT:   $vgpr2 = COPY [[C]](s32)
   ; GCN-NEXT:   $vgpr3 = COPY [[C]](s32)
   ; GCN-NEXT:   $vgpr4 = COPY [[C]](s32)
   ; GCN-NEXT:   $vgpr5 = COPY [[C]](s32)
   ; GCN-NEXT:   $vgpr6 = COPY [[C]](s32)
define fastcc i32 @no_sibling_call_callee_more_stack_space(i32 %a, i32 %b) #1 {
   ; GCN-NEXT:   $vgpr23 = COPY [[C]](s32)
   ; GCN-NEXT:   $vgpr24 = COPY [[C]](s32)
   ; GCN-NEXT:   $vgpr25 = COPY [[C]](s32)
   ; GCN-NEXT:   $vgpr26 = COPY [[C]](s32)
   ; GCN-NEXT:   $vgpr27 = COPY [[C]](s32)
   ; GCN-NEXT:   $vgpr28 = COPY [[C]](s32)
   ; GCN-NEXT:   $vgpr29 = COPY [[C]](s32)
   ; GCN-NEXT:   $vgpr30 = COPY [[C]](s32)
-  ; GCN-NEXT:   [[COPY4:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY4]](<4 x s32>)
+  ; GCN-NEXT:   [[COPY3:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY3]](<4 x s32>)
   ; GCN-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @i32_fastcc_i32_i32_a32i32, csr_amdgpu_highregs, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15, implicit $vgpr16, implicit $vgpr17, implicit $vgpr18, implicit $vgpr19, implicit $vgpr20, implicit $vgpr21, implicit $vgpr22, implicit $vgpr23, implicit $vgpr24, implicit $vgpr25, implicit $vgpr26, implicit $vgpr27, implicit $vgpr28, implicit $vgpr29, implicit $vgpr30, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit-def $vgpr0
-  ; GCN-NEXT:   [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; GCN-NEXT:   [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; GCN-NEXT:   ADJCALLSTACKDOWN 0, 12, implicit-def $scc
-  ; GCN-NEXT:   $vgpr0 = COPY [[COPY5]](s32)
-  ; GCN-NEXT:   [[COPY6:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; GCN-NEXT:   S_SETPC_B64_return [[COPY6]], implicit $vgpr0
+  ; GCN-NEXT:   $vgpr0 = COPY [[COPY4]](s32)
+  ; GCN-NEXT:   SI_RETURN implicit $vgpr0
 entry:
   %ret = tail call fastcc i32 @i32_fastcc_i32_i32_a32i32(i32 %a, i32 %b, [32 x i32] zeroinitializer)
   ret i32 %ret
 }
 
 ; Have another non-tail in the function
 define fastcc i32 @sibling_call_i32_fastcc_i32_i32_other_call(i32 %a, i32 %b, i32 %c) #1 {
   ; GCN-LABEL: name: sibling_call_i32_fastcc_i32_i32_other_call
   ; GCN: bb.1.entry:
-  ; GCN-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
+  ; GCN-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; GCN-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
   ; GCN-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
-  ; GCN-NEXT:   [[COPY3:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   ADJCALLSTACKUP 0, 0, implicit-def $scc
   ; GCN-NEXT:   [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @i32_fastcc_i32_i32
   ; GCN-NEXT:   $vgpr0 = COPY [[COPY]](s32)
   ; GCN-NEXT:   $vgpr1 = COPY [[COPY1]](s32)
-  ; GCN-NEXT:   [[COPY4:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY4]](<4 x s32>)
+  ; GCN-NEXT:   [[COPY3:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY3]](<4 x s32>)
   ; GCN-NEXT:   $sgpr30_sgpr31 = G_SI_CALL [[GV]](p0), @i32_fastcc_i32_i32, csr_amdgpu_highregs, implicit $vgpr0, implicit $vgpr1, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit-def $vgpr0
-  ; GCN-NEXT:   [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr0
+  ; GCN-NEXT:   [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; GCN-NEXT:   ADJCALLSTACKDOWN 0, 0, implicit-def $scc
   ; GCN-NEXT:   [[GV1:%[0-9]+]]:sreg_64(p0) = G_GLOBAL_VALUE @sibling_call_i32_fastcc_i32_i32
   ; GCN-NEXT:   $vgpr0 = COPY [[COPY]](s32)
   ; GCN-NEXT:   $vgpr1 = COPY [[COPY1]](s32)
-  ; GCN-NEXT:   $vgpr2 = COPY [[COPY5]](s32)
-  ; GCN-NEXT:   [[COPY6:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY6]](<4 x s32>)
+  ; GCN-NEXT:   $vgpr2 = COPY [[COPY4]](s32)
+  ; GCN-NEXT:   [[COPY5:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY5]](<4 x s32>)
   ; GCN-NEXT:   SI_TCRETURN [[GV1]](p0), @sibling_call_i32_fastcc_i32_i32, 0, csr_amdgpu_highregs, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $sgpr0_sgpr1_sgpr2_sgpr3
 entry:
   %other.call = tail call fastcc i32 @i32_fastcc_i32_i32(i32 %a, i32 %b)
   %ret = tail call fastcc i32 @sibling_call_i32_fastcc_i32_i32(i32 %a, i32 %b, i32 %other.call)
   ret i32 %ret
 }
 
 ; Have stack object in caller and stack passed arguments. SP should be
 ; in same place at function exit.
 define fastcc i32 @sibling_call_stack_objecti32_fastcc_i32_i32_a32i32(i32 %a, i32 %b, [32 x i32] %c) #1 {
   ; GCN-LABEL: name: sibling_call_stack_objecti32_fastcc_i32_i32_a32i32
   ; GCN: bb.1.entry:
-  ; GCN-NEXT:   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, $sgpr30_sgpr31
+  ; GCN-NEXT:   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
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; GCN-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
   ; GCN-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
   ; GCN-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
   ; GCN-NEXT:   [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
   ; GCN-NEXT:   [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
   ; GCN-NEXT:   [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
define fastcc i32 @sibling_call_stack_objecti32_fastcc_i32_i32_a32i32(i32 %a, i32 %b, [32 x i32] %c) #1 {
   ; GCN-NEXT:   [[COPY29:%[0-9]+]]:_(s32) = COPY $vgpr29
   ; GCN-NEXT:   [[COPY30:%[0-9]+]]:_(s32) = COPY $vgpr30
   ; GCN-NEXT:   [[FRAME_INDEX:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.5
   ; GCN-NEXT:   [[LOAD:%[0-9]+]]:_(s32) = G_LOAD [[FRAME_INDEX]](p5) :: (invariant load (s32) from %fixed-stack.5, align 16, addrspace 5)
   ; GCN-NEXT:   [[FRAME_INDEX1:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.4
   ; GCN-NEXT:   [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[FRAME_INDEX1]](p5) :: (invariant load (s32) from %fixed-stack.4, addrspace 5)
   ; GCN-NEXT:   [[FRAME_INDEX2:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.3
   ; GCN-NEXT:   [[LOAD2:%[0-9]+]]:_(s32) = G_LOAD [[FRAME_INDEX2]](p5) :: (invariant load (s32) from %fixed-stack.3, align 8, addrspace 5)
-  ; GCN-NEXT:   [[COPY31:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 9
   ; GCN-NEXT:   [[FRAME_INDEX3:%[0-9]+]]:_(p5) = G_FRAME_INDEX %stack.0.alloca
   ; GCN-NEXT:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 20
   ; GCN-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[FRAME_INDEX3]], [[C1]](s32)
   ; GCN-NEXT:   G_STORE [[C]](s32), [[PTR_ADD]](p5) :: (volatile store (s32) into %ir.gep, addrspace 5)
   ; GCN-NEXT:   [[GV:%[0-9]+]]:sreg_64(p0) = G_GLOBAL_VALUE @i32_fastcc_i32_i32_a32i32
   ; GCN-NEXT:   [[FRAME_INDEX4:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.2
   ; GCN-NEXT:   G_STORE [[LOAD]](s32), [[FRAME_INDEX4]](p5) :: (store (s32) into %fixed-stack.2, align 16, addrspace 5)
define fastcc i32 @sibling_call_stack_objecti32_fastcc_i32_i32_a32i32(i32 %a, i32 %b, [32 x i32] %c) #1 {
   ; GCN-NEXT:   $vgpr23 = COPY [[COPY23]](s32)
   ; GCN-NEXT:   $vgpr24 = COPY [[COPY24]](s32)
   ; GCN-NEXT:   $vgpr25 = COPY [[COPY25]](s32)
   ; GCN-NEXT:   $vgpr26 = COPY [[COPY26]](s32)
   ; GCN-NEXT:   $vgpr27 = COPY [[COPY27]](s32)
   ; GCN-NEXT:   $vgpr28 = COPY [[COPY28]](s32)
   ; GCN-NEXT:   $vgpr29 = COPY [[COPY29]](s32)
   ; GCN-NEXT:   $vgpr30 = COPY [[COPY30]](s32)
-  ; GCN-NEXT:   [[COPY32:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY32]](<4 x s32>)
+  ; GCN-NEXT:   [[COPY31:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY31]](<4 x s32>)
   ; GCN-NEXT:   SI_TCRETURN [[GV]](p0), @i32_fastcc_i32_i32_a32i32, 0, csr_amdgpu_highregs, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15, implicit $vgpr16, implicit $vgpr17, implicit $vgpr18, implicit $vgpr19, implicit $vgpr20, implicit $vgpr21, implicit $vgpr22, implicit $vgpr23, implicit $vgpr24, implicit $vgpr25, implicit $vgpr26, implicit $vgpr27, implicit $vgpr28, implicit $vgpr29, implicit $vgpr30, implicit $sgpr0_sgpr1_sgpr2_sgpr3
 entry:
   %alloca = alloca [16 x i32], align 4, addrspace(5)
   %gep = getelementptr inbounds [16 x i32], [16 x i32] addrspace(5)* %alloca, i32 0, i32 5
   store volatile i32 9, i32 addrspace(5)* %gep
   %ret = tail call fastcc i32 @i32_fastcc_i32_i32_a32i32(i32 %a, i32 %b, [32 x i32] %c)
   ret i32 %ret
 }
 
 define fastcc i32 @sibling_call_stack_objecti32_fastcc_i32_i32_a32i32_larger_arg_area(i32 %a, i32 %b, [36 x i32] %c) #1 {
   ; GCN-LABEL: name: sibling_call_stack_objecti32_fastcc_i32_i32_a32i32_larger_arg_area
   ; GCN: bb.1.entry:
-  ; GCN-NEXT:   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, $sgpr30_sgpr31
+  ; GCN-NEXT:   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
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; GCN-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
   ; GCN-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
   ; GCN-NEXT:   [[COPY3:%[0-9]+]]:_(s32) = COPY $vgpr3
   ; GCN-NEXT:   [[COPY4:%[0-9]+]]:_(s32) = COPY $vgpr4
   ; GCN-NEXT:   [[COPY5:%[0-9]+]]:_(s32) = COPY $vgpr5
   ; GCN-NEXT:   [[COPY6:%[0-9]+]]:_(s32) = COPY $vgpr6
define fastcc i32 @sibling_call_stack_objecti32_fastcc_i32_i32_a32i32_larger_arg_area(i32 %a, i32 %b, [36 x i32] %c) #1 {
   ; GCN-NEXT:   [[FRAME_INDEX3:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.6
   ; GCN-NEXT:   [[LOAD3:%[0-9]+]]:_(s32) = G_LOAD [[FRAME_INDEX3]](p5) :: (invariant load (s32) from %fixed-stack.6, addrspace 5)
   ; GCN-NEXT:   [[FRAME_INDEX4:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.5
   ; GCN-NEXT:   [[LOAD4:%[0-9]+]]:_(s32) = G_LOAD [[FRAME_INDEX4]](p5) :: (invariant load (s32) from %fixed-stack.5, align 16, addrspace 5)
   ; GCN-NEXT:   [[FRAME_INDEX5:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.4
   ; GCN-NEXT:   [[LOAD5:%[0-9]+]]:_(s32) = G_LOAD [[FRAME_INDEX5]](p5) :: (invariant load (s32) from %fixed-stack.4, addrspace 5)
   ; GCN-NEXT:   [[FRAME_INDEX6:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.3
   ; GCN-NEXT:   [[LOAD6:%[0-9]+]]:_(s32) = G_LOAD [[FRAME_INDEX6]](p5) :: (invariant load (s32) from %fixed-stack.3, align 8, addrspace 5)
-  ; GCN-NEXT:   [[COPY31:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 9
   ; GCN-NEXT:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
   ; GCN-NEXT:   [[FRAME_INDEX7:%[0-9]+]]:_(p5) = G_FRAME_INDEX %stack.0.alloca
   ; GCN-NEXT:   [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 20
   ; GCN-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[FRAME_INDEX7]], [[C2]](s32)
   ; GCN-NEXT:   G_STORE [[C]](s32), [[PTR_ADD]](p5) :: (volatile store (s32) into %ir.gep, addrspace 5)
   ; GCN-NEXT:   [[GV:%[0-9]+]]:sreg_64(p0) = G_GLOBAL_VALUE @i32_fastcc_i32_i32_a32i32
   ; GCN-NEXT:   [[FRAME_INDEX8:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.2
define fastcc i32 @sibling_call_stack_objecti32_fastcc_i32_i32_a32i32_larger_arg_area(i32 %a, i32 %b, [36 x i32] %c) #1 {
   ; GCN-NEXT:   $vgpr23 = COPY [[C1]](s32)
   ; GCN-NEXT:   $vgpr24 = COPY [[C1]](s32)
   ; GCN-NEXT:   $vgpr25 = COPY [[C1]](s32)
   ; GCN-NEXT:   $vgpr26 = COPY [[C1]](s32)
   ; GCN-NEXT:   $vgpr27 = COPY [[C1]](s32)
   ; GCN-NEXT:   $vgpr28 = COPY [[C1]](s32)
   ; GCN-NEXT:   $vgpr29 = COPY [[C1]](s32)
   ; GCN-NEXT:   $vgpr30 = COPY [[C1]](s32)
-  ; GCN-NEXT:   [[COPY32:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY32]](<4 x s32>)
+  ; GCN-NEXT:   [[COPY31:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY31]](<4 x s32>)
   ; GCN-NEXT:   SI_TCRETURN [[GV]](p0), @i32_fastcc_i32_i32_a32i32, 0, csr_amdgpu_highregs, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15, implicit $vgpr16, implicit $vgpr17, implicit $vgpr18, implicit $vgpr19, implicit $vgpr20, implicit $vgpr21, implicit $vgpr22, implicit $vgpr23, implicit $vgpr24, implicit $vgpr25, implicit $vgpr26, implicit $vgpr27, implicit $vgpr28, implicit $vgpr29, implicit $vgpr30, implicit $sgpr0_sgpr1_sgpr2_sgpr3
 entry:
   %alloca = alloca [16 x i32], align 4, addrspace(5)
   %gep = getelementptr inbounds [16 x i32], [16 x i32] addrspace(5)* %alloca, i32 0, i32 5
   store volatile i32 9, i32 addrspace(5)* %gep
   %ret = tail call fastcc i32 @i32_fastcc_i32_i32_a32i32(i32 %a, i32 %b, [32 x i32] zeroinitializer)
   ret i32 %ret
 }
 
 declare hidden void @void_fastcc_multi_byval(i32 %a, [3 x i32] addrspace(5)* byval([3 x i32]) align 16, [2 x i64] addrspace(5)* byval([2 x i64]))
 
 define fastcc void @sibling_call_fastcc_multi_byval(i32 %a, [64 x i32]) #1 {
   ; GCN-LABEL: name: sibling_call_fastcc_multi_byval
   ; GCN: bb.1.entry:
-  ; GCN-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $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, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11, $sgpr30_sgpr31
+  ; GCN-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $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, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT:   [[COPY:%[0-9]+]]:vgpr_32(s32) = COPY $vgpr31
   ; GCN-NEXT:   [[COPY1:%[0-9]+]]:sgpr_32 = COPY $sgpr14
   ; GCN-NEXT:   [[COPY2:%[0-9]+]]:sgpr_32 = COPY $sgpr13
   ; GCN-NEXT:   [[COPY3:%[0-9]+]]:sgpr_32 = COPY $sgpr12
   ; GCN-NEXT:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr10_sgpr11
   ; GCN-NEXT:   [[COPY5:%[0-9]+]]:sgpr_64 = COPY $sgpr8_sgpr9
   ; GCN-NEXT:   [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr6_sgpr7
define fastcc void @sibling_call_fastcc_multi_byval(i32 %a, [64 x i32]) #1 {
   ; GCN-NEXT:   [[FRAME_INDEX30:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.5
   ; GCN-NEXT:   [[LOAD30:%[0-9]+]]:_(s32) = G_LOAD [[FRAME_INDEX30]](p5) :: (invariant load (s32) from %fixed-stack.5, align 8, addrspace 5)
   ; GCN-NEXT:   [[FRAME_INDEX31:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.4
   ; GCN-NEXT:   [[LOAD31:%[0-9]+]]:_(s32) = G_LOAD [[FRAME_INDEX31]](p5) :: (invariant load (s32) from %fixed-stack.4, addrspace 5)
   ; GCN-NEXT:   [[FRAME_INDEX32:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.3
   ; GCN-NEXT:   [[LOAD32:%[0-9]+]]:_(s32) = G_LOAD [[FRAME_INDEX32]](p5) :: (invariant load (s32) from %fixed-stack.3, align 16, addrspace 5)
   ; GCN-NEXT:   [[FRAME_INDEX33:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.2
   ; GCN-NEXT:   [[LOAD33:%[0-9]+]]:_(s32) = G_LOAD [[FRAME_INDEX33]](p5) :: (invariant load (s32) from %fixed-stack.2, addrspace 5)
-  ; GCN-NEXT:   [[COPY39:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 9
   ; GCN-NEXT:   [[C1:%[0-9]+]]:_(s64) = G_CONSTANT i64 0
   ; GCN-NEXT:   [[FRAME_INDEX34:%[0-9]+]]:_(p5) = G_FRAME_INDEX %stack.0.alloca0
   ; GCN-NEXT:   [[FRAME_INDEX35:%[0-9]+]]:_(p5) = G_FRAME_INDEX %stack.1.alloca1
   ; GCN-NEXT:   G_STORE [[C]](s32), [[FRAME_INDEX34]](p5) :: (store (s32) into %ir.alloca0, addrspace 5)
   ; GCN-NEXT:   [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 4
   ; GCN-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[FRAME_INDEX34]], [[C2]](s32)
   ; GCN-NEXT:   G_STORE [[C]](s32), [[PTR_ADD]](p5) :: (store (s32) into %ir.alloca0 + 4, addrspace 5)
   ; GCN-NEXT:   [[C3:%[0-9]+]]:_(s32) = G_CONSTANT i32 8
   ; GCN-NEXT:   [[PTR_ADD1:%[0-9]+]]:_(p5) = G_PTR_ADD [[FRAME_INDEX34]], [[C3]](s32)
   ; GCN-NEXT:   G_STORE [[C]](s32), [[PTR_ADD1]](p5) :: (store (s32) into %ir.alloca0 + 8, addrspace 5)
   ; GCN-NEXT:   G_STORE [[C1]](s64), [[FRAME_INDEX35]](p5) :: (store (s64) into %ir.alloca1, addrspace 5)
   ; GCN-NEXT:   [[PTR_ADD2:%[0-9]+]]:_(p5) = G_PTR_ADD [[FRAME_INDEX35]], [[C3]](s32)
   ; GCN-NEXT:   G_STORE [[C1]](s64), [[PTR_ADD2]](p5) :: (store (s64) into %ir.alloca1 + 8, addrspace 5)
   ; GCN-NEXT:   [[GV:%[0-9]+]]:sreg_64(p0) = G_GLOBAL_VALUE @void_fastcc_multi_byval
-  ; GCN-NEXT:   [[COPY40:%[0-9]+]]:_(p4) = COPY [[COPY7]]
-  ; GCN-NEXT:   [[COPY41:%[0-9]+]]:_(p4) = COPY [[COPY6]]
-  ; GCN-NEXT:   [[COPY42:%[0-9]+]]:_(p4) = COPY [[COPY5]]
-  ; GCN-NEXT:   [[COPY43:%[0-9]+]]:_(s64) = COPY [[COPY4]]
-  ; GCN-NEXT:   [[COPY44:%[0-9]+]]:_(s32) = COPY [[COPY3]]
-  ; GCN-NEXT:   [[COPY45:%[0-9]+]]:_(s32) = COPY [[COPY2]]
-  ; GCN-NEXT:   [[COPY46:%[0-9]+]]:_(s32) = COPY [[COPY1]]
-  ; GCN-NEXT:   [[COPY47:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
+  ; GCN-NEXT:   [[COPY39:%[0-9]+]]:_(p4) = COPY [[COPY7]]
+  ; GCN-NEXT:   [[COPY40:%[0-9]+]]:_(p4) = COPY [[COPY6]]
+  ; GCN-NEXT:   [[COPY41:%[0-9]+]]:_(p4) = COPY [[COPY5]]
+  ; GCN-NEXT:   [[COPY42:%[0-9]+]]:_(s64) = COPY [[COPY4]]
+  ; GCN-NEXT:   [[COPY43:%[0-9]+]]:_(s32) = COPY [[COPY3]]
+  ; GCN-NEXT:   [[COPY44:%[0-9]+]]:_(s32) = COPY [[COPY2]]
+  ; GCN-NEXT:   [[COPY45:%[0-9]+]]:_(s32) = COPY [[COPY1]]
+  ; GCN-NEXT:   [[COPY46:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
   ; GCN-NEXT:   [[FRAME_INDEX36:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.1
   ; GCN-NEXT:   [[C4:%[0-9]+]]:_(s32) = G_CONSTANT i32 12
   ; GCN-NEXT:   G_MEMCPY [[FRAME_INDEX36]](p5), [[FRAME_INDEX34]](p5), [[C4]](s32), 0 :: (dereferenceable store (s96) into %fixed-stack.1, align 16, addrspace 5), (dereferenceable load (s96) from %ir.alloca0, align 16, addrspace 5)
   ; GCN-NEXT:   [[FRAME_INDEX37:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.0
   ; GCN-NEXT:   [[C5:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
   ; GCN-NEXT:   G_MEMCPY [[FRAME_INDEX37]](p5), [[FRAME_INDEX35]](p5), [[C5]](s32), 0 :: (dereferenceable store (s128) into %fixed-stack.0, addrspace 5), (dereferenceable load (s128) from %ir.alloca1, align 8, addrspace 5)
   ; GCN-NEXT:   $vgpr0 = COPY [[COPY8]](s32)
-  ; GCN-NEXT:   [[COPY48:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY48]](<4 x s32>)
-  ; GCN-NEXT:   $sgpr4_sgpr5 = COPY [[COPY40]](p4)
-  ; GCN-NEXT:   $sgpr6_sgpr7 = COPY [[COPY41]](p4)
-  ; GCN-NEXT:   $sgpr8_sgpr9 = COPY [[COPY42]](p4)
-  ; GCN-NEXT:   $sgpr10_sgpr11 = COPY [[COPY43]](s64)
-  ; GCN-NEXT:   $sgpr12 = COPY [[COPY44]](s32)
-  ; GCN-NEXT:   $sgpr13 = COPY [[COPY45]](s32)
-  ; GCN-NEXT:   $sgpr14 = COPY [[COPY46]](s32)
-  ; GCN-NEXT:   $vgpr31 = COPY [[COPY47]](s32)
+  ; GCN-NEXT:   [[COPY47:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY47]](<4 x s32>)
+  ; GCN-NEXT:   $sgpr4_sgpr5 = COPY [[COPY39]](p4)
+  ; GCN-NEXT:   $sgpr6_sgpr7 = COPY [[COPY40]](p4)
+  ; GCN-NEXT:   $sgpr8_sgpr9 = COPY [[COPY41]](p4)
+  ; GCN-NEXT:   $sgpr10_sgpr11 = COPY [[COPY42]](s64)
+  ; GCN-NEXT:   $sgpr12 = COPY [[COPY43]](s32)
+  ; GCN-NEXT:   $sgpr13 = COPY [[COPY44]](s32)
+  ; GCN-NEXT:   $sgpr14 = COPY [[COPY45]](s32)
+  ; GCN-NEXT:   $vgpr31 = COPY [[COPY46]](s32)
   ; GCN-NEXT:   SI_TCRETURN [[GV]](p0), @void_fastcc_multi_byval, 0, csr_amdgpu_highregs, implicit $vgpr0, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr4_sgpr5, implicit $sgpr6_sgpr7, implicit $sgpr8_sgpr9, implicit $sgpr10_sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $vgpr31
 entry:
   %alloca0 = alloca [3 x i32], align 16, addrspace(5)
   %alloca1 = alloca [2 x i64], align 8, addrspace(5)
   store [3 x i32] [i32 9, i32 9, i32 9], [3 x i32] addrspace(5)* %alloca0
   store [2 x i64] zeroinitializer, [2 x i64] addrspace(5)* %alloca1
   tail call fastcc void @void_fastcc_multi_byval(i32 %a, [3 x i32] addrspace(5)* byval([3 x i32]) %alloca0, [2 x i64] addrspace(5)* byval([2 x i64]) %alloca1)
   ret void
 }
 
 declare hidden void @void_fastcc_byval_and_stack_passed([3 x i32] addrspace(5)* byval([3 x i32]) align 16, [32 x i32], i32)
 
 ; Callee has a byval and non-byval stack passed argument
 define fastcc void @sibling_call_byval_and_stack_passed(i32 %stack.out.arg, [64 x i32]) #1 {
   ; GCN-LABEL: name: sibling_call_byval_and_stack_passed
   ; GCN: bb.1.entry:
-  ; GCN-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $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, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11, $sgpr30_sgpr31
+  ; GCN-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $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, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT:   [[COPY:%[0-9]+]]:vgpr_32(s32) = COPY $vgpr31
   ; GCN-NEXT:   [[COPY1:%[0-9]+]]:sgpr_32 = COPY $sgpr14
   ; GCN-NEXT:   [[COPY2:%[0-9]+]]:sgpr_32 = COPY $sgpr13
   ; GCN-NEXT:   [[COPY3:%[0-9]+]]:sgpr_32 = COPY $sgpr12
   ; GCN-NEXT:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr10_sgpr11
   ; GCN-NEXT:   [[COPY5:%[0-9]+]]:sgpr_64 = COPY $sgpr8_sgpr9
   ; GCN-NEXT:   [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr6_sgpr7
define fastcc void @sibling_call_byval_and_stack_passed(i32 %stack.out.arg, [64 x i32]) #1 {
   ; GCN-NEXT:   [[FRAME_INDEX30:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.6
   ; GCN-NEXT:   [[LOAD30:%[0-9]+]]:_(s32) = G_LOAD [[FRAME_INDEX30]](p5) :: (invariant load (s32) from %fixed-stack.6, align 8, addrspace 5)
   ; GCN-NEXT:   [[FRAME_INDEX31:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.5
   ; GCN-NEXT:   [[LOAD31:%[0-9]+]]:_(s32) = G_LOAD [[FRAME_INDEX31]](p5) :: (invariant load (s32) from %fixed-stack.5, addrspace 5)
   ; GCN-NEXT:   [[FRAME_INDEX32:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.4
   ; GCN-NEXT:   [[LOAD32:%[0-9]+]]:_(s32) = G_LOAD [[FRAME_INDEX32]](p5) :: (invariant load (s32) from %fixed-stack.4, align 16, addrspace 5)
   ; GCN-NEXT:   [[FRAME_INDEX33:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.3
   ; GCN-NEXT:   [[LOAD33:%[0-9]+]]:_(s32) = G_LOAD [[FRAME_INDEX33]](p5) :: (invariant load (s32) from %fixed-stack.3, addrspace 5)
-  ; GCN-NEXT:   [[COPY39:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 9
   ; GCN-NEXT:   [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
   ; GCN-NEXT:   [[FRAME_INDEX34:%[0-9]+]]:_(p5) = G_FRAME_INDEX %stack.0.alloca
   ; GCN-NEXT:   G_STORE [[C]](s32), [[FRAME_INDEX34]](p5) :: (store (s32) into %ir.alloca, addrspace 5)
   ; GCN-NEXT:   [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 4
   ; GCN-NEXT:   [[PTR_ADD:%[0-9]+]]:_(p5) = G_PTR_ADD [[FRAME_INDEX34]], [[C2]](s32)
   ; GCN-NEXT:   G_STORE [[C]](s32), [[PTR_ADD]](p5) :: (store (s32) into %ir.alloca + 4, addrspace 5)
   ; GCN-NEXT:   [[C3:%[0-9]+]]:_(s32) = G_CONSTANT i32 8
   ; GCN-NEXT:   [[PTR_ADD1:%[0-9]+]]:_(p5) = G_PTR_ADD [[FRAME_INDEX34]], [[C3]](s32)
   ; GCN-NEXT:   G_STORE [[C]](s32), [[PTR_ADD1]](p5) :: (store (s32) into %ir.alloca + 8, addrspace 5)
   ; GCN-NEXT:   [[GV:%[0-9]+]]:sreg_64(p0) = G_GLOBAL_VALUE @void_fastcc_byval_and_stack_passed
-  ; GCN-NEXT:   [[COPY40:%[0-9]+]]:_(p4) = COPY [[COPY7]]
-  ; GCN-NEXT:   [[COPY41:%[0-9]+]]:_(p4) = COPY [[COPY6]]
-  ; GCN-NEXT:   [[COPY42:%[0-9]+]]:_(p4) = COPY [[COPY5]]
-  ; GCN-NEXT:   [[COPY43:%[0-9]+]]:_(s64) = COPY [[COPY4]]
-  ; GCN-NEXT:   [[COPY44:%[0-9]+]]:_(s32) = COPY [[COPY3]]
-  ; GCN-NEXT:   [[COPY45:%[0-9]+]]:_(s32) = COPY [[COPY2]]
-  ; GCN-NEXT:   [[COPY46:%[0-9]+]]:_(s32) = COPY [[COPY1]]
-  ; GCN-NEXT:   [[COPY47:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
+  ; GCN-NEXT:   [[COPY39:%[0-9]+]]:_(p4) = COPY [[COPY7]]
+  ; GCN-NEXT:   [[COPY40:%[0-9]+]]:_(p4) = COPY [[COPY6]]
+  ; GCN-NEXT:   [[COPY41:%[0-9]+]]:_(p4) = COPY [[COPY5]]
+  ; GCN-NEXT:   [[COPY42:%[0-9]+]]:_(s64) = COPY [[COPY4]]
+  ; GCN-NEXT:   [[COPY43:%[0-9]+]]:_(s32) = COPY [[COPY3]]
+  ; GCN-NEXT:   [[COPY44:%[0-9]+]]:_(s32) = COPY [[COPY2]]
+  ; GCN-NEXT:   [[COPY45:%[0-9]+]]:_(s32) = COPY [[COPY1]]
+  ; GCN-NEXT:   [[COPY46:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
   ; GCN-NEXT:   [[FRAME_INDEX35:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.2
   ; GCN-NEXT:   [[C4:%[0-9]+]]:_(s32) = G_CONSTANT i32 12
   ; GCN-NEXT:   G_MEMCPY [[FRAME_INDEX35]](p5), [[FRAME_INDEX34]](p5), [[C4]](s32), 0 :: (dereferenceable store (s96) into %fixed-stack.2, align 16, addrspace 5), (dereferenceable load (s96) from %ir.alloca, align 16, addrspace 5)
   ; GCN-NEXT:   [[FRAME_INDEX36:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.1
   ; GCN-NEXT:   G_STORE [[C1]](s32), [[FRAME_INDEX36]](p5) :: (store (s32) into %fixed-stack.1, addrspace 5)
   ; GCN-NEXT:   [[FRAME_INDEX37:%[0-9]+]]:_(p5) = G_FRAME_INDEX %fixed-stack.0
   ; GCN-NEXT:   G_STORE [[COPY8]](s32), [[FRAME_INDEX37]](p5) :: (store (s32) into %fixed-stack.0, align 16, addrspace 5)
   ; GCN-NEXT:   $vgpr0 = COPY [[C1]](s32)
define fastcc void @sibling_call_byval_and_stack_passed(i32 %stack.out.arg, [64 x i32]) #1 {
   ; GCN-NEXT:   $vgpr23 = COPY [[C1]](s32)
   ; GCN-NEXT:   $vgpr24 = COPY [[C1]](s32)
   ; GCN-NEXT:   $vgpr25 = COPY [[C1]](s32)
   ; GCN-NEXT:   $vgpr26 = COPY [[C1]](s32)
   ; GCN-NEXT:   $vgpr27 = COPY [[C1]](s32)
   ; GCN-NEXT:   $vgpr28 = COPY [[C1]](s32)
   ; GCN-NEXT:   $vgpr29 = COPY [[C1]](s32)
   ; GCN-NEXT:   $vgpr30 = COPY [[C1]](s32)
-  ; GCN-NEXT:   [[COPY48:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY48]](<4 x s32>)
-  ; GCN-NEXT:   $sgpr4_sgpr5 = COPY [[COPY40]](p4)
-  ; GCN-NEXT:   $sgpr6_sgpr7 = COPY [[COPY41]](p4)
-  ; GCN-NEXT:   $sgpr8_sgpr9 = COPY [[COPY42]](p4)
-  ; GCN-NEXT:   $sgpr10_sgpr11 = COPY [[COPY43]](s64)
-  ; GCN-NEXT:   $sgpr12 = COPY [[COPY44]](s32)
-  ; GCN-NEXT:   $sgpr13 = COPY [[COPY45]](s32)
-  ; GCN-NEXT:   $sgpr14 = COPY [[COPY46]](s32)
-  ; GCN-NEXT:   $vgpr31 = COPY [[COPY47]](s32)
+  ; GCN-NEXT:   [[COPY47:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY47]](<4 x s32>)
+  ; GCN-NEXT:   $sgpr4_sgpr5 = COPY [[COPY39]](p4)
+  ; GCN-NEXT:   $sgpr6_sgpr7 = COPY [[COPY40]](p4)
+  ; GCN-NEXT:   $sgpr8_sgpr9 = COPY [[COPY41]](p4)
+  ; GCN-NEXT:   $sgpr10_sgpr11 = COPY [[COPY42]](s64)
+  ; GCN-NEXT:   $sgpr12 = COPY [[COPY43]](s32)
+  ; GCN-NEXT:   $sgpr13 = COPY [[COPY44]](s32)
+  ; GCN-NEXT:   $sgpr14 = COPY [[COPY45]](s32)
+  ; GCN-NEXT:   $vgpr31 = COPY [[COPY46]](s32)
   ; GCN-NEXT:   SI_TCRETURN [[GV]](p0), @void_fastcc_byval_and_stack_passed, 0, csr_amdgpu_highregs, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15, implicit $vgpr16, implicit $vgpr17, implicit $vgpr18, implicit $vgpr19, implicit $vgpr20, implicit $vgpr21, implicit $vgpr22, implicit $vgpr23, implicit $vgpr24, implicit $vgpr25, implicit $vgpr26, implicit $vgpr27, implicit $vgpr28, implicit $vgpr29, implicit $vgpr30, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr4_sgpr5, implicit $sgpr6_sgpr7, implicit $sgpr8_sgpr9, implicit $sgpr10_sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $vgpr31
 entry:
   %alloca = alloca [3 x i32], align 16, addrspace(5)
   store [3 x i32] [i32 9, i32 9, i32 9], [3 x i32] addrspace(5)* %alloca
   tail call fastcc void @void_fastcc_byval_and_stack_passed([3 x i32] addrspace(5)* byval([3 x i32]) %alloca, [32 x i32] zeroinitializer, i32 %stack.out.arg)
   ret void
 }
 
 declare hidden fastcc i64 @i64_fastcc_i64(i64 %arg0)
 
 define hidden fastcc i64 @sibling_call_i64_fastcc_i64(i64 %a) #1 {
   ; GCN-LABEL: name: sibling_call_i64_fastcc_i64
   ; GCN: bb.1.entry:
-  ; GCN-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr0, $vgpr1, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11, $sgpr30_sgpr31
+  ; GCN-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr0, $vgpr1, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT:   [[COPY:%[0-9]+]]:vgpr_32(s32) = COPY $vgpr31
   ; GCN-NEXT:   [[COPY1:%[0-9]+]]:sgpr_32 = COPY $sgpr14
   ; GCN-NEXT:   [[COPY2:%[0-9]+]]:sgpr_32 = COPY $sgpr13
   ; GCN-NEXT:   [[COPY3:%[0-9]+]]:sgpr_32 = COPY $sgpr12
   ; GCN-NEXT:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr10_sgpr11
   ; GCN-NEXT:   [[COPY5:%[0-9]+]]:sgpr_64 = COPY $sgpr8_sgpr9
   ; GCN-NEXT:   [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr6_sgpr7
   ; GCN-NEXT:   [[COPY7:%[0-9]+]]:sgpr_64 = COPY $sgpr4_sgpr5
   ; GCN-NEXT:   [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; GCN-NEXT:   [[COPY9:%[0-9]+]]:_(s32) = COPY $vgpr1
   ; GCN-NEXT:   [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY8]](s32), [[COPY9]](s32)
-  ; GCN-NEXT:   [[COPY10:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   [[GV:%[0-9]+]]:sreg_64(p0) = G_GLOBAL_VALUE @i64_fastcc_i64
-  ; GCN-NEXT:   [[COPY11:%[0-9]+]]:_(p4) = COPY [[COPY7]]
-  ; GCN-NEXT:   [[COPY12:%[0-9]+]]:_(p4) = COPY [[COPY6]]
-  ; GCN-NEXT:   [[COPY13:%[0-9]+]]:_(p4) = COPY [[COPY5]]
-  ; GCN-NEXT:   [[COPY14:%[0-9]+]]:_(s64) = COPY [[COPY4]]
-  ; GCN-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY3]]
-  ; GCN-NEXT:   [[COPY16:%[0-9]+]]:_(s32) = COPY [[COPY2]]
-  ; GCN-NEXT:   [[COPY17:%[0-9]+]]:_(s32) = COPY [[COPY1]]
-  ; GCN-NEXT:   [[COPY18:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
+  ; GCN-NEXT:   [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY7]]
+  ; GCN-NEXT:   [[COPY11:%[0-9]+]]:_(p4) = COPY [[COPY6]]
+  ; GCN-NEXT:   [[COPY12:%[0-9]+]]:_(p4) = COPY [[COPY5]]
+  ; GCN-NEXT:   [[COPY13:%[0-9]+]]:_(s64) = COPY [[COPY4]]
+  ; GCN-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY3]]
+  ; GCN-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY2]]
+  ; GCN-NEXT:   [[COPY16:%[0-9]+]]:_(s32) = COPY [[COPY1]]
+  ; GCN-NEXT:   [[COPY17:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
   ; GCN-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[MV]](s64)
   ; GCN-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; GCN-NEXT:   $vgpr1 = COPY [[UV1]](s32)
-  ; GCN-NEXT:   [[COPY19:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY19]](<4 x s32>)
-  ; GCN-NEXT:   $sgpr4_sgpr5 = COPY [[COPY11]](p4)
-  ; GCN-NEXT:   $sgpr6_sgpr7 = COPY [[COPY12]](p4)
-  ; GCN-NEXT:   $sgpr8_sgpr9 = COPY [[COPY13]](p4)
-  ; GCN-NEXT:   $sgpr10_sgpr11 = COPY [[COPY14]](s64)
-  ; GCN-NEXT:   $sgpr12 = COPY [[COPY15]](s32)
-  ; GCN-NEXT:   $sgpr13 = COPY [[COPY16]](s32)
-  ; GCN-NEXT:   $sgpr14 = COPY [[COPY17]](s32)
-  ; GCN-NEXT:   $vgpr31 = COPY [[COPY18]](s32)
+  ; GCN-NEXT:   [[COPY18:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY18]](<4 x s32>)
+  ; GCN-NEXT:   $sgpr4_sgpr5 = COPY [[COPY10]](p4)
+  ; GCN-NEXT:   $sgpr6_sgpr7 = COPY [[COPY11]](p4)
+  ; GCN-NEXT:   $sgpr8_sgpr9 = COPY [[COPY12]](p4)
+  ; GCN-NEXT:   $sgpr10_sgpr11 = COPY [[COPY13]](s64)
+  ; GCN-NEXT:   $sgpr12 = COPY [[COPY14]](s32)
+  ; GCN-NEXT:   $sgpr13 = COPY [[COPY15]](s32)
+  ; GCN-NEXT:   $sgpr14 = COPY [[COPY16]](s32)
+  ; GCN-NEXT:   $vgpr31 = COPY [[COPY17]](s32)
   ; GCN-NEXT:   SI_TCRETURN [[GV]](p0), @i64_fastcc_i64, 0, csr_amdgpu_highregs, implicit $vgpr0, implicit $vgpr1, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr4_sgpr5, implicit $sgpr6_sgpr7, implicit $sgpr8_sgpr9, implicit $sgpr10_sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $vgpr31
 entry:
   %ret = tail call fastcc i64 @i64_fastcc_i64(i64 %a)
   ret i64 %ret
 }
 
 declare hidden fastcc i8 addrspace(1)* @p1i8_fastcc_p1i8(i8 addrspace(1)* %arg0)
 
 define hidden fastcc i8 addrspace(1)* @sibling_call_p1i8_fastcc_p1i8(i8 addrspace(1)* %a) #1 {
   ; GCN-LABEL: name: sibling_call_p1i8_fastcc_p1i8
   ; GCN: bb.1.entry:
-  ; GCN-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr0, $vgpr1, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11, $sgpr30_sgpr31
+  ; GCN-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr0, $vgpr1, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT:   [[COPY:%[0-9]+]]:vgpr_32(s32) = COPY $vgpr31
   ; GCN-NEXT:   [[COPY1:%[0-9]+]]:sgpr_32 = COPY $sgpr14
   ; GCN-NEXT:   [[COPY2:%[0-9]+]]:sgpr_32 = COPY $sgpr13
   ; GCN-NEXT:   [[COPY3:%[0-9]+]]:sgpr_32 = COPY $sgpr12
   ; GCN-NEXT:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr10_sgpr11
   ; GCN-NEXT:   [[COPY5:%[0-9]+]]:sgpr_64 = COPY $sgpr8_sgpr9
   ; GCN-NEXT:   [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr6_sgpr7
   ; GCN-NEXT:   [[COPY7:%[0-9]+]]:sgpr_64 = COPY $sgpr4_sgpr5
   ; GCN-NEXT:   [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; GCN-NEXT:   [[COPY9:%[0-9]+]]:_(s32) = COPY $vgpr1
   ; GCN-NEXT:   [[MV:%[0-9]+]]:_(p1) = G_MERGE_VALUES [[COPY8]](s32), [[COPY9]](s32)
-  ; GCN-NEXT:   [[COPY10:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   [[GV:%[0-9]+]]:sreg_64(p0) = G_GLOBAL_VALUE @p1i8_fastcc_p1i8
-  ; GCN-NEXT:   [[COPY11:%[0-9]+]]:_(p4) = COPY [[COPY7]]
-  ; GCN-NEXT:   [[COPY12:%[0-9]+]]:_(p4) = COPY [[COPY6]]
-  ; GCN-NEXT:   [[COPY13:%[0-9]+]]:_(p4) = COPY [[COPY5]]
-  ; GCN-NEXT:   [[COPY14:%[0-9]+]]:_(s64) = COPY [[COPY4]]
-  ; GCN-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY3]]
-  ; GCN-NEXT:   [[COPY16:%[0-9]+]]:_(s32) = COPY [[COPY2]]
-  ; GCN-NEXT:   [[COPY17:%[0-9]+]]:_(s32) = COPY [[COPY1]]
-  ; GCN-NEXT:   [[COPY18:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
+  ; GCN-NEXT:   [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY7]]
+  ; GCN-NEXT:   [[COPY11:%[0-9]+]]:_(p4) = COPY [[COPY6]]
+  ; GCN-NEXT:   [[COPY12:%[0-9]+]]:_(p4) = COPY [[COPY5]]
+  ; GCN-NEXT:   [[COPY13:%[0-9]+]]:_(s64) = COPY [[COPY4]]
+  ; GCN-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY3]]
+  ; GCN-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY2]]
+  ; GCN-NEXT:   [[COPY16:%[0-9]+]]:_(s32) = COPY [[COPY1]]
+  ; GCN-NEXT:   [[COPY17:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
   ; GCN-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[MV]](p1)
   ; GCN-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; GCN-NEXT:   $vgpr1 = COPY [[UV1]](s32)
-  ; GCN-NEXT:   [[COPY19:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY19]](<4 x s32>)
-  ; GCN-NEXT:   $sgpr4_sgpr5 = COPY [[COPY11]](p4)
-  ; GCN-NEXT:   $sgpr6_sgpr7 = COPY [[COPY12]](p4)
-  ; GCN-NEXT:   $sgpr8_sgpr9 = COPY [[COPY13]](p4)
-  ; GCN-NEXT:   $sgpr10_sgpr11 = COPY [[COPY14]](s64)
-  ; GCN-NEXT:   $sgpr12 = COPY [[COPY15]](s32)
-  ; GCN-NEXT:   $sgpr13 = COPY [[COPY16]](s32)
-  ; GCN-NEXT:   $sgpr14 = COPY [[COPY17]](s32)
-  ; GCN-NEXT:   $vgpr31 = COPY [[COPY18]](s32)
+  ; GCN-NEXT:   [[COPY18:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY18]](<4 x s32>)
+  ; GCN-NEXT:   $sgpr4_sgpr5 = COPY [[COPY10]](p4)
+  ; GCN-NEXT:   $sgpr6_sgpr7 = COPY [[COPY11]](p4)
+  ; GCN-NEXT:   $sgpr8_sgpr9 = COPY [[COPY12]](p4)
+  ; GCN-NEXT:   $sgpr10_sgpr11 = COPY [[COPY13]](s64)
+  ; GCN-NEXT:   $sgpr12 = COPY [[COPY14]](s32)
+  ; GCN-NEXT:   $sgpr13 = COPY [[COPY15]](s32)
+  ; GCN-NEXT:   $sgpr14 = COPY [[COPY16]](s32)
+  ; GCN-NEXT:   $vgpr31 = COPY [[COPY17]](s32)
   ; GCN-NEXT:   SI_TCRETURN [[GV]](p0), @p1i8_fastcc_p1i8, 0, csr_amdgpu_highregs, implicit $vgpr0, implicit $vgpr1, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr4_sgpr5, implicit $sgpr6_sgpr7, implicit $sgpr8_sgpr9, implicit $sgpr10_sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $vgpr31
 entry:
   %ret = tail call fastcc i8 addrspace(1)* @p1i8_fastcc_p1i8(i8 addrspace(1)* %a)
   ret i8 addrspace(1)* %ret
 }
 
 declare hidden fastcc i16 @i16_fastcc_i16(i16 %arg0)
 
 define hidden fastcc i16 @sibling_call_i16_fastcc_i16(i16 %a) #1 {
   ; GCN-LABEL: name: sibling_call_i16_fastcc_i16
   ; GCN: bb.1.entry:
-  ; GCN-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr0, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11, $sgpr30_sgpr31
+  ; GCN-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr0, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT:   [[COPY:%[0-9]+]]:vgpr_32(s32) = COPY $vgpr31
   ; GCN-NEXT:   [[COPY1:%[0-9]+]]:sgpr_32 = COPY $sgpr14
   ; GCN-NEXT:   [[COPY2:%[0-9]+]]:sgpr_32 = COPY $sgpr13
   ; GCN-NEXT:   [[COPY3:%[0-9]+]]:sgpr_32 = COPY $sgpr12
   ; GCN-NEXT:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr10_sgpr11
   ; GCN-NEXT:   [[COPY5:%[0-9]+]]:sgpr_64 = COPY $sgpr8_sgpr9
   ; GCN-NEXT:   [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr6_sgpr7
   ; GCN-NEXT:   [[COPY7:%[0-9]+]]:sgpr_64 = COPY $sgpr4_sgpr5
   ; GCN-NEXT:   [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; GCN-NEXT:   [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY8]](s32)
-  ; GCN-NEXT:   [[COPY9:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   [[GV:%[0-9]+]]:sreg_64(p0) = G_GLOBAL_VALUE @i16_fastcc_i16
-  ; GCN-NEXT:   [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY7]]
-  ; GCN-NEXT:   [[COPY11:%[0-9]+]]:_(p4) = COPY [[COPY6]]
-  ; GCN-NEXT:   [[COPY12:%[0-9]+]]:_(p4) = COPY [[COPY5]]
-  ; GCN-NEXT:   [[COPY13:%[0-9]+]]:_(s64) = COPY [[COPY4]]
-  ; GCN-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY3]]
-  ; GCN-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY2]]
-  ; GCN-NEXT:   [[COPY16:%[0-9]+]]:_(s32) = COPY [[COPY1]]
-  ; GCN-NEXT:   [[COPY17:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
+  ; GCN-NEXT:   [[COPY9:%[0-9]+]]:_(p4) = COPY [[COPY7]]
+  ; GCN-NEXT:   [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY6]]
+  ; GCN-NEXT:   [[COPY11:%[0-9]+]]:_(p4) = COPY [[COPY5]]
+  ; GCN-NEXT:   [[COPY12:%[0-9]+]]:_(s64) = COPY [[COPY4]]
+  ; GCN-NEXT:   [[COPY13:%[0-9]+]]:_(s32) = COPY [[COPY3]]
+  ; GCN-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY2]]
+  ; GCN-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY1]]
+  ; GCN-NEXT:   [[COPY16:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
   ; GCN-NEXT:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[TRUNC]](s16)
   ; GCN-NEXT:   $vgpr0 = COPY [[ANYEXT]](s32)
-  ; GCN-NEXT:   [[COPY18:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY18]](<4 x s32>)
-  ; GCN-NEXT:   $sgpr4_sgpr5 = COPY [[COPY10]](p4)
-  ; GCN-NEXT:   $sgpr6_sgpr7 = COPY [[COPY11]](p4)
-  ; GCN-NEXT:   $sgpr8_sgpr9 = COPY [[COPY12]](p4)
-  ; GCN-NEXT:   $sgpr10_sgpr11 = COPY [[COPY13]](s64)
-  ; GCN-NEXT:   $sgpr12 = COPY [[COPY14]](s32)
-  ; GCN-NEXT:   $sgpr13 = COPY [[COPY15]](s32)
-  ; GCN-NEXT:   $sgpr14 = COPY [[COPY16]](s32)
-  ; GCN-NEXT:   $vgpr31 = COPY [[COPY17]](s32)
+  ; GCN-NEXT:   [[COPY17:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY17]](<4 x s32>)
+  ; GCN-NEXT:   $sgpr4_sgpr5 = COPY [[COPY9]](p4)
+  ; GCN-NEXT:   $sgpr6_sgpr7 = COPY [[COPY10]](p4)
+  ; GCN-NEXT:   $sgpr8_sgpr9 = COPY [[COPY11]](p4)
+  ; GCN-NEXT:   $sgpr10_sgpr11 = COPY [[COPY12]](s64)
+  ; GCN-NEXT:   $sgpr12 = COPY [[COPY13]](s32)
+  ; GCN-NEXT:   $sgpr13 = COPY [[COPY14]](s32)
+  ; GCN-NEXT:   $sgpr14 = COPY [[COPY15]](s32)
+  ; GCN-NEXT:   $vgpr31 = COPY [[COPY16]](s32)
   ; GCN-NEXT:   SI_TCRETURN [[GV]](p0), @i16_fastcc_i16, 0, csr_amdgpu_highregs, implicit $vgpr0, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr4_sgpr5, implicit $sgpr6_sgpr7, implicit $sgpr8_sgpr9, implicit $sgpr10_sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $vgpr31
 entry:
   %ret = tail call fastcc i16 @i16_fastcc_i16(i16 %a)
   ret i16 %ret
 }
 
 declare hidden fastcc half @f16_fastcc_f16(half %arg0)
 
 define hidden fastcc half @sibling_call_f16_fastcc_f16(half %a) #1 {
   ; GCN-LABEL: name: sibling_call_f16_fastcc_f16
   ; GCN: bb.1.entry:
-  ; GCN-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr0, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11, $sgpr30_sgpr31
+  ; GCN-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr0, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT:   [[COPY:%[0-9]+]]:vgpr_32(s32) = COPY $vgpr31
   ; GCN-NEXT:   [[COPY1:%[0-9]+]]:sgpr_32 = COPY $sgpr14
   ; GCN-NEXT:   [[COPY2:%[0-9]+]]:sgpr_32 = COPY $sgpr13
   ; GCN-NEXT:   [[COPY3:%[0-9]+]]:sgpr_32 = COPY $sgpr12
   ; GCN-NEXT:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr10_sgpr11
   ; GCN-NEXT:   [[COPY5:%[0-9]+]]:sgpr_64 = COPY $sgpr8_sgpr9
   ; GCN-NEXT:   [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr6_sgpr7
   ; GCN-NEXT:   [[COPY7:%[0-9]+]]:sgpr_64 = COPY $sgpr4_sgpr5
   ; GCN-NEXT:   [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; GCN-NEXT:   [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[COPY8]](s32)
-  ; GCN-NEXT:   [[COPY9:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   [[GV:%[0-9]+]]:sreg_64(p0) = G_GLOBAL_VALUE @f16_fastcc_f16
-  ; GCN-NEXT:   [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY7]]
-  ; GCN-NEXT:   [[COPY11:%[0-9]+]]:_(p4) = COPY [[COPY6]]
-  ; GCN-NEXT:   [[COPY12:%[0-9]+]]:_(p4) = COPY [[COPY5]]
-  ; GCN-NEXT:   [[COPY13:%[0-9]+]]:_(s64) = COPY [[COPY4]]
-  ; GCN-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY3]]
-  ; GCN-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY2]]
-  ; GCN-NEXT:   [[COPY16:%[0-9]+]]:_(s32) = COPY [[COPY1]]
-  ; GCN-NEXT:   [[COPY17:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
+  ; GCN-NEXT:   [[COPY9:%[0-9]+]]:_(p4) = COPY [[COPY7]]
+  ; GCN-NEXT:   [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY6]]
+  ; GCN-NEXT:   [[COPY11:%[0-9]+]]:_(p4) = COPY [[COPY5]]
+  ; GCN-NEXT:   [[COPY12:%[0-9]+]]:_(s64) = COPY [[COPY4]]
+  ; GCN-NEXT:   [[COPY13:%[0-9]+]]:_(s32) = COPY [[COPY3]]
+  ; GCN-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY2]]
+  ; GCN-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY1]]
+  ; GCN-NEXT:   [[COPY16:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
   ; GCN-NEXT:   [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[TRUNC]](s16)
   ; GCN-NEXT:   $vgpr0 = COPY [[ANYEXT]](s32)
-  ; GCN-NEXT:   [[COPY18:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY18]](<4 x s32>)
-  ; GCN-NEXT:   $sgpr4_sgpr5 = COPY [[COPY10]](p4)
-  ; GCN-NEXT:   $sgpr6_sgpr7 = COPY [[COPY11]](p4)
-  ; GCN-NEXT:   $sgpr8_sgpr9 = COPY [[COPY12]](p4)
-  ; GCN-NEXT:   $sgpr10_sgpr11 = COPY [[COPY13]](s64)
-  ; GCN-NEXT:   $sgpr12 = COPY [[COPY14]](s32)
-  ; GCN-NEXT:   $sgpr13 = COPY [[COPY15]](s32)
-  ; GCN-NEXT:   $sgpr14 = COPY [[COPY16]](s32)
-  ; GCN-NEXT:   $vgpr31 = COPY [[COPY17]](s32)
+  ; GCN-NEXT:   [[COPY17:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY17]](<4 x s32>)
+  ; GCN-NEXT:   $sgpr4_sgpr5 = COPY [[COPY9]](p4)
+  ; GCN-NEXT:   $sgpr6_sgpr7 = COPY [[COPY10]](p4)
+  ; GCN-NEXT:   $sgpr8_sgpr9 = COPY [[COPY11]](p4)
+  ; GCN-NEXT:   $sgpr10_sgpr11 = COPY [[COPY12]](s64)
+  ; GCN-NEXT:   $sgpr12 = COPY [[COPY13]](s32)
+  ; GCN-NEXT:   $sgpr13 = COPY [[COPY14]](s32)
+  ; GCN-NEXT:   $sgpr14 = COPY [[COPY15]](s32)
+  ; GCN-NEXT:   $vgpr31 = COPY [[COPY16]](s32)
   ; GCN-NEXT:   SI_TCRETURN [[GV]](p0), @f16_fastcc_f16, 0, csr_amdgpu_highregs, implicit $vgpr0, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr4_sgpr5, implicit $sgpr6_sgpr7, implicit $sgpr8_sgpr9, implicit $sgpr10_sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $vgpr31
 entry:
   %ret = tail call fastcc half @f16_fastcc_f16(half %a)
   ret half %ret
 }
 
 declare hidden fastcc <3 x i16> @v3i16_fastcc_v3i16(<3 x i16> %arg0)
 
 define hidden fastcc <3 x i16> @sibling_call_v3i16_fastcc_v3i16(<3 x i16> %a) #1 {
   ; GCN-LABEL: name: sibling_call_v3i16_fastcc_v3i16
   ; GCN: bb.1.entry:
-  ; GCN-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr0, $vgpr1, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11, $sgpr30_sgpr31
+  ; GCN-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr0, $vgpr1, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT:   [[COPY:%[0-9]+]]:vgpr_32(s32) = COPY $vgpr31
   ; GCN-NEXT:   [[COPY1:%[0-9]+]]:sgpr_32 = COPY $sgpr14
   ; GCN-NEXT:   [[COPY2:%[0-9]+]]:sgpr_32 = COPY $sgpr13
   ; GCN-NEXT:   [[COPY3:%[0-9]+]]:sgpr_32 = COPY $sgpr12
   ; GCN-NEXT:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr10_sgpr11
   ; GCN-NEXT:   [[COPY5:%[0-9]+]]:sgpr_64 = COPY $sgpr8_sgpr9
   ; GCN-NEXT:   [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr6_sgpr7
   ; GCN-NEXT:   [[COPY7:%[0-9]+]]:sgpr_64 = COPY $sgpr4_sgpr5
   ; GCN-NEXT:   [[COPY8:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
   ; GCN-NEXT:   [[COPY9:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
   ; GCN-NEXT:   [[DEF:%[0-9]+]]:_(<2 x s16>) = G_IMPLICIT_DEF
   ; GCN-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY8]](<2 x s16>), [[COPY9]](<2 x s16>), [[DEF]](<2 x s16>)
   ; GCN-NEXT:   [[UV:%[0-9]+]]:_(<3 x s16>), [[UV1:%[0-9]+]]:_(<3 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<6 x s16>)
-  ; GCN-NEXT:   [[COPY10:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   [[GV:%[0-9]+]]:sreg_64(p0) = G_GLOBAL_VALUE @v3i16_fastcc_v3i16
-  ; GCN-NEXT:   [[COPY11:%[0-9]+]]:_(p4) = COPY [[COPY7]]
-  ; GCN-NEXT:   [[COPY12:%[0-9]+]]:_(p4) = COPY [[COPY6]]
-  ; GCN-NEXT:   [[COPY13:%[0-9]+]]:_(p4) = COPY [[COPY5]]
-  ; GCN-NEXT:   [[COPY14:%[0-9]+]]:_(s64) = COPY [[COPY4]]
-  ; GCN-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY3]]
-  ; GCN-NEXT:   [[COPY16:%[0-9]+]]:_(s32) = COPY [[COPY2]]
-  ; GCN-NEXT:   [[COPY17:%[0-9]+]]:_(s32) = COPY [[COPY1]]
-  ; GCN-NEXT:   [[COPY18:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
+  ; GCN-NEXT:   [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY7]]
+  ; GCN-NEXT:   [[COPY11:%[0-9]+]]:_(p4) = COPY [[COPY6]]
+  ; GCN-NEXT:   [[COPY12:%[0-9]+]]:_(p4) = COPY [[COPY5]]
+  ; GCN-NEXT:   [[COPY13:%[0-9]+]]:_(s64) = COPY [[COPY4]]
+  ; GCN-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY3]]
+  ; GCN-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY2]]
+  ; GCN-NEXT:   [[COPY16:%[0-9]+]]:_(s32) = COPY [[COPY1]]
+  ; GCN-NEXT:   [[COPY17:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
   ; GCN-NEXT:   [[DEF1:%[0-9]+]]:_(<3 x s16>) = G_IMPLICIT_DEF
   ; GCN-NEXT:   [[CONCAT_VECTORS1:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[UV]](<3 x s16>), [[DEF1]](<3 x s16>)
   ; GCN-NEXT:   [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>), [[UV4:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS1]](<6 x s16>)
   ; GCN-NEXT:   $vgpr0 = COPY [[UV2]](<2 x s16>)
   ; GCN-NEXT:   $vgpr1 = COPY [[UV3]](<2 x s16>)
-  ; GCN-NEXT:   [[COPY19:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY19]](<4 x s32>)
-  ; GCN-NEXT:   $sgpr4_sgpr5 = COPY [[COPY11]](p4)
-  ; GCN-NEXT:   $sgpr6_sgpr7 = COPY [[COPY12]](p4)
-  ; GCN-NEXT:   $sgpr8_sgpr9 = COPY [[COPY13]](p4)
-  ; GCN-NEXT:   $sgpr10_sgpr11 = COPY [[COPY14]](s64)
-  ; GCN-NEXT:   $sgpr12 = COPY [[COPY15]](s32)
-  ; GCN-NEXT:   $sgpr13 = COPY [[COPY16]](s32)
-  ; GCN-NEXT:   $sgpr14 = COPY [[COPY17]](s32)
-  ; GCN-NEXT:   $vgpr31 = COPY [[COPY18]](s32)
+  ; GCN-NEXT:   [[COPY18:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY18]](<4 x s32>)
+  ; GCN-NEXT:   $sgpr4_sgpr5 = COPY [[COPY10]](p4)
+  ; GCN-NEXT:   $sgpr6_sgpr7 = COPY [[COPY11]](p4)
+  ; GCN-NEXT:   $sgpr8_sgpr9 = COPY [[COPY12]](p4)
+  ; GCN-NEXT:   $sgpr10_sgpr11 = COPY [[COPY13]](s64)
+  ; GCN-NEXT:   $sgpr12 = COPY [[COPY14]](s32)
+  ; GCN-NEXT:   $sgpr13 = COPY [[COPY15]](s32)
+  ; GCN-NEXT:   $sgpr14 = COPY [[COPY16]](s32)
+  ; GCN-NEXT:   $vgpr31 = COPY [[COPY17]](s32)
   ; GCN-NEXT:   SI_TCRETURN [[GV]](p0), @v3i16_fastcc_v3i16, 0, csr_amdgpu_highregs, implicit $vgpr0, implicit $vgpr1, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr4_sgpr5, implicit $sgpr6_sgpr7, implicit $sgpr8_sgpr9, implicit $sgpr10_sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $vgpr31
 entry:
   %ret = tail call fastcc <3 x i16> @v3i16_fastcc_v3i16(<3 x i16> %a)
   ret <3 x i16> %ret
 }
 
 declare hidden fastcc <4 x i16> @v4i16_fastcc_v4i16(<4 x i16> %arg0)
 
 define hidden fastcc <4 x i16> @sibling_call_v4i16_fastcc_v4i16(<4 x i16> %a) #1 {
   ; GCN-LABEL: name: sibling_call_v4i16_fastcc_v4i16
   ; GCN: bb.1.entry:
-  ; GCN-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr0, $vgpr1, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11, $sgpr30_sgpr31
+  ; GCN-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr0, $vgpr1, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT:   [[COPY:%[0-9]+]]:vgpr_32(s32) = COPY $vgpr31
   ; GCN-NEXT:   [[COPY1:%[0-9]+]]:sgpr_32 = COPY $sgpr14
   ; GCN-NEXT:   [[COPY2:%[0-9]+]]:sgpr_32 = COPY $sgpr13
   ; GCN-NEXT:   [[COPY3:%[0-9]+]]:sgpr_32 = COPY $sgpr12
   ; GCN-NEXT:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr10_sgpr11
   ; GCN-NEXT:   [[COPY5:%[0-9]+]]:sgpr_64 = COPY $sgpr8_sgpr9
   ; GCN-NEXT:   [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr6_sgpr7
   ; GCN-NEXT:   [[COPY7:%[0-9]+]]:sgpr_64 = COPY $sgpr4_sgpr5
   ; GCN-NEXT:   [[COPY8:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
   ; GCN-NEXT:   [[COPY9:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
   ; GCN-NEXT:   [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY8]](<2 x s16>), [[COPY9]](<2 x s16>)
-  ; GCN-NEXT:   [[COPY10:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   [[GV:%[0-9]+]]:sreg_64(p0) = G_GLOBAL_VALUE @v4i16_fastcc_v4i16
-  ; GCN-NEXT:   [[COPY11:%[0-9]+]]:_(p4) = COPY [[COPY7]]
-  ; GCN-NEXT:   [[COPY12:%[0-9]+]]:_(p4) = COPY [[COPY6]]
-  ; GCN-NEXT:   [[COPY13:%[0-9]+]]:_(p4) = COPY [[COPY5]]
-  ; GCN-NEXT:   [[COPY14:%[0-9]+]]:_(s64) = COPY [[COPY4]]
-  ; GCN-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY3]]
-  ; GCN-NEXT:   [[COPY16:%[0-9]+]]:_(s32) = COPY [[COPY2]]
-  ; GCN-NEXT:   [[COPY17:%[0-9]+]]:_(s32) = COPY [[COPY1]]
-  ; GCN-NEXT:   [[COPY18:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
+  ; GCN-NEXT:   [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY7]]
+  ; GCN-NEXT:   [[COPY11:%[0-9]+]]:_(p4) = COPY [[COPY6]]
+  ; GCN-NEXT:   [[COPY12:%[0-9]+]]:_(p4) = COPY [[COPY5]]
+  ; GCN-NEXT:   [[COPY13:%[0-9]+]]:_(s64) = COPY [[COPY4]]
+  ; GCN-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY3]]
+  ; GCN-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY2]]
+  ; GCN-NEXT:   [[COPY16:%[0-9]+]]:_(s32) = COPY [[COPY1]]
+  ; GCN-NEXT:   [[COPY17:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
   ; GCN-NEXT:   [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<4 x s16>)
   ; GCN-NEXT:   $vgpr0 = COPY [[UV]](<2 x s16>)
   ; GCN-NEXT:   $vgpr1 = COPY [[UV1]](<2 x s16>)
-  ; GCN-NEXT:   [[COPY19:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY19]](<4 x s32>)
-  ; GCN-NEXT:   $sgpr4_sgpr5 = COPY [[COPY11]](p4)
-  ; GCN-NEXT:   $sgpr6_sgpr7 = COPY [[COPY12]](p4)
-  ; GCN-NEXT:   $sgpr8_sgpr9 = COPY [[COPY13]](p4)
-  ; GCN-NEXT:   $sgpr10_sgpr11 = COPY [[COPY14]](s64)
-  ; GCN-NEXT:   $sgpr12 = COPY [[COPY15]](s32)
-  ; GCN-NEXT:   $sgpr13 = COPY [[COPY16]](s32)
-  ; GCN-NEXT:   $sgpr14 = COPY [[COPY17]](s32)
-  ; GCN-NEXT:   $vgpr31 = COPY [[COPY18]](s32)
+  ; GCN-NEXT:   [[COPY18:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY18]](<4 x s32>)
+  ; GCN-NEXT:   $sgpr4_sgpr5 = COPY [[COPY10]](p4)
+  ; GCN-NEXT:   $sgpr6_sgpr7 = COPY [[COPY11]](p4)
+  ; GCN-NEXT:   $sgpr8_sgpr9 = COPY [[COPY12]](p4)
+  ; GCN-NEXT:   $sgpr10_sgpr11 = COPY [[COPY13]](s64)
+  ; GCN-NEXT:   $sgpr12 = COPY [[COPY14]](s32)
+  ; GCN-NEXT:   $sgpr13 = COPY [[COPY15]](s32)
+  ; GCN-NEXT:   $sgpr14 = COPY [[COPY16]](s32)
+  ; GCN-NEXT:   $vgpr31 = COPY [[COPY17]](s32)
   ; GCN-NEXT:   SI_TCRETURN [[GV]](p0), @v4i16_fastcc_v4i16, 0, csr_amdgpu_highregs, implicit $vgpr0, implicit $vgpr1, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr4_sgpr5, implicit $sgpr6_sgpr7, implicit $sgpr8_sgpr9, implicit $sgpr10_sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $vgpr31
 entry:
   %ret = tail call fastcc <4 x i16> @v4i16_fastcc_v4i16(<4 x i16> %a)
   ret <4 x i16> %ret
 }
 
 declare hidden fastcc <2 x i64> @v2i64_fastcc_v2i64(<2 x i64> %arg0)
 
 define hidden fastcc <2 x i64> @sibling_call_v2i64_fastcc_v2i64(<2 x i64> %a) #1 {
   ; GCN-LABEL: name: sibling_call_v2i64_fastcc_v2i64
   ; GCN: bb.1.entry:
-  ; GCN-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11, $sgpr30_sgpr31
+  ; GCN-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT:   [[COPY:%[0-9]+]]:vgpr_32(s32) = COPY $vgpr31
   ; GCN-NEXT:   [[COPY1:%[0-9]+]]:sgpr_32 = COPY $sgpr14
   ; GCN-NEXT:   [[COPY2:%[0-9]+]]:sgpr_32 = COPY $sgpr13
   ; GCN-NEXT:   [[COPY3:%[0-9]+]]:sgpr_32 = COPY $sgpr12
   ; GCN-NEXT:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr10_sgpr11
   ; GCN-NEXT:   [[COPY5:%[0-9]+]]:sgpr_64 = COPY $sgpr8_sgpr9
   ; GCN-NEXT:   [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr6_sgpr7
   ; GCN-NEXT:   [[COPY7:%[0-9]+]]:sgpr_64 = COPY $sgpr4_sgpr5
   ; GCN-NEXT:   [[COPY8:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; GCN-NEXT:   [[COPY9:%[0-9]+]]:_(s32) = COPY $vgpr1
   ; GCN-NEXT:   [[COPY10:%[0-9]+]]:_(s32) = COPY $vgpr2
   ; GCN-NEXT:   [[COPY11:%[0-9]+]]:_(s32) = COPY $vgpr3
   ; GCN-NEXT:   [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY8]](s32), [[COPY9]](s32)
   ; GCN-NEXT:   [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[COPY10]](s32), [[COPY11]](s32)
   ; GCN-NEXT:   [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s64>) = G_BUILD_VECTOR [[MV]](s64), [[MV1]](s64)
-  ; GCN-NEXT:   [[COPY12:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GCN-NEXT:   [[GV:%[0-9]+]]:sreg_64(p0) = G_GLOBAL_VALUE @v2i64_fastcc_v2i64
-  ; GCN-NEXT:   [[COPY13:%[0-9]+]]:_(p4) = COPY [[COPY7]]
-  ; GCN-NEXT:   [[COPY14:%[0-9]+]]:_(p4) = COPY [[COPY6]]
-  ; GCN-NEXT:   [[COPY15:%[0-9]+]]:_(p4) = COPY [[COPY5]]
-  ; GCN-NEXT:   [[COPY16:%[0-9]+]]:_(s64) = COPY [[COPY4]]
-  ; GCN-NEXT:   [[COPY17:%[0-9]+]]:_(s32) = COPY [[COPY3]]
-  ; GCN-NEXT:   [[COPY18:%[0-9]+]]:_(s32) = COPY [[COPY2]]
-  ; GCN-NEXT:   [[COPY19:%[0-9]+]]:_(s32) = COPY [[COPY1]]
-  ; GCN-NEXT:   [[COPY20:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
+  ; GCN-NEXT:   [[COPY12:%[0-9]+]]:_(p4) = COPY [[COPY7]]
+  ; GCN-NEXT:   [[COPY13:%[0-9]+]]:_(p4) = COPY [[COPY6]]
+  ; GCN-NEXT:   [[COPY14:%[0-9]+]]:_(p4) = COPY [[COPY5]]
+  ; GCN-NEXT:   [[COPY15:%[0-9]+]]:_(s64) = COPY [[COPY4]]
+  ; GCN-NEXT:   [[COPY16:%[0-9]+]]:_(s32) = COPY [[COPY3]]
+  ; GCN-NEXT:   [[COPY17:%[0-9]+]]:_(s32) = COPY [[COPY2]]
+  ; GCN-NEXT:   [[COPY18:%[0-9]+]]:_(s32) = COPY [[COPY1]]
+  ; GCN-NEXT:   [[COPY19:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
   ; GCN-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32), [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[BUILD_VECTOR]](<2 x s64>)
   ; GCN-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; GCN-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; GCN-NEXT:   $vgpr2 = COPY [[UV2]](s32)
   ; GCN-NEXT:   $vgpr3 = COPY [[UV3]](s32)
-  ; GCN-NEXT:   [[COPY21:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY21]](<4 x s32>)
-  ; GCN-NEXT:   $sgpr4_sgpr5 = COPY [[COPY13]](p4)
-  ; GCN-NEXT:   $sgpr6_sgpr7 = COPY [[COPY14]](p4)
-  ; GCN-NEXT:   $sgpr8_sgpr9 = COPY [[COPY15]](p4)
-  ; GCN-NEXT:   $sgpr10_sgpr11 = COPY [[COPY16]](s64)
-  ; GCN-NEXT:   $sgpr12 = COPY [[COPY17]](s32)
-  ; GCN-NEXT:   $sgpr13 = COPY [[COPY18]](s32)
-  ; GCN-NEXT:   $sgpr14 = COPY [[COPY19]](s32)
-  ; GCN-NEXT:   $vgpr31 = COPY [[COPY20]](s32)
+  ; GCN-NEXT:   [[COPY20:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; GCN-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY20]](<4 x s32>)
+  ; GCN-NEXT:   $sgpr4_sgpr5 = COPY [[COPY12]](p4)
+  ; GCN-NEXT:   $sgpr6_sgpr7 = COPY [[COPY13]](p4)
+  ; GCN-NEXT:   $sgpr8_sgpr9 = COPY [[COPY14]](p4)
+  ; GCN-NEXT:   $sgpr10_sgpr11 = COPY [[COPY15]](s64)
+  ; GCN-NEXT:   $sgpr12 = COPY [[COPY16]](s32)
+  ; GCN-NEXT:   $sgpr13 = COPY [[COPY17]](s32)
+  ; GCN-NEXT:   $sgpr14 = COPY [[COPY18]](s32)
+  ; GCN-NEXT:   $vgpr31 = COPY [[COPY19]](s32)
   ; GCN-NEXT:   SI_TCRETURN [[GV]](p0), @v2i64_fastcc_v2i64, 0, csr_amdgpu_highregs, implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr4_sgpr5, implicit $sgpr6_sgpr7, implicit $sgpr8_sgpr9, implicit $sgpr10_sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $vgpr31
 entry:
   %ret = tail call fastcc <2 x i64> @v2i64_fastcc_v2i64(<2 x i64> %a)
   ret <2 x i64> %ret
 }
 
 attributes #0 = { nounwind }
 attributes #1 = { nounwind noinline }

llvm/test/CodeGen/AMDGPU/GlobalISel/irtranslator-tail-call.ll

 ; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
 ; RUN: llc -global-isel -amdgpu-fixed-function-abi -stop-after=irtranslator -mtriple=amdgcn-mesa-mesa3d -mcpu=gfx900 -verify-machineinstrs -o - %s | FileCheck -enable-var-scope %s
 
 declare hidden void @external_void_func_void()
 
 define void @tail_call_void_func_void() {
   ; CHECK-LABEL: name: tail_call_void_func_void
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11, $sgpr30_sgpr31
-  ; CHECK:   [[COPY:%[0-9]+]]:vgpr_32(s32) = COPY $vgpr31
-  ; CHECK:   [[COPY1:%[0-9]+]]:sgpr_32 = COPY $sgpr14
-  ; CHECK:   [[COPY2:%[0-9]+]]:sgpr_32 = COPY $sgpr13
-  ; CHECK:   [[COPY3:%[0-9]+]]:sgpr_32 = COPY $sgpr12
-  ; CHECK:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr10_sgpr11
-  ; CHECK:   [[COPY5:%[0-9]+]]:sgpr_64 = COPY $sgpr8_sgpr9
-  ; CHECK:   [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr6_sgpr7
-  ; CHECK:   [[COPY7:%[0-9]+]]:sgpr_64 = COPY $sgpr4_sgpr5
-  ; CHECK:   [[COPY8:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
-  ; CHECK:   [[GV:%[0-9]+]]:sreg_64(p0) = G_GLOBAL_VALUE @external_void_func_void
-  ; CHECK:   [[COPY9:%[0-9]+]]:_(p4) = COPY [[COPY7]]
-  ; CHECK:   [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY6]]
-  ; CHECK:   [[COPY11:%[0-9]+]]:_(p4) = COPY [[COPY5]]
-  ; CHECK:   [[COPY12:%[0-9]+]]:_(s64) = COPY [[COPY4]]
-  ; CHECK:   [[COPY13:%[0-9]+]]:_(s32) = COPY [[COPY3]]
-  ; CHECK:   [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY2]]
-  ; CHECK:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY1]]
-  ; CHECK:   [[COPY16:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
-  ; CHECK:   [[COPY17:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
-  ; CHECK:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY17]](<4 x s32>)
-  ; CHECK:   $sgpr4_sgpr5 = COPY [[COPY9]](p4)
-  ; CHECK:   $sgpr6_sgpr7 = COPY [[COPY10]](p4)
-  ; CHECK:   $sgpr8_sgpr9 = COPY [[COPY11]](p4)
-  ; CHECK:   $sgpr10_sgpr11 = COPY [[COPY12]](s64)
-  ; CHECK:   $sgpr12 = COPY [[COPY13]](s32)
-  ; CHECK:   $sgpr13 = COPY [[COPY14]](s32)
-  ; CHECK:   $sgpr14 = COPY [[COPY15]](s32)
-  ; CHECK:   $vgpr31 = COPY [[COPY16]](s32)
-  ; CHECK:   SI_TCRETURN [[GV]](p0), @external_void_func_void, 0, csr_amdgpu_highregs, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr4_sgpr5, implicit $sgpr6_sgpr7, implicit $sgpr8_sgpr9, implicit $sgpr10_sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $vgpr31
+  ; CHECK-NEXT:   liveins: $sgpr12, $sgpr13, $sgpr14, $vgpr31, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:vgpr_32(s32) = COPY $vgpr31
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sgpr_32 = COPY $sgpr14
+  ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sgpr_32 = COPY $sgpr13
+  ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sgpr_32 = COPY $sgpr12
+  ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:sgpr_64 = COPY $sgpr10_sgpr11
+  ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:sgpr_64 = COPY $sgpr8_sgpr9
+  ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr6_sgpr7
+  ; CHECK-NEXT:   [[COPY7:%[0-9]+]]:sgpr_64 = COPY $sgpr4_sgpr5
+  ; CHECK-NEXT:   [[GV:%[0-9]+]]:sreg_64(p0) = G_GLOBAL_VALUE @external_void_func_void
+  ; CHECK-NEXT:   [[COPY8:%[0-9]+]]:_(p4) = COPY [[COPY7]]
+  ; CHECK-NEXT:   [[COPY9:%[0-9]+]]:_(p4) = COPY [[COPY6]]
+  ; CHECK-NEXT:   [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY5]]
+  ; CHECK-NEXT:   [[COPY11:%[0-9]+]]:_(s64) = COPY [[COPY4]]
+  ; CHECK-NEXT:   [[COPY12:%[0-9]+]]:_(s32) = COPY [[COPY3]]
+  ; CHECK-NEXT:   [[COPY13:%[0-9]+]]:_(s32) = COPY [[COPY2]]
+  ; CHECK-NEXT:   [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY1]]
+  ; CHECK-NEXT:   [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
+  ; CHECK-NEXT:   [[COPY16:%[0-9]+]]:_(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
+  ; CHECK-NEXT:   $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY16]](<4 x s32>)
+  ; CHECK-NEXT:   $sgpr4_sgpr5 = COPY [[COPY8]](p4)
+  ; CHECK-NEXT:   $sgpr6_sgpr7 = COPY [[COPY9]](p4)
+  ; CHECK-NEXT:   $sgpr8_sgpr9 = COPY [[COPY10]](p4)
+  ; CHECK-NEXT:   $sgpr10_sgpr11 = COPY [[COPY11]](s64)
+  ; CHECK-NEXT:   $sgpr12 = COPY [[COPY12]](s32)
+  ; CHECK-NEXT:   $sgpr13 = COPY [[COPY13]](s32)
+  ; CHECK-NEXT:   $sgpr14 = COPY [[COPY14]](s32)
+  ; CHECK-NEXT:   $vgpr31 = COPY [[COPY15]](s32)
+  ; CHECK-NEXT:   SI_TCRETURN [[GV]](p0), @external_void_func_void, 0, csr_amdgpu_highregs, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr4_sgpr5, implicit $sgpr6_sgpr7, implicit $sgpr8_sgpr9, implicit $sgpr10_sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $vgpr31
   tail call void @external_void_func_void()
   ret void
 }

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.ds.fmax.ll

 ; GFX9-LABEL: ds_fmax_f32_vv:
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    ds_max_rtn_f32 v0, v0, v1
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
   ; GFX8-MIR-LABEL: name: ds_fmax_f32_vv
   ; GFX8-MIR: bb.1 (%ir-block.0):
-  ; GFX8-MIR:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+  ; GFX8-MIR:   liveins: $vgpr0, $vgpr1
   ; GFX8-MIR:   [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX8-MIR:   [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
-  ; GFX8-MIR:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GFX8-MIR:   $m0 = S_MOV_B32 -1
   ; GFX8-MIR:   [[DS_MAX_RTN_F32_:%[0-9]+]]:vgpr_32 = DS_MAX_RTN_F32 [[COPY]], [[COPY1]], 0, 0, implicit $m0, implicit $exec :: (load store (s32) on %ir.ptr, addrspace 3)
   ; GFX8-MIR:   $vgpr0 = COPY [[DS_MAX_RTN_F32_]]
-  ; GFX8-MIR:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; GFX8-MIR:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; GFX8-MIR:   SI_RETURN implicit $vgpr0
   ; GFX9-MIR-LABEL: name: ds_fmax_f32_vv
   ; GFX9-MIR: bb.1 (%ir-block.0):
-  ; GFX9-MIR:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+  ; GFX9-MIR:   liveins: $vgpr0, $vgpr1
   ; GFX9-MIR:   [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX9-MIR:   [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
-  ; GFX9-MIR:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GFX9-MIR:   [[DS_MAX_RTN_F32_gfx9_:%[0-9]+]]:vgpr_32 = DS_MAX_RTN_F32_gfx9 [[COPY]], [[COPY1]], 0, 0, implicit $exec :: (load store (s32) on %ir.ptr, addrspace 3)
   ; GFX9-MIR:   $vgpr0 = COPY [[DS_MAX_RTN_F32_gfx9_]]
-  ; GFX9-MIR:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; GFX9-MIR:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; GFX9-MIR:   SI_RETURN implicit $vgpr0
   %ret = call float @llvm.amdgcn.ds.fmax(float addrspace(3)* %ptr, float %val, i32 0, i32 0, i1 false)
   ret float %ret
 }
 
 define float @ds_fmax_f32_vv_offset(float addrspace(3)* %ptr, float %val) {
 ; GFX8-LABEL: ds_fmax_f32_vv_offset:
 ; GFX8:       ; %bb.0:
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX8-NEXT:    s_mov_b32 m0, -1
 ; GFX8-NEXT:    ds_max_rtn_f32 v0, v0, v1 offset:512
 ; GFX8-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX8-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX9-LABEL: ds_fmax_f32_vv_offset:
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    ds_max_rtn_f32 v0, v0, v1 offset:512
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
   ; GFX8-MIR-LABEL: name: ds_fmax_f32_vv_offset
   ; GFX8-MIR: bb.1 (%ir-block.0):
-  ; GFX8-MIR:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+  ; GFX8-MIR:   liveins: $vgpr0, $vgpr1
   ; GFX8-MIR:   [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX8-MIR:   [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
-  ; GFX8-MIR:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GFX8-MIR:   $m0 = S_MOV_B32 -1
   ; GFX8-MIR:   [[DS_MAX_RTN_F32_:%[0-9]+]]:vgpr_32 = DS_MAX_RTN_F32 [[COPY]], [[COPY1]], 512, 0, implicit $m0, implicit $exec :: (load store (s32) on %ir.gep, addrspace 3)
   ; GFX8-MIR:   $vgpr0 = COPY [[DS_MAX_RTN_F32_]]
-  ; GFX8-MIR:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; GFX8-MIR:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; GFX8-MIR:   SI_RETURN implicit $vgpr0
   ; GFX9-MIR-LABEL: name: ds_fmax_f32_vv_offset
   ; GFX9-MIR: bb.1 (%ir-block.0):
-  ; GFX9-MIR:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+  ; GFX9-MIR:   liveins: $vgpr0, $vgpr1
   ; GFX9-MIR:   [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX9-MIR:   [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
-  ; GFX9-MIR:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GFX9-MIR:   [[DS_MAX_RTN_F32_gfx9_:%[0-9]+]]:vgpr_32 = DS_MAX_RTN_F32_gfx9 [[COPY]], [[COPY1]], 512, 0, implicit $exec :: (load store (s32) on %ir.gep, addrspace 3)
   ; GFX9-MIR:   $vgpr0 = COPY [[DS_MAX_RTN_F32_gfx9_]]
-  ; GFX9-MIR:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; GFX9-MIR:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; GFX9-MIR:   SI_RETURN implicit $vgpr0
   %gep = getelementptr float, float addrspace(3)* %ptr, i32 128
   %ret = call float @llvm.amdgcn.ds.fmax(float addrspace(3)* %gep, float %val, i32 0, i32 0, i1 false)
   ret float %ret
 }
 
 define void @ds_fmax_f32_vv_nortn(float addrspace(3)* %ptr, float %val) {
 ; GFX8-LABEL: ds_fmax_f32_vv_nortn:
 ; GFX8:       ; %bb.0:
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX8-NEXT:    s_mov_b32 m0, -1
 ; GFX8-NEXT:    ds_max_rtn_f32 v0, v0, v1
 ; GFX8-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX8-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX9-LABEL: ds_fmax_f32_vv_nortn:
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    ds_max_rtn_f32 v0, v0, v1
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
   ; GFX8-MIR-LABEL: name: ds_fmax_f32_vv_nortn
   ; GFX8-MIR: bb.1 (%ir-block.0):
-  ; GFX8-MIR:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+  ; GFX8-MIR:   liveins: $vgpr0, $vgpr1
   ; GFX8-MIR:   [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX8-MIR:   [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
-  ; GFX8-MIR:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GFX8-MIR:   $m0 = S_MOV_B32 -1
   ; GFX8-MIR:   [[DS_MAX_RTN_F32_:%[0-9]+]]:vgpr_32 = DS_MAX_RTN_F32 [[COPY]], [[COPY1]], 0, 0, implicit $m0, implicit $exec :: (load store (s32) on %ir.ptr, addrspace 3)
-  ; GFX8-MIR:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; GFX8-MIR:   S_SETPC_B64_return [[COPY3]]
+  ; GFX8-MIR:   SI_RETURN
   ; GFX9-MIR-LABEL: name: ds_fmax_f32_vv_nortn
   ; GFX9-MIR: bb.1 (%ir-block.0):
-  ; GFX9-MIR:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+  ; GFX9-MIR:   liveins: $vgpr0, $vgpr1
   ; GFX9-MIR:   [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX9-MIR:   [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
-  ; GFX9-MIR:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GFX9-MIR:   [[DS_MAX_RTN_F32_gfx9_:%[0-9]+]]:vgpr_32 = DS_MAX_RTN_F32_gfx9 [[COPY]], [[COPY1]], 0, 0, implicit $exec :: (load store (s32) on %ir.ptr, addrspace 3)
-  ; GFX9-MIR:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; GFX9-MIR:   S_SETPC_B64_return [[COPY3]]
+  ; GFX9-MIR:   SI_RETURN
   %ret = call float @llvm.amdgcn.ds.fmax(float addrspace(3)* %ptr, float %val, i32 0, i32 0, i1 false)
   ret void
 }
 
 define void @ds_fmax_f32_vv_offset_nortn(float addrspace(3)* %ptr, float %val) {
 ; GFX8-LABEL: ds_fmax_f32_vv_offset_nortn:
 ; GFX8:       ; %bb.0:
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX8-NEXT:    s_mov_b32 m0, -1
 ; GFX8-NEXT:    ds_max_rtn_f32 v0, v0, v1 offset:512
 ; GFX8-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX8-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX9-LABEL: ds_fmax_f32_vv_offset_nortn:
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    ds_max_rtn_f32 v0, v0, v1 offset:512
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
   ; GFX8-MIR-LABEL: name: ds_fmax_f32_vv_offset_nortn
   ; GFX8-MIR: bb.1 (%ir-block.0):
-  ; GFX8-MIR:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+  ; GFX8-MIR:   liveins: $vgpr0, $vgpr1
   ; GFX8-MIR:   [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX8-MIR:   [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
-  ; GFX8-MIR:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GFX8-MIR:   $m0 = S_MOV_B32 -1
   ; GFX8-MIR:   [[DS_MAX_RTN_F32_:%[0-9]+]]:vgpr_32 = DS_MAX_RTN_F32 [[COPY]], [[COPY1]], 512, 0, implicit $m0, implicit $exec :: (load store (s32) on %ir.gep, addrspace 3)
-  ; GFX8-MIR:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; GFX8-MIR:   S_SETPC_B64_return [[COPY3]]
+  ; GFX8-MIR:   SI_RETURN
   ; GFX9-MIR-LABEL: name: ds_fmax_f32_vv_offset_nortn
   ; GFX9-MIR: bb.1 (%ir-block.0):
-  ; GFX9-MIR:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+  ; GFX9-MIR:   liveins: $vgpr0, $vgpr1
   ; GFX9-MIR:   [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX9-MIR:   [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
-  ; GFX9-MIR:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GFX9-MIR:   [[DS_MAX_RTN_F32_gfx9_:%[0-9]+]]:vgpr_32 = DS_MAX_RTN_F32_gfx9 [[COPY]], [[COPY1]], 512, 0, implicit $exec :: (load store (s32) on %ir.gep, addrspace 3)
-  ; GFX9-MIR:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; GFX9-MIR:   S_SETPC_B64_return [[COPY3]]
+  ; GFX9-MIR:   SI_RETURN
   %gep = getelementptr float, float addrspace(3)* %ptr, i32 128
   %ret = call float @llvm.amdgcn.ds.fmax(float addrspace(3)* %gep, float %val, i32 0, i32 0, i1 false)
   ret void
 }
 
 define float @ds_fmax_f32_vv_volatile(float addrspace(3)* %ptr, float %val) {
 ; GFX8-LABEL: ds_fmax_f32_vv_volatile:
 ; GFX8:       ; %bb.0:
 ; GFX8-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX8-NEXT:    s_mov_b32 m0, -1
 ; GFX8-NEXT:    ds_max_rtn_f32 v0, v0, v1
 ; GFX8-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX8-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX9-LABEL: ds_fmax_f32_vv_volatile:
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    ds_max_rtn_f32 v0, v0, v1
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
   ; GFX8-MIR-LABEL: name: ds_fmax_f32_vv_volatile
   ; GFX8-MIR: bb.1 (%ir-block.0):
-  ; GFX8-MIR:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+  ; GFX8-MIR:   liveins: $vgpr0, $vgpr1
   ; GFX8-MIR:   [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX8-MIR:   [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
-  ; GFX8-MIR:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GFX8-MIR:   $m0 = S_MOV_B32 -1
   ; GFX8-MIR:   [[DS_MAX_RTN_F32_:%[0-9]+]]:vgpr_32 = DS_MAX_RTN_F32 [[COPY]], [[COPY1]], 0, 0, implicit $m0, implicit $exec :: (volatile load store (s32) on %ir.ptr, addrspace 3)
   ; GFX8-MIR:   $vgpr0 = COPY [[DS_MAX_RTN_F32_]]
-  ; GFX8-MIR:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; GFX8-MIR:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; GFX8-MIR:   SI_RETURN implicit $vgpr0
   ; GFX9-MIR-LABEL: name: ds_fmax_f32_vv_volatile
   ; GFX9-MIR: bb.1 (%ir-block.0):
-  ; GFX9-MIR:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+  ; GFX9-MIR:   liveins: $vgpr0, $vgpr1
   ; GFX9-MIR:   [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
   ; GFX9-MIR:   [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
-  ; GFX9-MIR:   [[COPY2:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
   ; GFX9-MIR:   [[DS_MAX_RTN_F32_gfx9_:%[0-9]+]]:vgpr_32 = DS_MAX_RTN_F32_gfx9 [[COPY]], [[COPY1]], 0, 0, implicit $exec :: (volatile load store (s32) on %ir.ptr, addrspace 3)
   ; GFX9-MIR:   $vgpr0 = COPY [[DS_MAX_RTN_F32_gfx9_]]
-  ; GFX9-MIR:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY2]]
-  ; GFX9-MIR:   S_SETPC_B64_return [[COPY3]], implicit $vgpr0
+  ; GFX9-MIR:   SI_RETURN implicit $vgpr0
   %ret = call float @llvm.amdgcn.ds.fmax(float addrspace(3)* %ptr, float %val, i32 0, i32 0, i1 true)
   ret float %ret
 }
 
 declare float @llvm.amdgcn.ds.fmax(float addrspace(3)* nocapture, float, i32 immarg, i32 immarg, i1 immarg) #0
 
 attributes #0 = { argmemonly nounwind willreturn }

llvm/test/CodeGen/AMDGPU/GlobalISel/load-unaligned.ll

 ; GFX10-LABEL: store_lds_v3i32_align1:
 ; GFX10:       ; %bb.0:
 ; GFX10-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v4, 8, v1
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v5, 16, v1
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v6, 24, v1
 ; GFX10-NEXT:    ds_write_b8 v0, v1
+; GFX10-NEXT:    v_lshrrev_b32_e32 v1, 8, v2
 ; GFX10-NEXT:    ds_write_b8 v0, v4 offset:1
 ; GFX10-NEXT:    ds_write_b8 v0, v5 offset:2
-; GFX10-NEXT:    v_lshrrev_b32_e32 v1, 8, v2
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v4, 16, v2
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v5, 24, v2
 ; GFX10-NEXT:    ds_write_b8 v0, v6 offset:3
 ; GFX10-NEXT:    ds_write_b8 v0, v2 offset:4
 ; GFX10-NEXT:    ds_write_b8 v0, v1 offset:5
 ; GFX10-NEXT:    ds_write_b8 v0, v4 offset:6
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v1, 8, v3
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v2, 16, v3

llvm/test/CodeGen/AMDGPU/GlobalISel/regbankcombiner-smed3.mir

 # NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
 # RUN: llc -mtriple=amdgcn-amd-mesa3d -mcpu=gfx1010 -run-pass=amdgpu-regbank-combiner -verify-machineinstrs %s -o - | FileCheck %s
 
 ---
 name: test_min_max_ValK0_K1_i32
 legalized: true
 regBankSelected: true
 tracksRegLiveness: true
 body: |
   bb.1:
-    liveins: $vgpr0, $sgpr30_sgpr31
+    liveins: $vgpr0
 
     ; CHECK-LABEL: name: test_min_max_ValK0_K1_i32
-    ; CHECK: liveins: $vgpr0, $sgpr30_sgpr31
+    ; CHECK: liveins: $vgpr0
     ; CHECK: [[COPY:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
-    ; CHECK: [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; CHECK: [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 -12
     ; CHECK: [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 17
     ; CHECK: [[AMDGPU_SMED3_:%[0-9]+]]:vgpr(s32) = G_AMDGPU_SMED3 [[COPY]], [[C]], [[C1]]
     ; CHECK: $vgpr0 = COPY [[AMDGPU_SMED3_]](s32)
-    ; CHECK: [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-    ; CHECK: S_SETPC_B64_return [[COPY2]], implicit $vgpr0
+    ; CHECK: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     %0:vgpr(s32) = COPY $vgpr0
-    %1:sgpr_64 = COPY $sgpr30_sgpr31
     %2:sgpr(s32) = G_CONSTANT i32 -12
     %7:vgpr(s32) = COPY %2(s32)
     %3:vgpr(s32) = G_SMAX %0, %7
     %4:sgpr(s32) = G_CONSTANT i32 17
     %8:vgpr(s32) = COPY %4(s32)
     %5:vgpr(s32) = G_SMIN %3, %8
     $vgpr0 = COPY %5(s32)
-    %6:ccr_sgpr_64 = COPY %1
-    S_SETPC_B64_return %6, implicit $vgpr0
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
 ...
 
 ---
 name: min_max_ValK0_K1_i32
 legalized: true
 regBankSelected: true
 tracksRegLiveness: true
 body: |
   bb.1:
-    liveins: $vgpr0, $sgpr30_sgpr31
+    liveins: $vgpr0
 
     ; CHECK-LABEL: name: min_max_ValK0_K1_i32
-    ; CHECK: liveins: $vgpr0, $sgpr30_sgpr31
+    ; CHECK: liveins: $vgpr0
     ; CHECK: [[COPY:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
-    ; CHECK: [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; CHECK: [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 -12
     ; CHECK: [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 17
     ; CHECK: [[AMDGPU_SMED3_:%[0-9]+]]:vgpr(s32) = G_AMDGPU_SMED3 [[COPY]], [[C]], [[C1]]
     ; CHECK: $vgpr0 = COPY [[AMDGPU_SMED3_]](s32)
-    ; CHECK: [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-    ; CHECK: S_SETPC_B64_return [[COPY2]], implicit $vgpr0
+    ; CHECK: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     %0:vgpr(s32) = COPY $vgpr0
-    %1:sgpr_64 = COPY $sgpr30_sgpr31
     %2:sgpr(s32) = G_CONSTANT i32 -12
     %7:vgpr(s32) = COPY %2(s32)
     %3:vgpr(s32) = G_SMAX %7, %0
     %4:sgpr(s32) = G_CONSTANT i32 17
     %8:vgpr(s32) = COPY %4(s32)
     %5:vgpr(s32) = G_SMIN %3, %8
     $vgpr0 = COPY %5(s32)
-    %6:ccr_sgpr_64 = COPY %1
-    S_SETPC_B64_return %6, implicit $vgpr0
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
 ...
 
 ---
 name: test_min_K1max_ValK0__i32
 legalized: true
 regBankSelected: true
 tracksRegLiveness: true
 body: |
   bb.1:
-    liveins: $vgpr0, $sgpr30_sgpr31
+    liveins: $vgpr0
 
     ; CHECK-LABEL: name: test_min_K1max_ValK0__i32
-    ; CHECK: liveins: $vgpr0, $sgpr30_sgpr31
+    ; CHECK: liveins: $vgpr0
     ; CHECK: [[COPY:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
-    ; CHECK: [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; CHECK: [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 -12
     ; CHECK: [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 17
     ; CHECK: [[AMDGPU_SMED3_:%[0-9]+]]:vgpr(s32) = G_AMDGPU_SMED3 [[COPY]], [[C]], [[C1]]
     ; CHECK: $vgpr0 = COPY [[AMDGPU_SMED3_]](s32)
-    ; CHECK: [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-    ; CHECK: S_SETPC_B64_return [[COPY2]], implicit $vgpr0
+    ; CHECK: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     %0:vgpr(s32) = COPY $vgpr0
-    %1:sgpr_64 = COPY $sgpr30_sgpr31
     %2:sgpr(s32) = G_CONSTANT i32 -12
     %7:vgpr(s32) = COPY %2(s32)
     %3:vgpr(s32) = G_SMAX %0, %7
     %4:sgpr(s32) = G_CONSTANT i32 17
     %8:vgpr(s32) = COPY %4(s32)
     %5:vgpr(s32) = G_SMIN %8, %3
     $vgpr0 = COPY %5(s32)
-    %6:ccr_sgpr_64 = COPY %1
-    S_SETPC_B64_return %6, implicit $vgpr0
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
 ...
 
 ---
 name: test_min_K1max_K0Val__i32
 legalized: true
 regBankSelected: true
 tracksRegLiveness: true
 body: |
   bb.1:
-    liveins: $vgpr0, $sgpr30_sgpr31
+    liveins: $vgpr0
 
     ; CHECK-LABEL: name: test_min_K1max_K0Val__i32
-    ; CHECK: liveins: $vgpr0, $sgpr30_sgpr31
+    ; CHECK: liveins: $vgpr0
     ; CHECK: [[COPY:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
-    ; CHECK: [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; CHECK: [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 -12
     ; CHECK: [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 17
     ; CHECK: [[AMDGPU_SMED3_:%[0-9]+]]:vgpr(s32) = G_AMDGPU_SMED3 [[COPY]], [[C]], [[C1]]
     ; CHECK: $vgpr0 = COPY [[AMDGPU_SMED3_]](s32)
-    ; CHECK: [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-    ; CHECK: S_SETPC_B64_return [[COPY2]], implicit $vgpr0
+    ; CHECK: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     %0:vgpr(s32) = COPY $vgpr0
-    %1:sgpr_64 = COPY $sgpr30_sgpr31
     %2:sgpr(s32) = G_CONSTANT i32 -12
     %7:vgpr(s32) = COPY %2(s32)
     %3:vgpr(s32) = G_SMAX %7, %0
     %4:sgpr(s32) = G_CONSTANT i32 17
     %8:vgpr(s32) = COPY %4(s32)
     %5:vgpr(s32) = G_SMIN %8, %3
     $vgpr0 = COPY %5(s32)
-    %6:ccr_sgpr_64 = COPY %1
-    S_SETPC_B64_return %6, implicit $vgpr0
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
 ...
 
 ---
 name: test_max_min_ValK1_K0_i32
 legalized: true
 regBankSelected: true
 tracksRegLiveness: true
 body: |
   bb.1:
-    liveins: $vgpr0, $sgpr30_sgpr31
+    liveins: $vgpr0
 
     ; CHECK-LABEL: name: test_max_min_ValK1_K0_i32
-    ; CHECK: liveins: $vgpr0, $sgpr30_sgpr31
+    ; CHECK: liveins: $vgpr0
     ; CHECK: [[COPY:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
-    ; CHECK: [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; CHECK: [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 17
     ; CHECK: [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 -12
     ; CHECK: [[AMDGPU_SMED3_:%[0-9]+]]:vgpr(s32) = G_AMDGPU_SMED3 [[COPY]], [[C1]], [[C]]
     ; CHECK: $vgpr0 = COPY [[AMDGPU_SMED3_]](s32)
-    ; CHECK: [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-    ; CHECK: S_SETPC_B64_return [[COPY2]], implicit $vgpr0
+    ; CHECK: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     %0:vgpr(s32) = COPY $vgpr0
-    %1:sgpr_64 = COPY $sgpr30_sgpr31
     %2:sgpr(s32) = G_CONSTANT i32 17
     %7:vgpr(s32) = COPY %2(s32)
     %3:vgpr(s32) = G_SMIN %0, %7
     %4:sgpr(s32) = G_CONSTANT i32 -12
     %8:vgpr(s32) = COPY %4(s32)
     %5:vgpr(s32) = G_SMAX %3, %8
     $vgpr0 = COPY %5(s32)
-    %6:ccr_sgpr_64 = COPY %1
-    S_SETPC_B64_return %6, implicit $vgpr0
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
 ...
 
 ---
 name: test_max_min_K1Val_K0_i32
 legalized: true
 regBankSelected: true
 tracksRegLiveness: true
 body: |
   bb.1:
-    liveins: $vgpr0, $sgpr30_sgpr31
+    liveins: $vgpr0
 
     ; CHECK-LABEL: name: test_max_min_K1Val_K0_i32
-    ; CHECK: liveins: $vgpr0, $sgpr30_sgpr31
+    ; CHECK: liveins: $vgpr0
     ; CHECK: [[COPY:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
-    ; CHECK: [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; CHECK: [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 17
     ; CHECK: [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 -12
     ; CHECK: [[AMDGPU_SMED3_:%[0-9]+]]:vgpr(s32) = G_AMDGPU_SMED3 [[COPY]], [[C1]], [[C]]
     ; CHECK: $vgpr0 = COPY [[AMDGPU_SMED3_]](s32)
-    ; CHECK: [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-    ; CHECK: S_SETPC_B64_return [[COPY2]], implicit $vgpr0
+    ; CHECK: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     %0:vgpr(s32) = COPY $vgpr0
-    %1:sgpr_64 = COPY $sgpr30_sgpr31
     %2:sgpr(s32) = G_CONSTANT i32 17
     %7:vgpr(s32) = COPY %2(s32)
     %3:vgpr(s32) = G_SMIN %7, %0
     %4:sgpr(s32) = G_CONSTANT i32 -12
     %8:vgpr(s32) = COPY %4(s32)
     %5:vgpr(s32) = G_SMAX %3, %8
     $vgpr0 = COPY %5(s32)
-    %6:ccr_sgpr_64 = COPY %1
-    S_SETPC_B64_return %6, implicit $vgpr0
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
 ...
 
 ---
 name: test_max_K0min_ValK1__i32
 legalized: true
 regBankSelected: true
 tracksRegLiveness: true
 body: |
   bb.1:
-    liveins: $vgpr0, $sgpr30_sgpr31
+    liveins: $vgpr0
 
     ; CHECK-LABEL: name: test_max_K0min_ValK1__i32
-    ; CHECK: liveins: $vgpr0, $sgpr30_sgpr31
+    ; CHECK: liveins: $vgpr0
     ; CHECK: [[COPY:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
-    ; CHECK: [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; CHECK: [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 17
     ; CHECK: [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 -12
     ; CHECK: [[AMDGPU_SMED3_:%[0-9]+]]:vgpr(s32) = G_AMDGPU_SMED3 [[COPY]], [[C1]], [[C]]
     ; CHECK: $vgpr0 = COPY [[AMDGPU_SMED3_]](s32)
-    ; CHECK: [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-    ; CHECK: S_SETPC_B64_return [[COPY2]], implicit $vgpr0
+    ; CHECK: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     %0:vgpr(s32) = COPY $vgpr0
-    %1:sgpr_64 = COPY $sgpr30_sgpr31
     %2:sgpr(s32) = G_CONSTANT i32 17
     %7:vgpr(s32) = COPY %2(s32)
     %3:vgpr(s32) = G_SMIN %0, %7
     %4:sgpr(s32) = G_CONSTANT i32 -12
     %8:vgpr(s32) = COPY %4(s32)
     %5:vgpr(s32) = G_SMAX %8, %3
     $vgpr0 = COPY %5(s32)
-    %6:ccr_sgpr_64 = COPY %1
-    S_SETPC_B64_return %6, implicit $vgpr0
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
 ...
 
 ---
 name: test_max_K0min_K1Val__i32
 legalized: true
 regBankSelected: true
 tracksRegLiveness: true
 body: |
   bb.1:
-    liveins: $vgpr0, $sgpr30_sgpr31
+    liveins: $vgpr0
 
     ; CHECK-LABEL: name: test_max_K0min_K1Val__i32
-    ; CHECK: liveins: $vgpr0, $sgpr30_sgpr31
+    ; CHECK: liveins: $vgpr0
     ; CHECK: [[COPY:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
-    ; CHECK: [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; CHECK: [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 17
     ; CHECK: [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 -12
     ; CHECK: [[AMDGPU_SMED3_:%[0-9]+]]:vgpr(s32) = G_AMDGPU_SMED3 [[COPY]], [[C1]], [[C]]
     ; CHECK: $vgpr0 = COPY [[AMDGPU_SMED3_]](s32)
-    ; CHECK: [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-    ; CHECK: S_SETPC_B64_return [[COPY2]], implicit $vgpr0
+    ; CHECK: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     %0:vgpr(s32) = COPY $vgpr0
-    %1:sgpr_64 = COPY $sgpr30_sgpr31
     %2:sgpr(s32) = G_CONSTANT i32 17
     %7:vgpr(s32) = COPY %2(s32)
     %3:vgpr(s32) = G_SMIN %7, %0
     %4:sgpr(s32) = G_CONSTANT i32 -12
     %8:vgpr(s32) = COPY %4(s32)
     %5:vgpr(s32) = G_SMAX %8, %3
     $vgpr0 = COPY %5(s32)
-    %6:ccr_sgpr_64 = COPY %1
-    S_SETPC_B64_return %6, implicit $vgpr0
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
 ...
 
 ---
 name: test_max_K0min_K1Val__v2i16
 legalized: true
 regBankSelected: true
 tracksRegLiveness: true
 body: |
   bb.1:
-    liveins: $vgpr0, $sgpr30_sgpr31
+    liveins: $vgpr0
 
     ; CHECK-LABEL: name: test_max_K0min_K1Val__v2i16
-    ; CHECK: liveins: $vgpr0, $sgpr30_sgpr31
+    ; CHECK: liveins: $vgpr0
     ; CHECK: [[COPY:%[0-9]+]]:vgpr(<2 x s16>) = COPY $vgpr0
-    ; CHECK: [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; CHECK: [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 17
     ; CHECK: [[BUILD_VECTOR_TRUNC:%[0-9]+]]:sgpr(<2 x s16>) = G_BUILD_VECTOR_TRUNC [[C]](s32), [[C]](s32)
     ; CHECK: [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 -12
     ; CHECK: [[BUILD_VECTOR_TRUNC1:%[0-9]+]]:sgpr(<2 x s16>) = G_BUILD_VECTOR_TRUNC [[C1]](s32), [[C1]](s32)
     ; CHECK: [[COPY2:%[0-9]+]]:vgpr(<2 x s16>) = COPY [[BUILD_VECTOR_TRUNC]](<2 x s16>)
     ; CHECK: [[SMIN:%[0-9]+]]:vgpr(<2 x s16>) = G_SMIN [[COPY2]], [[COPY]]
     ; CHECK: [[COPY3:%[0-9]+]]:vgpr(<2 x s16>) = COPY [[BUILD_VECTOR_TRUNC1]](<2 x s16>)
     ; CHECK: [[SMAX:%[0-9]+]]:vgpr(<2 x s16>) = G_SMAX [[COPY3]], [[SMIN]]
     ; CHECK: $vgpr0 = COPY [[SMAX]](<2 x s16>)
-    ; CHECK: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-    ; CHECK: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; CHECK: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     %0:vgpr(<2 x s16>) = COPY $vgpr0
-    %1:sgpr_64 = COPY $sgpr30_sgpr31
     %9:sgpr(s32) = G_CONSTANT i32 17
     %2:sgpr(<2 x s16>) = G_BUILD_VECTOR_TRUNC %9(s32), %9(s32)
     %10:sgpr(s32) = G_CONSTANT i32 -12
     %5:sgpr(<2 x s16>) = G_BUILD_VECTOR_TRUNC %10(s32), %10(s32)
     %11:vgpr(<2 x s16>) = COPY %2(<2 x s16>)
     %4:vgpr(<2 x s16>) = G_SMIN %11, %0
     %12:vgpr(<2 x s16>) = COPY %5(<2 x s16>)
     %7:vgpr(<2 x s16>) = G_SMAX %12, %4
     $vgpr0 = COPY %7(<2 x s16>)
-    %8:ccr_sgpr_64 = COPY %1
-    S_SETPC_B64_return %8, implicit $vgpr0
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
 ...
 
 ---
 name: test_uniform_min_max
 legalized: true
 regBankSelected: true
 tracksRegLiveness: true
 body: |

llvm/test/CodeGen/AMDGPU/GlobalISel/regbankcombiner-umed3.mir

 # NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
 # RUN: llc -mtriple=amdgcn-amd-mesa3d -mcpu=gfx1010 -run-pass=amdgpu-regbank-combiner -verify-machineinstrs %s -o - | FileCheck %s
 
 ---
 name: test_min_max_ValK0_K1_u32
 legalized: true
 regBankSelected: true
 tracksRegLiveness: true
 body: |
   bb.1:
-    liveins: $vgpr0, $sgpr30_sgpr31
+    liveins: $vgpr0
 
     ; CHECK-LABEL: name: test_min_max_ValK0_K1_u32
-    ; CHECK: liveins: $vgpr0, $sgpr30_sgpr31
+    ; CHECK: liveins: $vgpr0
     ; CHECK: [[COPY:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
-    ; CHECK: [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; CHECK: [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 12
     ; CHECK: [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 17
     ; CHECK: [[AMDGPU_UMED3_:%[0-9]+]]:vgpr(s32) = G_AMDGPU_UMED3 [[COPY]], [[C]], [[C1]]
     ; CHECK: $vgpr0 = COPY [[AMDGPU_UMED3_]](s32)
-    ; CHECK: [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-    ; CHECK: S_SETPC_B64_return [[COPY2]], implicit $vgpr0
+    ; CHECK: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     %0:vgpr(s32) = COPY $vgpr0
-    %1:sgpr_64 = COPY $sgpr30_sgpr31
     %2:sgpr(s32) = G_CONSTANT i32 12
     %7:vgpr(s32) = COPY %2(s32)
     %3:vgpr(s32) = G_UMAX %0, %7
     %4:sgpr(s32) = G_CONSTANT i32 17
     %8:vgpr(s32) = COPY %4(s32)
     %5:vgpr(s32) = G_UMIN %3, %8
     $vgpr0 = COPY %5(s32)
-    %6:ccr_sgpr_64 = COPY %1
-    S_SETPC_B64_return %6, implicit $vgpr0
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
 ...
 
 ---
 name: min_max_ValK0_K1_i32
 legalized: true
 regBankSelected: true
 tracksRegLiveness: true
 body: |
   bb.1:
-    liveins: $vgpr0, $sgpr30_sgpr31
+    liveins: $vgpr0
 
     ; CHECK-LABEL: name: min_max_ValK0_K1_i32
-    ; CHECK: liveins: $vgpr0, $sgpr30_sgpr31
+    ; CHECK: liveins: $vgpr0
     ; CHECK: [[COPY:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
-    ; CHECK: [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; CHECK: [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 12
     ; CHECK: [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 17
     ; CHECK: [[AMDGPU_UMED3_:%[0-9]+]]:vgpr(s32) = G_AMDGPU_UMED3 [[COPY]], [[C]], [[C1]]
     ; CHECK: $vgpr0 = COPY [[AMDGPU_UMED3_]](s32)
-    ; CHECK: [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-    ; CHECK: S_SETPC_B64_return [[COPY2]], implicit $vgpr0
+    ; CHECK: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     %0:vgpr(s32) = COPY $vgpr0
-    %1:sgpr_64 = COPY $sgpr30_sgpr31
     %2:sgpr(s32) = G_CONSTANT i32 12
     %7:vgpr(s32) = COPY %2(s32)
     %3:vgpr(s32) = G_UMAX %7, %0
     %4:sgpr(s32) = G_CONSTANT i32 17
     %8:vgpr(s32) = COPY %4(s32)
     %5:vgpr(s32) = G_UMIN %3, %8
     $vgpr0 = COPY %5(s32)
-    %6:ccr_sgpr_64 = COPY %1
-    S_SETPC_B64_return %6, implicit $vgpr0
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
 ...
 
 ---
 name: test_min_K1max_ValK0__u32
 legalized: true
 regBankSelected: true
 tracksRegLiveness: true
 body: |
   bb.1:
-    liveins: $vgpr0, $sgpr30_sgpr31
+    liveins: $vgpr0
 
     ; CHECK-LABEL: name: test_min_K1max_ValK0__u32
-    ; CHECK: liveins: $vgpr0, $sgpr30_sgpr31
+    ; CHECK: liveins: $vgpr0
     ; CHECK: [[COPY:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
-    ; CHECK: [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; CHECK: [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 12
     ; CHECK: [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 17
     ; CHECK: [[AMDGPU_UMED3_:%[0-9]+]]:vgpr(s32) = G_AMDGPU_UMED3 [[COPY]], [[C]], [[C1]]
     ; CHECK: $vgpr0 = COPY [[AMDGPU_UMED3_]](s32)
-    ; CHECK: [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-    ; CHECK: S_SETPC_B64_return [[COPY2]], implicit $vgpr0
+    ; CHECK: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     %0:vgpr(s32) = COPY $vgpr0
-    %1:sgpr_64 = COPY $sgpr30_sgpr31
     %2:sgpr(s32) = G_CONSTANT i32 12
     %7:vgpr(s32) = COPY %2(s32)
     %3:vgpr(s32) = G_UMAX %0, %7
     %4:sgpr(s32) = G_CONSTANT i32 17
     %8:vgpr(s32) = COPY %4(s32)
     %5:vgpr(s32) = G_UMIN %8, %3
     $vgpr0 = COPY %5(s32)
-    %6:ccr_sgpr_64 = COPY %1
-    S_SETPC_B64_return %6, implicit $vgpr0
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
 ...
 
 ---
 name: test_min_K1max_K0Val__u32
 legalized: true
 regBankSelected: true
 tracksRegLiveness: true
 body: |
   bb.1:
-    liveins: $vgpr0, $sgpr30_sgpr31
+    liveins: $vgpr0
 
     ; CHECK-LABEL: name: test_min_K1max_K0Val__u32
-    ; CHECK: liveins: $vgpr0, $sgpr30_sgpr31
+    ; CHECK: liveins: $vgpr0
     ; CHECK: [[COPY:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
-    ; CHECK: [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; CHECK: [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 12
     ; CHECK: [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 17
     ; CHECK: [[AMDGPU_UMED3_:%[0-9]+]]:vgpr(s32) = G_AMDGPU_UMED3 [[COPY]], [[C]], [[C1]]
     ; CHECK: $vgpr0 = COPY [[AMDGPU_UMED3_]](s32)
-    ; CHECK: [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-    ; CHECK: S_SETPC_B64_return [[COPY2]], implicit $vgpr0
+    ; CHECK: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     %0:vgpr(s32) = COPY $vgpr0
-    %1:sgpr_64 = COPY $sgpr30_sgpr31
     %2:sgpr(s32) = G_CONSTANT i32 12
     %7:vgpr(s32) = COPY %2(s32)
     %3:vgpr(s32) = G_UMAX %7, %0
     %4:sgpr(s32) = G_CONSTANT i32 17
     %8:vgpr(s32) = COPY %4(s32)
     %5:vgpr(s32) = G_UMIN %8, %3
     $vgpr0 = COPY %5(s32)
-    %6:ccr_sgpr_64 = COPY %1
-    S_SETPC_B64_return %6, implicit $vgpr0
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
 ...
 
 ---
 name: test_max_min_ValK1_K0_u32
 legalized: true
 regBankSelected: true
 tracksRegLiveness: true
 body: |
   bb.1:
-    liveins: $vgpr0, $sgpr30_sgpr31
+    liveins: $vgpr0
 
     ; CHECK-LABEL: name: test_max_min_ValK1_K0_u32
-    ; CHECK: liveins: $vgpr0, $sgpr30_sgpr31
+    ; CHECK: liveins: $vgpr0
     ; CHECK: [[COPY:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
-    ; CHECK: [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; CHECK: [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 17
     ; CHECK: [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 12
     ; CHECK: [[AMDGPU_UMED3_:%[0-9]+]]:vgpr(s32) = G_AMDGPU_UMED3 [[COPY]], [[C1]], [[C]]
     ; CHECK: $vgpr0 = COPY [[AMDGPU_UMED3_]](s32)
-    ; CHECK: [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-    ; CHECK: S_SETPC_B64_return [[COPY2]], implicit $vgpr0
+    ; CHECK: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     %0:vgpr(s32) = COPY $vgpr0
-    %1:sgpr_64 = COPY $sgpr30_sgpr31
     %2:sgpr(s32) = G_CONSTANT i32 17
     %7:vgpr(s32) = COPY %2(s32)
     %3:vgpr(s32) = G_UMIN %0, %7
     %4:sgpr(s32) = G_CONSTANT i32 12
     %8:vgpr(s32) = COPY %4(s32)
     %5:vgpr(s32) = G_UMAX %3, %8
     $vgpr0 = COPY %5(s32)
-    %6:ccr_sgpr_64 = COPY %1
-    S_SETPC_B64_return %6, implicit $vgpr0
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
 ...
 
 ---
 name: test_max_min_K1Val_K0_u32
 legalized: true
 regBankSelected: true
 tracksRegLiveness: true
 body: |
   bb.1:
-    liveins: $vgpr0, $sgpr30_sgpr31
+    liveins: $vgpr0
 
     ; CHECK-LABEL: name: test_max_min_K1Val_K0_u32
-    ; CHECK: liveins: $vgpr0, $sgpr30_sgpr31
+    ; CHECK: liveins: $vgpr0
     ; CHECK: [[COPY:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
-    ; CHECK: [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; CHECK: [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 17
     ; CHECK: [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 12
     ; CHECK: [[AMDGPU_UMED3_:%[0-9]+]]:vgpr(s32) = G_AMDGPU_UMED3 [[COPY]], [[C1]], [[C]]
     ; CHECK: $vgpr0 = COPY [[AMDGPU_UMED3_]](s32)
-    ; CHECK: [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-    ; CHECK: S_SETPC_B64_return [[COPY2]], implicit $vgpr0
+    ; CHECK: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     %0:vgpr(s32) = COPY $vgpr0
-    %1:sgpr_64 = COPY $sgpr30_sgpr31
     %2:sgpr(s32) = G_CONSTANT i32 17
     %7:vgpr(s32) = COPY %2(s32)
     %3:vgpr(s32) = G_UMIN %7, %0
     %4:sgpr(s32) = G_CONSTANT i32 12
     %8:vgpr(s32) = COPY %4(s32)
     %5:vgpr(s32) = G_UMAX %3, %8
     $vgpr0 = COPY %5(s32)
-    %6:ccr_sgpr_64 = COPY %1
-    S_SETPC_B64_return %6, implicit $vgpr0
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
 ...
 
 ---
 name: test_max_K0min_ValK1__u32
 legalized: true
 regBankSelected: true
 tracksRegLiveness: true
 body: |
   bb.1:
-    liveins: $vgpr0, $sgpr30_sgpr31
+    liveins: $vgpr0
 
     ; CHECK-LABEL: name: test_max_K0min_ValK1__u32
-    ; CHECK: liveins: $vgpr0, $sgpr30_sgpr31
+    ; CHECK: liveins: $vgpr0
     ; CHECK: [[COPY:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
-    ; CHECK: [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; CHECK: [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 17
     ; CHECK: [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 12
     ; CHECK: [[AMDGPU_UMED3_:%[0-9]+]]:vgpr(s32) = G_AMDGPU_UMED3 [[COPY]], [[C1]], [[C]]
     ; CHECK: $vgpr0 = COPY [[AMDGPU_UMED3_]](s32)
-    ; CHECK: [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-    ; CHECK: S_SETPC_B64_return [[COPY2]], implicit $vgpr0
+    ; CHECK: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     %0:vgpr(s32) = COPY $vgpr0
-    %1:sgpr_64 = COPY $sgpr30_sgpr31
     %2:sgpr(s32) = G_CONSTANT i32 17
     %7:vgpr(s32) = COPY %2(s32)
     %3:vgpr(s32) = G_UMIN %0, %7
     %4:sgpr(s32) = G_CONSTANT i32 12
     %8:vgpr(s32) = COPY %4(s32)
     %5:vgpr(s32) = G_UMAX %8, %3
     $vgpr0 = COPY %5(s32)
-    %6:ccr_sgpr_64 = COPY %1
-    S_SETPC_B64_return %6, implicit $vgpr0
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
 ...
 
 ---
 name: test_max_K0min_K1Val__u32
 legalized: true
 regBankSelected: true
 tracksRegLiveness: true
 body: |
   bb.1:
-    liveins: $vgpr0, $sgpr30_sgpr31
+    liveins: $vgpr0
 
     ; CHECK-LABEL: name: test_max_K0min_K1Val__u32
-    ; CHECK: liveins: $vgpr0, $sgpr30_sgpr31
+    ; CHECK: liveins: $vgpr0
     ; CHECK: [[COPY:%[0-9]+]]:vgpr(s32) = COPY $vgpr0
-    ; CHECK: [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; CHECK: [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 17
     ; CHECK: [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 12
     ; CHECK: [[AMDGPU_UMED3_:%[0-9]+]]:vgpr(s32) = G_AMDGPU_UMED3 [[COPY]], [[C1]], [[C]]
     ; CHECK: $vgpr0 = COPY [[AMDGPU_UMED3_]](s32)
-    ; CHECK: [[COPY2:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-    ; CHECK: S_SETPC_B64_return [[COPY2]], implicit $vgpr0
+    ; CHECK: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     %0:vgpr(s32) = COPY $vgpr0
-    %1:sgpr_64 = COPY $sgpr30_sgpr31
     %2:sgpr(s32) = G_CONSTANT i32 17
     %7:vgpr(s32) = COPY %2(s32)
     %3:vgpr(s32) = G_UMIN %7, %0
     %4:sgpr(s32) = G_CONSTANT i32 12
     %8:vgpr(s32) = COPY %4(s32)
     %5:vgpr(s32) = G_UMAX %8, %3
     $vgpr0 = COPY %5(s32)
-    %6:ccr_sgpr_64 = COPY %1
-    S_SETPC_B64_return %6, implicit $vgpr0
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
 ...
 
 ---
 name: test_max_K0min_K1Val__v2u16
 legalized: true
 regBankSelected: true
 tracksRegLiveness: true
 body: |
   bb.1:
-    liveins: $vgpr0, $sgpr30_sgpr31
+    liveins: $vgpr0
 
     ; CHECK-LABEL: name: test_max_K0min_K1Val__v2u16
-    ; CHECK: liveins: $vgpr0, $sgpr30_sgpr31
+    ; CHECK: liveins: $vgpr0
     ; CHECK: [[COPY:%[0-9]+]]:vgpr(<2 x s16>) = COPY $vgpr0
-    ; CHECK: [[COPY1:%[0-9]+]]:sgpr_64 = COPY $sgpr30_sgpr31
     ; CHECK: [[C:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 17
     ; CHECK: [[BUILD_VECTOR_TRUNC:%[0-9]+]]:sgpr(<2 x s16>) = G_BUILD_VECTOR_TRUNC [[C]](s32), [[C]](s32)
     ; CHECK: [[C1:%[0-9]+]]:sgpr(s32) = G_CONSTANT i32 12
     ; CHECK: [[BUILD_VECTOR_TRUNC1:%[0-9]+]]:sgpr(<2 x s16>) = G_BUILD_VECTOR_TRUNC [[C1]](s32), [[C1]](s32)
     ; CHECK: [[COPY2:%[0-9]+]]:vgpr(<2 x s16>) = COPY [[BUILD_VECTOR_TRUNC]](<2 x s16>)
     ; CHECK: [[UMIN:%[0-9]+]]:vgpr(<2 x s16>) = G_UMIN [[COPY2]], [[COPY]]
     ; CHECK: [[COPY3:%[0-9]+]]:vgpr(<2 x s16>) = COPY [[BUILD_VECTOR_TRUNC1]](<2 x s16>)
     ; CHECK: [[UMAX:%[0-9]+]]:vgpr(<2 x s16>) = G_UMAX [[COPY3]], [[UMIN]]
     ; CHECK: $vgpr0 = COPY [[UMAX]](<2 x s16>)
-    ; CHECK: [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY1]]
-    ; CHECK: S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+    ; CHECK: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     %0:vgpr(<2 x s16>) = COPY $vgpr0
-    %1:sgpr_64 = COPY $sgpr30_sgpr31
     %9:sgpr(s32) = G_CONSTANT i32 17
     %2:sgpr(<2 x s16>) = G_BUILD_VECTOR_TRUNC %9(s32), %9(s32)
     %10:sgpr(s32) = G_CONSTANT i32 12
     %5:sgpr(<2 x s16>) = G_BUILD_VECTOR_TRUNC %10(s32), %10(s32)
     %11:vgpr(<2 x s16>) = COPY %2(<2 x s16>)
     %4:vgpr(<2 x s16>) = G_UMIN %11, %0
     %12:vgpr(<2 x s16>) = COPY %5(<2 x s16>)
     %7:vgpr(<2 x s16>) = G_UMAX %12, %4
     $vgpr0 = COPY %7(<2 x s16>)
-    %8:ccr_sgpr_64 = COPY %1
-    S_SETPC_B64_return %8, implicit $vgpr0
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
 ...
 
 ---
 name: test_uniform_min_max
 legalized: true
 regBankSelected: true
 tracksRegLiveness: true
 body: |

llvm/test/CodeGen/AMDGPU/GlobalISel/roundeven.ll

 ; GFX9-NEXT:    v_and_or_b32 v0, v2, v4, v0
 ; GFX9-NEXT:    v_and_or_b32 v1, v3, v4, v1
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: v_roundeven_v4f16:
 ; GFX10:       ; %bb.0:
 ; GFX10-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
-; GFX10-NEXT:    v_rndne_f16_e32 v2, v0
+; GFX10-NEXT:    v_mov_b32_e32 v2, 0xffff
+; GFX10-NEXT:    v_rndne_f16_e32 v3, v0
 ; GFX10-NEXT:    v_rndne_f16_sdwa v0, v0 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1
-; GFX10-NEXT:    v_rndne_f16_e32 v3, v1
-; GFX10-NEXT:    v_mov_b32_e32 v4, 0xffff
+; GFX10-NEXT:    v_rndne_f16_e32 v4, v1
 ; GFX10-NEXT:    v_rndne_f16_sdwa v1, v1 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1
-; GFX10-NEXT:    v_and_or_b32 v0, v2, v4, v0
-; GFX10-NEXT:    v_and_or_b32 v1, v3, v4, v1
+; GFX10-NEXT:    v_and_or_b32 v0, v3, v2, v0
+; GFX10-NEXT:    v_and_or_b32 v1, v4, v2, v1
 ; GFX10-NEXT:    s_setpc_b64 s[30:31]
   %roundeven = call <4 x half> @llvm.roundeven.v4f16(<4 x half> %x)
   ret <4 x half> %roundeven
 }
 
 
 define float @v_roundeven_f32_fabs(float %x) {
 ; GFX6-LABEL: v_roundeven_f32_fabs:

llvm/test/CodeGen/AMDGPU/GlobalISel/saddsat.ll

 ;
 ; GFX10-LABEL: v_saddsat_v4i8:
 ; GFX10:       ; %bb.0:
 ; GFX10-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v4, 24, v0
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v5, 24, v1
 ; GFX10-NEXT:    s_mov_b32 s4, 8
-; GFX10-NEXT:    v_lshrrev_b32_e32 v3, 16, v0
-; GFX10-NEXT:    v_lshrrev_b32_sdwa v2, s4, v0 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
-; GFX10-NEXT:    v_lshrrev_b32_sdwa v6, s4, v1 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
-; GFX10-NEXT:    v_mov_b32_e32 v7, 0xffff
+; GFX10-NEXT:    v_mov_b32_e32 v2, 0xffff
+; GFX10-NEXT:    v_lshrrev_b32_sdwa v3, s4, v0 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
+; GFX10-NEXT:    v_lshrrev_b32_e32 v6, 16, v0
+; GFX10-NEXT:    v_lshrrev_b32_sdwa v7, s4, v1 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v8, 16, v1
 ; GFX10-NEXT:    v_lshlrev_b32_e32 v4, 16, v4
 ; GFX10-NEXT:    v_lshlrev_b32_e32 v5, 16, v5
+; GFX10-NEXT:    v_and_or_b32 v0, v0, v2, v3
+; GFX10-NEXT:    v_and_or_b32 v1, v1, v2, v7
 ; GFX10-NEXT:    s_movk_i32 s4, 0xff
-; GFX10-NEXT:    v_and_or_b32 v0, v0, v7, v2
-; GFX10-NEXT:    v_and_or_b32 v1, v1, v7, v6
-; GFX10-NEXT:    v_and_or_b32 v2, v3, v7, v4
-; GFX10-NEXT:    v_and_or_b32 v3, v8, v7, v5
-; GFX10-NEXT:    v_mov_b32_e32 v4, 24
+; GFX10-NEXT:    v_and_or_b32 v3, v6, v2, v4
+; GFX10-NEXT:    v_and_or_b32 v2, v8, v2, v5
 ; GFX10-NEXT:    v_pk_lshlrev_b16 v0, 8, v0 op_sel_hi:[0,1]
 ; GFX10-NEXT:    v_pk_lshlrev_b16 v1, 8, v1 op_sel_hi:[0,1]
-; GFX10-NEXT:    v_pk_lshlrev_b16 v2, 8, v2 op_sel_hi:[0,1]
+; GFX10-NEXT:    v_mov_b32_e32 v4, 24
 ; GFX10-NEXT:    v_pk_lshlrev_b16 v3, 8, v3 op_sel_hi:[0,1]
+; GFX10-NEXT:    v_pk_lshlrev_b16 v2, 8, v2 op_sel_hi:[0,1]
 ; GFX10-NEXT:    v_pk_add_i16 v0, v0, v1 clamp
-; GFX10-NEXT:    v_pk_add_i16 v1, v2, v3 clamp
+; GFX10-NEXT:    v_pk_add_i16 v1, v3, v2 clamp
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 8
 ; GFX10-NEXT:    v_pk_ashrrev_i16 v0, 8, v0 op_sel_hi:[0,1]
 ; GFX10-NEXT:    v_pk_ashrrev_i16 v1, 8, v1 op_sel_hi:[0,1]
 ; GFX10-NEXT:    v_lshlrev_b32_sdwa v2, v2, v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:BYTE_2
 ; GFX10-NEXT:    v_and_b32_e32 v3, s4, v1
 ; GFX10-NEXT:    v_lshlrev_b32_sdwa v1, v4, v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:BYTE_2
 ; GFX10-NEXT:    v_and_or_b32 v0, v0, s4, v2
 ; GFX10-NEXT:    v_lshlrev_b32_e32 v2, 16, v3

llvm/test/CodeGen/AMDGPU/GlobalISel/ssubsat.ll

 ;
 ; GFX10-LABEL: v_ssubsat_v4i8:
 ; GFX10:       ; %bb.0:
 ; GFX10-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v4, 24, v0
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v5, 24, v1
 ; GFX10-NEXT:    s_mov_b32 s4, 8
-; GFX10-NEXT:    v_lshrrev_b32_e32 v3, 16, v0
-; GFX10-NEXT:    v_lshrrev_b32_sdwa v2, s4, v0 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
-; GFX10-NEXT:    v_lshrrev_b32_sdwa v6, s4, v1 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
-; GFX10-NEXT:    v_mov_b32_e32 v7, 0xffff
+; GFX10-NEXT:    v_mov_b32_e32 v2, 0xffff
+; GFX10-NEXT:    v_lshrrev_b32_sdwa v3, s4, v0 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
+; GFX10-NEXT:    v_lshrrev_b32_e32 v6, 16, v0
+; GFX10-NEXT:    v_lshrrev_b32_sdwa v7, s4, v1 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v8, 16, v1
 ; GFX10-NEXT:    v_lshlrev_b32_e32 v4, 16, v4
 ; GFX10-NEXT:    v_lshlrev_b32_e32 v5, 16, v5
+; GFX10-NEXT:    v_and_or_b32 v0, v0, v2, v3
+; GFX10-NEXT:    v_and_or_b32 v1, v1, v2, v7
 ; GFX10-NEXT:    s_movk_i32 s4, 0xff
-; GFX10-NEXT:    v_and_or_b32 v0, v0, v7, v2
-; GFX10-NEXT:    v_and_or_b32 v1, v1, v7, v6
-; GFX10-NEXT:    v_and_or_b32 v2, v3, v7, v4
-; GFX10-NEXT:    v_and_or_b32 v3, v8, v7, v5
-; GFX10-NEXT:    v_mov_b32_e32 v4, 24
+; GFX10-NEXT:    v_and_or_b32 v3, v6, v2, v4
+; GFX10-NEXT:    v_and_or_b32 v2, v8, v2, v5
 ; GFX10-NEXT:    v_pk_lshlrev_b16 v0, 8, v0 op_sel_hi:[0,1]
 ; GFX10-NEXT:    v_pk_lshlrev_b16 v1, 8, v1 op_sel_hi:[0,1]
-; GFX10-NEXT:    v_pk_lshlrev_b16 v2, 8, v2 op_sel_hi:[0,1]
+; GFX10-NEXT:    v_mov_b32_e32 v4, 24
 ; GFX10-NEXT:    v_pk_lshlrev_b16 v3, 8, v3 op_sel_hi:[0,1]
+; GFX10-NEXT:    v_pk_lshlrev_b16 v2, 8, v2 op_sel_hi:[0,1]
 ; GFX10-NEXT:    v_pk_sub_i16 v0, v0, v1 clamp
-; GFX10-NEXT:    v_pk_sub_i16 v1, v2, v3 clamp
+; GFX10-NEXT:    v_pk_sub_i16 v1, v3, v2 clamp
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 8
 ; GFX10-NEXT:    v_pk_ashrrev_i16 v0, 8, v0 op_sel_hi:[0,1]
 ; GFX10-NEXT:    v_pk_ashrrev_i16 v1, 8, v1 op_sel_hi:[0,1]
 ; GFX10-NEXT:    v_lshlrev_b32_sdwa v2, v2, v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:BYTE_2
 ; GFX10-NEXT:    v_and_b32_e32 v3, s4, v1
 ; GFX10-NEXT:    v_lshlrev_b32_sdwa v1, v4, v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:BYTE_2
 ; GFX10-NEXT:    v_and_or_b32 v0, v0, s4, v2
 ; GFX10-NEXT:    v_lshlrev_b32_e32 v2, 16, v3

llvm/test/CodeGen/AMDGPU/GlobalISel/uaddsat.ll

 ;
 ; GFX10-LABEL: v_uaddsat_v4i8:
 ; GFX10:       ; %bb.0:
 ; GFX10-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v4, 24, v0
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v5, 24, v1
 ; GFX10-NEXT:    s_mov_b32 s4, 8
-; GFX10-NEXT:    v_lshrrev_b32_e32 v3, 16, v0
-; GFX10-NEXT:    v_lshrrev_b32_sdwa v2, s4, v0 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
-; GFX10-NEXT:    v_lshrrev_b32_sdwa v6, s4, v1 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
-; GFX10-NEXT:    v_mov_b32_e32 v7, 0xffff
+; GFX10-NEXT:    v_mov_b32_e32 v2, 0xffff
+; GFX10-NEXT:    v_lshrrev_b32_sdwa v3, s4, v0 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
+; GFX10-NEXT:    v_lshrrev_b32_e32 v6, 16, v0
+; GFX10-NEXT:    v_lshrrev_b32_sdwa v7, s4, v1 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v8, 16, v1
 ; GFX10-NEXT:    v_lshlrev_b32_e32 v4, 16, v4
 ; GFX10-NEXT:    v_lshlrev_b32_e32 v5, 16, v5
+; GFX10-NEXT:    v_and_or_b32 v0, v0, v2, v3
+; GFX10-NEXT:    v_and_or_b32 v1, v1, v2, v7
 ; GFX10-NEXT:    s_movk_i32 s4, 0xff
-; GFX10-NEXT:    v_and_or_b32 v0, v0, v7, v2
-; GFX10-NEXT:    v_and_or_b32 v1, v1, v7, v6
-; GFX10-NEXT:    v_and_or_b32 v2, v3, v7, v4
-; GFX10-NEXT:    v_and_or_b32 v3, v8, v7, v5
-; GFX10-NEXT:    v_mov_b32_e32 v4, 24
+; GFX10-NEXT:    v_and_or_b32 v3, v6, v2, v4
+; GFX10-NEXT:    v_and_or_b32 v2, v8, v2, v5
 ; GFX10-NEXT:    v_pk_lshlrev_b16 v0, 8, v0 op_sel_hi:[0,1]
 ; GFX10-NEXT:    v_pk_lshlrev_b16 v1, 8, v1 op_sel_hi:[0,1]
-; GFX10-NEXT:    v_pk_lshlrev_b16 v2, 8, v2 op_sel_hi:[0,1]
+; GFX10-NEXT:    v_mov_b32_e32 v4, 24
 ; GFX10-NEXT:    v_pk_lshlrev_b16 v3, 8, v3 op_sel_hi:[0,1]
+; GFX10-NEXT:    v_pk_lshlrev_b16 v2, 8, v2 op_sel_hi:[0,1]
 ; GFX10-NEXT:    v_pk_add_u16 v0, v0, v1 clamp
-; GFX10-NEXT:    v_pk_add_u16 v1, v2, v3 clamp
+; GFX10-NEXT:    v_pk_add_u16 v1, v3, v2 clamp
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 8
 ; GFX10-NEXT:    v_pk_lshrrev_b16 v0, 8, v0 op_sel_hi:[0,1]
 ; GFX10-NEXT:    v_pk_lshrrev_b16 v1, 8, v1 op_sel_hi:[0,1]
 ; GFX10-NEXT:    v_lshlrev_b32_sdwa v2, v2, v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:BYTE_2
 ; GFX10-NEXT:    v_and_b32_e32 v3, s4, v1
 ; GFX10-NEXT:    v_lshlrev_b32_sdwa v1, v4, v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:BYTE_2
 ; GFX10-NEXT:    v_and_or_b32 v0, v0, s4, v2
 ; GFX10-NEXT:    v_lshlrev_b32_e32 v2, 16, v3

llvm/test/CodeGen/AMDGPU/GlobalISel/usubsat.ll

 ;
 ; GFX10-LABEL: v_usubsat_v4i8:
 ; GFX10:       ; %bb.0:
 ; GFX10-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v4, 24, v0
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v5, 24, v1
 ; GFX10-NEXT:    s_mov_b32 s4, 8
-; GFX10-NEXT:    v_lshrrev_b32_e32 v3, 16, v0
-; GFX10-NEXT:    v_lshrrev_b32_sdwa v2, s4, v0 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
-; GFX10-NEXT:    v_lshrrev_b32_sdwa v6, s4, v1 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
-; GFX10-NEXT:    v_mov_b32_e32 v7, 0xffff
+; GFX10-NEXT:    v_mov_b32_e32 v2, 0xffff
+; GFX10-NEXT:    v_lshrrev_b32_sdwa v3, s4, v0 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
+; GFX10-NEXT:    v_lshrrev_b32_e32 v6, 16, v0
+; GFX10-NEXT:    v_lshrrev_b32_sdwa v7, s4, v1 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v8, 16, v1
 ; GFX10-NEXT:    v_lshlrev_b32_e32 v4, 16, v4
 ; GFX10-NEXT:    v_lshlrev_b32_e32 v5, 16, v5
+; GFX10-NEXT:    v_and_or_b32 v0, v0, v2, v3
+; GFX10-NEXT:    v_and_or_b32 v1, v1, v2, v7
 ; GFX10-NEXT:    s_movk_i32 s4, 0xff
-; GFX10-NEXT:    v_and_or_b32 v0, v0, v7, v2
-; GFX10-NEXT:    v_and_or_b32 v1, v1, v7, v6
-; GFX10-NEXT:    v_and_or_b32 v2, v3, v7, v4
-; GFX10-NEXT:    v_and_or_b32 v3, v8, v7, v5
-; GFX10-NEXT:    v_mov_b32_e32 v4, 24
+; GFX10-NEXT:    v_and_or_b32 v3, v6, v2, v4
+; GFX10-NEXT:    v_and_or_b32 v2, v8, v2, v5
 ; GFX10-NEXT:    v_pk_lshlrev_b16 v0, 8, v0 op_sel_hi:[0,1]
 ; GFX10-NEXT:    v_pk_lshlrev_b16 v1, 8, v1 op_sel_hi:[0,1]
-; GFX10-NEXT:    v_pk_lshlrev_b16 v2, 8, v2 op_sel_hi:[0,1]
+; GFX10-NEXT:    v_mov_b32_e32 v4, 24
 ; GFX10-NEXT:    v_pk_lshlrev_b16 v3, 8, v3 op_sel_hi:[0,1]
+; GFX10-NEXT:    v_pk_lshlrev_b16 v2, 8, v2 op_sel_hi:[0,1]
 ; GFX10-NEXT:    v_pk_sub_u16 v0, v0, v1 clamp
-; GFX10-NEXT:    v_pk_sub_u16 v1, v2, v3 clamp
+; GFX10-NEXT:    v_pk_sub_u16 v1, v3, v2 clamp
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 8
 ; GFX10-NEXT:    v_pk_lshrrev_b16 v0, 8, v0 op_sel_hi:[0,1]
 ; GFX10-NEXT:    v_pk_lshrrev_b16 v1, 8, v1 op_sel_hi:[0,1]
 ; GFX10-NEXT:    v_lshlrev_b32_sdwa v2, v2, v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:BYTE_2
 ; GFX10-NEXT:    v_and_b32_e32 v3, s4, v1
 ; GFX10-NEXT:    v_lshlrev_b32_sdwa v1, v4, v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:BYTE_2
 ; GFX10-NEXT:    v_and_or_b32 v0, v0, s4, v2
 ; GFX10-NEXT:    v_lshlrev_b32_e32 v2, 16, v3

llvm/test/CodeGen/AMDGPU/abi-attribute-hints-undefined-behavior.ll

 ; VARABI-NEXT:    s_mov_b32 s33, s32
 ; VARABI-NEXT:    s_addk_i32 s32, 0x400
 ; VARABI-NEXT:    v_writelane_b32 v40, s30, 0
 ; VARABI-NEXT:    v_writelane_b32 v40, s31, 1
 ; VARABI-NEXT:    s_getpc_b64 s[4:5]
 ; VARABI-NEXT:    s_add_u32 s4, s4, requires_all_inputs@rel32@lo+4
 ; VARABI-NEXT:    s_addc_u32 s5, s5, requires_all_inputs@rel32@hi+12
 ; VARABI-NEXT:    s_swappc_b64 s[30:31], s[4:5]
-; VARABI-NEXT:    v_readlane_b32 s4, v40, 0
-; VARABI-NEXT:    v_readlane_b32 s5, v40, 1
+; VARABI-NEXT:    v_readlane_b32 s30, v40, 0
+; VARABI-NEXT:    v_readlane_b32 s31, v40, 1
 ; VARABI-NEXT:    s_addk_i32 s32, 0xfc00
 ; VARABI-NEXT:    v_readlane_b32 s33, v40, 2
-; VARABI-NEXT:    s_or_saveexec_b64 s[6:7], -1
+; VARABI-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; VARABI-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
-; VARABI-NEXT:    s_mov_b64 exec, s[6:7]
+; VARABI-NEXT:    s_mov_b64 exec, s[4:5]
 ; VARABI-NEXT:    s_waitcnt vmcnt(0)
-; VARABI-NEXT:    s_setpc_b64 s[4:5]
+; VARABI-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; FIXEDABI-LABEL: parent_func_missing_inputs:
 ; FIXEDABI:       ; %bb.0:
 ; FIXEDABI-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; FIXEDABI-NEXT:    s_or_saveexec_b64 s[16:17], -1
 ; FIXEDABI-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; FIXEDABI-NEXT:    s_mov_b64 exec, s[16:17]
 ; FIXEDABI-NEXT:    v_writelane_b32 v40, s33, 2
 ; FIXEDABI-NEXT:    s_mov_b32 s33, s32
 ; FIXEDABI-NEXT:    s_addk_i32 s32, 0x400
 ; FIXEDABI-NEXT:    v_writelane_b32 v40, s30, 0
 ; FIXEDABI-NEXT:    v_writelane_b32 v40, s31, 1
 ; FIXEDABI-NEXT:    s_getpc_b64 s[16:17]
 ; FIXEDABI-NEXT:    s_add_u32 s16, s16, requires_all_inputs@rel32@lo+4
 ; FIXEDABI-NEXT:    s_addc_u32 s17, s17, requires_all_inputs@rel32@hi+12
 ; FIXEDABI-NEXT:    s_swappc_b64 s[30:31], s[16:17]
-; FIXEDABI-NEXT:    v_readlane_b32 s4, v40, 0
-; FIXEDABI-NEXT:    v_readlane_b32 s5, v40, 1
+; FIXEDABI-NEXT:    v_readlane_b32 s30, v40, 0
+; FIXEDABI-NEXT:    v_readlane_b32 s31, v40, 1
 ; FIXEDABI-NEXT:    s_addk_i32 s32, 0xfc00
 ; FIXEDABI-NEXT:    v_readlane_b32 s33, v40, 2
-; FIXEDABI-NEXT:    s_or_saveexec_b64 s[6:7], -1
+; FIXEDABI-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; FIXEDABI-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
-; FIXEDABI-NEXT:    s_mov_b64 exec, s[6:7]
+; FIXEDABI-NEXT:    s_mov_b64 exec, s[4:5]
 ; FIXEDABI-NEXT:    s_waitcnt vmcnt(0)
-; FIXEDABI-NEXT:    s_setpc_b64 s[4:5]
+; FIXEDABI-NEXT:    s_setpc_b64 s[30:31]
   call void @requires_all_inputs()
   ret void
 }
 
 define amdgpu_kernel void @parent_kernel_missing_inputs() #0 {
 ; VARABI-LABEL: parent_kernel_missing_inputs:
 ; VARABI:       ; %bb.0:
 ; VARABI-NEXT:    s_add_i32 s4, s4, s9

llvm/test/CodeGen/AMDGPU/amdpal-callable.ll

 ; GCN-NEXT:        .vgpr_count:     0x3{{$}}
 ; GCN-NEXT:      no_stack:
 ; GCN-NEXT:        .lds_size:       0{{$}}
 ; GCN-NEXT:        .sgpr_count:     0x20{{$}}
 ; GCN-NEXT:        .stack_frame_size_in_bytes: 0{{$}}
 ; GCN-NEXT:        .vgpr_count:     0x1{{$}}
 ; GCN-NEXT:      no_stack_call:
 ; GCN-NEXT:        .lds_size:       0{{$}}
-; GCN-NEXT:        .sgpr_count:     0x26{{$}}
-; GCN-NEXT:        .stack_frame_size_in_bytes: 0{{$}}
-; GCN-NEXT:        .vgpr_count:     0x2{{$}}
+; GCN-NEXT:        .sgpr_count:     0x24{{$}}
+; GCN-NEXT:        .stack_frame_size_in_bytes: 0x10{{$}}
+; GCN-NEXT:        .vgpr_count:     0x3{{$}}
 ; GCN-NEXT:      no_stack_extern_call:
 ; GCN-NEXT:        .lds_size:       0{{$}}
 ; GFX8-NEXT:        .sgpr_count:     0x28{{$}}
 ; GFX9-NEXT:        .sgpr_count:     0x2c{{$}}
 ; GCN-NEXT:        .stack_frame_size_in_bytes: 0x10{{$}}
 ; GCN-NEXT:        .vgpr_count:     0x29{{$}}
 ; GCN-NEXT:      no_stack_extern_call_many_args:
 ; GCN-NEXT:        .lds_size:       0{{$}}

 ; GCN-NEXT:        .vgpr_count:     0x29{{$}}
 ; GCN-NEXT:      simple_stack:
 ; GCN-NEXT:        .lds_size:       0{{$}}
 ; GCN-NEXT:        .sgpr_count:     0x21{{$}}
 ; GCN-NEXT:        .stack_frame_size_in_bytes: 0x14{{$}}
 ; GCN-NEXT:        .vgpr_count:     0x2{{$}}
 ; GCN-NEXT:      simple_stack_call:
 ; GCN-NEXT:        .lds_size:       0{{$}}
-; GCN-NEXT:        .sgpr_count:     0x26{{$}}
+; GCN-NEXT:        .sgpr_count:     0x24{{$}}
 ; GCN-NEXT:        .stack_frame_size_in_bytes: 0x20{{$}}
-; GCN-NEXT:        .vgpr_count:     0x3{{$}}
+; GCN-NEXT:        .vgpr_count:     0x4{{$}}
 ; GCN-NEXT:      simple_stack_extern_call:
 ; GCN-NEXT:        .lds_size:       0{{$}}
 ; GFX8-NEXT:        .sgpr_count:     0x28{{$}}
 ; GFX9-NEXT:        .sgpr_count:     0x2c{{$}}
 ; GCN-NEXT:        .stack_frame_size_in_bytes: 0x20{{$}}
 ; GCN-NEXT:        .vgpr_count:     0x2a{{$}}
 ; GCN-NEXT:      simple_stack_indirect_call:
 ; GCN-NEXT:        .lds_size:       0{{$}}

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

bb3:
   tail call void asm sideeffect "; reg use $0", "{s99}"(i32 %sgpr99) #0
   tail call void asm sideeffect "; reg use $0", "{s100}"(i32 %sgpr100) #0
   tail call void asm sideeffect "; reg use $0", "{s101}"(i32 %sgpr101) #0
   tail call void asm sideeffect "; reg use $0", "{vcc_lo}"(i32 %vcc_lo) #0
   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[6:7], -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[6:7]
 ; CHECK-NEXT:    s_waitcnt expcnt(1)

 ; CHECK-NEXT:    v_writelane_b32 v0, s83, 50
 ; CHECK-NEXT:    v_writelane_b32 v0, s84, 51
 ; CHECK-NEXT:    v_writelane_b32 v0, s85, 52
 ; CHECK-NEXT:    v_writelane_b32 v0, s86, 53
 ; CHECK-NEXT:    v_writelane_b32 v0, s87, 54
 ; CHECK-NEXT:    v_writelane_b32 v0, s88, 55
 ; CHECK-NEXT:    v_writelane_b32 v0, s89, 56
 ; CHECK-NEXT:    v_writelane_b32 v0, s90, 57
-; CHECK-NEXT:    s_waitcnt expcnt(0)
-; CHECK-NEXT:    v_writelane_b32 v1, s97, 0
 ; CHECK-NEXT:    v_writelane_b32 v0, s91, 58
-; CHECK-NEXT:    v_writelane_b32 v1, s98, 1
 ; CHECK-NEXT:    v_writelane_b32 v0, s92, 59
-; CHECK-NEXT:    v_writelane_b32 v1, s99, 2
+; CHECK-NEXT:    s_waitcnt expcnt(0)
+; CHECK-NEXT:    v_writelane_b32 v1, s97, 0
 ; CHECK-NEXT:    v_writelane_b32 v0, s93, 60
-; CHECK-NEXT:    v_writelane_b32 v1, s100, 3
+; CHECK-NEXT:    v_writelane_b32 v1, s98, 1
 ; CHECK-NEXT:    v_writelane_b32 v0, s94, 61
-; CHECK-NEXT:    v_writelane_b32 v1, s101, 4
+; CHECK-NEXT:    v_writelane_b32 v1, s99, 2
 ; CHECK-NEXT:    v_writelane_b32 v0, s95, 62
-; CHECK-NEXT:    v_writelane_b32 v1, s30, 5
-; CHECK-NEXT:    s_mov_b32 s29, s4
+; CHECK-NEXT:    v_writelane_b32 v1, s100, 3
+; CHECK-NEXT:    s_mov_b32 s35, s4
 ; CHECK-NEXT:    v_writelane_b32 v0, s96, 63
-; CHECK-NEXT:    v_writelane_b32 v1, s31, 6
-; CHECK-NEXT:    s_cmp_eq_u32 s29, 0
+; CHECK-NEXT:    v_writelane_b32 v1, s101, 4
+; CHECK-NEXT:    s_cmp_eq_u32 s35, 0
 ; CHECK-NEXT:    ;;#ASMSTART
 ; CHECK-NEXT:    s_mov_b32 s0, 0
 ; CHECK-NEXT:    ;;#ASMEND
 ; CHECK-NEXT:    ;;#ASMSTART
 ; CHECK-NEXT:    s_mov_b32 s1, 0
 ; CHECK-NEXT:    ;;#ASMEND
 ; CHECK-NEXT:    ;;#ASMSTART
 ; CHECK-NEXT:    s_mov_b32 s2, 0

 ; CHECK-NEXT:    ; reg use s3
 ; CHECK-NEXT:    ;;#ASMEND
 ; CHECK-NEXT:    ;;#ASMSTART
 ; CHECK-NEXT:    ; reg use s4
 ; CHECK-NEXT:    ;;#ASMEND
 ; CHECK-NEXT:    ;;#ASMSTART
 ; CHECK-NEXT:    ; reg use s5
 ; CHECK-NEXT:    ;;#ASMEND
-; CHECK-NEXT:    v_readlane_b32 s4, v1, 5
 ; CHECK-NEXT:    ;;#ASMSTART
 ; CHECK-NEXT:    ; reg use s6
 ; CHECK-NEXT:    ;;#ASMEND
 ; CHECK-NEXT:    ;;#ASMSTART
 ; CHECK-NEXT:    ; reg use s7
 ; CHECK-NEXT:    ;;#ASMEND
 ; CHECK-NEXT:    ;;#ASMSTART
 ; CHECK-NEXT:    ; reg use s8

 ; CHECK-NEXT:    ; reg use s101
 ; CHECK-NEXT:    ;;#ASMEND
 ; CHECK-NEXT:    ;;#ASMSTART
 ; CHECK-NEXT:    ; reg use vcc_lo
 ; CHECK-NEXT:    ;;#ASMEND
 ; CHECK-NEXT:    ;;#ASMSTART
 ; CHECK-NEXT:    ; reg use vcc_hi
 ; CHECK-NEXT:    ;;#ASMEND
-; CHECK-NEXT:    v_readlane_b32 s5, v1, 6
 ; CHECK-NEXT:    v_readlane_b32 s101, v1, 4
 ; CHECK-NEXT:    v_readlane_b32 s100, v1, 3
 ; CHECK-NEXT:    v_readlane_b32 s99, v1, 2
 ; CHECK-NEXT:    v_readlane_b32 s98, v1, 1
 ; CHECK-NEXT:    v_readlane_b32 s97, v1, 0
 ; CHECK-NEXT:    v_readlane_b32 s96, v0, 63
 ; CHECK-NEXT:    v_readlane_b32 s95, v0, 62
 ; CHECK-NEXT:    v_readlane_b32 s94, v0, 61

 ; CHECK-NEXT:    v_readlane_b32 s40, v0, 7
 ; CHECK-NEXT:    v_readlane_b32 s39, v0, 6
 ; CHECK-NEXT:    v_readlane_b32 s38, v0, 5
 ; CHECK-NEXT:    v_readlane_b32 s37, v0, 4
 ; CHECK-NEXT:    v_readlane_b32 s36, v0, 3
 ; CHECK-NEXT:    v_readlane_b32 s35, v0, 2
 ; CHECK-NEXT:    v_readlane_b32 s34, v0, 1
 ; CHECK-NEXT:    v_readlane_b32 s33, v0, 0
-; CHECK-NEXT:    s_or_saveexec_b64 s[6:7], -1
+; CHECK-NEXT:    s_or_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[6:7]
+; CHECK-NEXT:    s_mov_b64 exec, s[4:5]
 ; CHECK-NEXT:    s_waitcnt vmcnt(0)
-; CHECK-NEXT:    s_setpc_b64 s[4:5]
+; 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
   %sgpr1 = tail call i32 asm sideeffect "s_mov_b32 s1, 0", "={s1}"() #0
   %sgpr2 = tail call i32 asm sideeffect "s_mov_b32 s2, 0", "={s2}"() #0
   %sgpr3 = tail call i32 asm sideeffect "s_mov_b32 s3, 0", "={s3}"() #0
   %sgpr4 = tail call i32 asm sideeffect "s_mov_b32 s4, 0", "={s4}"() #0
   %sgpr5 = tail call i32 asm sideeffect "s_mov_b32 s5, 0", "={s5}"() #0

llvm/test/CodeGen/AMDGPU/call-graph-register-usage.ll

define void @use_vcc() #1 {
   ret void
 }
 
 ; GCN-LABEL: {{^}}indirect_use_vcc:
 ; GCN: v_writelane_b32 v40, s33, 2
 ; GCN: v_writelane_b32 v40, s30, 0
 ; GCN: v_writelane_b32 v40, s31, 1
 ; GCN: s_swappc_b64
-; GCN: v_readlane_b32 s4, v40, 0
-; GCN: v_readlane_b32 s5, v40, 1
+; GCN: v_readlane_b32 s30, v40, 0
+; GCN: v_readlane_b32 s31, v40, 1
 ; GCN: v_readlane_b32 s33, v40, 2
+; GCN: s_setpc_b64 s[30:31]
 ; GCN: ; NumSgprs: 36
 ; GCN: ; NumVgprs: 41
 define void @indirect_use_vcc() #1 {
   call void @use_vcc()
   ret void
 }
 
 ; GCN-LABEL: {{^}}indirect_2level_use_vcc_kernel:

llvm/test/CodeGen/AMDGPU/call-preserved-registers.ll

 ; GCN: v_writelane_b32 v40, s35, 1
 ; GCN: v_writelane_b32 v40, s30, 2
 ; GCN: v_writelane_b32 v40, s31, 3
 
 ; GCN: s_swappc_b64
 ; GCN-NEXT: ;;#ASMSTART
 ; GCN-NEXT: ;;#ASMEND
 ; GCN-NEXT: s_swappc_b64
-; MUBUF-DAG:   v_readlane_b32 s4, v40, 2
-; MUBUF-DAG:   v_readlane_b32 s5, v40, 3
-; FLATSCR-DAG: v_readlane_b32 s0, v40, 2
-; FLATSCR-DAG: v_readlane_b32 s1, v40, 3
+; MUBUF-DAG:   v_readlane_b32 s30, v40, 2
+; MUBUF-DAG:   v_readlane_b32 s31, v40, 3
+; FLATSCR-DAG: v_readlane_b32 s30, v40, 2
+; FLATSCR-DAG: v_readlane_b32 s31, v40, 3
 ; GCN: v_readlane_b32 s35, v40, 1
 ; GCN: v_readlane_b32 s34, v40, 0
 
 ; GCN: v_readlane_b32 s33, v40, 4
 ; MUBUF:   buffer_load_dword
 ; FLATSCR: scratch_load_dword
-; GCN: s_setpc_b64
+; GCN: s_setpc_b64 s[30:31]
 define void @test_func_call_external_void_func_void_clobber_s30_s31_call_external_void_func_void() #0 {
   call void @external_void_func_void()
   call void asm sideeffect "", ""() #0
   call void @external_void_func_void()
   ret void
 }
 
 ; GCN-LABEL: {{^}}test_func_call_external_void_funcx2:

 ; MUBUF:   buffer_load_dword v40
 ; FLATSCR: scratch_load_dword v40
 define void @test_func_call_external_void_funcx2() #0 {
   call void @external_void_func_void()
   call void @external_void_func_void()
   ret void
 }
 
-; GCN-LABEL: {{^}}void_func_void_clobber_s30_s31:
+; GCN-LABEL: {{^}}void_func_void_clobber_s28_s29:
 ; GCN: s_waitcnt
-; GCN-NEXT: s_mov_b64 [[SAVEPC:s\[[0-9]+:[0-9]+\]]], s[30:31]
 ; GCN-NEXT: #ASMSTART
 ; GCN: ; clobber
 ; GCN-NEXT: #ASMEND
-; GCN-NEXT: s_setpc_b64 [[SAVEPC]]
-define void @void_func_void_clobber_s30_s31() #2 {
-  call void asm sideeffect "; clobber", "~{s[30:31]}"() #0
+; GCN-NEXT: s_setpc_b64 s[30:31]
+define void @void_func_void_clobber_s28_s29() #2 {
+  call void asm sideeffect "; clobber", "~{s[28:29]}"() #0
   ret void
 }
 
 ; GCN-LABEL: {{^}}void_func_void_clobber_vcc:
 ; GCN: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GCN-NEXT: ;;#ASMSTART
 ; GCN-NEXT: ;;#ASMEND
 ; GCN-NEXT: s_setpc_b64 s[30:31]
 define hidden void @void_func_void_clobber_vcc() #2 {
   call void asm sideeffect "", "~{vcc}"() #0

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

 ; GCN-DAG: v_writelane_b32 [[CSR_VGPR]], s30,
 ; GCN-DAG: v_writelane_b32 [[CSR_VGPR]], s31,
 
 ; MUBUF-DAG:   buffer_store_dword [[ZERO]], off, s[0:3], s33{{$}}
 ; FLATSCR-DAG: scratch_store_dword off, [[ZERO]], s33{{$}}
 
 ; GCN: s_swappc_b64
 
-; MUBUF-DAG: v_readlane_b32 s5, [[CSR_VGPR]]
-; MUBUF-DAG: v_readlane_b32 s4, [[CSR_VGPR]]
-; FLATSCR-DAG: v_readlane_b32 s0, [[CSR_VGPR]]
-; FLATSCR-DAG: v_readlane_b32 s1, [[CSR_VGPR]]
+; GCN-DAG: v_readlane_b32 s30, [[CSR_VGPR]]
+; GCN-DAG: v_readlane_b32 s31, [[CSR_VGPR]]
 
 ; MUBUF:    s_addk_i32 s32, 0xfc00{{$}}
 ; FLATSCR:  s_add_i32 s32, s32, -16{{$}}
 ; GCN-NEXT: v_readlane_b32 s33, [[CSR_VGPR]], 2
 ; GCN-NEXT: s_or_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
+; GCN-NEXT: s_setpc_b64 s[30:31]
 define void @callee_with_stack_and_call() #0 {
   %alloca = alloca i32, addrspace(5)
   store volatile i32 0, i32 addrspace(5)* %alloca
   call void @external_void_func_void()
   ret void
 }
 
 ; Should be able to copy incoming stack pointer directly to inner

 ; MUBUF-DAG:   s_addk_i32 s32, 0x400
 ; FLATSCR-DAG: s_add_i32 s32, s32, 16
 ; GCN-DAG: v_writelane_b32 [[CSR_VGPR]], s33, [[FP_SPILL_LANE:[0-9]+]]
 
 ; GCN-DAG: v_writelane_b32 [[CSR_VGPR]], s30, 0
 ; GCN-DAG: v_writelane_b32 [[CSR_VGPR]], s31, 1
 ; GCN: s_swappc_b64
 
-; MUBUF-DAG: v_readlane_b32 s4, v40, 0
-; MUBUF-DAG: v_readlane_b32 s5, v40, 1
-; FLATSCR-DAG: v_readlane_b32 s0, v40, 0
-; FLATSCR-DAG: v_readlane_b32 s1, v40, 1
+; GCN-DAG: v_readlane_b32 s30, [[CSR_VGPR]], 0
+; GCN-DAG: v_readlane_b32 s31, [[CSR_VGPR]], 1
 
 ; MUBUF:   s_addk_i32 s32, 0xfc00
 ; FLATSCR: s_add_i32 s32, s32, -16
 ; GCN-NEXT: v_readlane_b32 s33, [[CSR_VGPR]], [[FP_SPILL_LANE]]
 ; GCN-NEXT: s_or_saveexec_b64 [[COPY_EXEC1:s\[[0-9]+:[0-9]+\]]], -1{{$}}
 ; MUBUF-NEXT:   buffer_load_dword [[CSR_VGPR]], off, s[0:3], s32 ; 4-byte Folded Reload
 ; FLATSCR-NEXT: scratch_load_dword [[CSR_VGPR]], off, s32 ; 4-byte Folded Reload
 ; GCN-NEXT: s_mov_b64 exec, [[COPY_EXEC1]]
 ; GCN-NEXT: s_waitcnt vmcnt(0)
-; GCN-NEXT: s_setpc_b64
+; GCN-NEXT: s_setpc_b64 s[30:31]
 define void @callee_no_stack_with_call() #0 {
   call void @external_void_func_void()
   ret void
 }
 
 declare hidden void @external_void_func_void() #0
 
 ; Make sure if a CSR vgpr is used for SGPR spilling, it is saved and

 }
 
 ; 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{{$}}
 ; MUBUF-NEXT:   buffer_store_dword [[CSR_VGPR:v[0-9]+]], off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; FLATSCR-NEXT: scratch_store_dword off, [[CSR_VGPR:v[0-9]+]], s32 offset:4 ; 4-byte Folded Spill
 ; GCN-NEXT: s_mov_b64 exec, [[COPY_EXEC0]]
-; GCN-NEXT: v_writelane_b32 v0, s33, 2
+; GCN-NEXT: v_writelane_b32 [[CSR_VGPR]], s33, 0
 ; GCN-NEXT: s_mov_b32 s33, s32
-; GCN-NEXT: v_writelane_b32 v0, s30, 0
+; MUBUF:        s_addk_i32 s32, 0x300
+; FLATSCR:      s_add_i32 s32, s32, 12
 ; GCN: v_mov_b32_e32 [[ZERO:v[0-9]+]], 0
-; GCN: v_writelane_b32 v0, s31, 1
 ; MUBUF:   buffer_store_dword [[ZERO]], off, s[0:3], s33{{$}}
 ; FLATSCR: scratch_store_dword off, [[ZERO]], s33{{$}}
 ; GCN-NEXT:     s_waitcnt vmcnt(0)
 ; GCN: ;;#ASMSTART
-; MUBUF:        s_addk_i32 s32, 0x300
-; MUBUF-NEXT:   v_readlane_b32 s4, v0, 0
-; MUBUF-NEXT:   v_readlane_b32 s5, v0, 1
-; FLATSCR:      s_add_i32 s32, s32, 12
-; FLATSCR-NEXT: v_readlane_b32 s0, v0, 0
-; FLATSCR-NEXT: v_readlane_b32 s1, v0, 1
-; MUBUF-NEXT:   s_addk_i32 s32, 0xfd00
-; FLATSCR-NEXT: s_add_i32 s32, s32, -12
-; GCN-NEXT:     v_readlane_b32 s33, v0, 2
+; MUBUF:   s_addk_i32 s32, 0xfd00
+; FLATSCR: s_add_i32 s32, s32, -12
+; GCN-NEXT:     v_readlane_b32 s33, [[CSR_VGPR]], 0
 ; GCN-NEXT: s_or_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)
-; MUBUF-NEXT:   s_setpc_b64 s[4:5]
-; FLATSCR-NEXT: s_setpc_b64 s[0:1]
+; 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, i32 addrspace(5)* %alloca
 
   ; Use all clobberable registers, so FP has to spill to a VGPR.
   call void asm sideeffect "",
     "~{s0},~{s1},~{s2},~{s3},~{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}
-    ,~{s30},~{s31}"() #0
+    ,~{s20},~{s21},~{s22},~{s23},~{s24},~{s25},~{s26},~{s27},~{s28},~{s29},~{vcc}"() #0
 
   ret void
 }
 
 ; Need a new CSR VGPR to satisfy the FP spill.
 ; GCN-LABEL: {{^}}no_unused_non_csr_sgpr_for_fp_no_scratch_vgpr:
 ; GCN: s_waitcnt
 ; GCN-NEXT: s_or_saveexec_b64 [[COPY_EXEC0:s\[[0-9]+:[0-9]+\]]], -1{{$}}
 ; MUBUF-NEXT:   buffer_store_dword [[CSR_VGPR:v[0-9]+]], off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; FLATSCR-NEXT: scratch_store_dword off, [[CSR_VGPR:v[0-9]+]], s32 offset:4 ; 4-byte Folded Spill
 ; GCN-NEXT: s_mov_b64 exec, [[COPY_EXEC0]]
-; GCN-NEXT: v_writelane_b32 [[CSR_VGPR]], s33, 2
+; GCN-NEXT: v_writelane_b32 [[CSR_VGPR]], s33, 0
 ; GCN-NEXT: s_mov_b32 s33, s32
-; GCN-NEXT: v_writelane_b32 [[CSR_VGPR]], s30, 0
+; MUBUF:       s_addk_i32 s32, 0x300{{$}}
+; FLATSCR:     s_add_i32 s32, s32, 12{{$}}
 
-; GCN-DAG: v_writelane_b32 [[CSR_VGPR]], s31, 1
 ; MUBUF-DAG:   buffer_store_dword
 ; FLATSCR-DAG: scratch_store_dword
-; MUBUF:       s_addk_i32 s32, 0x300{{$}}
-; FLATSCR:     s_add_i32 s32, s32, 12{{$}}
 
-; MUBUF:        v_readlane_b32 s4, [[CSR_VGPR]], 0
-; FLATSCR:      v_readlane_b32 s0, [[CSR_VGPR]], 0
 ; GCN: ;;#ASMSTART
-; MUBUF:        v_readlane_b32 s5, [[CSR_VGPR]], 1
-; FLATSCR:      v_readlane_b32 s1, [[CSR_VGPR]], 1
-; MUBUF-NEXT:   s_addk_i32 s32, 0xfd00{{$}}
-; FLATSCR-NEXT: s_add_i32 s32, s32, -12{{$}}
-; GCN-NEXT: v_readlane_b32 s33, [[CSR_VGPR]], 2
+; MUBUF:   s_addk_i32 s32, 0xfd00{{$}}
+; FLATSCR: s_add_i32 s32, s32, -12{{$}}
+; GCN-NEXT: v_readlane_b32 s33, [[CSR_VGPR]], 0
 ; GCN-NEXT: s_or_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
+; GCN-NEXT: s_setpc_b64 s[30:31]
 define void @no_unused_non_csr_sgpr_for_fp_no_scratch_vgpr() #1 {
   %alloca = alloca i32, addrspace(5)
   store volatile i32 0, i32 addrspace(5)* %alloca
 
   ; Use all clobberable registers, so FP has to spill to a VGPR.
   call void asm sideeffect "",
     "~{s0},~{s1},~{s2},~{s3},~{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}
-    ,~{s30},~{s31}"() #0
+    ,~{s20},~{s21},~{s22},~{s23},~{s24},~{s25},~{s26},~{s27},~{s28},~{s29},~{vcc}"() #0
 
   call void asm sideeffect "; clobber nonpreserved initial VGPRs",
     "~{v0},~{v1},~{v2},~{v3},~{v4},~{v5},~{v6},~{v7},~{v8},~{v9}
     ,~{v10},~{v11},~{v12},~{v13},~{v14},~{v15},~{v16},~{v17},~{v18},~{v19}
     ,~{v20},~{v21},~{v22},~{v23},~{v24},~{v25},~{v26},~{v27},~{v28},~{v29}
     ,~{v30},~{v31},~{v32},~{v33},~{v34},~{v35},~{v36},~{v37},~{v38},~{v39}"() #1
 
   ret void
 }
 
 ; The byval argument exceeds the MUBUF constant offset, so a scratch
 ; register is needed to access the CSR VGPR slot.
 ; GCN-LABEL: {{^}}scratch_reg_needed_mubuf_offset:
 ; GCN: s_waitcnt
 ; GCN-NEXT: s_or_saveexec_b64 [[COPY_EXEC0:s\[[0-9]+:[0-9]+\]]], -1{{$}}
 ; MUBUF-NEXT: s_add_i32 [[SCRATCH_SGPR:s[0-9]+]], s32, 0x40100
 ; MUBUF-NEXT: buffer_store_dword [[CSR_VGPR:v[0-9]+]], off, s[0:3], [[SCRATCH_SGPR]] ; 4-byte Folded Spill
 ; FLATSCR-NEXT: s_add_i32 [[SCRATCH_SGPR:s[0-9]+]], s32, 0x1004
 ; FLATSCR-NEXT: scratch_store_dword off, [[CSR_VGPR:v[0-9]+]], [[SCRATCH_SGPR]] ; 4-byte Folded Spill
 ; GCN-NEXT: s_mov_b64 exec, [[COPY_EXEC0]]
-; GCN-NEXT: v_writelane_b32 [[CSR_VGPR]], s33, 2
-; GCN-DAG:  v_writelane_b32 [[CSR_VGPR]], s30, 0
+; GCN-NEXT: v_writelane_b32 [[CSR_VGPR]], s33, 0
 ; GCN-DAG:  s_mov_b32 s33, s32
-; GCN-DAG:  v_writelane_b32 [[CSR_VGPR]], s31, 1
 ; MUBUF-DAG:   s_add_i32 s32, s32, 0x40300{{$}}
 ; FLATSCR-DAG: s_addk_i32 s32, 0x100c{{$}}
 ; MUBUF-DAG:   buffer_store_dword
 ; FLATSCR-DAG: scratch_store_dword
 
-; MUBUF:   v_readlane_b32 s4, [[CSR_VGPR]], 0
-; FLATSCR: v_readlane_b32 s0, [[CSR_VGPR]], 0
 ; GCN: ;;#ASMSTART
-; MUBUF:   v_readlane_b32 s5, [[CSR_VGPR]], 1
-; FLATSCR: v_readlane_b32 s1, [[CSR_VGPR]], 1
-; MUBUF-NEXT:   s_add_i32 s32, s32, 0xfffbfd00{{$}}
-; FLATSCR-NEXT: s_addk_i32 s32, 0xeff4{{$}}
-; GCN-NEXT: v_readlane_b32 s33, [[CSR_VGPR]], 2
+; MUBUF:   s_add_i32 s32, s32, 0xfffbfd00{{$}}
+; FLATSCR: s_addk_i32 s32, 0xeff4{{$}}
+; GCN-NEXT: v_readlane_b32 s33, [[CSR_VGPR]], 0
 ; GCN-NEXT: s_or_saveexec_b64 [[COPY_EXEC1:s\[[0-9]+:[0-9]+\]]], -1{{$}}
 ; MUBUF-NEXT: s_add_i32 [[SCRATCH_SGPR:s[0-9]+]], s32, 0x40100
 ; MUBUF-NEXT: buffer_load_dword [[CSR_VGPR]], off, s[0:3], [[SCRATCH_SGPR]] ; 4-byte Folded Reload
 ; FLATSCR-NEXT: s_add_i32 [[SCRATCH_SGPR:s[0-9]+]], s32, 0x1004
 ; FLATSCR-NEXT: scratch_load_dword [[CSR_VGPR]], off, [[SCRATCH_SGPR]] ; 4-byte Folded Reload
 ; GCN-NEXT: s_mov_b64 exec, [[COPY_EXEC1]]
 ; GCN-NEXT: s_waitcnt vmcnt(0)
-; GCN-NEXT: s_setpc_b64
+; GCN-NEXT: s_setpc_b64 s[30:31]
 define void @scratch_reg_needed_mubuf_offset([4096 x i8] addrspace(5)* byval([4096 x i8]) align 4 %arg) #1 {
   %alloca = alloca i32, addrspace(5)
   store volatile i32 0, i32 addrspace(5)* %alloca
 
   ; Use all clobberable registers, so FP has to spill to a VGPR.
   call void asm sideeffect "; clobber nonpreserved SGPRs",
     "~{s0},~{s1},~{s2},~{s3},~{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}
-    ,~{s30},~{s31}"() #0
+    ,~{s20},~{s21},~{s22},~{s23},~{s24},~{s25},~{s26},~{s27},~{s28},~{s29},~{vcc}"() #0
 
   ; Use all clobberable VGPRs, so a CSR spill is needed for the VGPR
   call void asm sideeffect "; clobber nonpreserved VGPRs",
     "~{v0},~{v1},~{v2},~{v3},~{v4},~{v5},~{v6},~{v7},~{v8},~{v9}
     ,~{v10},~{v11},~{v12},~{v13},~{v14},~{v15},~{v16},~{v17},~{v18},~{v19}
     ,~{v20},~{v21},~{v22},~{v23},~{v24},~{v25},~{v26},~{v27},~{v28},~{v29}
     ,~{v30},~{v31},~{v32},~{v33},~{v34},~{v35},~{v36},~{v37},~{v38},~{v39}"() #1
 
   ret void
 }
 
 ; GCN-LABEL: {{^}}local_empty_func:
 ; GCN: s_waitcnt
 ; GCN-NEXT: s_setpc_b64
 define internal void @local_empty_func() #0 {
   ret void
 }
 
 ; An FP is needed, despite not needing any spills
 ; TODO: Ccould see callee does not use stack and omit FP.
 ; GCN-LABEL: {{^}}ipra_call_with_stack:
-; GCN: s_mov_b32 [[FP_COPY:s[0-9]+]], s33
+; GCN: v_writelane_b32 v0, s33, 2
 ; GCN: s_mov_b32 s33, s32
 ; MUBUF:   s_addk_i32 s32, 0x400
 ; FLATSCR: s_add_i32 s32, s32, 16
 ; MUBUF:   buffer_store_dword v{{[0-9]+}}, off, s[0:3], s33{{$}}
 ; FLATSCR: scratch_store_dword off, v{{[0-9]+}}, s33{{$}}
 ; GCN:     s_swappc_b64
 ; MUBUF:   s_addk_i32 s32, 0xfc00
 ; FLATSCR: s_add_i32 s32, s32, -16
-; GCN: s_mov_b32 s33, [[FP_COPY:s[0-9]+]]
+; GCN: v_readlane_b32 s33, v0, 2
 define void @ipra_call_with_stack() #0 {
   %alloca = alloca i32, addrspace(5)
   store volatile i32 0, i32 addrspace(5)* %alloca
   call void @local_empty_func()
   ret void
 }
 
 ; With no free registers, we must spill the FP to memory.

llvm/test/CodeGen/AMDGPU/callee-special-input-sgprs-fixed-abi.ll

define hidden void @use_workgroup_id_yz() #1 {
   ret void
 }
 
 ; Argument is in right place already
 ; GCN-LABEL: {{^}}func_indirect_use_workgroup_id_x:
 ; GCN-NOT: s12
 ; GCN-NOT: s13
 ; GCN-NOT: s14
-; GCN: v_readlane_b32 s4, v40, 0
+; GCN: v_readlane_b32 s30, v40, 0
 define hidden void @func_indirect_use_workgroup_id_x() #1 {
   call void @use_workgroup_id_x()
   ret void
 }
 
 ; Argument is in right place already. We are free to clobber other
 ; SGPR arguments
 ; GCN-LABEL: {{^}}func_indirect_use_workgroup_id_y:

llvm/test/CodeGen/AMDGPU/callee-special-input-sgprs.ll

 define amdgpu_kernel void @kern_indirect_use_workgroup_id_yz() #1 {
   call void @use_workgroup_id_yz()
   ret void
 }
 
 ; Argument is in right place already
 ; GCN-LABEL: {{^}}func_indirect_use_workgroup_id_x:
 ; GCN-NOT: s4
-; GCN: v_readlane_b32 s4, v40, 0
+; GCN: v_readlane_b32 s30, v40, 0
 define hidden void @func_indirect_use_workgroup_id_x() #1 {
   call void @use_workgroup_id_x()
   ret void
 }
 
 ; GCN-LABEL: {{^}}func_indirect_use_workgroup_id_y:
 ; GCN-NOT: s4
-; GCN: v_readlane_b32 s4, v40, 0
+; GCN: v_readlane_b32 s30, v40, 0
 define hidden void @func_indirect_use_workgroup_id_y() #1 {
   call void @use_workgroup_id_y()
   ret void
 }
 
 ; GCN-LABEL: {{^}}func_indirect_use_workgroup_id_z:
 ; GCN-NOT: s4
-; GCN: v_readlane_b32 s4, v40, 0
+; GCN: v_readlane_b32 s30, v40, 0
 define hidden void @func_indirect_use_workgroup_id_z() #1 {
   call void @use_workgroup_id_z()
   ret void
 }
 
 ; GCN-LABEL: {{^}}other_arg_use_workgroup_id_x:
 ; GCN: {{flat|global}}_store_dword v{{\[[0-9]+:[0-9]+\]}}, v0
 ; GCN: ; use s4

llvm/test/CodeGen/AMDGPU/cross-block-use-is-not-abi-copy.ll

 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
 ; GCN-NEXT:    v_writelane_b32 v40, s30, 0
 ; GCN-NEXT:    v_writelane_b32 v40, s31, 1
 ; GCN-NEXT:    s_getpc_b64 s[4:5]
 ; GCN-NEXT:    s_add_u32 s4, s4, func_v2f32@rel32@lo+4
 ; GCN-NEXT:    s_addc_u32 s5, s5, func_v2f32@rel32@hi+12
 ; GCN-NEXT:    s_swappc_b64 s[30:31], s[4:5]
-; GCN-NEXT:    v_readlane_b32 s4, v40, 0
-; GCN-NEXT:    v_readlane_b32 s5, v40, 1
+; GCN-NEXT:    v_readlane_b32 s30, v40, 0
+; GCN-NEXT:    v_readlane_b32 s31, v40, 1
 ; GCN-NEXT:    s_addk_i32 s32, 0xfc00
 ; GCN-NEXT:    v_readlane_b32 s33, v40, 2
-; GCN-NEXT:    s_or_saveexec_b64 s[6:7], -1
+; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
-; GCN-NEXT:    s_mov_b64 exec, s[6:7]
+; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
-; GCN-NEXT:    s_setpc_b64 s[4:5]
+; GCN-NEXT:    s_setpc_b64 s[30:31]
 bb0:
   %split.ret.type = call <2 x float> @func_v2f32()
   br label %bb1
 
 bb1:
   %extract = extractelement <2 x float> %split.ret.type, i32 0
   ret float %extract
 }

 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
 ; GCN-NEXT:    v_writelane_b32 v40, s30, 0
 ; GCN-NEXT:    v_writelane_b32 v40, s31, 1
 ; GCN-NEXT:    s_getpc_b64 s[4:5]
 ; GCN-NEXT:    s_add_u32 s4, s4, func_v3f32@rel32@lo+4
 ; GCN-NEXT:    s_addc_u32 s5, s5, func_v3f32@rel32@hi+12
 ; GCN-NEXT:    s_swappc_b64 s[30:31], s[4:5]
-; GCN-NEXT:    v_readlane_b32 s4, v40, 0
-; GCN-NEXT:    v_readlane_b32 s5, v40, 1
+; GCN-NEXT:    v_readlane_b32 s30, v40, 0
+; GCN-NEXT:    v_readlane_b32 s31, v40, 1
 ; GCN-NEXT:    s_addk_i32 s32, 0xfc00
 ; GCN-NEXT:    v_readlane_b32 s33, v40, 2
-; GCN-NEXT:    s_or_saveexec_b64 s[6:7], -1
+; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
-; GCN-NEXT:    s_mov_b64 exec, s[6:7]
+; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
-; GCN-NEXT:    s_setpc_b64 s[4:5]
+; GCN-NEXT:    s_setpc_b64 s[30:31]
 bb0:
   %split.ret.type = call <3 x float> @func_v3f32()
   br label %bb1
 
 bb1:
   %extract = extractelement <3 x float> %split.ret.type, i32 0
   ret float %extract
 }

 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
 ; GCN-NEXT:    v_writelane_b32 v40, s30, 0
 ; GCN-NEXT:    v_writelane_b32 v40, s31, 1
 ; GCN-NEXT:    s_getpc_b64 s[4:5]
 ; GCN-NEXT:    s_add_u32 s4, s4, func_v4f16@rel32@lo+4
 ; GCN-NEXT:    s_addc_u32 s5, s5, func_v4f16@rel32@hi+12
 ; GCN-NEXT:    s_swappc_b64 s[30:31], s[4:5]
-; GCN-NEXT:    v_readlane_b32 s4, v40, 0
-; GCN-NEXT:    v_readlane_b32 s5, v40, 1
+; GCN-NEXT:    v_readlane_b32 s30, v40, 0
+; GCN-NEXT:    v_readlane_b32 s31, v40, 1
 ; GCN-NEXT:    s_addk_i32 s32, 0xfc00
 ; GCN-NEXT:    v_readlane_b32 s33, v40, 2
-; GCN-NEXT:    s_or_saveexec_b64 s[6:7], -1
+; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
-; GCN-NEXT:    s_mov_b64 exec, s[6:7]
+; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
-; GCN-NEXT:    s_setpc_b64 s[4:5]
+; GCN-NEXT:    s_setpc_b64 s[30:31]
 bb0:
   %split.ret.type = call <4 x half> @func_v4f16()
   br label %bb1
 
 bb1:
   %extract = extractelement <4 x half> %split.ret.type, i32 0
   ret half %extract
 }

 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
 ; GCN-NEXT:    v_writelane_b32 v40, s30, 0
 ; GCN-NEXT:    v_writelane_b32 v40, s31, 1
 ; GCN-NEXT:    s_getpc_b64 s[4:5]
 ; GCN-NEXT:    s_add_u32 s4, s4, func_struct@rel32@lo+4
 ; GCN-NEXT:    s_addc_u32 s5, s5, func_struct@rel32@hi+12
 ; GCN-NEXT:    s_swappc_b64 s[30:31], s[4:5]
-; GCN-NEXT:    v_readlane_b32 s4, v40, 0
+; GCN-NEXT:    v_readlane_b32 s30, v40, 0
 ; GCN-NEXT:    v_mov_b32_e32 v1, v4
-; GCN-NEXT:    v_readlane_b32 s5, v40, 1
+; GCN-NEXT:    v_readlane_b32 s31, v40, 1
 ; GCN-NEXT:    s_addk_i32 s32, 0xfc00
 ; GCN-NEXT:    v_readlane_b32 s33, v40, 2
-; GCN-NEXT:    s_or_saveexec_b64 s[6:7], -1
+; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
-; GCN-NEXT:    s_mov_b64 exec, s[6:7]
+; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
-; GCN-NEXT:    s_setpc_b64 s[4:5]
+; GCN-NEXT:    s_setpc_b64 s[30:31]
 bb0:
   %split.ret.type = call { <4 x i32>, <4 x half> } @func_struct()
   br label %bb1
 
 bb1:
   %val0 = extractvalue { <4 x i32>, <4 x half> } %split.ret.type, 0
   %val1 = extractvalue { <4 x i32>, <4 x half> } %split.ret.type, 1
   %extract0 = extractelement <4 x i32> %val0, i32 0

llvm/test/CodeGen/AMDGPU/fdiv-nofpexcept.ll

 ; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
 ; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 -stop-after=finalize-isel -o - %s | FileCheck -check-prefix=GCN %s
 
 ; Make sure nofpexcept flags are emitted when lowering a
 ; non-constrained fdiv.
 
 define float @fdiv_f32(float %a, float %b) #0 {
   ; GCN-LABEL: name: fdiv_f32
   ; GCN: bb.0.entry:
-  ; GCN-NEXT:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+  ; GCN-NEXT:   liveins: $vgpr0, $vgpr1
   ; GCN-NEXT: {{  $}}
-  ; GCN-NEXT:   [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr30_sgpr31
-  ; GCN-NEXT:   [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
-  ; GCN-NEXT:   [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0
-  ; GCN-NEXT:   %6:vgpr_32, %7:sreg_64 = nofpexcept V_DIV_SCALE_F32_e64 0, [[COPY2]], 0, [[COPY1]], 0, [[COPY2]], 0, 0, implicit $mode, implicit $exec
-  ; GCN-NEXT:   %8:vgpr_32, %9:sreg_64 = nofpexcept V_DIV_SCALE_F32_e64 0, [[COPY1]], 0, [[COPY1]], 0, [[COPY2]], 0, 0, implicit $mode, implicit $exec
-  ; GCN-NEXT:   %10:vgpr_32 = nofpexcept V_RCP_F32_e64 0, %8, 0, 0, implicit $mode, implicit $exec
+  ; GCN-NEXT:   [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+  ; GCN-NEXT:   [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+  ; GCN-NEXT:   %4:vgpr_32, %5:sreg_64 = nofpexcept V_DIV_SCALE_F32_e64 0, [[COPY1]], 0, [[COPY]], 0, [[COPY1]], 0, 0, implicit $mode, implicit $exec
+  ; GCN-NEXT:   %6:vgpr_32, %7:sreg_64 = nofpexcept V_DIV_SCALE_F32_e64 0, [[COPY]], 0, [[COPY]], 0, [[COPY1]], 0, 0, implicit $mode, implicit $exec
+  ; GCN-NEXT:   %8:vgpr_32 = nofpexcept V_RCP_F32_e64 0, %6, 0, 0, implicit $mode, implicit $exec
   ; GCN-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 3
   ; GCN-NEXT:   [[S_MOV_B32_1:%[0-9]+]]:sgpr_32 = S_MOV_B32 1065353216
   ; GCN-NEXT:   [[S_MOV_B32_2:%[0-9]+]]:sreg_32 = S_MOV_B32 0
   ; GCN-NEXT:   S_SETREG_B32_mode killed [[S_MOV_B32_]], 2305, implicit-def $mode, implicit $mode
-  ; GCN-NEXT:   %14:vgpr_32 = nofpexcept V_FMA_F32_e64 1, %8, 0, %10, 0, killed [[S_MOV_B32_1]], 0, 0, implicit $mode, implicit $exec
-  ; GCN-NEXT:   %15:vgpr_32 = nofpexcept V_FMA_F32_e64 0, killed %14, 0, %10, 0, %10, 0, 0, implicit $mode, implicit $exec
-  ; GCN-NEXT:   %16:vgpr_32 = nofpexcept V_MUL_F32_e64 0, %6, 0, %15, 0, 0, implicit $mode, implicit $exec
-  ; GCN-NEXT:   %17:vgpr_32 = nofpexcept V_FMA_F32_e64 1, %8, 0, %16, 0, %6, 0, 0, implicit $mode, implicit $exec
-  ; GCN-NEXT:   %18:vgpr_32 = nofpexcept V_FMA_F32_e64 0, killed %17, 0, %15, 0, %16, 0, 0, implicit $mode, implicit $exec
-  ; GCN-NEXT:   %19:vgpr_32 = nofpexcept V_FMA_F32_e64 1, %8, 0, %18, 0, %6, 0, 0, implicit $mode, implicit $exec
+  ; GCN-NEXT:   %12:vgpr_32 = nofpexcept V_FMA_F32_e64 1, %6, 0, %8, 0, killed [[S_MOV_B32_1]], 0, 0, implicit $mode, implicit $exec
+  ; GCN-NEXT:   %13:vgpr_32 = nofpexcept V_FMA_F32_e64 0, killed %12, 0, %8, 0, %8, 0, 0, implicit $mode, implicit $exec
+  ; GCN-NEXT:   %14:vgpr_32 = nofpexcept V_MUL_F32_e64 0, %4, 0, %13, 0, 0, implicit $mode, implicit $exec
+  ; GCN-NEXT:   %15:vgpr_32 = nofpexcept V_FMA_F32_e64 1, %6, 0, %14, 0, %4, 0, 0, implicit $mode, implicit $exec
+  ; GCN-NEXT:   %16:vgpr_32 = nofpexcept V_FMA_F32_e64 0, killed %15, 0, %13, 0, %14, 0, 0, implicit $mode, implicit $exec
+  ; GCN-NEXT:   %17:vgpr_32 = nofpexcept V_FMA_F32_e64 1, %6, 0, %16, 0, %4, 0, 0, implicit $mode, implicit $exec
   ; GCN-NEXT:   S_SETREG_B32_mode killed [[S_MOV_B32_2]], 2305, implicit-def dead $mode, implicit $mode
-  ; GCN-NEXT:   $vcc = COPY %7
-  ; GCN-NEXT:   %20:vgpr_32 = nofpexcept V_DIV_FMAS_F32_e64 0, killed %19, 0, %15, 0, %18, 0, 0, implicit $mode, implicit $vcc, implicit $exec
-  ; GCN-NEXT:   %21:vgpr_32 = nofpexcept V_DIV_FIXUP_F32_e64 0, killed %20, 0, [[COPY1]], 0, [[COPY2]], 0, 0, implicit $mode, implicit $exec
-  ; GCN-NEXT:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; GCN-NEXT:   $vgpr0 = COPY %21
-  ; GCN-NEXT:   [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-  ; GCN-NEXT:   S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+  ; GCN-NEXT:   $vcc = COPY %5
+  ; GCN-NEXT:   %18:vgpr_32 = nofpexcept V_DIV_FMAS_F32_e64 0, killed %17, 0, %13, 0, %16, 0, 0, implicit $mode, implicit $vcc, implicit $exec
+  ; GCN-NEXT:   %19:vgpr_32 = nofpexcept V_DIV_FIXUP_F32_e64 0, killed %18, 0, [[COPY]], 0, [[COPY1]], 0, 0, implicit $mode, implicit $exec
+  ; GCN-NEXT:   $vgpr0 = COPY %19
+  ; GCN-NEXT:   SI_RETURN implicit $vgpr0
 entry:
   %fdiv = fdiv float %a, %b
   ret float %fdiv
 }
 
 define float @fdiv_nnan_f32(float %a, float %b) #0 {
   ; GCN-LABEL: name: fdiv_nnan_f32
   ; GCN: bb.0.entry:
-  ; GCN-NEXT:   liveins: $vgpr0, $vgpr1, $sgpr30_sgpr31
+  ; GCN-NEXT:   liveins: $vgpr0, $vgpr1
   ; GCN-NEXT: {{  $}}
-  ; GCN-NEXT:   [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr30_sgpr31
-  ; GCN-NEXT:   [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
-  ; GCN-NEXT:   [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0
-  ; GCN-NEXT:   %6:vgpr_32, %7:sreg_64 = nnan nofpexcept V_DIV_SCALE_F32_e64 0, [[COPY2]], 0, [[COPY1]], 0, [[COPY2]], 0, 0, implicit $mode, implicit $exec
-  ; GCN-NEXT:   %8:vgpr_32, %9:sreg_64 = nnan nofpexcept V_DIV_SCALE_F32_e64 0, [[COPY1]], 0, [[COPY1]], 0, [[COPY2]], 0, 0, implicit $mode, implicit $exec
-  ; GCN-NEXT:   %10:vgpr_32 = nnan nofpexcept V_RCP_F32_e64 0, %8, 0, 0, implicit $mode, implicit $exec
+  ; GCN-NEXT:   [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+  ; GCN-NEXT:   [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+  ; GCN-NEXT:   %4:vgpr_32, %5:sreg_64 = nnan nofpexcept V_DIV_SCALE_F32_e64 0, [[COPY1]], 0, [[COPY]], 0, [[COPY1]], 0, 0, implicit $mode, implicit $exec
+  ; GCN-NEXT:   %6:vgpr_32, %7:sreg_64 = nnan nofpexcept V_DIV_SCALE_F32_e64 0, [[COPY]], 0, [[COPY]], 0, [[COPY1]], 0, 0, implicit $mode, implicit $exec
+  ; GCN-NEXT:   %8:vgpr_32 = nnan nofpexcept V_RCP_F32_e64 0, %6, 0, 0, implicit $mode, implicit $exec
   ; GCN-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 3
   ; GCN-NEXT:   [[S_MOV_B32_1:%[0-9]+]]:sgpr_32 = S_MOV_B32 1065353216
   ; GCN-NEXT:   [[S_MOV_B32_2:%[0-9]+]]:sreg_32 = S_MOV_B32 0
   ; GCN-NEXT:   S_SETREG_B32_mode killed [[S_MOV_B32_]], 2305, implicit-def $mode, implicit $mode
-  ; GCN-NEXT:   %14:vgpr_32 = nnan nofpexcept V_FMA_F32_e64 1, %8, 0, %10, 0, killed [[S_MOV_B32_1]], 0, 0, implicit $mode, implicit $exec
-  ; GCN-NEXT:   %15:vgpr_32 = nnan nofpexcept V_FMA_F32_e64 0, killed %14, 0, %10, 0, %10, 0, 0, implicit $mode, implicit $exec
-  ; GCN-NEXT:   %16:vgpr_32 = nnan nofpexcept V_MUL_F32_e64 0, %6, 0, %15, 0, 0, implicit $mode, implicit $exec
-  ; GCN-NEXT:   %17:vgpr_32 = nnan nofpexcept V_FMA_F32_e64 1, %8, 0, %16, 0, %6, 0, 0, implicit $mode, implicit $exec
-  ; GCN-NEXT:   %18:vgpr_32 = nnan nofpexcept V_FMA_F32_e64 0, killed %17, 0, %15, 0, %16, 0, 0, implicit $mode, implicit $exec
-  ; GCN-NEXT:   %19:vgpr_32 = nnan nofpexcept V_FMA_F32_e64 1, %8, 0, %18, 0, %6, 0, 0, implicit $mode, implicit $exec
+  ; GCN-NEXT:   %12:vgpr_32 = nnan nofpexcept V_FMA_F32_e64 1, %6, 0, %8, 0, killed [[S_MOV_B32_1]], 0, 0, implicit $mode, implicit $exec
+  ; GCN-NEXT:   %13:vgpr_32 = nnan nofpexcept V_FMA_F32_e64 0, killed %12, 0, %8, 0, %8, 0, 0, implicit $mode, implicit $exec
+  ; GCN-NEXT:   %14:vgpr_32 = nnan nofpexcept V_MUL_F32_e64 0, %4, 0, %13, 0, 0, implicit $mode, implicit $exec
+  ; GCN-NEXT:   %15:vgpr_32 = nnan nofpexcept V_FMA_F32_e64 1, %6, 0, %14, 0, %4, 0, 0, implicit $mode, implicit $exec
+  ; GCN-NEXT:   %16:vgpr_32 = nnan nofpexcept V_FMA_F32_e64 0, killed %15, 0, %13, 0, %14, 0, 0, implicit $mode, implicit $exec
+  ; GCN-NEXT:   %17:vgpr_32 = nnan nofpexcept V_FMA_F32_e64 1, %6, 0, %16, 0, %4, 0, 0, implicit $mode, implicit $exec
   ; GCN-NEXT:   S_SETREG_B32_mode killed [[S_MOV_B32_2]], 2305, implicit-def dead $mode, implicit $mode
-  ; GCN-NEXT:   $vcc = COPY %7
-  ; GCN-NEXT:   %20:vgpr_32 = nnan nofpexcept V_DIV_FMAS_F32_e64 0, killed %19, 0, %15, 0, %18, 0, 0, implicit $mode, implicit $vcc, implicit $exec
-  ; GCN-NEXT:   %21:vgpr_32 = nnan nofpexcept V_DIV_FIXUP_F32_e64 0, killed %20, 0, [[COPY1]], 0, [[COPY2]], 0, 0, implicit $mode, implicit $exec
-  ; GCN-NEXT:   [[COPY3:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
-  ; GCN-NEXT:   $vgpr0 = COPY %21
-  ; GCN-NEXT:   [[COPY4:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY3]]
-  ; GCN-NEXT:   S_SETPC_B64_return [[COPY4]], implicit $vgpr0
+  ; GCN-NEXT:   $vcc = COPY %5
+  ; GCN-NEXT:   %18:vgpr_32 = nnan nofpexcept V_DIV_FMAS_F32_e64 0, killed %17, 0, %13, 0, %16, 0, 0, implicit $mode, implicit $vcc, implicit $exec
+  ; GCN-NEXT:   %19:vgpr_32 = nnan nofpexcept V_DIV_FIXUP_F32_e64 0, killed %18, 0, [[COPY]], 0, [[COPY1]], 0, 0, implicit $mode, implicit $exec
+  ; GCN-NEXT:   $vgpr0 = COPY %19
+  ; GCN-NEXT:   SI_RETURN implicit $vgpr0
 entry:
   %fdiv = fdiv nnan float %a, %b
   ret float %fdiv
 }
 
 attributes #0 = { nounwind "denormal-fp-math-f32"="preserve-sign,preserve-sign" }

llvm/test/CodeGen/AMDGPU/fix-frame-ptr-reg-copy-livein.ll

 ; RUN: llc -march=amdgcn -mcpu=gfx900 -stop-after=prologepilog -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
 
 ; It is a small loop test that iterates over the array member of the structure argument  passed byval to the function.
 ; The loop code will keep the prologue and epilogue blocks apart.
 ; The test is primarily to check the temp register used to preserve the earlier FP value
 ; is live-in at every BB in the function.
 
 %struct.Data = type { [20 x i32] }
 
 define i32 @fp_save_restore_in_temp_sgpr(%struct.Data addrspace(5)* nocapture readonly byval(%struct.Data) align 4 %arg) #0 {
   ; GCN-LABEL: name: fp_save_restore_in_temp_sgpr
   ; GCN: bb.0.begin:
-  ; GCN:   liveins: $sgpr11, $sgpr30_sgpr31
+  ; GCN:   liveins: $sgpr11
   ; GCN:   $sgpr11 = frame-setup COPY $sgpr33
   ; GCN:   $sgpr33 = frame-setup COPY $sgpr32
   ; GCN: bb.1.lp_end:
-  ; GCN:   liveins: $sgpr10, $sgpr11, $vgpr1, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr30_sgpr31
+  ; GCN:   liveins: $sgpr10, $sgpr11, $vgpr1, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9
   ; GCN: bb.2.lp_begin:
-  ; GCN:   liveins: $sgpr10, $sgpr11, $vgpr1, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr30_sgpr31
+  ; GCN:   liveins: $sgpr10, $sgpr11, $vgpr1, $sgpr4_sgpr5, $sgpr6_sgpr7
   ; GCN: bb.3.Flow:
-  ; GCN:   liveins: $sgpr10, $sgpr11, $vgpr0, $vgpr1, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr30_sgpr31
+  ; GCN:   liveins: $sgpr10, $sgpr11, $vgpr0, $vgpr1, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9
   ; GCN: bb.4.end:
-  ; GCN:   liveins: $sgpr11, $vgpr0, $sgpr4_sgpr5, $sgpr30_sgpr31
+  ; GCN:   liveins: $sgpr11, $vgpr0, $sgpr4_sgpr5
   ; GCN:   $sgpr33 = frame-destroy COPY $sgpr11
 begin:
   br label %lp_begin
 
 lp_end:                                                ; preds = %lp_begin
   %cur_idx = add nuw nsw i32 %idx, 1
   %lp_term_cond = icmp eq i32 %cur_idx, 20
   br i1 %lp_term_cond, label %end, label %lp_begin

llvm/test/CodeGen/AMDGPU/fpow.ll

 ; GFX8-NEXT:    v_cvt_f32_f16_sdwa v2, v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX8-NEXT:    v_cvt_f32_f16_e32 v0, v0
 ; GFX8-NEXT:    v_cvt_f32_f16_sdwa v3, v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX8-NEXT:    v_cvt_f32_f16_e32 v1, v1
 ; GFX8-NEXT:    v_log_f32_e32 v2, v2
 ; GFX8-NEXT:    v_log_f32_e32 v0, v0
 ; GFX8-NEXT:    v_mul_legacy_f32_e32 v2, v3, v2
 ; GFX8-NEXT:    v_mul_legacy_f32_e32 v0, v1, v0
-; GFX8-NEXT:    v_exp_f32_e32 v2, v2
+; GFX8-NEXT:    v_exp_f32_e32 v1, v2
 ; GFX8-NEXT:    v_exp_f32_e32 v0, v0
-; GFX8-NEXT:    v_cvt_f16_f32_sdwa v1, v2 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD
+; GFX8-NEXT:    v_cvt_f16_f32_sdwa v1, v1 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD
 ; GFX8-NEXT:    v_cvt_f16_f32_e32 v0, v0
 ; GFX8-NEXT:    v_or_b32_e32 v0, v0, v1
 ; GFX8-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX9-LABEL: v_pow_v2f16:
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    v_cvt_f32_f16_sdwa v2, v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX9-NEXT:    v_cvt_f32_f16_e32 v0, v0
 ; GFX9-NEXT:    v_cvt_f32_f16_sdwa v3, v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX9-NEXT:    v_cvt_f32_f16_e32 v1, v1
 ; GFX9-NEXT:    v_log_f32_e32 v2, v2
 ; GFX9-NEXT:    v_log_f32_e32 v0, v0
 ; GFX9-NEXT:    v_mul_legacy_f32_e32 v2, v3, v2
 ; GFX9-NEXT:    v_mul_legacy_f32_e32 v0, v1, v0
-; GFX9-NEXT:    v_exp_f32_e32 v2, v2
+; GFX9-NEXT:    v_exp_f32_e32 v1, v2
 ; GFX9-NEXT:    v_exp_f32_e32 v0, v0
-; GFX9-NEXT:    v_cvt_f16_f32_e32 v1, v2
+; GFX9-NEXT:    v_cvt_f16_f32_e32 v1, v1
 ; GFX9-NEXT:    v_cvt_f16_f32_e32 v0, v0
 ; GFX9-NEXT:    v_pack_b32_f16 v0, v0, v1
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX90A-LABEL: v_pow_v2f16:
 ; GFX90A:       ; %bb.0:
 ; GFX90A-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX90A-NEXT:    v_cvt_f32_f16_sdwa v2, v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX90A-NEXT:    v_cvt_f32_f16_e32 v0, v0
 ; GFX90A-NEXT:    v_cvt_f32_f16_sdwa v3, v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX90A-NEXT:    v_cvt_f32_f16_e32 v1, v1
 ; GFX90A-NEXT:    v_log_f32_e32 v2, v2
 ; GFX90A-NEXT:    v_log_f32_e32 v0, v0
 ; GFX90A-NEXT:    v_mul_legacy_f32 v2, v3, v2
 ; GFX90A-NEXT:    v_mul_legacy_f32 v0, v1, v0
-; GFX90A-NEXT:    v_exp_f32_e32 v2, v2
+; GFX90A-NEXT:    v_exp_f32_e32 v1, v2
 ; GFX90A-NEXT:    v_exp_f32_e32 v0, v0
-; GFX90A-NEXT:    v_cvt_f16_f32_e32 v1, v2
+; GFX90A-NEXT:    v_cvt_f16_f32_e32 v1, v1
 ; GFX90A-NEXT:    v_cvt_f16_f32_e32 v0, v0
 ; GFX90A-NEXT:    v_pack_b32_f16 v0, v0, v1
 ; GFX90A-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: v_pow_v2f16:
 ; GFX10:       ; %bb.0:
 ; GFX10-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0

 ; GFX8-NEXT:    v_cvt_f32_f16_sdwa v2, -v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX8-NEXT:    v_cvt_f32_f16_e64 v0, -v0
 ; GFX8-NEXT:    v_cvt_f32_f16_sdwa v3, v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX8-NEXT:    v_cvt_f32_f16_e32 v1, v1
 ; GFX8-NEXT:    v_log_f32_e32 v2, v2
 ; GFX8-NEXT:    v_log_f32_e32 v0, v0
 ; GFX8-NEXT:    v_mul_legacy_f32_e32 v2, v3, v2
 ; GFX8-NEXT:    v_mul_legacy_f32_e32 v0, v1, v0
-; GFX8-NEXT:    v_exp_f32_e32 v2, v2
+; GFX8-NEXT:    v_exp_f32_e32 v1, v2
 ; GFX8-NEXT:    v_exp_f32_e32 v0, v0
-; GFX8-NEXT:    v_cvt_f16_f32_sdwa v1, v2 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD
+; GFX8-NEXT:    v_cvt_f16_f32_sdwa v1, v1 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD
 ; GFX8-NEXT:    v_cvt_f16_f32_e32 v0, v0
 ; GFX8-NEXT:    v_or_b32_e32 v0, v0, v1
 ; GFX8-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX9-LABEL: v_pow_v2f16_fneg_lhs:
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    v_cvt_f32_f16_sdwa v2, -v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX9-NEXT:    v_cvt_f32_f16_e64 v0, -v0
 ; GFX9-NEXT:    v_cvt_f32_f16_sdwa v3, v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX9-NEXT:    v_cvt_f32_f16_e32 v1, v1
 ; GFX9-NEXT:    v_log_f32_e32 v2, v2
 ; GFX9-NEXT:    v_log_f32_e32 v0, v0
 ; GFX9-NEXT:    v_mul_legacy_f32_e32 v2, v3, v2
 ; GFX9-NEXT:    v_mul_legacy_f32_e32 v0, v1, v0
-; GFX9-NEXT:    v_exp_f32_e32 v2, v2
+; GFX9-NEXT:    v_exp_f32_e32 v1, v2
 ; GFX9-NEXT:    v_exp_f32_e32 v0, v0
-; GFX9-NEXT:    v_cvt_f16_f32_e32 v1, v2
+; GFX9-NEXT:    v_cvt_f16_f32_e32 v1, v1
 ; GFX9-NEXT:    v_cvt_f16_f32_e32 v0, v0
 ; GFX9-NEXT:    v_pack_b32_f16 v0, v0, v1
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX90A-LABEL: v_pow_v2f16_fneg_lhs:
 ; GFX90A:       ; %bb.0:
 ; GFX90A-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX90A-NEXT:    v_cvt_f32_f16_sdwa v2, -v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX90A-NEXT:    v_cvt_f32_f16_e64 v0, -v0
 ; GFX90A-NEXT:    v_cvt_f32_f16_sdwa v3, v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX90A-NEXT:    v_cvt_f32_f16_e32 v1, v1
 ; GFX90A-NEXT:    v_log_f32_e32 v2, v2
 ; GFX90A-NEXT:    v_log_f32_e32 v0, v0
 ; GFX90A-NEXT:    v_mul_legacy_f32 v2, v3, v2
 ; GFX90A-NEXT:    v_mul_legacy_f32 v0, v1, v0
-; GFX90A-NEXT:    v_exp_f32_e32 v2, v2
+; GFX90A-NEXT:    v_exp_f32_e32 v1, v2
 ; GFX90A-NEXT:    v_exp_f32_e32 v0, v0
-; GFX90A-NEXT:    v_cvt_f16_f32_e32 v1, v2
+; GFX90A-NEXT:    v_cvt_f16_f32_e32 v1, v1
 ; GFX90A-NEXT:    v_cvt_f16_f32_e32 v0, v0
 ; GFX90A-NEXT:    v_pack_b32_f16 v0, v0, v1
 ; GFX90A-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: v_pow_v2f16_fneg_lhs:
 ; GFX10:       ; %bb.0:
 ; GFX10-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0

 ; GFX8-NEXT:    v_cvt_f32_f16_sdwa v2, v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX8-NEXT:    v_cvt_f32_f16_e32 v0, v0
 ; GFX8-NEXT:    v_cvt_f32_f16_sdwa v3, -v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX8-NEXT:    v_cvt_f32_f16_e64 v1, -v1
 ; GFX8-NEXT:    v_log_f32_e32 v2, v2
 ; GFX8-NEXT:    v_log_f32_e32 v0, v0
 ; GFX8-NEXT:    v_mul_legacy_f32_e32 v2, v3, v2
 ; GFX8-NEXT:    v_mul_legacy_f32_e32 v0, v1, v0
-; GFX8-NEXT:    v_exp_f32_e32 v2, v2
+; GFX8-NEXT:    v_exp_f32_e32 v1, v2
 ; GFX8-NEXT:    v_exp_f32_e32 v0, v0
-; GFX8-NEXT:    v_cvt_f16_f32_sdwa v1, v2 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD
+; GFX8-NEXT:    v_cvt_f16_f32_sdwa v1, v1 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD
 ; GFX8-NEXT:    v_cvt_f16_f32_e32 v0, v0
 ; GFX8-NEXT:    v_or_b32_e32 v0, v0, v1
 ; GFX8-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX9-LABEL: v_pow_v2f16_fneg_rhs:
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    v_cvt_f32_f16_sdwa v2, v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX9-NEXT:    v_cvt_f32_f16_e32 v0, v0
 ; GFX9-NEXT:    v_cvt_f32_f16_sdwa v3, -v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX9-NEXT:    v_cvt_f32_f16_e64 v1, -v1
 ; GFX9-NEXT:    v_log_f32_e32 v2, v2
 ; GFX9-NEXT:    v_log_f32_e32 v0, v0
 ; GFX9-NEXT:    v_mul_legacy_f32_e32 v2, v3, v2
 ; GFX9-NEXT:    v_mul_legacy_f32_e32 v0, v1, v0
-; GFX9-NEXT:    v_exp_f32_e32 v2, v2
+; GFX9-NEXT:    v_exp_f32_e32 v1, v2
 ; GFX9-NEXT:    v_exp_f32_e32 v0, v0
-; GFX9-NEXT:    v_cvt_f16_f32_e32 v1, v2
+; GFX9-NEXT:    v_cvt_f16_f32_e32 v1, v1
 ; GFX9-NEXT:    v_cvt_f16_f32_e32 v0, v0
 ; GFX9-NEXT:    v_pack_b32_f16 v0, v0, v1
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX90A-LABEL: v_pow_v2f16_fneg_rhs:
 ; GFX90A:       ; %bb.0:
 ; GFX90A-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX90A-NEXT:    v_cvt_f32_f16_sdwa v2, v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX90A-NEXT:    v_cvt_f32_f16_e32 v0, v0
 ; GFX90A-NEXT:    v_cvt_f32_f16_sdwa v3, -v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX90A-NEXT:    v_cvt_f32_f16_e64 v1, -v1
 ; GFX90A-NEXT:    v_log_f32_e32 v2, v2
 ; GFX90A-NEXT:    v_log_f32_e32 v0, v0
 ; GFX90A-NEXT:    v_mul_legacy_f32 v2, v3, v2
 ; GFX90A-NEXT:    v_mul_legacy_f32 v0, v1, v0
-; GFX90A-NEXT:    v_exp_f32_e32 v2, v2
+; GFX90A-NEXT:    v_exp_f32_e32 v1, v2
 ; GFX90A-NEXT:    v_exp_f32_e32 v0, v0
-; GFX90A-NEXT:    v_cvt_f16_f32_e32 v1, v2
+; GFX90A-NEXT:    v_cvt_f16_f32_e32 v1, v1
 ; GFX90A-NEXT:    v_cvt_f16_f32_e32 v0, v0
 ; GFX90A-NEXT:    v_pack_b32_f16 v0, v0, v1
 ; GFX90A-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: v_pow_v2f16_fneg_rhs:
 ; GFX10:       ; %bb.0:
 ; GFX10-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0

 ; GFX8-NEXT:    v_cvt_f32_f16_sdwa v2, -v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX8-NEXT:    v_cvt_f32_f16_e64 v0, -v0
 ; GFX8-NEXT:    v_cvt_f32_f16_sdwa v3, -v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX8-NEXT:    v_cvt_f32_f16_e64 v1, -v1
 ; GFX8-NEXT:    v_log_f32_e32 v2, v2
 ; GFX8-NEXT:    v_log_f32_e32 v0, v0
 ; GFX8-NEXT:    v_mul_legacy_f32_e32 v2, v3, v2
 ; GFX8-NEXT:    v_mul_legacy_f32_e32 v0, v1, v0
-; GFX8-NEXT:    v_exp_f32_e32 v2, v2
+; GFX8-NEXT:    v_exp_f32_e32 v1, v2
 ; GFX8-NEXT:    v_exp_f32_e32 v0, v0
-; GFX8-NEXT:    v_cvt_f16_f32_sdwa v1, v2 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD
+; GFX8-NEXT:    v_cvt_f16_f32_sdwa v1, v1 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD
 ; GFX8-NEXT:    v_cvt_f16_f32_e32 v0, v0
 ; GFX8-NEXT:    v_or_b32_e32 v0, v0, v1
 ; GFX8-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX9-LABEL: v_pow_v2f16_fneg_lhs_rhs:
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    v_cvt_f32_f16_sdwa v2, -v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX9-NEXT:    v_cvt_f32_f16_e64 v0, -v0
 ; GFX9-NEXT:    v_cvt_f32_f16_sdwa v3, -v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX9-NEXT:    v_cvt_f32_f16_e64 v1, -v1
 ; GFX9-NEXT:    v_log_f32_e32 v2, v2
 ; GFX9-NEXT:    v_log_f32_e32 v0, v0
 ; GFX9-NEXT:    v_mul_legacy_f32_e32 v2, v3, v2
 ; GFX9-NEXT:    v_mul_legacy_f32_e32 v0, v1, v0
-; GFX9-NEXT:    v_exp_f32_e32 v2, v2
+; GFX9-NEXT:    v_exp_f32_e32 v1, v2
 ; GFX9-NEXT:    v_exp_f32_e32 v0, v0
-; GFX9-NEXT:    v_cvt_f16_f32_e32 v1, v2
+; GFX9-NEXT:    v_cvt_f16_f32_e32 v1, v1
 ; GFX9-NEXT:    v_cvt_f16_f32_e32 v0, v0
 ; GFX9-NEXT:    v_pack_b32_f16 v0, v0, v1
 ; GFX9-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX90A-LABEL: v_pow_v2f16_fneg_lhs_rhs:
 ; GFX90A:       ; %bb.0:
 ; GFX90A-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX90A-NEXT:    v_cvt_f32_f16_sdwa v2, -v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX90A-NEXT:    v_cvt_f32_f16_e64 v0, -v0
 ; GFX90A-NEXT:    v_cvt_f32_f16_sdwa v3, -v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
 ; GFX90A-NEXT:    v_cvt_f32_f16_e64 v1, -v1
 ; GFX90A-NEXT:    v_log_f32_e32 v2, v2
 ; GFX90A-NEXT:    v_log_f32_e32 v0, v0
 ; GFX90A-NEXT:    v_mul_legacy_f32 v2, v3, v2
 ; GFX90A-NEXT:    v_mul_legacy_f32 v0, v1, v0
-; GFX90A-NEXT:    v_exp_f32_e32 v2, v2
+; GFX90A-NEXT:    v_exp_f32_e32 v1, v2
 ; GFX90A-NEXT:    v_exp_f32_e32 v0, v0
-; GFX90A-NEXT:    v_cvt_f16_f32_e32 v1, v2
+; GFX90A-NEXT:    v_cvt_f16_f32_e32 v1, v1
 ; GFX90A-NEXT:    v_cvt_f16_f32_e32 v0, v0
 ; GFX90A-NEXT:    v_pack_b32_f16 v0, v0, v1
 ; GFX90A-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: v_pow_v2f16_fneg_lhs_rhs:
 ; GFX10:       ; %bb.0:
 ; GFX10-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0

llvm/test/CodeGen/AMDGPU/frame-setup-without-sgpr-to-vgpr-spills.ll

 ; SPILL-TO-VGPR-NEXT:    v_mov_b32_e32 v0, 0
 ; SPILL-TO-VGPR-NEXT:    v_writelane_b32 v40, s31, 1
 ; SPILL-TO-VGPR-NEXT:    buffer_store_dword v0, off, s[0:3], s33
 ; SPILL-TO-VGPR-NEXT:    s_waitcnt vmcnt(0)
 ; SPILL-TO-VGPR-NEXT:    s_getpc_b64 s[4:5]
 ; SPILL-TO-VGPR-NEXT:    s_add_u32 s4, s4, external_void_func_void@rel32@lo+4
 ; SPILL-TO-VGPR-NEXT:    s_addc_u32 s5, s5, external_void_func_void@rel32@hi+12
 ; SPILL-TO-VGPR-NEXT:    s_swappc_b64 s[30:31], s[4:5]
-; SPILL-TO-VGPR-NEXT:    v_readlane_b32 s4, v40, 0
-; SPILL-TO-VGPR-NEXT:    v_readlane_b32 s5, v40, 1
+; SPILL-TO-VGPR-NEXT:    v_readlane_b32 s30, v40, 0
+; SPILL-TO-VGPR-NEXT:    v_readlane_b32 s31, v40, 1
 ; SPILL-TO-VGPR-NEXT:    s_addk_i32 s32, 0xfc00
 ; SPILL-TO-VGPR-NEXT:    v_readlane_b32 s33, v40, 2
-; SPILL-TO-VGPR-NEXT:    s_or_saveexec_b64 s[6:7], -1
+; SPILL-TO-VGPR-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; SPILL-TO-VGPR-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
-; SPILL-TO-VGPR-NEXT:    s_mov_b64 exec, s[6:7]
+; SPILL-TO-VGPR-NEXT:    s_mov_b64 exec, s[4:5]
 ; SPILL-TO-VGPR-NEXT:    s_waitcnt vmcnt(0)
-; SPILL-TO-VGPR-NEXT:    s_setpc_b64 s[4:5]
+; SPILL-TO-VGPR-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; NO-SPILL-TO-VGPR-LABEL: callee_with_stack_and_call:
 ; NO-SPILL-TO-VGPR:       ; %bb.0:
 ; NO-SPILL-TO-VGPR-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; NO-SPILL-TO-VGPR-NEXT:    v_mov_b32_e32 v0, s33
 ; NO-SPILL-TO-VGPR-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:12 ; 4-byte Folded Spill
 ; NO-SPILL-TO-VGPR-NEXT:    s_mov_b32 s33, s32
 ; NO-SPILL-TO-VGPR-NEXT:    s_addk_i32 s32, 0x800
 ; NO-SPILL-TO-VGPR-NEXT:    s_mov_b64 s[6:7], exec
 ; NO-SPILL-TO-VGPR-NEXT:    s_mov_b64 exec, 3
 ; NO-SPILL-TO-VGPR-NEXT:    buffer_store_dword v1, off, s[0:3], s33 offset:16
 ; NO-SPILL-TO-VGPR-NEXT:    v_writelane_b32 v1, s30, 0
 ; NO-SPILL-TO-VGPR-NEXT:    v_writelane_b32 v1, s31, 1
 ; NO-SPILL-TO-VGPR-NEXT:    buffer_store_dword v1, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; NO-SPILL-TO-VGPR-NEXT:    buffer_load_dword v1, off, s[0:3], s33 offset:16
 ; NO-SPILL-TO-VGPR-NEXT:    s_waitcnt vmcnt(0)
 ; NO-SPILL-TO-VGPR-NEXT:    s_mov_b64 exec, s[6:7]
 ; NO-SPILL-TO-VGPR-NEXT:    v_mov_b32_e32 v0, 0
 ; NO-SPILL-TO-VGPR-NEXT:    buffer_store_dword v0, off, s[0:3], s33
 ; NO-SPILL-TO-VGPR-NEXT:    s_waitcnt vmcnt(0)
 ; NO-SPILL-TO-VGPR-NEXT:    s_getpc_b64 s[4:5]
 ; NO-SPILL-TO-VGPR-NEXT:    s_add_u32 s4, s4, external_void_func_void@rel32@lo+4
 ; NO-SPILL-TO-VGPR-NEXT:    s_addc_u32 s5, s5, external_void_func_void@rel32@hi+12
 ; NO-SPILL-TO-VGPR-NEXT:    s_swappc_b64 s[30:31], s[4:5]
-; NO-SPILL-TO-VGPR-NEXT:    s_mov_b64 s[6:7], exec
+; NO-SPILL-TO-VGPR-NEXT:    s_mov_b64 s[4:5], exec
 ; NO-SPILL-TO-VGPR-NEXT:    s_mov_b64 exec, 3
 ; NO-SPILL-TO-VGPR-NEXT:    buffer_store_dword v1, off, s[0:3], s33 offset:16
 ; NO-SPILL-TO-VGPR-NEXT:    buffer_load_dword v1, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
 ; NO-SPILL-TO-VGPR-NEXT:    s_waitcnt vmcnt(0)
-; NO-SPILL-TO-VGPR-NEXT:    v_readlane_b32 s4, v1, 0
-; NO-SPILL-TO-VGPR-NEXT:    v_readlane_b32 s5, v1, 1
+; NO-SPILL-TO-VGPR-NEXT:    v_readlane_b32 s30, v1, 0
+; NO-SPILL-TO-VGPR-NEXT:    v_readlane_b32 s31, v1, 1
 ; NO-SPILL-TO-VGPR-NEXT:    buffer_load_dword v1, off, s[0:3], s33 offset:16
 ; NO-SPILL-TO-VGPR-NEXT:    s_waitcnt vmcnt(0)
-; NO-SPILL-TO-VGPR-NEXT:    s_mov_b64 exec, s[6:7]
+; NO-SPILL-TO-VGPR-NEXT:    s_mov_b64 exec, s[4:5]
 ; NO-SPILL-TO-VGPR-NEXT:    s_addk_i32 s32, 0xf800
 ; NO-SPILL-TO-VGPR-NEXT:    buffer_load_dword v0, off, s[0:3], s32 offset:12 ; 4-byte Folded Reload
 ; NO-SPILL-TO-VGPR-NEXT:    s_waitcnt vmcnt(0)
 ; NO-SPILL-TO-VGPR-NEXT:    v_readfirstlane_b32 s33, v0
-; NO-SPILL-TO-VGPR-NEXT:    s_setpc_b64 s[4:5]
+; NO-SPILL-TO-VGPR-NEXT:    s_setpc_b64 s[30:31]
   %alloca = alloca i32, addrspace(5)
   store volatile i32 0, i32 addrspace(5)* %alloca
   call void @external_void_func_void()
   ret void
 }
 
 attributes #0 = { nounwind }

llvm/test/CodeGen/AMDGPU/fshr.ll

 ; GFX10-LABEL: v_fshr_v4i16:
 ; GFX10:       ; %bb.0:
 ; GFX10-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v6, 16, v1
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v7, 16, v5
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v8, 16, v3
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v11, 16, v4
-; GFX10-NEXT:    v_lshlrev_b16 v1, 1, v1
+; GFX10-NEXT:    v_lshrrev_b32_e32 v12, 16, v0
 ; GFX10-NEXT:    v_lshlrev_b16 v6, 1, v6
 ; GFX10-NEXT:    v_xor_b32_e32 v9, -1, v7
 ; GFX10-NEXT:    v_lshrrev_b16 v7, v7, v8
-; GFX10-NEXT:    v_lshrrev_b32_e32 v8, 16, v0
 ; GFX10-NEXT:    v_lshlrev_b16 v0, 1, v0
-; GFX10-NEXT:    v_xor_b32_e32 v12, -1, v5
+; GFX10-NEXT:    v_xor_b32_e32 v8, -1, v4
+; GFX10-NEXT:    v_lshlrev_b16 v1, 1, v1
 ; GFX10-NEXT:    v_lshlrev_b16 v6, v9, v6
-; GFX10-NEXT:    v_xor_b32_e32 v9, -1, v4
+; GFX10-NEXT:    v_xor_b32_e32 v9, -1, v5
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v10, 16, v2
-; GFX10-NEXT:    v_lshlrev_b16 v8, 1, v8
+; GFX10-NEXT:    v_lshlrev_b16 v12, 1, v12
 ; GFX10-NEXT:    v_xor_b32_e32 v13, -1, v11
+; GFX10-NEXT:    v_lshlrev_b16 v0, v8, v0
 ; GFX10-NEXT:    v_lshrrev_b16 v2, v4, v2
-; GFX10-NEXT:    v_lshlrev_b16 v0, v9, v0
-; GFX10-NEXT:    v_lshlrev_b16 v1, v12, v1
+; GFX10-NEXT:    v_lshlrev_b16 v1, v9, v1
 ; GFX10-NEXT:    v_lshrrev_b16 v3, v5, v3
 ; GFX10-NEXT:    v_lshrrev_b16 v4, v11, v10
-; GFX10-NEXT:    v_lshlrev_b16 v5, v13, v8
+; GFX10-NEXT:    v_lshlrev_b16 v5, v13, v12
 ; GFX10-NEXT:    v_or_b32_e32 v0, v0, v2
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 0xffff
 ; GFX10-NEXT:    v_or_b32_e32 v1, v1, v3
 ; GFX10-NEXT:    v_or_b32_e32 v3, v6, v7
 ; GFX10-NEXT:    v_or_b32_e32 v4, v5, v4
 ; GFX10-NEXT:    v_and_b32_e32 v0, v2, v0
 ; GFX10-NEXT:    v_and_b32_e32 v1, v2, v1
 ; GFX10-NEXT:    v_lshl_or_b32 v0, v4, 16, v0

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

 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_i1@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_i1@rel32@hi+12
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_i1@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_i1@rel32@hi+12
 ; GFX9-NEXT:    buffer_store_byte v0, off, s[0:3], s32
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_i1_imm:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 1
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i1@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_i1@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    buffer_store_byte v0, off, s[0:3], s32
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_i1@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_i1@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    global_load_ubyte v0, v[0:1], off glc
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_i1_signext@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_i1_signext@rel32@hi+12
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_i1_signext@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_i1_signext@rel32@hi+12
 ; GFX9-NEXT:    v_and_b32_e32 v0, 1, v0
 ; GFX9-NEXT:    buffer_store_byte v0, off, s[0:3], s32
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_i1_signext:
 ; 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:    buffer_store_dword v40, 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:    global_load_ubyte v0, v[0:1], off glc dlc
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i1_signext@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_i1_signext@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_i1_signext@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_i1_signext@rel32@hi+12
 ; GFX10-NEXT:    v_and_b32_e32 v0, 1, v0
 ; GFX10-NEXT:    buffer_store_byte v0, off, s[0:3], s32
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    global_load_ubyte v0, v[0:1], off glc
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_i1_zeroext@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_i1_zeroext@rel32@hi+12
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_i1_zeroext@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_i1_zeroext@rel32@hi+12
 ; GFX9-NEXT:    v_and_b32_e32 v0, 1, v0
 ; GFX9-NEXT:    buffer_store_byte v0, off, s[0:3], s32
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_i1_zeroext:
 ; 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:    buffer_store_dword v40, 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:    global_load_ubyte v0, v[0:1], off glc dlc
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i1_zeroext@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_i1_zeroext@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_i1_zeroext@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_i1_zeroext@rel32@hi+12
 ; GFX10-NEXT:    v_and_b32_e32 v0, 1, v0
 ; GFX10-NEXT:    buffer_store_byte v0, off, s[0:3], s32
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0x7b
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_i8@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_i8@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_i8@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_i8@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_i8_imm:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 0x7b
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i8@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_i8@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_i8@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_i8@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    global_load_sbyte v0, v[0:1], off glc
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_i8_signext@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_i8_signext@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_i8_signext@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_i8_signext@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_i8_signext:
 ; 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:    buffer_store_dword v40, 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:    global_load_sbyte v0, v[0:1], off glc dlc
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i8_signext@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_i8_signext@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_i8_signext@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_i8_signext@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    global_load_ubyte v0, v[0:1], off glc
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_i8_zeroext@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_i8_zeroext@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_i8_zeroext@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_i8_zeroext@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_i8_zeroext:
 ; 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:    buffer_store_dword v40, 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:    global_load_ubyte v0, v[0:1], off glc dlc
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i8_zeroext@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_i8_zeroext@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_i8_zeroext@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_i8_zeroext@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0x7b
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_i16@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_i16@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_i16@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_i16@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_i16_imm:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 0x7b
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i16@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_i16@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_i16@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_i16@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    global_load_ushort v0, v[0:1], off glc
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_i16_signext@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_i16_signext@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_i16_signext@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_i16_signext@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_i16_signext:
 ; 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:    buffer_store_dword v40, 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:    global_load_ushort v0, v[0:1], off glc dlc
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i16_signext@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_i16_signext@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_i16_signext@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_i16_signext@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    global_load_ushort v0, v[0:1], off glc
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_i16_zeroext@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_i16_zeroext@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_i16_zeroext@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_i16_zeroext@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_i16_zeroext:
 ; 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:    buffer_store_dword v40, 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:    global_load_ushort v0, v[0:1], off glc dlc
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i16_zeroext@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_i16_zeroext@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_i16_zeroext@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_i16_zeroext@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 42
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_i32@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_i32@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_i32@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_i32@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_i32_imm:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 42
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i32@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_i32@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_i32@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_i32@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0x7b
 ; GFX9-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_i64@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_i64@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_i64@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_i64@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_i64_imm:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 0x7b
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i64@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_i64@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_i64@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_i64@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX9-NEXT:    global_load_dwordx4 v[0:3], v[0:1], off
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v2i64@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v2i64@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v2i64@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v2i64@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_v2i64:
 ; 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:    buffer_store_dword v40, 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_mov_b32_e32 v0, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2i64@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v2i64@rel32@hi+12
 ; GFX10-NEXT:    global_load_dwordx4 v[0:3], v[0:1], off
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v2i64@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v2i64@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 1
 ; GFX9-NEXT:    v_mov_b32_e32 v1, 2
 ; GFX9-NEXT:    v_mov_b32_e32 v2, 3
 ; GFX9-NEXT:    v_mov_b32_e32 v3, 4
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v2i64@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v2i64@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v2i64@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v2i64@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)

 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 1
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 3
 ; GFX10-NEXT:    v_mov_b32_e32 v3, 4
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2i64@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v2i64@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v2i64@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v2i64@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    global_load_dwordx4 v[0:3], v[0:1], off
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v4, 1
 ; GFX9-NEXT:    v_mov_b32_e32 v5, 2
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v3i64@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v3i64@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v3i64@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v3i64@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)

 ; GFX10-NEXT:    v_mov_b32_e32 v1, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v4, 1
 ; GFX10-NEXT:    v_mov_b32_e32 v5, 2
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    global_load_dwordx4 v[0:3], v[0:1], off
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3i64@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v3i64@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v3i64@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v3i64@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v4, 1
 ; GFX9-NEXT:    v_mov_b32_e32 v5, 2
 ; GFX9-NEXT:    v_mov_b32_e32 v6, 3
 ; GFX9-NEXT:    v_mov_b32_e32 v7, 4
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v4i64@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v4i64@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v4i64@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v4i64@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)

 ; GFX10-NEXT:    v_mov_b32_e32 v4, 1
 ; GFX10-NEXT:    v_mov_b32_e32 v5, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v6, 3
 ; GFX10-NEXT:    global_load_dwordx4 v[0:3], v[0:1], off
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v7, 4
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v4i64@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v4i64@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v4i64@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v4i64@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0x4400
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_f16@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_f16@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_f16@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_f16@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_f16_imm:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 0x4400
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_f16@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_f16@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_f16@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_f16@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 4.0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_f32@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_f32@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_f32@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_f32@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_f32_imm:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 4.0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_f32@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_f32@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_f32@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_f32@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 1.0
 ; GFX9-NEXT:    v_mov_b32_e32 v1, 2.0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v2f32@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v2f32@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v2f32@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v2f32@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_v2f32_imm:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 1.0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 2.0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2f32@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v2f32@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v2f32@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v2f32@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 1.0
 ; GFX9-NEXT:    v_mov_b32_e32 v1, 2.0
 ; GFX9-NEXT:    v_mov_b32_e32 v2, 4.0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v3f32@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v3f32@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v3f32@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v3f32@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)

 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 1.0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 2.0
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 4.0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3f32@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v3f32@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v3f32@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v3f32@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 1.0
 ; GFX9-NEXT:    v_mov_b32_e32 v1, 2.0
 ; GFX9-NEXT:    v_mov_b32_e32 v2, 4.0
 ; GFX9-NEXT:    v_mov_b32_e32 v3, -1.0
 ; GFX9-NEXT:    v_mov_b32_e32 v4, 0.5
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v5f32@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v5f32@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v5f32@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v5f32@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)

 ; GFX10-NEXT:    v_mov_b32_e32 v0, 1.0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 2.0
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 4.0
 ; GFX10-NEXT:    v_mov_b32_e32 v3, -1.0
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v4, 0.5
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v5f32@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v5f32@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v5f32@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v5f32@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v1, 0x40100000
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_f64@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_f64@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_f64@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_f64@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_f64_imm:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 0x40100000
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_f64@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_f64@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_f64@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_f64@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v1, 2.0
 ; GFX9-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v3, 0x40100000
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v2f64@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v2f64@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v2f64@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v2f64@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)

 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 2.0
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v3, 0x40100000
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2f64@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v2f64@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v2f64@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v2f64@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v1, 2.0
 ; GFX9-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v3, 0x40100000
 ; GFX9-NEXT:    v_mov_b32_e32 v4, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v5, 0x40200000
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v3f64@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v3f64@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v3f64@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v3f64@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)

 ; GFX10-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v3, 0x40100000
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v4, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v5, 0x40200000
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v3f64@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v3f64@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3f64@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v3f64@rel32@hi+12
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    global_load_dword v0, v[0:1], off
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v2i16@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v2i16@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v2i16@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v2i16@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_v2i16:
 ; 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:    buffer_store_dword v40, 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:    global_load_dword v0, v[0:1], off
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2i16@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v2i16@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v2i16@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v2i16@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    global_load_dwordx2 v[0:1], v[0:1], off
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v3i16@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v3i16@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v3i16@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v3i16@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_v3i16:
 ; 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:    buffer_store_dword v40, 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:    global_load_dwordx2 v[0:1], v[0:1], off
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3i16@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v3i16@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v3i16@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v3i16@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    global_load_dwordx2 v[0:1], v[0:1], off
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v3f16@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v3f16@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v3f16@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v3f16@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_v3f16:
 ; 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:    buffer_store_dword v40, 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:    global_load_dwordx2 v[0:1], v[0:1], off
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3f16@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v3f16@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v3f16@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v3f16@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0x20001
 ; GFX9-NEXT:    v_mov_b32_e32 v1, 3
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v3i16@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v3i16@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v3i16@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v3i16@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_v3i16_imm:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 0x20001
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 3
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3i16@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v3i16@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v3i16@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v3i16@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0x40003c00
 ; GFX9-NEXT:    v_mov_b32_e32 v1, 0x4400
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v3f16@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v3f16@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v3f16@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v3f16@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_v3f16_imm:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 0x40003c00
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 0x4400
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3f16@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v3f16@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v3f16@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v3f16@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    global_load_dwordx2 v[0:1], v[0:1], off
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v4i16@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v4i16@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v4i16@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v4i16@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_v4i16:
 ; 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:    buffer_store_dword v40, 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:    global_load_dwordx2 v[0:1], v[0:1], off
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v4i16@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v4i16@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v4i16@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v4i16@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0x20001
 ; GFX9-NEXT:    v_mov_b32_e32 v1, 0x40003
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v4i16@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v4i16@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v4i16@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v4i16@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_v4i16_imm:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 0x20001
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 0x40003
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v4i16@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v4i16@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v4i16@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v4i16@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    global_load_dword v0, v[0:1], off
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v2f16@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v2f16@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v2f16@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v2f16@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_v2f16:
 ; 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:    buffer_store_dword v40, 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:    global_load_dword v0, v[0:1], off
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2f16@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v2f16@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v2f16@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v2f16@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    global_load_dwordx2 v[0:1], v[0:1], off
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v2i32@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v2i32@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v2i32@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v2i32@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_v2i32:
 ; 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:    buffer_store_dword v40, 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:    global_load_dwordx2 v[0:1], v[0:1], off
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2i32@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v2i32@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v2i32@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v2i32@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 1
 ; GFX9-NEXT:    v_mov_b32_e32 v1, 2
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v2i32@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v2i32@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v2i32@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v2i32@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_v2i32_imm:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 1
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 2
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2i32@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v2i32@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v2i32@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v2i32@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 3
 ; GFX9-NEXT:    v_mov_b32_e32 v1, 4
 ; GFX9-NEXT:    v_mov_b32_e32 v2, 5
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v3i32@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v3i32@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v3i32@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v3i32@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)

 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 3
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 4
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 5
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3i32@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v3i32@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v3i32@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v3i32@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 3
 ; GFX9-NEXT:    v_mov_b32_e32 v1, 4
 ; GFX9-NEXT:    v_mov_b32_e32 v2, 5
 ; GFX9-NEXT:    v_mov_b32_e32 v3, 6
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v3i32_i32@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v3i32_i32@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v3i32_i32@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v3i32_i32@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)

 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 3
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 4
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 5
 ; GFX10-NEXT:    v_mov_b32_e32 v3, 6
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3i32_i32@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v3i32_i32@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v3i32_i32@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v3i32_i32@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    global_load_dwordx4 v[0:3], v[0:1], off
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v4i32@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v4i32@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v4i32@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v4i32@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_v4i32:
 ; 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:    buffer_store_dword v40, 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:    global_load_dwordx4 v[0:3], v[0:1], off
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v4i32@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v4i32@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v4i32@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v4i32@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 1
 ; GFX9-NEXT:    v_mov_b32_e32 v1, 2
 ; GFX9-NEXT:    v_mov_b32_e32 v2, 3
 ; GFX9-NEXT:    v_mov_b32_e32 v3, 4
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v4i32@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v4i32@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v4i32@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v4i32@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)

 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 1
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 3
 ; GFX10-NEXT:    v_mov_b32_e32 v3, 4
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v4i32@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v4i32@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v4i32@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v4i32@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 1
 ; GFX9-NEXT:    v_mov_b32_e32 v1, 2
 ; GFX9-NEXT:    v_mov_b32_e32 v2, 3
 ; GFX9-NEXT:    v_mov_b32_e32 v3, 4
 ; GFX9-NEXT:    v_mov_b32_e32 v4, 5
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v5i32@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v5i32@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v5i32@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v5i32@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)

 ; GFX10-NEXT:    v_mov_b32_e32 v0, 1
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 3
 ; GFX10-NEXT:    v_mov_b32_e32 v3, 4
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v4, 5
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v5i32@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v5i32@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v5i32@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v5i32@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 
 define amdgpu_gfx void @test_call_external_void_func_v8i32() #0 {
 ; GFX9-LABEL: test_call_external_void_func_v8i32:
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
-; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
-; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
+; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX9-NEXT:    v_mov_b32_e32 v8, 0
+; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX9-NEXT:    global_load_dwordx4 v[0:3], v8, s[30:31]
-; GFX9-NEXT:    global_load_dwordx4 v[4:7], v8, s[30:31] offset:16
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v8i32@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v8i32@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    global_load_dwordx4 v[0:3], v8, s[34:35]
+; GFX9-NEXT:    global_load_dwordx4 v[4:7], v8, s[34:35] offset:16
+; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v8i32@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v8i32@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_v8i32:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 2
+; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX10-NEXT:    v_mov_b32_e32 v8, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
-; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX10-NEXT:    s_clause 0x1
-; GFX10-NEXT:    global_load_dwordx4 v[0:3], v8, s[30:31]
-; GFX10-NEXT:    global_load_dwordx4 v[4:7], v8, s[30:31] offset:16
-; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v8i32@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v8i32@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    global_load_dwordx4 v[0:3], v8, s[34:35]
+; GFX10-NEXT:    global_load_dwordx4 v[4:7], v8, s[34:35] offset:16
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v8i32@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v8i32@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    v_mov_b32_e32 v1, 2
 ; GFX9-NEXT:    v_mov_b32_e32 v2, 3
 ; GFX9-NEXT:    v_mov_b32_e32 v3, 4
 ; GFX9-NEXT:    v_mov_b32_e32 v4, 5
 ; GFX9-NEXT:    v_mov_b32_e32 v5, 6
 ; GFX9-NEXT:    v_mov_b32_e32 v6, 7
 ; GFX9-NEXT:    v_mov_b32_e32 v7, 8
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v8i32@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v8i32@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v8i32@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v8i32@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)

 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v4, 5
 ; GFX10-NEXT:    v_mov_b32_e32 v5, 6
 ; GFX10-NEXT:    v_mov_b32_e32 v6, 7
 ; GFX10-NEXT:    v_mov_b32_e32 v7, 8
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v8i32@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v8i32@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v8i32@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v8i32@rel32@hi+12
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 
 define amdgpu_gfx void @test_call_external_void_func_v16i32() #0 {
 ; GFX9-LABEL: test_call_external_void_func_v16i32:
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
-; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
-; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
+; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX9-NEXT:    v_mov_b32_e32 v16, 0
+; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX9-NEXT:    global_load_dwordx4 v[0:3], v16, s[30:31]
-; GFX9-NEXT:    global_load_dwordx4 v[4:7], v16, s[30:31] offset:16
-; GFX9-NEXT:    global_load_dwordx4 v[8:11], v16, s[30:31] offset:32
-; GFX9-NEXT:    global_load_dwordx4 v[12:15], v16, s[30:31] offset:48
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v16i32@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v16i32@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    global_load_dwordx4 v[0:3], v16, s[34:35]
+; GFX9-NEXT:    global_load_dwordx4 v[4:7], v16, s[34:35] offset:16
+; GFX9-NEXT:    global_load_dwordx4 v[8:11], v16, s[34:35] offset:32
+; GFX9-NEXT:    global_load_dwordx4 v[12:15], v16, s[34:35] offset:48
+; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v16i32@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v16i32@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_v16i32:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 2
+; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX10-NEXT:    v_mov_b32_e32 v16, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
-; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX10-NEXT:    s_clause 0x3
-; GFX10-NEXT:    global_load_dwordx4 v[0:3], v16, s[30:31]
-; GFX10-NEXT:    global_load_dwordx4 v[4:7], v16, s[30:31] offset:16
-; GFX10-NEXT:    global_load_dwordx4 v[8:11], v16, s[30:31] offset:32
-; GFX10-NEXT:    global_load_dwordx4 v[12:15], v16, s[30:31] offset:48
-; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v16i32@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v16i32@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    global_load_dwordx4 v[0:3], v16, s[34:35]
+; GFX10-NEXT:    global_load_dwordx4 v[4:7], v16, s[34:35] offset:16
+; GFX10-NEXT:    global_load_dwordx4 v[8:11], v16, s[34:35] offset:32
+; GFX10-NEXT:    global_load_dwordx4 v[12:15], v16, s[34:35] offset:48
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v16i32@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v16i32@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 
 define amdgpu_gfx void @test_call_external_void_func_v32i32() #0 {
 ; GFX9-LABEL: test_call_external_void_func_v32i32:
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
-; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
-; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
+; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX9-NEXT:    v_mov_b32_e32 v28, 0
+; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX9-NEXT:    global_load_dwordx4 v[0:3], v28, s[30:31]
-; GFX9-NEXT:    global_load_dwordx4 v[4:7], v28, s[30:31] offset:16
-; GFX9-NEXT:    global_load_dwordx4 v[8:11], v28, s[30:31] offset:32
-; GFX9-NEXT:    global_load_dwordx4 v[12:15], v28, s[30:31] offset:48
-; GFX9-NEXT:    global_load_dwordx4 v[16:19], v28, s[30:31] offset:64
-; GFX9-NEXT:    global_load_dwordx4 v[20:23], v28, s[30:31] offset:80
-; GFX9-NEXT:    global_load_dwordx4 v[24:27], v28, s[30:31] offset:96
+; GFX9-NEXT:    global_load_dwordx4 v[0:3], v28, s[34:35]
+; GFX9-NEXT:    global_load_dwordx4 v[4:7], v28, s[34:35] offset:16
+; GFX9-NEXT:    global_load_dwordx4 v[8:11], v28, s[34:35] offset:32
+; GFX9-NEXT:    global_load_dwordx4 v[12:15], v28, s[34:35] offset:48
+; GFX9-NEXT:    global_load_dwordx4 v[16:19], v28, s[34:35] offset:64
+; GFX9-NEXT:    global_load_dwordx4 v[20:23], v28, s[34:35] offset:80
+; GFX9-NEXT:    global_load_dwordx4 v[24:27], v28, s[34:35] offset:96
 ; GFX9-NEXT:    s_nop 0
-; GFX9-NEXT:    global_load_dwordx4 v[28:31], v28, s[30:31] offset:112
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v32i32@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v32i32@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    global_load_dwordx4 v[28:31], v28, s[34:35] offset:112
+; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v32i32@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v32i32@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_v32i32:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 2
+; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX10-NEXT:    v_mov_b32_e32 v32, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
-; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX10-NEXT:    s_clause 0x7
-; GFX10-NEXT:    global_load_dwordx4 v[0:3], v32, s[30:31]
-; GFX10-NEXT:    global_load_dwordx4 v[4:7], v32, s[30:31] offset:16
-; GFX10-NEXT:    global_load_dwordx4 v[8:11], v32, s[30:31] offset:32
-; GFX10-NEXT:    global_load_dwordx4 v[12:15], v32, s[30:31] offset:48
-; GFX10-NEXT:    global_load_dwordx4 v[16:19], v32, s[30:31] offset:64
-; GFX10-NEXT:    global_load_dwordx4 v[20:23], v32, s[30:31] offset:80
-; GFX10-NEXT:    global_load_dwordx4 v[24:27], v32, s[30:31] offset:96
-; GFX10-NEXT:    global_load_dwordx4 v[28:31], v32, s[30:31] offset:112
-; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v32i32@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v32i32@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    global_load_dwordx4 v[0:3], v32, s[34:35]
+; GFX10-NEXT:    global_load_dwordx4 v[4:7], v32, s[34:35] offset:16
+; GFX10-NEXT:    global_load_dwordx4 v[8:11], v32, s[34:35] offset:32
+; GFX10-NEXT:    global_load_dwordx4 v[12:15], v32, s[34:35] offset:48
+; GFX10-NEXT:    global_load_dwordx4 v[16:19], v32, s[34:35] offset:64
+; GFX10-NEXT:    global_load_dwordx4 v[20:23], v32, s[34:35] offset:80
+; GFX10-NEXT:    global_load_dwordx4 v[24:27], v32, s[34:35] offset:96
+; GFX10-NEXT:    global_load_dwordx4 v[28:31], v32, s[34:35] offset:112
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v32i32@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v32i32@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 
 define amdgpu_gfx void @test_call_external_void_func_v32i32_i32(i32) #0 {
 ; GFX9-LABEL: test_call_external_void_func_v32i32_i32:
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
-; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
-; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
+; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX9-NEXT:    v_mov_b32_e32 v28, 0
+; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX9-NEXT:    global_load_dwordx4 v[0:3], v28, s[30:31]
-; GFX9-NEXT:    global_load_dwordx4 v[4:7], v28, s[30:31] offset:16
-; GFX9-NEXT:    global_load_dwordx4 v[8:11], v28, s[30:31] offset:32
-; GFX9-NEXT:    global_load_dwordx4 v[12:15], v28, s[30:31] offset:48
-; GFX9-NEXT:    global_load_dwordx4 v[16:19], v28, s[30:31] offset:64
-; GFX9-NEXT:    global_load_dwordx4 v[20:23], v28, s[30:31] offset:80
-; GFX9-NEXT:    global_load_dwordx4 v[24:27], v28, s[30:31] offset:96
+; GFX9-NEXT:    global_load_dwordx4 v[0:3], v28, s[34:35]
+; GFX9-NEXT:    global_load_dwordx4 v[4:7], v28, s[34:35] offset:16
+; GFX9-NEXT:    global_load_dwordx4 v[8:11], v28, s[34:35] offset:32
+; GFX9-NEXT:    global_load_dwordx4 v[12:15], v28, s[34:35] offset:48
+; GFX9-NEXT:    global_load_dwordx4 v[16:19], v28, s[34:35] offset:64
+; GFX9-NEXT:    global_load_dwordx4 v[20:23], v28, s[34:35] offset:80
+; GFX9-NEXT:    global_load_dwordx4 v[24:27], v28, s[34:35] offset:96
 ; GFX9-NEXT:    s_nop 0
-; GFX9-NEXT:    global_load_dwordx4 v[28:31], v28, s[30:31] offset:112
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v32i32_i32@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v32i32_i32@rel32@hi+12
+; GFX9-NEXT:    global_load_dwordx4 v[28:31], v28, s[34:35] offset:112
+; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v32i32_i32@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v32i32_i32@rel32@hi+12
 ; GFX9-NEXT:    s_waitcnt vmcnt(7)
 ; GFX9-NEXT:    global_load_dword v32, v[0:1], off
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    buffer_store_dword v32, off, s[0:3], s32
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_v32i32_i32:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 2
+; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX10-NEXT:    v_mov_b32_e32 v32, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
-; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
 ; GFX10-NEXT:    global_load_dword v33, v[0:1], off
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX10-NEXT:    s_clause 0x7
-; GFX10-NEXT:    global_load_dwordx4 v[0:3], v32, s[30:31]
-; GFX10-NEXT:    global_load_dwordx4 v[4:7], v32, s[30:31] offset:16
-; GFX10-NEXT:    global_load_dwordx4 v[8:11], v32, s[30:31] offset:32
-; GFX10-NEXT:    global_load_dwordx4 v[12:15], v32, s[30:31] offset:48
-; GFX10-NEXT:    global_load_dwordx4 v[16:19], v32, s[30:31] offset:64
-; GFX10-NEXT:    global_load_dwordx4 v[20:23], v32, s[30:31] offset:80
-; GFX10-NEXT:    global_load_dwordx4 v[24:27], v32, s[30:31] offset:96
-; GFX10-NEXT:    global_load_dwordx4 v[28:31], v32, s[30:31] offset:112
-; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v32i32_i32@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v32i32_i32@rel32@hi+12
+; GFX10-NEXT:    global_load_dwordx4 v[0:3], v32, s[34:35]
+; GFX10-NEXT:    global_load_dwordx4 v[4:7], v32, s[34:35] offset:16
+; GFX10-NEXT:    global_load_dwordx4 v[8:11], v32, s[34:35] offset:32
+; GFX10-NEXT:    global_load_dwordx4 v[12:15], v32, s[34:35] offset:48
+; GFX10-NEXT:    global_load_dwordx4 v[16:19], v32, s[34:35] offset:64
+; GFX10-NEXT:    global_load_dwordx4 v[20:23], v32, s[34:35] offset:80
+; GFX10-NEXT:    global_load_dwordx4 v[24:27], v32, s[34:35] offset:96
+; GFX10-NEXT:    global_load_dwordx4 v[28:31], v32, s[34:35] offset:112
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v32i32_i32@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v32i32_i32@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX10-NEXT:    s_waitcnt vmcnt(8)
 ; GFX10-NEXT:    buffer_store_dword v33, off, s[0:3], s32
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX9-NEXT:    buffer_store_dword v42, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v41, v0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 42
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX9-NEXT:    v_mov_b32_e32 v42, v1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_i32_func_i32@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_i32_func_i32@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_i32_func_i32@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_i32_func_i32@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    global_store_dword v[41:42], v0, off
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    buffer_load_dword v42, off, s[0:3], s33 ; 4-byte Folded Reload
 ; GFX9-NEXT:    buffer_load_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2

 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT:    buffer_store_dword v42, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX10-NEXT:    v_mov_b32_e32 v41, v0
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 42
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_mov_b32_e32 v42, v1
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_i32_func_i32@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_i32_func_i32@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_i32_func_i32@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_i32_func_i32@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    global_store_dword v[41:42], v0, off
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    s_clause 0x1
 ; GFX10-NEXT:    buffer_load_dword v42, off, s[0:3], s33
 ; GFX10-NEXT:    buffer_load_dword v41, off, s[0:3], s33 offset:4
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00

 
 define amdgpu_gfx void @test_call_external_void_func_struct_i8_i32() #0 {
 ; GFX9-LABEL: test_call_external_void_func_struct_i8_i32:
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
-; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
-; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
+; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX9-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX9-NEXT:    global_load_dword v1, v2, s[30:31] offset:4
-; GFX9-NEXT:    global_load_ubyte v0, v2, s[30:31]
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_struct_i8_i32@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_struct_i8_i32@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    global_load_dword v1, v2, s[34:35] offset:4
+; GFX9-NEXT:    global_load_ubyte v0, v2, s[34:35]
+; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_struct_i8_i32@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_struct_i8_i32@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_struct_i8_i32:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 2
+; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX10-NEXT:    v_mov_b32_e32 v2, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
-; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX10-NEXT:    s_clause 0x1
-; GFX10-NEXT:    global_load_ubyte v0, v2, s[30:31]
-; GFX10-NEXT:    global_load_dword v1, v2, s[30:31] offset:4
-; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_struct_i8_i32@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_struct_i8_i32@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    global_load_ubyte v0, v2, s[34:35]
+; GFX10-NEXT:    global_load_dword v1, v2, s[34:35] offset:4
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_struct_i8_i32@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_struct_i8_i32@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    v_mov_b32_e32 v0, 3
 ; GFX9-NEXT:    buffer_store_byte v0, off, s[0:3], s33
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 8
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], s33 offset:4
 ; GFX9-NEXT:    v_lshrrev_b32_e64 v0, 6, s33
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_byval_struct_i8_i32@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_byval_struct_i8_i32@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_byval_struct_i8_i32@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_byval_struct_i8_i32@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)

 ; GFX10-NEXT:    v_mov_b32_e32 v0, 3
 ; GFX10-NEXT:    v_mov_b32_e32 v1, 8
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    buffer_store_byte v0, off, s[0:3], s33
 ; GFX10-NEXT:    buffer_store_dword v1, off, s[0:3], s33 offset:4
 ; GFX10-NEXT:    v_lshrrev_b32_e64 v0, 5, s33
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_byval_struct_i8_i32@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_byval_struct_i8_i32@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_byval_struct_i8_i32@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_byval_struct_i8_i32@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    v_mov_b32_e32 v0, 8
 ; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], s33 offset:4
 ; GFX9-NEXT:    v_lshrrev_b32_e64 v0, 6, s33
 ; GFX9-NEXT:    s_addk_i32 s32, 0x800
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_add_u32_e32 v0, 8, v0
 ; GFX9-NEXT:    v_lshrrev_b32_e64 v1, 6, s33
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_sret_struct_i8_i32_byval_struct_i8_i32@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_sret_struct_i8_i32_byval_struct_i8_i32@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_sret_struct_i8_i32_byval_struct_i8_i32@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_sret_struct_i8_i32_byval_struct_i8_i32@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    buffer_load_ubyte v0, off, s[0:3], s33 offset:8
 ; GFX9-NEXT:    buffer_load_dword v1, off, s[0:3], s33 offset:12
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xf800
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    global_store_byte v[0:1], v0, off

 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x400
 ; GFX10-NEXT:    buffer_store_byte v0, off, s[0:3], s33
 ; GFX10-NEXT:    buffer_store_dword v1, off, s[0:3], s33 offset:4
 ; GFX10-NEXT:    v_lshrrev_b32_e64 v0, 5, s33
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_lshrrev_b32_e64 v1, 5, s33
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_sret_struct_i8_i32_byval_struct_i8_i32@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_sret_struct_i8_i32_byval_struct_i8_i32@rel32@hi+12
 ; GFX10-NEXT:    v_add_nc_u32_e32 v0, 8, v0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_sret_struct_i8_i32_byval_struct_i8_i32@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_sret_struct_i8_i32_byval_struct_i8_i32@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    s_clause 0x1
 ; GFX10-NEXT:    buffer_load_ubyte v0, off, s[0:3], s33 offset:8
 ; GFX10-NEXT:    buffer_load_dword v1, off, s[0:3], s33 offset:12
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)

 
 define amdgpu_gfx void @test_call_external_void_func_v16i8() #0 {
 ; GFX9-LABEL: test_call_external_void_func_v16i8:
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
-; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
-; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
+; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 0
+; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX9-NEXT:    global_load_dwordx4 v[0:3], v0, s[30:31]
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v16i8@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v16i8@rel32@hi+12
+; GFX9-NEXT:    global_load_dwordx4 v[0:3], v0, s[34:35]
+; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v16i8@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v16i8@rel32@hi+12
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    v_lshrrev_b32_e32 v16, 8, v0
 ; GFX9-NEXT:    v_lshrrev_b32_e32 v17, 16, v0
 ; GFX9-NEXT:    v_lshrrev_b32_e32 v18, 24, v0
 ; GFX9-NEXT:    v_lshrrev_b32_e32 v5, 8, v1
 ; GFX9-NEXT:    v_lshrrev_b32_e32 v6, 16, v1
 ; GFX9-NEXT:    v_lshrrev_b32_e32 v7, 24, v1
 ; GFX9-NEXT:    v_lshrrev_b32_e32 v9, 8, v2
 ; GFX9-NEXT:    v_lshrrev_b32_e32 v10, 16, v2
 ; GFX9-NEXT:    v_lshrrev_b32_e32 v11, 24, v2
 ; GFX9-NEXT:    v_lshrrev_b32_e32 v13, 8, v3
 ; GFX9-NEXT:    v_lshrrev_b32_e32 v14, 16, v3
 ; GFX9-NEXT:    v_lshrrev_b32_e32 v15, 24, v3
 ; GFX9-NEXT:    v_mov_b32_e32 v4, v1
 ; GFX9-NEXT:    v_mov_b32_e32 v8, v2
 ; GFX9-NEXT:    v_mov_b32_e32 v12, v3
 ; GFX9-NEXT:    v_mov_b32_e32 v1, v16
 ; GFX9-NEXT:    v_mov_b32_e32 v2, v17
 ; GFX9-NEXT:    v_mov_b32_e32 v3, v18
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_v16i8:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 2
+; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 0
+; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX10-NEXT:    global_load_dwordx4 v[0:3], v0, s[30:31]
+; GFX10-NEXT:    global_load_dwordx4 v[0:3], v0, s[34:35]
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v16i8@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v16i8@rel32@hi+12
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v16i8@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v16i8@rel32@hi+12
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v16, 8, v0
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v17, 16, v0
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v18, 24, v0
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v5, 8, v1
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v6, 16, v1
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v7, 24, v1
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v9, 8, v2
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v10, 16, v2
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v11, 24, v2
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v13, 8, v3
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v14, 16, v3
 ; GFX10-NEXT:    v_lshrrev_b32_e32 v15, 24, v3
 ; GFX10-NEXT:    v_mov_b32_e32 v4, v1
 ; GFX10-NEXT:    v_mov_b32_e32 v8, v2
 ; GFX10-NEXT:    v_mov_b32_e32 v12, v3
 ; GFX10-NEXT:    v_mov_b32_e32 v1, v16
 ; GFX10-NEXT:    v_mov_b32_e32 v2, v17
 ; GFX10-NEXT:    v_mov_b32_e32 v3, v18
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 
 define void @tail_call_byval_align16(<32 x i32> %val, double %tmp) #0 {
 ; GFX9-LABEL: tail_call_byval_align16:
 ; 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:    buffer_store_dword v40, off, s[0:3], s32 offset:16 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[4:5]
-; GFX9-NEXT:    v_writelane_b32 v40, s33, 30
+; GFX9-NEXT:    v_writelane_b32 v40, s33, 32
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    buffer_load_dword v32, off, s[0:3], s33 offset:8
 ; GFX9-NEXT:    buffer_load_dword v33, off, s[0:3], s33 offset:12
 ; GFX9-NEXT:    v_writelane_b32 v40, s34, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s35, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s36, 2
 ; GFX9-NEXT:    v_writelane_b32 v40, s37, 3
 ; GFX9-NEXT:    v_writelane_b32 v40, s38, 4

 ; GFX9-NEXT:    v_writelane_b32 v40, s54, 20
 ; GFX9-NEXT:    v_writelane_b32 v40, s55, 21
 ; GFX9-NEXT:    v_writelane_b32 v40, s56, 22
 ; GFX9-NEXT:    v_writelane_b32 v40, s57, 23
 ; GFX9-NEXT:    v_writelane_b32 v40, s58, 24
 ; GFX9-NEXT:    v_writelane_b32 v40, s59, 25
 ; GFX9-NEXT:    v_writelane_b32 v40, s60, 26
 ; GFX9-NEXT:    v_writelane_b32 v40, s61, 27
-; GFX9-NEXT:    s_addk_i32 s32, 0x800
 ; GFX9-NEXT:    v_writelane_b32 v40, s62, 28
 ; GFX9-NEXT:    v_writelane_b32 v40, s63, 29
-; GFX9-NEXT:    s_mov_b64 s[4:5], s[30:31]
-; GFX9-NEXT:    s_getpc_b64 s[6:7]
-; GFX9-NEXT:    s_add_u32 s6, s6, byval_align16_f64_arg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s7, s7, byval_align16_f64_arg@rel32@hi+12
+; GFX9-NEXT:    s_addk_i32 s32, 0x800
+; GFX9-NEXT:    v_writelane_b32 v40, s30, 30
+; GFX9-NEXT:    v_writelane_b32 v40, s31, 31
+; GFX9-NEXT:    s_getpc_b64 s[4:5]
+; GFX9-NEXT:    s_add_u32 s4, s4, byval_align16_f64_arg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s5, s5, byval_align16_f64_arg@rel32@hi+12
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-NEXT:    buffer_store_dword v33, off, s[0:3], s32 offset:4
 ; GFX9-NEXT:    buffer_store_dword v32, off, s[0:3], s32
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[6:7]
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[4:5]
+; GFX9-NEXT:    v_readlane_b32 s30, v40, 30
+; GFX9-NEXT:    v_readlane_b32 s31, v40, 31
 ; GFX9-NEXT:    v_readlane_b32 s63, v40, 29
 ; GFX9-NEXT:    v_readlane_b32 s62, v40, 28
 ; GFX9-NEXT:    v_readlane_b32 s61, v40, 27
 ; GFX9-NEXT:    v_readlane_b32 s60, v40, 26
 ; GFX9-NEXT:    v_readlane_b32 s59, v40, 25
 ; GFX9-NEXT:    v_readlane_b32 s58, v40, 24
 ; GFX9-NEXT:    v_readlane_b32 s57, v40, 23
 ; GFX9-NEXT:    v_readlane_b32 s56, v40, 22

 ; GFX9-NEXT:    v_readlane_b32 s40, v40, 6
 ; GFX9-NEXT:    v_readlane_b32 s39, v40, 5
 ; GFX9-NEXT:    v_readlane_b32 s38, v40, 4
 ; GFX9-NEXT:    v_readlane_b32 s37, v40, 3
 ; GFX9-NEXT:    v_readlane_b32 s36, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s35, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s34, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xf800
-; GFX9-NEXT:    v_readlane_b32 s33, v40, 30
-; GFX9-NEXT:    s_or_saveexec_b64 s[6:7], -1
+; GFX9-NEXT:    v_readlane_b32 s33, v40, 32
+; GFX9-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:16 ; 4-byte Folded Reload
-; GFX9-NEXT:    s_mov_b64 exec, s[6:7]
+; GFX9-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
-; GFX9-NEXT:    s_setpc_b64 s[4:5]
+; GFX9-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: tail_call_byval_align16:
 ; 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:    buffer_store_dword v40, off, s[0:3], s32 offset:16 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s4
-; GFX10-NEXT:    v_writelane_b32 v40, s33, 30
+; GFX10-NEXT:    s_mov_b32 s6, s33
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_clause 0x1
 ; GFX10-NEXT:    buffer_load_dword v32, off, s[0:3], s33 offset:12
 ; GFX10-NEXT:    buffer_load_dword v33, off, s[0:3], s33 offset:8
-; GFX10-NEXT:    s_addk_i32 s32, 0x400
 ; GFX10-NEXT:    v_writelane_b32 v40, s34, 0
-; GFX10-NEXT:    s_mov_b64 s[4:5], s[30:31]
-; GFX10-NEXT:    s_getpc_b64 s[6:7]
-; GFX10-NEXT:    s_add_u32 s6, s6, byval_align16_f64_arg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s7, s7, byval_align16_f64_arg@rel32@hi+12
+; GFX10-NEXT:    s_addk_i32 s32, 0x400
+; GFX10-NEXT:    s_getpc_b64 s[4:5]
+; GFX10-NEXT:    s_add_u32 s4, s4, byval_align16_f64_arg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s5, s5, byval_align16_f64_arg@rel32@hi+12
 ; GFX10-NEXT:    s_waitcnt vmcnt(1)
 ; GFX10-NEXT:    buffer_store_dword v32, off, s[0:3], s32 offset:4
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    buffer_store_dword v33, off, s[0:3], s32
 ; GFX10-NEXT:    v_writelane_b32 v40, s35, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s36, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s37, 3
 ; GFX10-NEXT:    v_writelane_b32 v40, s38, 4

 ; GFX10-NEXT:    v_writelane_b32 v40, s56, 22
 ; GFX10-NEXT:    v_writelane_b32 v40, s57, 23
 ; GFX10-NEXT:    v_writelane_b32 v40, s58, 24
 ; GFX10-NEXT:    v_writelane_b32 v40, s59, 25
 ; GFX10-NEXT:    v_writelane_b32 v40, s60, 26
 ; GFX10-NEXT:    v_writelane_b32 v40, s61, 27
 ; GFX10-NEXT:    v_writelane_b32 v40, s62, 28
 ; GFX10-NEXT:    v_writelane_b32 v40, s63, 29
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[6:7]
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 30
+; GFX10-NEXT:    v_writelane_b32 v40, s31, 31
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[4:5]
+; GFX10-NEXT:    v_readlane_b32 s30, v40, 30
+; GFX10-NEXT:    v_readlane_b32 s31, v40, 31
 ; GFX10-NEXT:    v_readlane_b32 s63, v40, 29
 ; GFX10-NEXT:    v_readlane_b32 s62, v40, 28
 ; GFX10-NEXT:    v_readlane_b32 s61, v40, 27
 ; GFX10-NEXT:    v_readlane_b32 s60, v40, 26
 ; GFX10-NEXT:    v_readlane_b32 s59, v40, 25
 ; GFX10-NEXT:    v_readlane_b32 s58, v40, 24
 ; GFX10-NEXT:    v_readlane_b32 s57, v40, 23
 ; GFX10-NEXT:    v_readlane_b32 s56, v40, 22

 ; GFX10-NEXT:    v_readlane_b32 s40, v40, 6
 ; GFX10-NEXT:    v_readlane_b32 s39, v40, 5
 ; GFX10-NEXT:    v_readlane_b32 s38, v40, 4
 ; GFX10-NEXT:    v_readlane_b32 s37, v40, 3
 ; GFX10-NEXT:    v_readlane_b32 s36, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s35, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s34, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfc00
-; GFX10-NEXT:    v_readlane_b32 s33, v40, 30
-; GFX10-NEXT:    s_or_saveexec_b32 s6, -1
+; GFX10-NEXT:    s_mov_b32 s33, s6
+; GFX10-NEXT:    s_or_saveexec_b32 s4, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:16 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
-; GFX10-NEXT:    s_mov_b32 exec_lo, s6
+; GFX10-NEXT:    s_mov_b32 exec_lo, s4
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-NEXT:    s_setpc_b64 s[4:5]
+; GFX10-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-SCRATCH-LABEL: tail_call_byval_align16:
 ; GFX10-SCRATCH:       ; %bb.0: ; %entry
 ; GFX10-SCRATCH-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-SCRATCH-NEXT:    s_or_saveexec_b32 s0, -1
 ; GFX10-SCRATCH-NEXT:    scratch_store_dword off, v40, s32 offset:16 ; 4-byte Folded Spill
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
-; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s33, 30
+; GFX10-SCRATCH-NEXT:    s_mov_b32 s4, s33
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s32
-; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 32
-; GFX10-SCRATCH-NEXT:    scratch_load_dwordx2 v[32:33], off, s33 offset:8
-; GFX10-SCRATCH-NEXT:    s_mov_b64 s[4:5], s[30:31]
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s34, 0
+; GFX10-SCRATCH-NEXT:    scratch_load_dwordx2 v[32:33], off, s33 offset:8
+; GFX10-SCRATCH-NEXT:    s_add_i32 s32, s32, 32
 ; GFX10-SCRATCH-NEXT:    s_getpc_b64 s[0:1]
 ; GFX10-SCRATCH-NEXT:    s_add_u32 s0, s0, byval_align16_f64_arg@rel32@lo+4
 ; GFX10-SCRATCH-NEXT:    s_addc_u32 s1, s1, byval_align16_f64_arg@rel32@hi+12
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s35, 1
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s36, 2
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s37, 3
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s38, 4
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s39, 5

 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s56, 22
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s57, 23
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s58, 24
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s59, 25
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s60, 26
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s61, 27
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s62, 28
 ; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s63, 29
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s30, 30
+; GFX10-SCRATCH-NEXT:    v_writelane_b32 v40, s31, 31
 ; GFX10-SCRATCH-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-SCRATCH-NEXT:    scratch_store_dwordx2 off, v[32:33], s32
 ; GFX10-SCRATCH-NEXT:    s_swappc_b64 s[30:31], s[0:1]
+; GFX10-SCRATCH-NEXT:    v_readlane_b32 s30, v40, 30
+; GFX10-SCRATCH-NEXT:    v_readlane_b32 s31, v40, 31
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s63, v40, 29
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s62, v40, 28
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s61, v40, 27
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s60, v40, 26
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s59, v40, 25
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s58, v40, 24
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s57, v40, 23
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s56, v40, 22

 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s40, v40, 6
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s39, v40, 5
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s38, v40, 4
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s37, v40, 3
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s36, v40, 2
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s35, v40, 1
 ; GFX10-SCRATCH-NEXT:    v_readlane_b32 s34, v40, 0
 ; GFX10-SCRATCH-NEXT:    s_addk_i32 s32, 0xffe0
-; GFX10-SCRATCH-NEXT:    v_readlane_b32 s33, v40, 30
+; GFX10-SCRATCH-NEXT:    s_mov_b32 s33, s4
 ; GFX10-SCRATCH-NEXT:    s_or_saveexec_b32 s0, -1
 ; GFX10-SCRATCH-NEXT:    scratch_load_dword v40, off, s32 offset:16 ; 4-byte Folded Reload
 ; GFX10-SCRATCH-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-SCRATCH-NEXT:    s_mov_b32 exec_lo, s0
 ; GFX10-SCRATCH-NEXT:    s_waitcnt vmcnt(0)
-; GFX10-SCRATCH-NEXT:    s_setpc_b64 s[4:5]
+; GFX10-SCRATCH-NEXT:    s_setpc_b64 s[30:31]
 entry:
   %alloca = alloca double, align 8, addrspace(5)
   tail call amdgpu_gfx void @byval_align16_f64_arg(<32 x i32> %val, double addrspace(5)* byval(double) align 16 %alloca)
   ret void
 }
 
 ; inreg arguments are put in sgprs
 define amdgpu_gfx void @test_call_external_void_func_i1_imm_inreg() #0 {
 ; GFX9-LABEL: test_call_external_void_func_i1_imm_inreg:
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_mov_b32_e32 v0, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_i1_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_i1_inreg@rel32@hi+12
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_i1_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_i1_inreg@rel32@hi+12
 ; GFX9-NEXT:    buffer_store_byte v0, off, s[0:3], s32
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_i1_imm_inreg:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 2
 ; GFX10-NEXT:    v_mov_b32_e32 v0, 1
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i1_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_i1_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    buffer_store_byte v0, off, s[0:3], s32
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_i1_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_i1_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 3
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX9-NEXT:    s_movk_i32 s4, 0x7b
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 2
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_i8_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_i8_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_i8_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_i8_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 3
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_i8_imm_inreg:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 3
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i8_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_i8_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_movk_i32 s4, 0x7b
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 2
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_i8_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_i8_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 3
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3

 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 3
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX9-NEXT:    s_movk_i32 s4, 0x7b
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 2
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_i16_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_i16_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_i16_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_i16_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 3
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_i16_imm_inreg:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 3
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i16_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_i16_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_movk_i32 s4, 0x7b
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 2
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_i16_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_i16_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 3
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3

 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 3
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX9-NEXT:    s_mov_b32 s4, 42
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 2
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_i32_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_i32_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_i32_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_i32_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 3
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_i32_imm_inreg:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 3
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i32_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_i32_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 42
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 2
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_i32_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_i32_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 3
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3

 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX9-NEXT:    s_movk_i32 s4, 0x7b
 ; GFX9-NEXT:    s_mov_b32 s5, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 3
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_i64_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_i64_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_i64_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_i64_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 3
 ; GFX9-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 4
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_i64_imm_inreg:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 4
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_i64_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_i64_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_movk_i32 s4, 0x7b
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    s_mov_b32 s5, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 3
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_i64_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_i64_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 3
 ; GFX10-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 4
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload

 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 6
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s6, 2
+; GFX9-NEXT:    s_mov_b64 s[34:35], 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s7, 3
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 4
-; GFX9-NEXT:    v_writelane_b32 v40, s31, 5
-; GFX9-NEXT:    s_mov_b64 s[30:31], 0
-; GFX9-NEXT:    s_load_dwordx4 s[4:7], s[30:31], 0x0
+; GFX9-NEXT:    s_load_dwordx4 s[4:7], s[34:35], 0x0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v2i64_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v2i64_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    v_writelane_b32 v40, s30, 4
+; GFX9-NEXT:    v_writelane_b32 v40, s31, 5
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v2i64_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v2i64_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 4
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 5
 ; GFX9-NEXT:    v_readlane_b32 s7, v40, 3
 ; GFX9-NEXT:    v_readlane_b32 s6, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 6
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_v2i64_inreg:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 6
+; GFX10-NEXT:    s_mov_b64 s[34:35], 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s7, 3
+; GFX10-NEXT:    s_load_dwordx4 s[4:7], s[34:35], 0x0
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2i64_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v2i64_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 4
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 5
-; GFX10-NEXT:    s_mov_b64 s[30:31], 0
-; GFX10-NEXT:    s_load_dwordx4 s[4:7], s[30:31], 0x0
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v2i64_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v2i64_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 4
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 5
 ; GFX10-NEXT:    v_readlane_b32 s7, v40, 3
 ; GFX10-NEXT:    v_readlane_b32 s6, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 6

 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 4
 ; GFX9-NEXT:    s_mov_b32 s4, 1
 ; GFX9-NEXT:    s_mov_b32 s5, 2
 ; GFX9-NEXT:    s_mov_b32 s6, 3
 ; GFX9-NEXT:    s_mov_b32 s7, 4
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 5
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v2i64_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v2i64_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v2i64_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v2i64_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 4
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 5
 ; GFX9-NEXT:    v_readlane_b32 s7, v40, 3
 ; GFX9-NEXT:    v_readlane_b32 s6, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 6

 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 6
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2i64_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v2i64_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    s_mov_b32 s5, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-NEXT:    s_mov_b32 s6, 3
 ; GFX10-NEXT:    v_writelane_b32 v40, s7, 3
 ; GFX10-NEXT:    s_mov_b32 s7, 4
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 4
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 5
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v2i64_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v2i64_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 4
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 5
 ; GFX10-NEXT:    v_readlane_b32 s7, v40, 3
 ; GFX10-NEXT:    v_readlane_b32 s6, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 6

 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 8
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s6, 2
+; GFX9-NEXT:    s_mov_b64 s[34:35], 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s7, 3
+; GFX9-NEXT:    s_load_dwordx4 s[4:7], s[34:35], 0x0
 ; GFX9-NEXT:    v_writelane_b32 v40, s8, 4
 ; GFX9-NEXT:    v_writelane_b32 v40, s9, 5
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 6
-; GFX9-NEXT:    v_writelane_b32 v40, s31, 7
-; GFX9-NEXT:    s_mov_b64 s[30:31], 0
-; GFX9-NEXT:    s_load_dwordx4 s[4:7], s[30:31], 0x0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
+; GFX9-NEXT:    v_writelane_b32 v40, s30, 6
 ; GFX9-NEXT:    s_mov_b32 s8, 1
 ; GFX9-NEXT:    s_mov_b32 s9, 2
-; GFX9-NEXT:    s_getpc_b64 s[34:35]
-; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v3i64_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v3i64_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
+; GFX9-NEXT:    v_writelane_b32 v40, s31, 7
+; GFX9-NEXT:    s_getpc_b64 s[36:37]
+; GFX9-NEXT:    s_add_u32 s36, s36, external_void_func_v3i64_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s37, s37, external_void_func_v3i64_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[36:37]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 6
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 7
 ; GFX9-NEXT:    v_readlane_b32 s9, v40, 5
 ; GFX9-NEXT:    v_readlane_b32 s8, v40, 4
 ; GFX9-NEXT:    v_readlane_b32 s7, v40, 3
 ; GFX9-NEXT:    v_readlane_b32 s6, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0

 ; 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:    buffer_store_dword v40, 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 v40, s33, 8
+; GFX10-NEXT:    s_mov_b64 s[34:35], 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s7, 3
+; GFX10-NEXT:    s_load_dwordx4 s[4:7], s[34:35], 0x0
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3i64_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v3i64_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s8, 4
 ; GFX10-NEXT:    s_mov_b32 s8, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s9, 5
 ; GFX10-NEXT:    s_mov_b32 s9, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 6
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 7
-; GFX10-NEXT:    s_mov_b64 s[30:31], 0
-; GFX10-NEXT:    s_load_dwordx4 s[4:7], s[30:31], 0x0
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v3i64_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v3i64_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 6
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 7
 ; GFX10-NEXT:    v_readlane_b32 s9, v40, 5
 ; GFX10-NEXT:    v_readlane_b32 s8, v40, 4
 ; GFX10-NEXT:    v_readlane_b32 s7, v40, 3
 ; GFX10-NEXT:    v_readlane_b32 s6, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0

 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 10
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX9-NEXT:    v_writelane_b32 v40, s7, 3
+; GFX9-NEXT:    s_mov_b64 s[34:35], 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s8, 4
+; GFX9-NEXT:    s_load_dwordx4 s[4:7], s[34:35], 0x0
 ; GFX9-NEXT:    v_writelane_b32 v40, s9, 5
 ; GFX9-NEXT:    v_writelane_b32 v40, s10, 6
 ; GFX9-NEXT:    v_writelane_b32 v40, s11, 7
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 8
-; GFX9-NEXT:    v_writelane_b32 v40, s31, 9
-; GFX9-NEXT:    s_mov_b64 s[30:31], 0
-; GFX9-NEXT:    s_load_dwordx4 s[4:7], s[30:31], 0x0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
+; GFX9-NEXT:    v_writelane_b32 v40, s30, 8
 ; GFX9-NEXT:    s_mov_b32 s8, 1
 ; GFX9-NEXT:    s_mov_b32 s9, 2
 ; GFX9-NEXT:    s_mov_b32 s10, 3
 ; GFX9-NEXT:    s_mov_b32 s11, 4
-; GFX9-NEXT:    s_getpc_b64 s[34:35]
-; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v4i64_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v4i64_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
+; GFX9-NEXT:    v_writelane_b32 v40, s31, 9
+; GFX9-NEXT:    s_getpc_b64 s[36:37]
+; GFX9-NEXT:    s_add_u32 s36, s36, external_void_func_v4i64_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s37, s37, external_void_func_v4i64_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[36:37]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 8
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 9
 ; GFX9-NEXT:    v_readlane_b32 s11, v40, 7
 ; GFX9-NEXT:    v_readlane_b32 s10, v40, 6
 ; GFX9-NEXT:    v_readlane_b32 s9, v40, 5
 ; GFX9-NEXT:    v_readlane_b32 s8, v40, 4
 ; GFX9-NEXT:    v_readlane_b32 s7, v40, 3
 ; GFX9-NEXT:    v_readlane_b32 s6, v40, 2

 ; 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:    buffer_store_dword v40, 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 v40, s33, 10
+; GFX10-NEXT:    s_mov_b64 s[34:35], 0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s7, 3
+; GFX10-NEXT:    s_load_dwordx4 s[4:7], s[34:35], 0x0
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v4i64_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v4i64_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s8, 4
 ; GFX10-NEXT:    s_mov_b32 s8, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s9, 5
 ; GFX10-NEXT:    s_mov_b32 s9, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s10, 6
 ; GFX10-NEXT:    s_mov_b32 s10, 3
 ; GFX10-NEXT:    v_writelane_b32 v40, s11, 7
 ; GFX10-NEXT:    s_mov_b32 s11, 4
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 8
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 9
-; GFX10-NEXT:    s_mov_b64 s[30:31], 0
-; GFX10-NEXT:    s_load_dwordx4 s[4:7], s[30:31], 0x0
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v4i64_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v4i64_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 8
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 9
 ; GFX10-NEXT:    v_readlane_b32 s11, v40, 7
 ; GFX10-NEXT:    v_readlane_b32 s10, v40, 6
 ; GFX10-NEXT:    v_readlane_b32 s9, v40, 5
 ; GFX10-NEXT:    v_readlane_b32 s8, v40, 4
 ; GFX10-NEXT:    v_readlane_b32 s7, v40, 3
 ; GFX10-NEXT:    v_readlane_b32 s6, v40, 2

 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 3
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX9-NEXT:    s_movk_i32 s4, 0x4400
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 2
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_f16_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_f16_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_f16_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_f16_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 3
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_f16_imm_inreg:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 3
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_f16_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_f16_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_movk_i32 s4, 0x4400
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 2
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_f16_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_f16_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 3
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3

 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 3
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX9-NEXT:    s_mov_b32 s4, 4.0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 2
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_f32_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_f32_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_f32_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_f32_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 3
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_f32_imm_inreg:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 3
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_f32_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_f32_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 4.0
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 2
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_f32_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_f32_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 3
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3

 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX9-NEXT:    s_mov_b32 s4, 1.0
 ; GFX9-NEXT:    s_mov_b32 s5, 2.0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 3
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v2f32_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v2f32_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v2f32_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v2f32_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 3
 ; GFX9-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 4
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_v2f32_imm_inreg:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 4
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2f32_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v2f32_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 1.0
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    s_mov_b32 s5, 2.0
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 3
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v2f32_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v2f32_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 3
 ; GFX10-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 4
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload

 ; GFX9-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 3
 ; GFX9-NEXT:    s_mov_b32 s4, 1.0
 ; GFX9-NEXT:    s_mov_b32 s5, 2.0
 ; GFX9-NEXT:    s_mov_b32 s6, 4.0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 4
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v3f32_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v3f32_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v3f32_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v3f32_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 3
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 4
 ; GFX9-NEXT:    v_readlane_b32 s6, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 5
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_v3f32_imm_inreg:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 5
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3f32_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v3f32_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 1.0
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    s_mov_b32 s5, 2.0
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-NEXT:    s_mov_b32 s6, 4.0
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 3
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 4
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v3f32_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v3f32_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 3
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 4
 ; GFX10-NEXT:    v_readlane_b32 s6, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 5
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1

 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 5
 ; GFX9-NEXT:    s_mov_b32 s4, 1.0
 ; GFX9-NEXT:    s_mov_b32 s5, 2.0
 ; GFX9-NEXT:    s_mov_b32 s6, 4.0
 ; GFX9-NEXT:    s_mov_b32 s7, -1.0
 ; GFX9-NEXT:    s_mov_b32 s8, 0.5
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 6
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v5f32_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v5f32_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v5f32_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v5f32_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 5
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 6
 ; GFX9-NEXT:    v_readlane_b32 s8, v40, 4
 ; GFX9-NEXT:    v_readlane_b32 s7, v40, 3
 ; GFX9-NEXT:    v_readlane_b32 s6, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00

 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 7
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v5f32_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v5f32_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 1.0
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    s_mov_b32 s5, 2.0
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-NEXT:    s_mov_b32 s6, 4.0
 ; GFX10-NEXT:    v_writelane_b32 v40, s7, 3
 ; GFX10-NEXT:    s_mov_b32 s7, -1.0
 ; GFX10-NEXT:    v_writelane_b32 v40, s8, 4
 ; GFX10-NEXT:    s_mov_b32 s8, 0.5
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 5
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 6
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v5f32_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v5f32_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 5
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 6
 ; GFX10-NEXT:    v_readlane_b32 s8, v40, 4
 ; GFX10-NEXT:    v_readlane_b32 s7, v40, 3
 ; GFX10-NEXT:    v_readlane_b32 s6, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00

 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX9-NEXT:    s_mov_b32 s4, 0
 ; GFX9-NEXT:    s_mov_b32 s5, 0x40100000
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 3
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_f64_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_f64_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_f64_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_f64_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 3
 ; GFX9-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 4
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_f64_imm_inreg:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 4
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_f64_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_f64_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    s_mov_b32 s5, 0x40100000
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 3
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_f64_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_f64_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 3
 ; GFX10-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 4
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload

 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 4
 ; GFX9-NEXT:    s_mov_b32 s4, 0
 ; GFX9-NEXT:    s_mov_b32 s5, 2.0
 ; GFX9-NEXT:    s_mov_b32 s6, 0
 ; GFX9-NEXT:    s_mov_b32 s7, 0x40100000
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 5
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v2f64_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v2f64_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v2f64_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v2f64_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 4
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 5
 ; GFX9-NEXT:    v_readlane_b32 s7, v40, 3
 ; GFX9-NEXT:    v_readlane_b32 s6, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 6

 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 6
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2f64_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v2f64_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    s_mov_b32 s5, 2.0
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-NEXT:    s_mov_b32 s6, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s7, 3
 ; GFX10-NEXT:    s_mov_b32 s7, 0x40100000
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 4
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 5
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v2f64_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v2f64_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 4
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 5
 ; GFX10-NEXT:    v_readlane_b32 s7, v40, 3
 ; GFX10-NEXT:    v_readlane_b32 s6, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 6

 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 6
 ; GFX9-NEXT:    s_mov_b32 s4, 0
 ; GFX9-NEXT:    s_mov_b32 s5, 2.0
 ; GFX9-NEXT:    s_mov_b32 s6, 0
 ; GFX9-NEXT:    s_mov_b32 s7, 0x40100000
 ; GFX9-NEXT:    s_mov_b32 s8, 0
 ; GFX9-NEXT:    s_mov_b32 s9, 0x40200000
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 7
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v3f64_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v3f64_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v3f64_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v3f64_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 6
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 7
 ; GFX9-NEXT:    v_readlane_b32 s9, v40, 5
 ; GFX9-NEXT:    v_readlane_b32 s8, v40, 4
 ; GFX9-NEXT:    v_readlane_b32 s7, v40, 3
 ; GFX9-NEXT:    v_readlane_b32 s6, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0

 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 8
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3f64_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v3f64_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    s_mov_b32 s5, 2.0
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-NEXT:    s_mov_b32 s6, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s7, 3
 ; GFX10-NEXT:    s_mov_b32 s7, 0x40100000
 ; GFX10-NEXT:    v_writelane_b32 v40, s8, 4
 ; GFX10-NEXT:    s_mov_b32 s8, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s9, 5
 ; GFX10-NEXT:    s_mov_b32 s9, 0x40200000
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 6
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 7
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v3f64_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v3f64_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 6
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 7
 ; GFX10-NEXT:    v_readlane_b32 s9, v40, 5
 ; GFX10-NEXT:    v_readlane_b32 s8, v40, 4
 ; GFX10-NEXT:    v_readlane_b32 s7, v40, 3
 ; GFX10-NEXT:    v_readlane_b32 s6, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0

 ; GFX9-LABEL: test_call_external_void_func_v2i16_inreg:
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 3
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
-; GFX9-NEXT:    s_load_dword s4, s[30:31], 0x0
+; GFX9-NEXT:    s_load_dword s4, s[34:35], 0x0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
+; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 2
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v2i16_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v2i16_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v2i16_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v2i16_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 3
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_v2i16_inreg:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 3
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    s_load_dword s4, s[34:35], 0x0
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2i16_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v2i16_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 1
-; GFX10-NEXT:    s_load_dword s4, s[30:31], 0x0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 2
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v2i16_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v2i16_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 3
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3

 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 4
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
-; GFX9-NEXT:    s_load_dwordx2 s[4:5], s[30:31], 0x0
+; GFX9-NEXT:    s_load_dwordx2 s[4:5], s[34:35], 0x0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
+; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 3
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v3i16_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v3i16_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v3i16_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v3i16_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 3
 ; GFX9-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 4
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload

 ; GFX10-NEXT:    buffer_store_dword v40, 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 v40, s33, 4
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-NEXT:    s_load_dwordx2 s[4:5], s[34:35], 0x0
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3i16_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v3i16_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 2
-; GFX10-NEXT:    s_load_dwordx2 s[4:5], s[30:31], 0x0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 3
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v3i16_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v3i16_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 3
 ; GFX10-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 4
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload

 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 4
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
-; GFX9-NEXT:    s_load_dwordx2 s[4:5], s[30:31], 0x0
+; GFX9-NEXT:    s_load_dwordx2 s[4:5], s[34:35], 0x0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
+; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 3
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v3f16_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v3f16_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v3f16_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v3f16_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 3
 ; GFX9-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 4
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload

 ; GFX10-NEXT:    buffer_store_dword v40, 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 v40, s33, 4
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-NEXT:    s_load_dwordx2 s[4:5], s[34:35], 0x0
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3f16_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v3f16_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 2
-; GFX10-NEXT:    s_load_dwordx2 s[4:5], s[30:31], 0x0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 3
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v3f16_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v3f16_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 3
 ; GFX10-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 4
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload

 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX9-NEXT:    s_mov_b32 s4, 0x20001
 ; GFX9-NEXT:    s_mov_b32 s5, 3
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 3
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v3i16_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v3i16_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v3i16_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v3i16_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 3
 ; GFX9-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 4
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_v3i16_imm_inreg:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 4
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3i16_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v3i16_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 0x20001
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    s_mov_b32 s5, 3
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 3
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v3i16_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v3i16_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 3
 ; GFX10-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 4
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload

 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX9-NEXT:    s_mov_b32 s4, 0x40003c00
 ; GFX9-NEXT:    s_movk_i32 s5, 0x4400
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 3
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v3f16_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v3f16_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v3f16_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v3f16_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 3
 ; GFX9-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 4
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_v3f16_imm_inreg:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 4
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3f16_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v3f16_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 0x40003c00
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    s_movk_i32 s5, 0x4400
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 3
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v3f16_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v3f16_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 3
 ; GFX10-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 4
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload

 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 4
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
-; GFX9-NEXT:    s_load_dwordx2 s[4:5], s[30:31], 0x0
+; GFX9-NEXT:    s_load_dwordx2 s[4:5], s[34:35], 0x0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
+; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 3
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v4i16_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v4i16_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v4i16_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v4i16_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 3
 ; GFX9-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 4
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload

 ; GFX10-NEXT:    buffer_store_dword v40, 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 v40, s33, 4
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-NEXT:    s_load_dwordx2 s[4:5], s[34:35], 0x0
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v4i16_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v4i16_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 2
-; GFX10-NEXT:    s_load_dwordx2 s[4:5], s[30:31], 0x0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 3
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v4i16_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v4i16_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 3
 ; GFX10-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 4
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload

 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX9-NEXT:    s_mov_b32 s4, 0x20001
 ; GFX9-NEXT:    s_mov_b32 s5, 0x40003
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 3
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v4i16_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v4i16_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v4i16_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v4i16_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 3
 ; GFX9-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 4
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_v4i16_imm_inreg:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 4
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v4i16_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v4i16_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 0x20001
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    s_mov_b32 s5, 0x40003
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 3
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v4i16_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v4i16_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 3
 ; GFX10-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 4
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload

 ; GFX9-LABEL: test_call_external_void_func_v2f16_inreg:
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 3
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
-; GFX9-NEXT:    s_load_dword s4, s[30:31], 0x0
+; GFX9-NEXT:    s_load_dword s4, s[34:35], 0x0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
+; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 2
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v2f16_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v2f16_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v2f16_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v2f16_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 3
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_v2f16_inreg:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 3
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
+; GFX10-NEXT:    s_load_dword s4, s[34:35], 0x0
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2f16_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v2f16_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 1
-; GFX10-NEXT:    s_load_dword s4, s[30:31], 0x0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 2
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v2f16_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v2f16_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 3
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3

 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 4
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
-; GFX9-NEXT:    s_load_dwordx2 s[4:5], s[30:31], 0x0
+; GFX9-NEXT:    s_load_dwordx2 s[4:5], s[34:35], 0x0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
+; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 3
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v2i32_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v2i32_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v2i32_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v2i32_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 3
 ; GFX9-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 4
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload

 ; GFX10-NEXT:    buffer_store_dword v40, 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 v40, s33, 4
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
+; GFX10-NEXT:    s_load_dwordx2 s[4:5], s[34:35], 0x0
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2i32_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v2i32_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 2
-; GFX10-NEXT:    s_load_dwordx2 s[4:5], s[30:31], 0x0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 3
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v2i32_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v2i32_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 3
 ; GFX10-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 4
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload

 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX9-NEXT:    s_mov_b32 s4, 1
 ; GFX9-NEXT:    s_mov_b32 s5, 2
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 3
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v2i32_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v2i32_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v2i32_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v2i32_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 3
 ; GFX9-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 4
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_v2i32_imm_inreg:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 4
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v2i32_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v2i32_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    s_mov_b32 s5, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 3
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v2i32_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v2i32_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 3
 ; GFX10-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 4
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload

 ; GFX9-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 3
 ; GFX9-NEXT:    s_mov_b32 s4, 3
 ; GFX9-NEXT:    s_mov_b32 s5, 4
 ; GFX9-NEXT:    s_mov_b32 s6, 5
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 4
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v3i32_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v3i32_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v3i32_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v3i32_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 3
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 4
 ; GFX9-NEXT:    v_readlane_b32 s6, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 5
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_external_void_func_v3i32_imm_inreg:
 ; 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:    buffer_store_dword v40, 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 v40, s33, 5
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3i32_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v3i32_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 3
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    s_mov_b32 s5, 4
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-NEXT:    s_mov_b32 s6, 5
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 3
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 4
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v3i32_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v3i32_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 3
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 4
 ; GFX10-NEXT:    v_readlane_b32 s6, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 5
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1

 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 4
 ; GFX9-NEXT:    s_mov_b32 s4, 3
 ; GFX9-NEXT:    s_mov_b32 s5, 4
 ; GFX9-NEXT:    s_mov_b32 s6, 5
 ; GFX9-NEXT:    s_mov_b32 s7, 6
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 5
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v3i32_i32_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v3i32_i32_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v3i32_i32_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v3i32_i32_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 4
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 5
 ; GFX9-NEXT:    v_readlane_b32 s7, v40, 3
 ; GFX9-NEXT:    v_readlane_b32 s6, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 6

 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 6
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v3i32_i32_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v3i32_i32_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 3
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    s_mov_b32 s5, 4
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-NEXT:    s_mov_b32 s6, 5
 ; GFX10-NEXT:    v_writelane_b32 v40, s7, 3
 ; GFX10-NEXT:    s_mov_b32 s7, 6
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 4
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 5
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v3i32_i32_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v3i32_i32_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 4
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 5
 ; GFX10-NEXT:    v_readlane_b32 s7, v40, 3
 ; GFX10-NEXT:    v_readlane_b32 s6, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 6

 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 6
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX9-NEXT:    v_writelane_b32 v40, s7, 3
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 4
-; GFX9-NEXT:    s_load_dwordx4 s[4:7], s[30:31], 0x0
+; GFX9-NEXT:    s_load_dwordx4 s[4:7], s[34:35], 0x0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
+; GFX9-NEXT:    v_writelane_b32 v40, s30, 4
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 5
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v4i32_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v4i32_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v4i32_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v4i32_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 4
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 5
 ; GFX9-NEXT:    v_readlane_b32 s7, v40, 3
 ; GFX9-NEXT:    v_readlane_b32 s6, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 6

 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 6
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s7, 3
+; GFX10-NEXT:    s_load_dwordx4 s[4:7], s[34:35], 0x0
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v4i32_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v4i32_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 4
-; GFX10-NEXT:    s_load_dwordx4 s[4:7], s[30:31], 0x0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 5
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v4i32_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v4i32_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 4
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 5
 ; GFX10-NEXT:    v_readlane_b32 s7, v40, 3
 ; GFX10-NEXT:    v_readlane_b32 s6, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 6

 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 4
 ; GFX9-NEXT:    s_mov_b32 s4, 1
 ; GFX9-NEXT:    s_mov_b32 s5, 2
 ; GFX9-NEXT:    s_mov_b32 s6, 3
 ; GFX9-NEXT:    s_mov_b32 s7, 4
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 5
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v4i32_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v4i32_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v4i32_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v4i32_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 4
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 5
 ; GFX9-NEXT:    v_readlane_b32 s7, v40, 3
 ; GFX9-NEXT:    v_readlane_b32 s6, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 6

 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 6
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v4i32_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v4i32_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    s_mov_b32 s5, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-NEXT:    s_mov_b32 s6, 3
 ; GFX10-NEXT:    v_writelane_b32 v40, s7, 3
 ; GFX10-NEXT:    s_mov_b32 s7, 4
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 4
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 5
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v4i32_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v4i32_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 4
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 5
 ; GFX10-NEXT:    v_readlane_b32 s7, v40, 3
 ; GFX10-NEXT:    v_readlane_b32 s6, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 6

 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 5
 ; GFX9-NEXT:    s_mov_b32 s4, 1
 ; GFX9-NEXT:    s_mov_b32 s5, 2
 ; GFX9-NEXT:    s_mov_b32 s6, 3
 ; GFX9-NEXT:    s_mov_b32 s7, 4
 ; GFX9-NEXT:    s_mov_b32 s8, 5
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 6
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v5i32_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v5i32_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v5i32_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v5i32_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 5
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 6
 ; GFX9-NEXT:    v_readlane_b32 s8, v40, 4
 ; GFX9-NEXT:    v_readlane_b32 s7, v40, 3
 ; GFX9-NEXT:    v_readlane_b32 s6, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00

 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 7
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v5i32_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v5i32_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    s_mov_b32 s5, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-NEXT:    s_mov_b32 s6, 3
 ; GFX10-NEXT:    v_writelane_b32 v40, s7, 3
 ; GFX10-NEXT:    s_mov_b32 s7, 4
 ; GFX10-NEXT:    v_writelane_b32 v40, s8, 4
 ; GFX10-NEXT:    s_mov_b32 s8, 5
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 5
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 6
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v5i32_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v5i32_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 5
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 6
 ; GFX10-NEXT:    v_readlane_b32 s8, v40, 4
 ; GFX10-NEXT:    v_readlane_b32 s7, v40, 3
 ; GFX10-NEXT:    v_readlane_b32 s6, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s5, v40, 1
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00

 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 10
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s6, 2
+; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX9-NEXT:    v_writelane_b32 v40, s7, 3
 ; GFX9-NEXT:    v_writelane_b32 v40, s8, 4
 ; GFX9-NEXT:    v_writelane_b32 v40, s9, 5
 ; GFX9-NEXT:    v_writelane_b32 v40, s10, 6
 ; GFX9-NEXT:    v_writelane_b32 v40, s11, 7
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 8
-; GFX9-NEXT:    v_writelane_b32 v40, s31, 9
-; GFX9-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
+; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9-NEXT:    s_load_dwordx8 s[4:11], s[34:35], 0x0
 ; 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, external_void_func_v8i32_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v8i32_inreg@rel32@hi+12
-; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX9-NEXT:    s_load_dwordx8 s[4:11], s[30:31], 0x0
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
+; GFX9-NEXT:    v_writelane_b32 v40, s30, 8
+; GFX9-NEXT:    v_writelane_b32 v40, s31, 9
+; GFX9-NEXT:    s_getpc_b64 s[36:37]
+; GFX9-NEXT:    s_add_u32 s36, s36, external_void_func_v8i32_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s37, s37, external_void_func_v8i32_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[36:37]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 8
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 9
 ; GFX9-NEXT:    v_readlane_b32 s11, v40, 7
 ; GFX9-NEXT:    v_readlane_b32 s10, v40, 6
 ; GFX9-NEXT:    v_readlane_b32 s9, v40, 5
 ; GFX9-NEXT:    v_readlane_b32 s8, v40, 4
 ; GFX9-NEXT:    v_readlane_b32 s7, v40, 3
 ; GFX9-NEXT:    v_readlane_b32 s6, v40, 2

 ; 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:    buffer_store_dword v40, 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 v40, s33, 10
+; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s7, 3
 ; GFX10-NEXT:    v_writelane_b32 v40, s8, 4
 ; GFX10-NEXT:    v_writelane_b32 v40, s9, 5
 ; GFX10-NEXT:    v_writelane_b32 v40, s10, 6
 ; GFX10-NEXT:    v_writelane_b32 v40, s11, 7
+; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10-NEXT:    s_load_dwordx8 s[4:11], s[34:35], 0x0
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v8i32_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v8i32_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 8
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 9
-; GFX10-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
-; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX10-NEXT:    s_load_dwordx8 s[4:11], s[30:31], 0x0
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v8i32_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v8i32_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 8
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 9
 ; GFX10-NEXT:    v_readlane_b32 s11, v40, 7
 ; GFX10-NEXT:    v_readlane_b32 s10, v40, 6
 ; GFX10-NEXT:    v_readlane_b32 s9, v40, 5
 ; GFX10-NEXT:    v_readlane_b32 s8, v40, 4
 ; GFX10-NEXT:    v_readlane_b32 s7, v40, 3
 ; GFX10-NEXT:    v_readlane_b32 s6, v40, 2

 ; GFX9-NEXT:    s_mov_b32 s5, 2
 ; GFX9-NEXT:    s_mov_b32 s6, 3
 ; GFX9-NEXT:    s_mov_b32 s7, 4
 ; GFX9-NEXT:    s_mov_b32 s8, 5
 ; GFX9-NEXT:    s_mov_b32 s9, 6
 ; GFX9-NEXT:    s_mov_b32 s10, 7
 ; GFX9-NEXT:    s_mov_b32 s11, 8
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 9
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v8i32_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v8i32_inreg@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v8i32_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v8i32_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 8
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 9
 ; GFX9-NEXT:    v_readlane_b32 s11, v40, 7
 ; GFX9-NEXT:    v_readlane_b32 s10, v40, 6
 ; GFX9-NEXT:    v_readlane_b32 s9, v40, 5
 ; GFX9-NEXT:    v_readlane_b32 s8, v40, 4
 ; GFX9-NEXT:    v_readlane_b32 s7, v40, 3
 ; GFX9-NEXT:    v_readlane_b32 s6, v40, 2

 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 10
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v8i32_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v8i32_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    s_mov_b32 s5, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-NEXT:    s_mov_b32 s6, 3
 ; GFX10-NEXT:    v_writelane_b32 v40, s7, 3
 ; GFX10-NEXT:    s_mov_b32 s7, 4
 ; GFX10-NEXT:    v_writelane_b32 v40, s8, 4
 ; GFX10-NEXT:    s_mov_b32 s8, 5
 ; GFX10-NEXT:    v_writelane_b32 v40, s9, 5
 ; GFX10-NEXT:    s_mov_b32 s9, 6
 ; GFX10-NEXT:    v_writelane_b32 v40, s10, 6
 ; GFX10-NEXT:    s_mov_b32 s10, 7
 ; GFX10-NEXT:    v_writelane_b32 v40, s11, 7
 ; GFX10-NEXT:    s_mov_b32 s11, 8
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 8
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 9
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v8i32_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v8i32_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 8
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 9
 ; GFX10-NEXT:    v_readlane_b32 s11, v40, 7
 ; GFX10-NEXT:    v_readlane_b32 s10, v40, 6
 ; GFX10-NEXT:    v_readlane_b32 s9, v40, 5
 ; GFX10-NEXT:    v_readlane_b32 s8, v40, 4
 ; GFX10-NEXT:    v_readlane_b32 s7, v40, 3
 ; GFX10-NEXT:    v_readlane_b32 s6, v40, 2

 ; GFX9-NEXT:    v_writelane_b32 v40, s7, 3
 ; GFX9-NEXT:    v_writelane_b32 v40, s8, 4
 ; GFX9-NEXT:    v_writelane_b32 v40, s9, 5
 ; GFX9-NEXT:    v_writelane_b32 v40, s10, 6
 ; GFX9-NEXT:    v_writelane_b32 v40, s11, 7
 ; GFX9-NEXT:    v_writelane_b32 v40, s12, 8
 ; GFX9-NEXT:    v_writelane_b32 v40, s13, 9
 ; GFX9-NEXT:    v_writelane_b32 v40, s14, 10
+; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX9-NEXT:    v_writelane_b32 v40, s15, 11
 ; GFX9-NEXT:    v_writelane_b32 v40, s16, 12
 ; GFX9-NEXT:    v_writelane_b32 v40, s17, 13
 ; GFX9-NEXT:    v_writelane_b32 v40, s18, 14
 ; GFX9-NEXT:    v_writelane_b32 v40, s19, 15
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 16
-; GFX9-NEXT:    v_writelane_b32 v40, s31, 17
-; GFX9-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
+; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9-NEXT:    s_load_dwordx16 s[4:19], s[34:35], 0x0
 ; 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, external_void_func_v16i32_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v16i32_inreg@rel32@hi+12
-; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX9-NEXT:    s_load_dwordx16 s[4:19], s[30:31], 0x0
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
+; GFX9-NEXT:    v_writelane_b32 v40, s30, 16
+; GFX9-NEXT:    v_writelane_b32 v40, s31, 17
+; GFX9-NEXT:    s_getpc_b64 s[36:37]
+; GFX9-NEXT:    s_add_u32 s36, s36, external_void_func_v16i32_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s37, s37, external_void_func_v16i32_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[36:37]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 16
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 17
 ; GFX9-NEXT:    v_readlane_b32 s19, v40, 15
 ; GFX9-NEXT:    v_readlane_b32 s18, v40, 14
 ; GFX9-NEXT:    v_readlane_b32 s17, v40, 13
 ; GFX9-NEXT:    v_readlane_b32 s16, v40, 12
 ; GFX9-NEXT:    v_readlane_b32 s15, v40, 11
 ; GFX9-NEXT:    v_readlane_b32 s14, v40, 10

 ; 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:    buffer_store_dword v40, 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 v40, s33, 18
+; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s7, 3
 ; GFX10-NEXT:    v_writelane_b32 v40, s8, 4
 ; GFX10-NEXT:    v_writelane_b32 v40, s9, 5
 ; GFX10-NEXT:    v_writelane_b32 v40, s10, 6
 ; GFX10-NEXT:    v_writelane_b32 v40, s11, 7
 ; GFX10-NEXT:    v_writelane_b32 v40, s12, 8
 ; GFX10-NEXT:    v_writelane_b32 v40, s13, 9
 ; GFX10-NEXT:    v_writelane_b32 v40, s14, 10
 ; GFX10-NEXT:    v_writelane_b32 v40, s15, 11
 ; GFX10-NEXT:    v_writelane_b32 v40, s16, 12
 ; GFX10-NEXT:    v_writelane_b32 v40, s17, 13
 ; GFX10-NEXT:    v_writelane_b32 v40, s18, 14
 ; GFX10-NEXT:    v_writelane_b32 v40, s19, 15
+; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10-NEXT:    s_load_dwordx16 s[4:19], s[34:35], 0x0
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v16i32_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v16i32_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 16
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 17
-; GFX10-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
-; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX10-NEXT:    s_load_dwordx16 s[4:19], s[30:31], 0x0
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v16i32_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v16i32_inreg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 16
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 17
 ; GFX10-NEXT:    v_readlane_b32 s19, v40, 15
 ; GFX10-NEXT:    v_readlane_b32 s18, v40, 14
 ; GFX10-NEXT:    v_readlane_b32 s17, v40, 13
 ; GFX10-NEXT:    v_readlane_b32 s16, v40, 12
 ; GFX10-NEXT:    v_readlane_b32 s15, v40, 11
 ; GFX10-NEXT:    v_readlane_b32 s14, v40, 10

 ; GFX9-NEXT:    v_writelane_b32 v40, s11, 7
 ; GFX9-NEXT:    v_writelane_b32 v40, s12, 8
 ; GFX9-NEXT:    v_writelane_b32 v40, s13, 9
 ; GFX9-NEXT:    v_writelane_b32 v40, s14, 10
 ; GFX9-NEXT:    v_writelane_b32 v40, s15, 11
 ; GFX9-NEXT:    v_writelane_b32 v40, s16, 12
 ; GFX9-NEXT:    v_writelane_b32 v40, s17, 13
 ; GFX9-NEXT:    v_writelane_b32 v40, s18, 14
+; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX9-NEXT:    v_writelane_b32 v40, s19, 15
 ; GFX9-NEXT:    v_writelane_b32 v40, s20, 16
 ; GFX9-NEXT:    v_writelane_b32 v40, s21, 17
 ; GFX9-NEXT:    v_writelane_b32 v40, s22, 18
 ; GFX9-NEXT:    v_writelane_b32 v40, s23, 19
+; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9-NEXT:    s_load_dwordx16 s[4:19], s[34:35], 0x0
+; GFX9-NEXT:    s_load_dwordx16 s[36:51], s[34:35], 0x40
 ; GFX9-NEXT:    v_writelane_b32 v40, s24, 20
 ; GFX9-NEXT:    v_writelane_b32 v40, s25, 21
 ; GFX9-NEXT:    v_writelane_b32 v40, s26, 22
 ; GFX9-NEXT:    v_writelane_b32 v40, s27, 23
-; GFX9-NEXT:    v_writelane_b32 v40, s28, 24
-; GFX9-NEXT:    v_writelane_b32 v40, s29, 25
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 26
-; GFX9-NEXT:    v_writelane_b32 v40, s31, 27
-; GFX9-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
-; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX9-NEXT:    s_load_dwordx16 s[4:19], s[30:31], 0x0
-; GFX9-NEXT:    s_load_dwordx16 s[36:51], s[30:31], 0x40
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v32i32_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v32i32_inreg@rel32@hi+12
+; GFX9-NEXT:    v_writelane_b32 v40, s28, 24
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX9-NEXT:    v_mov_b32_e32 v0, s46
+; GFX9-NEXT:    v_writelane_b32 v40, s29, 25
 ; GFX9-NEXT:    v_mov_b32_e32 v1, s47
 ; GFX9-NEXT:    v_mov_b32_e32 v2, s48
 ; GFX9-NEXT:    v_mov_b32_e32 v3, s49
 ; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], s32
 ; GFX9-NEXT:    buffer_store_dword v1, off, s[0:3], s32 offset:4
 ; GFX9-NEXT:    buffer_store_dword v2, off, s[0:3], s32 offset:8
 ; GFX9-NEXT:    buffer_store_dword v3, off, s[0:3], s32 offset:12
 ; GFX9-NEXT:    v_mov_b32_e32 v0, s50
+; GFX9-NEXT:    v_writelane_b32 v40, s30, 26
 ; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:16
 ; GFX9-NEXT:    v_mov_b32_e32 v0, s51
 ; GFX9-NEXT:    s_mov_b32 s20, s36
 ; GFX9-NEXT:    s_mov_b32 s21, s37
 ; GFX9-NEXT:    s_mov_b32 s22, s38
 ; GFX9-NEXT:    s_mov_b32 s23, s39
 ; GFX9-NEXT:    s_mov_b32 s24, s40
 ; GFX9-NEXT:    s_mov_b32 s25, s41
 ; GFX9-NEXT:    s_mov_b32 s26, s42
 ; GFX9-NEXT:    s_mov_b32 s27, s43
 ; GFX9-NEXT:    s_mov_b32 s28, s44
 ; GFX9-NEXT:    s_mov_b32 s29, s45
+; GFX9-NEXT:    v_writelane_b32 v40, s31, 27
 ; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:20
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v32i32_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v32i32_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 26
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 27
 ; GFX9-NEXT:    v_readlane_b32 s29, v40, 25
 ; GFX9-NEXT:    v_readlane_b32 s28, v40, 24
 ; GFX9-NEXT:    v_readlane_b32 s27, v40, 23
 ; GFX9-NEXT:    v_readlane_b32 s26, v40, 22
 ; GFX9-NEXT:    v_readlane_b32 s25, v40, 21
 ; GFX9-NEXT:    v_readlane_b32 s24, v40, 20

 ; 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:    buffer_store_dword v40, 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 v40, s33, 28
+; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s7, 3
 ; GFX10-NEXT:    v_writelane_b32 v40, s8, 4
 ; GFX10-NEXT:    v_writelane_b32 v40, s9, 5
 ; GFX10-NEXT:    v_writelane_b32 v40, s10, 6
 ; GFX10-NEXT:    v_writelane_b32 v40, s11, 7
 ; GFX10-NEXT:    v_writelane_b32 v40, s12, 8
 ; GFX10-NEXT:    v_writelane_b32 v40, s13, 9
 ; GFX10-NEXT:    v_writelane_b32 v40, s14, 10
 ; GFX10-NEXT:    v_writelane_b32 v40, s15, 11
 ; GFX10-NEXT:    v_writelane_b32 v40, s16, 12
 ; GFX10-NEXT:    v_writelane_b32 v40, s17, 13
 ; GFX10-NEXT:    v_writelane_b32 v40, s18, 14
 ; GFX10-NEXT:    v_writelane_b32 v40, s19, 15
+; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10-NEXT:    s_clause 0x1
+; GFX10-NEXT:    s_load_dwordx16 s[36:51], s[34:35], 0x40
+; GFX10-NEXT:    s_load_dwordx16 s[4:19], s[34:35], 0x0
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v32i32_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v32i32_inreg@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s20, 16
 ; GFX10-NEXT:    v_writelane_b32 v40, s21, 17
 ; GFX10-NEXT:    v_writelane_b32 v40, s22, 18
-; GFX10-NEXT:    v_writelane_b32 v40, s23, 19
-; GFX10-NEXT:    v_writelane_b32 v40, s24, 20
-; GFX10-NEXT:    v_writelane_b32 v40, s25, 21
-; GFX10-NEXT:    v_writelane_b32 v40, s26, 22
-; GFX10-NEXT:    v_writelane_b32 v40, s27, 23
-; GFX10-NEXT:    v_writelane_b32 v40, s28, 24
-; GFX10-NEXT:    v_writelane_b32 v40, s29, 25
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 26
-; GFX10-NEXT:    v_writelane_b32 v40, s31, 27
-; GFX10-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
-; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX10-NEXT:    s_clause 0x1
-; GFX10-NEXT:    s_load_dwordx16 s[36:51], s[30:31], 0x40
-; GFX10-NEXT:    s_load_dwordx16 s[4:19], s[30:31], 0x0
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v32i32_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v32i32_inreg@rel32@hi+12
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX10-NEXT:    v_mov_b32_e32 v0, s46
+; GFX10-NEXT:    v_writelane_b32 v40, s23, 19
 ; GFX10-NEXT:    v_mov_b32_e32 v1, s47
 ; GFX10-NEXT:    v_mov_b32_e32 v2, s48
 ; GFX10-NEXT:    v_mov_b32_e32 v3, s49
 ; GFX10-NEXT:    s_mov_b32 s20, s36
+; GFX10-NEXT:    v_writelane_b32 v40, s24, 20
 ; GFX10-NEXT:    s_mov_b32 s21, s37
 ; GFX10-NEXT:    s_mov_b32 s22, s38
 ; GFX10-NEXT:    s_mov_b32 s23, s39
 ; GFX10-NEXT:    s_mov_b32 s24, s40
+; GFX10-NEXT:    v_writelane_b32 v40, s25, 21
 ; GFX10-NEXT:    s_mov_b32 s25, s41
-; GFX10-NEXT:    s_mov_b32 s26, s42
-; GFX10-NEXT:    s_mov_b32 s27, s43
-; GFX10-NEXT:    s_mov_b32 s28, s44
-; GFX10-NEXT:    s_mov_b32 s29, s45
 ; GFX10-NEXT:    v_mov_b32_e32 v4, s50
 ; GFX10-NEXT:    v_mov_b32_e32 v5, s51
 ; GFX10-NEXT:    buffer_store_dword v0, off, s[0:3], s32
 ; GFX10-NEXT:    buffer_store_dword v1, off, s[0:3], s32 offset:4
 ; GFX10-NEXT:    buffer_store_dword v2, off, s[0:3], s32 offset:8
 ; GFX10-NEXT:    buffer_store_dword v3, off, s[0:3], s32 offset:12
 ; GFX10-NEXT:    buffer_store_dword v4, off, s[0:3], s32 offset:16
 ; GFX10-NEXT:    buffer_store_dword v5, off, s[0:3], s32 offset:20
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    v_writelane_b32 v40, s26, 22
+; GFX10-NEXT:    s_mov_b32 s26, s42
+; GFX10-NEXT:    v_writelane_b32 v40, s27, 23
+; GFX10-NEXT:    s_mov_b32 s27, s43
+; GFX10-NEXT:    v_writelane_b32 v40, s28, 24
+; GFX10-NEXT:    s_mov_b32 s28, s44
+; GFX10-NEXT:    v_writelane_b32 v40, s29, 25
+; GFX10-NEXT:    s_mov_b32 s29, s45
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 26
+; GFX10-NEXT:    v_writelane_b32 v40, s31, 27
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 26
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 27
 ; GFX10-NEXT:    v_readlane_b32 s29, v40, 25
 ; GFX10-NEXT:    v_readlane_b32 s28, v40, 24
 ; GFX10-NEXT:    v_readlane_b32 s27, v40, 23
 ; GFX10-NEXT:    v_readlane_b32 s26, v40, 22
 ; GFX10-NEXT:    v_readlane_b32 s25, v40, 21
 ; GFX10-NEXT:    v_readlane_b32 s24, v40, 20

 ; GFX9-NEXT:    v_writelane_b32 v40, s10, 6
 ; GFX9-NEXT:    v_writelane_b32 v40, s11, 7
 ; GFX9-NEXT:    v_writelane_b32 v40, s12, 8
 ; GFX9-NEXT:    v_writelane_b32 v40, s13, 9
 ; GFX9-NEXT:    v_writelane_b32 v40, s14, 10
 ; GFX9-NEXT:    v_writelane_b32 v40, s15, 11
 ; GFX9-NEXT:    v_writelane_b32 v40, s16, 12
 ; GFX9-NEXT:    v_writelane_b32 v40, s17, 13
+; GFX9-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX9-NEXT:    v_writelane_b32 v40, s18, 14
 ; GFX9-NEXT:    v_writelane_b32 v40, s19, 15
 ; GFX9-NEXT:    v_writelane_b32 v40, s20, 16
 ; GFX9-NEXT:    v_writelane_b32 v40, s21, 17
 ; GFX9-NEXT:    v_writelane_b32 v40, s22, 18
+; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9-NEXT:    s_load_dword s52, s[34:35], 0x0
+; GFX9-NEXT:    ; kill: killed $sgpr34_sgpr35
+; GFX9-NEXT:    ; kill: killed $sgpr34_sgpr35
+; GFX9-NEXT:    s_load_dwordx16 s[4:19], s[34:35], 0x0
+; GFX9-NEXT:    s_load_dwordx16 s[36:51], s[34:35], 0x40
 ; GFX9-NEXT:    v_writelane_b32 v40, s23, 19
 ; GFX9-NEXT:    v_writelane_b32 v40, s24, 20
 ; GFX9-NEXT:    v_writelane_b32 v40, s25, 21
-; GFX9-NEXT:    v_writelane_b32 v40, s26, 22
-; GFX9-NEXT:    v_writelane_b32 v40, s27, 23
-; GFX9-NEXT:    v_writelane_b32 v40, s28, 24
-; GFX9-NEXT:    v_writelane_b32 v40, s29, 25
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 26
-; GFX9-NEXT:    v_writelane_b32 v40, s31, 27
-; GFX9-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
+; GFX9-NEXT:    v_writelane_b32 v40, s26, 22
 ; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX9-NEXT:    s_load_dword s34, s[30:31], 0x0
-; GFX9-NEXT:    ; kill: killed $sgpr30_sgpr31
-; GFX9-NEXT:    ; kill: killed $sgpr30_sgpr31
-; GFX9-NEXT:    s_load_dwordx16 s[4:19], s[30:31], 0x0
-; GFX9-NEXT:    s_load_dwordx16 s[36:51], s[30:31], 0x40
-; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX9-NEXT:    v_mov_b32_e32 v0, s34
+; GFX9-NEXT:    v_mov_b32_e32 v0, s52
+; GFX9-NEXT:    v_writelane_b32 v40, s27, 23
 ; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:24
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_v32i32_i32_inreg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_v32i32_i32_inreg@rel32@hi+12
 ; GFX9-NEXT:    v_mov_b32_e32 v0, s46
+; GFX9-NEXT:    v_writelane_b32 v40, s28, 24
 ; GFX9-NEXT:    v_mov_b32_e32 v1, s47
+; GFX9-NEXT:    v_mov_b32_e32 v2, s48
 ; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], s32
 ; GFX9-NEXT:    buffer_store_dword v1, off, s[0:3], s32 offset:4
-; GFX9-NEXT:    v_mov_b32_e32 v0, s48
-; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:8
+; GFX9-NEXT:    buffer_store_dword v2, off, s[0:3], s32 offset:8
 ; GFX9-NEXT:    v_mov_b32_e32 v0, s49
+; GFX9-NEXT:    v_writelane_b32 v40, s29, 25
 ; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:12
 ; GFX9-NEXT:    v_mov_b32_e32 v0, s50
+; GFX9-NEXT:    v_writelane_b32 v40, s30, 26
 ; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:16
 ; GFX9-NEXT:    v_mov_b32_e32 v0, s51
 ; GFX9-NEXT:    s_mov_b32 s20, s36
 ; GFX9-NEXT:    s_mov_b32 s21, s37
 ; GFX9-NEXT:    s_mov_b32 s22, s38
 ; GFX9-NEXT:    s_mov_b32 s23, s39
 ; GFX9-NEXT:    s_mov_b32 s24, s40
 ; GFX9-NEXT:    s_mov_b32 s25, s41
 ; GFX9-NEXT:    s_mov_b32 s26, s42
 ; GFX9-NEXT:    s_mov_b32 s27, s43
 ; GFX9-NEXT:    s_mov_b32 s28, s44
 ; GFX9-NEXT:    s_mov_b32 s29, s45
+; GFX9-NEXT:    v_writelane_b32 v40, s31, 27
 ; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:20
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_v32i32_i32_inreg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_v32i32_i32_inreg@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 26
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 27
 ; GFX9-NEXT:    v_readlane_b32 s29, v40, 25
 ; GFX9-NEXT:    v_readlane_b32 s28, v40, 24
 ; GFX9-NEXT:    v_readlane_b32 s27, v40, 23
 ; GFX9-NEXT:    v_readlane_b32 s26, v40, 22
 ; GFX9-NEXT:    v_readlane_b32 s25, v40, 21
 ; GFX9-NEXT:    v_readlane_b32 s24, v40, 20

 ; 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:    buffer_store_dword v40, 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 v40, s33, 28
+; GFX10-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s5, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s6, 2
 ; GFX10-NEXT:    v_writelane_b32 v40, s7, 3
 ; GFX10-NEXT:    v_writelane_b32 v40, s8, 4
 ; GFX10-NEXT:    v_writelane_b32 v40, s9, 5
 ; GFX10-NEXT:    v_writelane_b32 v40, s10, 6
 ; GFX10-NEXT:    v_writelane_b32 v40, s11, 7
 ; GFX10-NEXT:    v_writelane_b32 v40, s12, 8
 ; GFX10-NEXT:    v_writelane_b32 v40, s13, 9
 ; GFX10-NEXT:    v_writelane_b32 v40, s14, 10
 ; GFX10-NEXT:    v_writelane_b32 v40, s15, 11
 ; GFX10-NEXT:    v_writelane_b32 v40, s16, 12
 ; GFX10-NEXT:    v_writelane_b32 v40, s17, 13
 ; GFX10-NEXT:    v_writelane_b32 v40, s18, 14
 ; GFX10-NEXT:    v_writelane_b32 v40, s19, 15
-; GFX10-NEXT:    v_writelane_b32 v40, s20, 16
-; GFX10-NEXT:    v_writelane_b32 v40, s21, 17
-; GFX10-NEXT:    v_writelane_b32 v40, s22, 18
-; GFX10-NEXT:    v_writelane_b32 v40, s23, 19
-; GFX10-NEXT:    v_writelane_b32 v40, s24, 20
-; GFX10-NEXT:    v_writelane_b32 v40, s25, 21
-; GFX10-NEXT:    v_writelane_b32 v40, s26, 22
-; GFX10-NEXT:    v_writelane_b32 v40, s27, 23
-; GFX10-NEXT:    v_writelane_b32 v40, s28, 24
-; GFX10-NEXT:    v_writelane_b32 v40, s29, 25
-; GFX10-NEXT:    v_writelane_b32 v40, s30, 26
-; GFX10-NEXT:    v_writelane_b32 v40, s31, 27
-; GFX10-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
 ; GFX10-NEXT:    s_clause 0x2
-; GFX10-NEXT:    s_load_dword s34, s[30:31], 0x0
+; GFX10-NEXT:    s_load_dword s52, s[34:35], 0x0
 ; GFX10-NEXT:    ; meta instruction
 ; GFX10-NEXT:    ; meta instruction
-; GFX10-NEXT:    s_load_dwordx16 s[36:51], s[30:31], 0x40
-; GFX10-NEXT:    s_load_dwordx16 s[4:19], s[30:31], 0x0
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_v32i32_i32_inreg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_v32i32_i32_inreg@rel32@hi+12
+; GFX10-NEXT:    s_load_dwordx16 s[36:51], s[34:35], 0x40
+; GFX10-NEXT:    s_load_dwordx16 s[4:19], s[34:35], 0x0
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_v32i32_i32_inreg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_v32i32_i32_inreg@rel32@hi+12
+; GFX10-NEXT:    v_writelane_b32 v40, s20, 16
+; GFX10-NEXT:    v_writelane_b32 v40, s21, 17
+; GFX10-NEXT:    v_writelane_b32 v40, s22, 18
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX10-NEXT:    v_mov_b32_e32 v0, s34
+; GFX10-NEXT:    v_mov_b32_e32 v0, s52
 ; GFX10-NEXT:    v_mov_b32_e32 v1, s47
-; GFX10-NEXT:    v_mov_b32_e32 v2, s48
+; GFX10-NEXT:    v_writelane_b32 v40, s23, 19
 ; GFX10-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:24
 ; GFX10-NEXT:    v_mov_b32_e32 v0, s46
+; GFX10-NEXT:    v_mov_b32_e32 v2, s48
 ; GFX10-NEXT:    v_mov_b32_e32 v3, s49
+; GFX10-NEXT:    v_writelane_b32 v40, s24, 20
 ; GFX10-NEXT:    s_mov_b32 s20, s36
 ; GFX10-NEXT:    s_mov_b32 s21, s37
 ; GFX10-NEXT:    s_mov_b32 s22, s38
 ; GFX10-NEXT:    s_mov_b32 s23, s39
+; GFX10-NEXT:    v_writelane_b32 v40, s25, 21
 ; GFX10-NEXT:    s_mov_b32 s24, s40
 ; GFX10-NEXT:    s_mov_b32 s25, s41
-; GFX10-NEXT:    s_mov_b32 s26, s42
-; GFX10-NEXT:    s_mov_b32 s27, s43
-; GFX10-NEXT:    s_mov_b32 s28, s44
-; GFX10-NEXT:    s_mov_b32 s29, s45
 ; GFX10-NEXT:    v_mov_b32_e32 v4, s50
 ; GFX10-NEXT:    v_mov_b32_e32 v5, s51
+; GFX10-NEXT:    v_writelane_b32 v40, s26, 22
+; GFX10-NEXT:    s_mov_b32 s26, s42
 ; GFX10-NEXT:    buffer_store_dword v0, off, s[0:3], s32
 ; GFX10-NEXT:    buffer_store_dword v1, off, s[0:3], s32 offset:4
 ; GFX10-NEXT:    buffer_store_dword v2, off, s[0:3], s32 offset:8
 ; GFX10-NEXT:    buffer_store_dword v3, off, s[0:3], s32 offset:12
 ; GFX10-NEXT:    buffer_store_dword v4, off, s[0:3], s32 offset:16
 ; GFX10-NEXT:    buffer_store_dword v5, off, s[0:3], s32 offset:20
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    v_writelane_b32 v40, s27, 23
+; GFX10-NEXT:    s_mov_b32 s27, s43
+; GFX10-NEXT:    v_writelane_b32 v40, s28, 24
+; GFX10-NEXT:    s_mov_b32 s28, s44
+; GFX10-NEXT:    v_writelane_b32 v40, s29, 25
+; GFX10-NEXT:    s_mov_b32 s29, s45
+; GFX10-NEXT:    v_writelane_b32 v40, s30, 26
+; GFX10-NEXT:    v_writelane_b32 v40, s31, 27
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 26
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 27
 ; GFX10-NEXT:    v_readlane_b32 s29, v40, 25
 ; GFX10-NEXT:    v_readlane_b32 s28, v40, 24
 ; GFX10-NEXT:    v_readlane_b32 s27, v40, 23
 ; GFX10-NEXT:    v_readlane_b32 s26, v40, 22
 ; GFX10-NEXT:    v_readlane_b32 s25, v40, 21
 ; GFX10-NEXT:    v_readlane_b32 s24, v40, 20

 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    buffer_load_dword v32, off, s[0:3], s33
 ; GFX9-NEXT:    buffer_load_dword v33, off, s[0:3], s33 offset:4
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, stack_passed_f64_arg@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, stack_passed_f64_arg@rel32@hi+12
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, stack_passed_f64_arg@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, stack_passed_f64_arg@rel32@hi+12
 ; GFX9-NEXT:    s_waitcnt vmcnt(1)
 ; GFX9-NEXT:    buffer_store_dword v32, off, s[0:3], s32
 ; GFX9-NEXT:    s_waitcnt vmcnt(1)
 ; GFX9-NEXT:    buffer_store_dword v33, off, s[0:3], s32 offset:4
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)

 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_clause 0x1
 ; GFX10-NEXT:    buffer_load_dword v32, off, s[0:3], s33
 ; GFX10-NEXT:    buffer_load_dword v33, off, s[0:3], s33 offset:4
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, stack_passed_f64_arg@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, stack_passed_f64_arg@rel32@hi+12
 ; GFX10-NEXT:    s_waitcnt vmcnt(1)
 ; GFX10-NEXT:    buffer_store_dword v32, off, s[0:3], s32
 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    buffer_store_dword v33, off, s[0:3], s32 offset:4
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, stack_passed_f64_arg@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, stack_passed_f64_arg@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    v_mov_b32_e32 v25, 8
 ; GFX9-NEXT:    v_mov_b32_e32 v26, 8
 ; GFX9-NEXT:    v_mov_b32_e32 v27, 9
 ; GFX9-NEXT:    v_mov_b32_e32 v28, 9
 ; GFX9-NEXT:    v_mov_b32_e32 v29, 9
 ; GFX9-NEXT:    v_mov_b32_e32 v30, 10
 ; GFX9-NEXT:    v_mov_b32_e32 v31, 11
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_12xv3i32@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_12xv3i32@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_12xv3i32@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_12xv3i32@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)

 ; GFX10-NEXT:    v_mov_b32_e32 v25, 8
 ; GFX10-NEXT:    v_mov_b32_e32 v26, 8
 ; GFX10-NEXT:    v_mov_b32_e32 v27, 9
 ; GFX10-NEXT:    v_mov_b32_e32 v28, 9
 ; GFX10-NEXT:    v_mov_b32_e32 v29, 9
 ; GFX10-NEXT:    v_mov_b32_e32 v30, 10
 ; GFX10-NEXT:    v_mov_b32_e32 v31, 11
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_12xv3i32@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_12xv3i32@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_12xv3i32@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_12xv3i32@rel32@hi+12
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    v_mov_b32_e32 v25, 5
 ; GFX9-NEXT:    v_mov_b32_e32 v26, 5
 ; GFX9-NEXT:    v_mov_b32_e32 v27, 5
 ; GFX9-NEXT:    v_mov_b32_e32 v28, 5
 ; GFX9-NEXT:    v_mov_b32_e32 v29, 5
 ; GFX9-NEXT:    v_mov_b32_e32 v30, 6
 ; GFX9-NEXT:    v_mov_b32_e32 v31, 7
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_8xv5i32@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_8xv5i32@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_8xv5i32@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_8xv5i32@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)

 ; GFX10-NEXT:    v_mov_b32_e32 v25, 5
 ; GFX10-NEXT:    v_mov_b32_e32 v26, 5
 ; GFX10-NEXT:    v_mov_b32_e32 v27, 5
 ; GFX10-NEXT:    v_mov_b32_e32 v28, 5
 ; GFX10-NEXT:    v_mov_b32_e32 v29, 5
 ; GFX10-NEXT:    v_mov_b32_e32 v30, 6
 ; GFX10-NEXT:    v_mov_b32_e32 v31, 7
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_8xv5i32@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_8xv5i32@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_8xv5i32@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_8xv5i32@rel32@hi+12
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    v_mov_b32_e32 v25, 0x40a00000
 ; GFX9-NEXT:    v_mov_b32_e32 v26, 0x40a00000
 ; GFX9-NEXT:    v_mov_b32_e32 v27, 0x40a00000
 ; GFX9-NEXT:    v_mov_b32_e32 v28, 0x40a00000
 ; GFX9-NEXT:    v_mov_b32_e32 v29, 0x40a00000
 ; GFX9-NEXT:    v_mov_b32_e32 v30, 0x40c00000
 ; GFX9-NEXT:    v_mov_b32_e32 v31, 0x40e00000
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_8xv5f32@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_8xv5f32@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_8xv5f32@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_8xv5f32@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)

 ; GFX10-NEXT:    v_mov_b32_e32 v25, 0x40a00000
 ; GFX10-NEXT:    v_mov_b32_e32 v26, 0x40a00000
 ; GFX10-NEXT:    v_mov_b32_e32 v27, 0x40a00000
 ; GFX10-NEXT:    v_mov_b32_e32 v28, 0x40a00000
 ; GFX10-NEXT:    v_mov_b32_e32 v29, 0x40a00000
 ; GFX10-NEXT:    v_mov_b32_e32 v30, 0x40c00000
 ; GFX10-NEXT:    v_mov_b32_e32 v31, 0x40e00000
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_8xv5f32@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_8xv5f32@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_8xv5f32@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_8xv5f32@rel32@hi+12
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

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

 ; GFX10-NEXT:    s_waitcnt vmcnt(0)
 ; GFX10-NEXT:    s_setpc_b64 s[30:31]
   call amdgpu_gfx void @external_void_func_void()
   call void asm sideeffect "", ""() #0
   call amdgpu_gfx void @external_void_func_void()
   ret void
 }
 
-define amdgpu_gfx void @void_func_void_clobber_s30_s31() #1 {
-; GFX9-LABEL: void_func_void_clobber_s30_s31:
+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_mov_b64 s[36:37], s[30:31]
+; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-NEXT:    buffer_store_dword v0, off, s[0:3], s32 ; 4-byte Folded Spill
+; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-NEXT:    v_writelane_b32 v0, s28, 0
+; GFX9-NEXT:    v_writelane_b32 v0, s29, 1
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; clobber
 ; GFX9-NEXT:    ;;#ASMEND
-; GFX9-NEXT:    s_setpc_b64 s[36:37]
+; 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:    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_s30_s31:
+; 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_mov_b64 s[36:37], s[30:31]
+; GFX10-NEXT:    s_or_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:    ;;#ASMSTART
 ; GFX10-NEXT:    ; clobber
 ; GFX10-NEXT:    ;;#ASMEND
-; GFX10-NEXT:    s_setpc_b64 s[36:37]
-  call void asm sideeffect "; clobber", "~{s[30:31]}"() #0
+; 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:    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]
+; GCN:          v_writelane_b32 v0, s28, 0
+; GCN:          v_writelane_b32 v0, s29, 1
+
+; GCN:          v_readlane_b32 s28, v0, 0
+; GCN:          v_readlane_b32 s29, v0, 1
+  call void asm sideeffect "; clobber", "~{s[28:29]}"() #0
   ret void
 }
 
 define amdgpu_gfx void @test_call_void_func_void_mayclobber_s31(i32 addrspace(1)* %out) #0 {
 ; GFX9-LABEL: test_call_void_func_void_mayclobber_s31:
 ; GFX9:       ; %bb.0:
 ; GFX9-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 3
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
-; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
+; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 2
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; def s31
 ; GFX9-NEXT:    ;;#ASMEND
 ; GFX9-NEXT:    s_mov_b32 s4, s31
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_void@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_void@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_void@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_void@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    s_mov_b32 s31, s4
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; use s31
 ; GFX9-NEXT:    ;;#ASMEND
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00

 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    buffer_store_dword v41, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; def v31
 ; GFX9-NEXT:    ;;#ASMEND
 ; GFX9-NEXT:    v_mov_b32_e32 v41, v31
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_void@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_void@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_void@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_void@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_mov_b32_e32 v31, v41
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; use v31
 ; GFX9-NEXT:    ;;#ASMEND
 ; GFX9-NEXT:    buffer_load_dword v41, off, s[0:3], s33 ; 4-byte Folded Reload
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00

 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    buffer_store_dword v41, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; def v31
 ; GFX10-NEXT:    ;;#ASMEND
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_mov_b32_e32 v41, v31
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_void@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_void@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_void@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_void@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_mov_b32_e32 v31, v41
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; use v31
 ; GFX10-NEXT:    ;;#ASMEND
 ; GFX10-NEXT:    buffer_load_dword v41, off, s[0:3], s33 ; 4-byte Folded Reload
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00

 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 2
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; def s33
 ; GFX9-NEXT:    ;;#ASMEND
 ; GFX9-NEXT:    s_mov_b32 s4, s33
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_void@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_void@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_void@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_void@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    s_mov_b32 s33, s4
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 1
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; use s33
 ; GFX9-NEXT:    ;;#ASMEND
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00

 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 3
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_void@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_void@rel32@hi+12
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; def s33
 ; GFX10-NEXT:    ;;#ASMEND
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, s33
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 2
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_void@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_void@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    s_mov_b32 s33, s4
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 1
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; use s33
 ; GFX10-NEXT:    ;;#ASMEND
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00

 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 3
 ; GFX9-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
-; GFX9-NEXT:    v_writelane_b32 v40, s31, 2
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; def s34
 ; GFX9-NEXT:    ;;#ASMEND
+; GFX9-NEXT:    v_writelane_b32 v40, s31, 2
 ; GFX9-NEXT:    s_mov_b32 s4, s34
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_void@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_void@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_void@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_void@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 1
 ; GFX9-NEXT:    s_mov_b32 s34, s4
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; use s34
 ; GFX9-NEXT:    ;;#ASMEND
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00

 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 3
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[36:37]
+; GFX10-NEXT:    s_add_u32 s36, s36, external_void_func_void@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s37, s37, external_void_func_void@rel32@hi+12
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; def s34
 ; GFX10-NEXT:    ;;#ASMEND
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, s34
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 2
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_void@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_void@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[36:37]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 1
 ; GFX10-NEXT:    s_mov_b32 s34, s4
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; use s34
 ; GFX10-NEXT:    ;;#ASMEND
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00

 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v41, s30, 0
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX9-NEXT:    v_writelane_b32 v41, s31, 1
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; def v40
 ; GFX9-NEXT:    ;;#ASMEND
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_void@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_void@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_void@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_void@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; use v40
 ; GFX9-NEXT:    ;;#ASMEND
 ; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s33 ; 4-byte Folded Reload
 ; GFX9-NEXT:    v_readlane_b32 s30, v41, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v41, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v41, 2

 ; GFX10-NEXT:    v_writelane_b32 v41, s33, 2
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
 ; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; def v40
 ; GFX10-NEXT:    ;;#ASMEND
 ; GFX10-NEXT:    v_writelane_b32 v41, s30, 0
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_void@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_void@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v41, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_void@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_void@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; use v40
 ; GFX10-NEXT:    ;;#ASMEND
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s33 ; 4-byte Folded Reload
 ; GFX10-NEXT:    v_readlane_b32 s30, v41, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v41, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v41, 2

 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, void_func_void_clobber_s33@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, void_func_void_clobber_s33@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, void_func_void_clobber_s33@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, void_func_void_clobber_s33@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_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 s34, -1
 ; GFX10-NEXT:    buffer_store_dword v40, 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 v40, s33, 2
 ; 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, void_func_void_clobber_s33@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, void_func_void_clobber_s33@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, void_func_void_clobber_s33@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, void_func_void_clobber_s33@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-NEXT:    v_writelane_b32 v40, s33, 2
 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, void_func_void_clobber_s34@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, void_func_void_clobber_s34@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, void_func_void_clobber_s34@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, void_func_void_clobber_s34@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX9-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-NEXT:    buffer_load_dword v40, 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: test_call_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 s34, -1
 ; GFX10-NEXT:    buffer_store_dword v40, 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 v40, s33, 2
 ; 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, void_func_void_clobber_s34@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, void_func_void_clobber_s34@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 0
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 1
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, void_func_void_clobber_s34@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, void_func_void_clobber_s34@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 0
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 1
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 2
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34

 ; GFX9-NEXT:    s_mov_b32 s33, s32
 ; GFX9-NEXT:    s_addk_i32 s32, 0x400
 ; GFX9-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX9-NEXT:    v_writelane_b32 v40, s31, 2
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; def s40
 ; GFX9-NEXT:    ;;#ASMEND
 ; GFX9-NEXT:    s_mov_b32 s4, s40
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_void@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_void@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_void@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_void@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 1
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; use s4
 ; GFX9-NEXT:    ;;#ASMEND
 ; GFX9-NEXT:    v_readlane_b32 s31, v40, 2
 ; GFX9-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 3

 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX10-NEXT:    s_or_saveexec_b32 s34, -1
 ; GFX10-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX10-NEXT:    s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT:    s_mov_b32 exec_lo, s34
 ; GFX10-NEXT:    v_writelane_b32 v40, s33, 3
 ; GFX10-NEXT:    s_mov_b32 s33, s32
 ; GFX10-NEXT:    s_addk_i32 s32, 0x200
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_void@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_void@rel32@hi+12
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; def s40
 ; GFX10-NEXT:    ;;#ASMEND
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, s40
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 2
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_void@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_void@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 1
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; use s4
 ; GFX10-NEXT:    ;;#ASMEND
 ; GFX10-NEXT:    v_readlane_b32 s31, v40, 2
 ; GFX10-NEXT:    v_readlane_b32 s4, v40, 0
 ; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
 ; GFX10-NEXT:    v_readlane_b32 s33, v40, 3

 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; def s40
 ; GFX9-NEXT:    ;;#ASMEND
 ; GFX9-NEXT:    s_mov_b32 s4, s40
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; def v32
 ; GFX9-NEXT:    ;;#ASMEND
 ; GFX9-NEXT:    v_mov_b32_e32 v41, v32
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, external_void_func_void@rel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, external_void_func_void@rel32@hi+12
-; GFX9-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-NEXT:    s_add_u32 s34, s34, external_void_func_void@rel32@lo+4
+; GFX9-NEXT:    s_addc_u32 s35, s35, external_void_func_void@rel32@hi+12
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; use s4
 ; GFX9-NEXT:    ;;#ASMEND
 ; GFX9-NEXT:    ;;#ASMSTART
 ; GFX9-NEXT:    ; use v41
 ; GFX9-NEXT:    ;;#ASMEND
 ; GFX9-NEXT:    buffer_load_dword v41, off, s[0:3], s33 ; 4-byte Folded Reload
 ; GFX9-NEXT:    v_readlane_b32 s30, v40, 1

 ; GFX10-NEXT:    ; def s40
 ; GFX10-NEXT:    ;;#ASMEND
 ; GFX10-NEXT:    v_writelane_b32 v40, s4, 0
 ; GFX10-NEXT:    s_mov_b32 s4, s40
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; def v32
 ; GFX10-NEXT:    ;;#ASMEND
 ; GFX10-NEXT:    v_mov_b32_e32 v41, v32
+; GFX10-NEXT:    s_getpc_b64 s[34:35]
+; GFX10-NEXT:    s_add_u32 s34, s34, external_void_func_void@rel32@lo+4
+; GFX10-NEXT:    s_addc_u32 s35, s35, external_void_func_void@rel32@hi+12
 ; GFX10-NEXT:    v_writelane_b32 v40, s30, 1
 ; GFX10-NEXT:    v_writelane_b32 v40, s31, 2
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, external_void_func_void@rel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, external_void_func_void@rel32@hi+12
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; use s4
 ; GFX10-NEXT:    ;;#ASMEND
 ; GFX10-NEXT:    ;;#ASMSTART
 ; GFX10-NEXT:    ; use v41
 ; GFX10-NEXT:    ;;#ASMEND
 ; GFX10-NEXT:    buffer_load_dword v41, off, s[0:3], s33 ; 4-byte Folded Reload
 ; GFX10-NEXT:    v_readlane_b32 s30, v40, 1

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_mov_b64 s[36:37], s[30:31]
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, return_i1@gotpcrel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, return_i1@gotpcrel32@hi+12
-; GFX9-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
+; GFX9-NEXT:    s_or_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:    v_writelane_b32 v1, s33, 2
+; 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[30:31]
-; GFX9-NEXT:    s_setpc_b64 s[36:37]
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
+; GFX9-NEXT:    v_readlane_b32 s30, v1, 0
+; GFX9-NEXT:    v_readlane_b32 s31, v1, 1
+; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
+; GFX9-NEXT:    v_readlane_b32 s33, v1, 2
+; GFX9-NEXT:    s_or_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_mov_b64 s[36:37], s[30:31]
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, return_i1@gotpcrel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, return_i1@gotpcrel32@hi+12
-; GFX10-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
+; GFX10-NEXT:    s_or_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:    v_writelane_b32 v1, s33, 2
+; 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:    s_load_dwordx2 s[34:35], s[34:35], 0x0
+; GFX10-NEXT:    v_writelane_b32 v1, s30, 0
+; GFX10-NEXT:    v_writelane_b32 v1, s31, 1
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
-; GFX10-NEXT:    s_setpc_b64 s[36:37]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
+; GFX10-NEXT:    v_readlane_b32 s30, v1, 0
+; GFX10-NEXT:    v_readlane_b32 s31, v1, 1
+; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
+; GFX10-NEXT:    v_readlane_b32 s33, v1, 2
+; GFX10-NEXT:    s_or_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]
 entry:
   call amdgpu_gfx i1 @return_i1()
   ret void
 }
 
 define amdgpu_gfx i16 @return_i16() #0 {
 ; GFX9-LABEL: return_i16:
 ; GFX9:       ; %bb.0: ; %entry

 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_mov_b64 s[36:37], s[30:31]
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, return_i16@gotpcrel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, return_i16@gotpcrel32@hi+12
-; GFX9-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
+; GFX9-NEXT:    s_or_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:    v_writelane_b32 v1, s33, 2
+; 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[30:31]
-; GFX9-NEXT:    s_setpc_b64 s[36:37]
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
+; GFX9-NEXT:    v_readlane_b32 s30, v1, 0
+; GFX9-NEXT:    v_readlane_b32 s31, v1, 1
+; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
+; GFX9-NEXT:    v_readlane_b32 s33, v1, 2
+; GFX9-NEXT:    s_or_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_mov_b64 s[36:37], s[30:31]
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, return_i16@gotpcrel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, return_i16@gotpcrel32@hi+12
-; GFX10-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
+; GFX10-NEXT:    s_or_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:    v_writelane_b32 v1, s33, 2
+; 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:    s_load_dwordx2 s[34:35], s[34:35], 0x0
+; GFX10-NEXT:    v_writelane_b32 v1, s30, 0
+; GFX10-NEXT:    v_writelane_b32 v1, s31, 1
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
-; GFX10-NEXT:    s_setpc_b64 s[36:37]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
+; GFX10-NEXT:    v_readlane_b32 s30, v1, 0
+; GFX10-NEXT:    v_readlane_b32 s31, v1, 1
+; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
+; GFX10-NEXT:    v_readlane_b32 s33, v1, 2
+; GFX10-NEXT:    s_or_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]
 entry:
   call amdgpu_gfx i16 @return_i16()
   ret void
 }
 
 define amdgpu_gfx <2 x i16> @return_2xi16() #0 {
 ; GFX9-LABEL: return_2xi16:
 ; GFX9:       ; %bb.0: ; %entry

 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_mov_b64 s[36:37], s[30:31]
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, return_2xi16@gotpcrel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, return_2xi16@gotpcrel32@hi+12
-; GFX9-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
+; GFX9-NEXT:    s_or_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:    v_writelane_b32 v1, s33, 2
+; 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[30:31]
-; GFX9-NEXT:    s_setpc_b64 s[36:37]
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
+; GFX9-NEXT:    v_readlane_b32 s30, v1, 0
+; GFX9-NEXT:    v_readlane_b32 s31, v1, 1
+; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
+; GFX9-NEXT:    v_readlane_b32 s33, v1, 2
+; GFX9-NEXT:    s_or_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_mov_b64 s[36:37], s[30:31]
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, return_2xi16@gotpcrel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, return_2xi16@gotpcrel32@hi+12
-; GFX10-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
+; GFX10-NEXT:    s_or_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:    v_writelane_b32 v1, s33, 2
+; 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:    s_load_dwordx2 s[34:35], s[34:35], 0x0
+; GFX10-NEXT:    v_writelane_b32 v1, s30, 0
+; GFX10-NEXT:    v_writelane_b32 v1, s31, 1
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
-; GFX10-NEXT:    s_setpc_b64 s[36:37]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
+; GFX10-NEXT:    v_readlane_b32 s30, v1, 0
+; GFX10-NEXT:    v_readlane_b32 s31, v1, 1
+; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
+; GFX10-NEXT:    v_readlane_b32 s33, v1, 2
+; GFX10-NEXT:    s_or_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]
 entry:
   call amdgpu_gfx <2 x i16> @return_2xi16()
   ret void
 }
 
 define amdgpu_gfx <3 x i16> @return_3xi16() #0 {
 ; GFX9-LABEL: return_3xi16:
 ; GFX9:       ; %bb.0: ; %entry

 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_mov_b64 s[36:37], s[30:31]
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, return_3xi16@gotpcrel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, return_3xi16@gotpcrel32@hi+12
-; GFX9-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
+; GFX9-NEXT:    s_or_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:    v_writelane_b32 v2, s33, 2
+; 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[30:31]
-; GFX9-NEXT:    s_setpc_b64 s[36:37]
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
+; GFX9-NEXT:    v_readlane_b32 s30, v2, 0
+; GFX9-NEXT:    v_readlane_b32 s31, v2, 1
+; GFX9-NEXT:    s_addk_i32 s32, 0xfc00
+; GFX9-NEXT:    v_readlane_b32 s33, v2, 2
+; GFX9-NEXT:    s_or_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_mov_b64 s[36:37], s[30:31]
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, return_3xi16@gotpcrel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, return_3xi16@gotpcrel32@hi+12
-; GFX10-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
+; GFX10-NEXT:    s_or_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:    v_writelane_b32 v2, s33, 2
+; 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:    s_load_dwordx2 s[34:35], s[34:35], 0x0
+; GFX10-NEXT:    v_writelane_b32 v2, s30, 0
+; GFX10-NEXT:    v_writelane_b32 v2, s31, 1
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
-; GFX10-NEXT:    s_setpc_b64 s[36:37]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
+; GFX10-NEXT:    v_readlane_b32 s30, v2, 0
+; GFX10-NEXT:    v_readlane_b32 s31, v2, 1
+; GFX10-NEXT:    s_addk_i32 s32, 0xfe00
+; GFX10-NEXT:    v_readlane_b32 s33, v2, 2
+; GFX10-NEXT:    s_or_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]
 entry:
   call amdgpu_gfx <3 x i16> @return_3xi16()
   ret void
 }
 
 ; Check that return values that do not fit in registers do not crash
 
 define amdgpu_gfx <512 x i32> @return_512xi32() #0 {

 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_mov_b32 s34, s33
+; GFX9-NEXT:    s_or_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:    v_writelane_b32 v2, s33, 2
 ; 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_mov_b64 s[36:37], s[30:31]
-; GFX9-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-NEXT:    s_add_u32 s30, s30, return_512xi32@gotpcrel32@lo+4
-; GFX9-NEXT:    s_addc_u32 s31, s31, return_512xi32@gotpcrel32@hi+12
-; GFX9-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
+; 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[30:31]
+; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
+; GFX9-NEXT:    v_readlane_b32 s30, v2, 0
+; GFX9-NEXT:    v_readlane_b32 s31, v2, 1
 ; GFX9-NEXT:    s_add_i32 s32, s32, 0xfffa0000
-; GFX9-NEXT:    s_mov_b32 s33, s34
-; GFX9-NEXT:    s_setpc_b64 s[36:37]
+; GFX9-NEXT:    v_readlane_b32 s33, v2, 2
+; GFX9-NEXT:    s_or_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_mov_b32 s34, s33
+; GFX10-NEXT:    s_or_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:    v_writelane_b32 v2, s33, 2
 ; 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_mov_b64 s[36:37], s[30:31]
-; GFX10-NEXT:    s_getpc_b64 s[30:31]
-; GFX10-NEXT:    s_add_u32 s30, s30, return_512xi32@gotpcrel32@lo+4
-; GFX10-NEXT:    s_addc_u32 s31, s31, return_512xi32@gotpcrel32@hi+12
+; 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:    s_load_dwordx2 s[30:31], s[30:31], 0x0
+; GFX10-NEXT:    v_writelane_b32 v2, s31, 1
 ; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX10-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX10-NEXT:    s_swappc_b64 s[30:31], s[34:35]
+; GFX10-NEXT:    v_readlane_b32 s30, v2, 0
+; GFX10-NEXT:    v_readlane_b32 s31, v2, 1
 ; GFX10-NEXT:    s_add_i32 s32, s32, 0xfffd0000
-; GFX10-NEXT:    s_mov_b32 s33, s34
-; GFX10-NEXT:    s_setpc_b64 s[36:37]
+; GFX10-NEXT:    v_readlane_b32 s33, v2, 2
+; GFX10-NEXT:    s_or_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]
 entry:
   call amdgpu_gfx <512 x i32> @return_512xi32()
   ret void
 }
 
 attributes #0 = { nounwind }

llvm/test/CodeGen/AMDGPU/indirect-call.ll

 ; GCN-NEXT:    s_mov_b32 s14, s42
 ; GCN-NEXT:    s_swappc_b64 s[30:31], s[16:17]
 ; GCN-NEXT:    ; implicit-def: $vgpr0_vgpr1
 ; GCN-NEXT:    ; implicit-def: $vgpr31
 ; GCN-NEXT:    s_xor_b64 exec, exec, s[48:49]
 ; GCN-NEXT:    s_cbranch_execnz .LBB2_1
 ; GCN-NEXT:  ; %bb.2:
 ; GCN-NEXT:    s_mov_b64 exec, s[46:47]
-; GCN-NEXT:    v_readlane_b32 s4, v40, 15
-; GCN-NEXT:    v_readlane_b32 s5, v40, 16
+; GCN-NEXT:    v_readlane_b32 s30, v40, 15
+; GCN-NEXT:    v_readlane_b32 s31, v40, 16
 ; GCN-NEXT:    v_readlane_b32 s49, v40, 14
 ; GCN-NEXT:    v_readlane_b32 s48, v40, 13
 ; GCN-NEXT:    v_readlane_b32 s47, v40, 12
 ; GCN-NEXT:    v_readlane_b32 s46, v40, 11
 ; GCN-NEXT:    v_readlane_b32 s44, v40, 10
 ; GCN-NEXT:    v_readlane_b32 s43, v40, 9
 ; GCN-NEXT:    v_readlane_b32 s42, v40, 8
 ; GCN-NEXT:    v_readlane_b32 s41, v40, 7
 ; GCN-NEXT:    v_readlane_b32 s40, v40, 6
 ; GCN-NEXT:    v_readlane_b32 s39, v40, 5
 ; GCN-NEXT:    v_readlane_b32 s38, v40, 4
 ; GCN-NEXT:    v_readlane_b32 s37, v40, 3
 ; GCN-NEXT:    v_readlane_b32 s36, v40, 2
 ; GCN-NEXT:    v_readlane_b32 s35, v40, 1
 ; GCN-NEXT:    v_readlane_b32 s34, v40, 0
 ; GCN-NEXT:    s_addk_i32 s32, 0xfc00
 ; GCN-NEXT:    v_readlane_b32 s33, v40, 17
-; GCN-NEXT:    s_or_saveexec_b64 s[6:7], -1
+; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
-; GCN-NEXT:    s_mov_b64 exec, s[6:7]
+; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
-; GCN-NEXT:    s_setpc_b64 s[4:5]
+; GCN-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GISEL-LABEL: test_indirect_call_vgpr_ptr:
 ; GISEL:       ; %bb.0:
 ; GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GISEL-NEXT:    s_or_saveexec_b64 s[16:17], -1
 ; GISEL-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GISEL-NEXT:    s_mov_b64 exec, s[16:17]
 ; GISEL-NEXT:    v_writelane_b32 v40, s33, 17

 ; GISEL-NEXT:    v_writelane_b32 v40, s41, 7
 ; GISEL-NEXT:    v_writelane_b32 v40, s42, 8
 ; GISEL-NEXT:    v_writelane_b32 v40, s43, 9
 ; GISEL-NEXT:    v_writelane_b32 v40, s44, 10
 ; GISEL-NEXT:    v_writelane_b32 v40, s46, 11
 ; GISEL-NEXT:    v_writelane_b32 v40, s47, 12
 ; GISEL-NEXT:    v_writelane_b32 v40, s48, 13
 ; GISEL-NEXT:    v_writelane_b32 v40, s49, 14
+; GISEL-NEXT:    v_writelane_b32 v40, s30, 15
+; GISEL-NEXT:    v_writelane_b32 v40, s31, 16
 ; GISEL-NEXT:    s_mov_b32 s42, s14
 ; GISEL-NEXT:    s_mov_b32 s43, s13
 ; GISEL-NEXT:    s_mov_b32 s44, s12
 ; GISEL-NEXT:    s_mov_b64 s[34:35], s[10:11]
 ; GISEL-NEXT:    s_mov_b64 s[36:37], s[8:9]
 ; GISEL-NEXT:    s_mov_b64 s[38:39], s[6:7]
 ; GISEL-NEXT:    s_mov_b64 s[40:41], s[4:5]
-; GISEL-NEXT:    v_writelane_b32 v40, s30, 15
-; GISEL-NEXT:    v_writelane_b32 v40, s31, 16
 ; GISEL-NEXT:    s_mov_b64 s[46:47], exec
 ; GISEL-NEXT:  .LBB2_1: ; =>This Inner Loop Header: Depth=1
 ; GISEL-NEXT:    v_readfirstlane_b32 s16, v0
 ; GISEL-NEXT:    v_readfirstlane_b32 s17, v1
 ; GISEL-NEXT:    v_cmp_eq_u64_e32 vcc, s[16:17], v[0:1]
 ; GISEL-NEXT:    s_and_saveexec_b64 s[48:49], vcc
 ; GISEL-NEXT:    s_mov_b64 s[4:5], s[40:41]
 ; GISEL-NEXT:    s_mov_b64 s[6:7], s[38:39]
 ; GISEL-NEXT:    s_mov_b64 s[8:9], s[36:37]
 ; GISEL-NEXT:    s_mov_b64 s[10:11], s[34:35]
 ; GISEL-NEXT:    s_mov_b32 s12, s44
 ; GISEL-NEXT:    s_mov_b32 s13, s43
 ; GISEL-NEXT:    s_mov_b32 s14, s42
 ; GISEL-NEXT:    s_swappc_b64 s[30:31], s[16:17]
 ; GISEL-NEXT:    ; implicit-def: $vgpr0_vgpr1
 ; GISEL-NEXT:    ; implicit-def: $vgpr31
 ; GISEL-NEXT:    s_xor_b64 exec, exec, s[48:49]
 ; GISEL-NEXT:    s_cbranch_execnz .LBB2_1
 ; GISEL-NEXT:  ; %bb.2:
 ; GISEL-NEXT:    s_mov_b64 exec, s[46:47]
-; GISEL-NEXT:    v_readlane_b32 s4, v40, 15
-; GISEL-NEXT:    v_readlane_b32 s5, v40, 16
+; GISEL-NEXT:    v_readlane_b32 s30, v40, 15
+; GISEL-NEXT:    v_readlane_b32 s31, v40, 16
 ; GISEL-NEXT:    v_readlane_b32 s49, v40, 14
 ; GISEL-NEXT:    v_readlane_b32 s48, v40, 13
 ; GISEL-NEXT:    v_readlane_b32 s47, v40, 12
 ; GISEL-NEXT:    v_readlane_b32 s46, v40, 11
 ; GISEL-NEXT:    v_readlane_b32 s44, v40, 10
 ; GISEL-NEXT:    v_readlane_b32 s43, v40, 9
 ; GISEL-NEXT:    v_readlane_b32 s42, v40, 8
 ; GISEL-NEXT:    v_readlane_b32 s41, v40, 7
 ; GISEL-NEXT:    v_readlane_b32 s40, v40, 6
 ; GISEL-NEXT:    v_readlane_b32 s39, v40, 5
 ; GISEL-NEXT:    v_readlane_b32 s38, v40, 4
 ; GISEL-NEXT:    v_readlane_b32 s37, v40, 3
 ; GISEL-NEXT:    v_readlane_b32 s36, v40, 2
 ; GISEL-NEXT:    v_readlane_b32 s35, v40, 1
 ; GISEL-NEXT:    v_readlane_b32 s34, v40, 0
 ; GISEL-NEXT:    s_addk_i32 s32, 0xfc00
 ; GISEL-NEXT:    v_readlane_b32 s33, v40, 17
-; GISEL-NEXT:    s_or_saveexec_b64 s[6:7], -1
+; GISEL-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GISEL-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
-; GISEL-NEXT:    s_mov_b64 exec, s[6:7]
+; GISEL-NEXT:    s_mov_b64 exec, s[4:5]
 ; GISEL-NEXT:    s_waitcnt vmcnt(0)
-; GISEL-NEXT:    s_setpc_b64 s[4:5]
+; GISEL-NEXT:    s_setpc_b64 s[30:31]
   call void %fptr()
   ret void
 }
 
 define void @test_indirect_call_vgpr_ptr_arg(void(i32)* %fptr) {
 ; GCN-LABEL: test_indirect_call_vgpr_ptr_arg:
 ; GCN:       ; %bb.0:
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)

 ; GCN-NEXT:    s_swappc_b64 s[30:31], s[16:17]
 ; GCN-NEXT:    ; implicit-def: $vgpr0_vgpr1
 ; GCN-NEXT:    ; implicit-def: $vgpr31
 ; GCN-NEXT:    ; implicit-def: $vgpr2
 ; GCN-NEXT:    s_xor_b64 exec, exec, s[48:49]
 ; GCN-NEXT:    s_cbranch_execnz .LBB3_1
 ; GCN-NEXT:  ; %bb.2:
 ; GCN-NEXT:    s_mov_b64 exec, s[46:47]
-; GCN-NEXT:    v_readlane_b32 s4, v40, 15
-; GCN-NEXT:    v_readlane_b32 s5, v40, 16
+; GCN-NEXT:    v_readlane_b32 s30, v40, 15
+; GCN-NEXT:    v_readlane_b32 s31, v40, 16
 ; GCN-NEXT:    v_readlane_b32 s49, v40, 14
 ; GCN-NEXT:    v_readlane_b32 s48, v40, 13
 ; GCN-NEXT:    v_readlane_b32 s47, v40, 12
 ; GCN-NEXT:    v_readlane_b32 s46, v40, 11
 ; GCN-NEXT:    v_readlane_b32 s44, v40, 10
 ; GCN-NEXT:    v_readlane_b32 s43, v40, 9
 ; GCN-NEXT:    v_readlane_b32 s42, v40, 8
 ; GCN-NEXT:    v_readlane_b32 s41, v40, 7
 ; GCN-NEXT:    v_readlane_b32 s40, v40, 6
 ; GCN-NEXT:    v_readlane_b32 s39, v40, 5
 ; GCN-NEXT:    v_readlane_b32 s38, v40, 4
 ; GCN-NEXT:    v_readlane_b32 s37, v40, 3
 ; GCN-NEXT:    v_readlane_b32 s36, v40, 2
 ; GCN-NEXT:    v_readlane_b32 s35, v40, 1
 ; GCN-NEXT:    v_readlane_b32 s34, v40, 0
 ; GCN-NEXT:    s_addk_i32 s32, 0xfc00
 ; GCN-NEXT:    v_readlane_b32 s33, v40, 17
-; GCN-NEXT:    s_or_saveexec_b64 s[6:7], -1
+; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
-; GCN-NEXT:    s_mov_b64 exec, s[6:7]
+; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
-; GCN-NEXT:    s_setpc_b64 s[4:5]
+; GCN-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GISEL-LABEL: test_indirect_call_vgpr_ptr_arg:
 ; GISEL:       ; %bb.0:
 ; GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GISEL-NEXT:    s_or_saveexec_b64 s[16:17], -1
 ; GISEL-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GISEL-NEXT:    s_mov_b64 exec, s[16:17]
 ; GISEL-NEXT:    v_writelane_b32 v40, s33, 17

 ; GISEL-NEXT:    v_writelane_b32 v40, s41, 7
 ; GISEL-NEXT:    v_writelane_b32 v40, s42, 8
 ; GISEL-NEXT:    v_writelane_b32 v40, s43, 9
 ; GISEL-NEXT:    v_writelane_b32 v40, s44, 10
 ; GISEL-NEXT:    v_writelane_b32 v40, s46, 11
 ; GISEL-NEXT:    v_writelane_b32 v40, s47, 12
 ; GISEL-NEXT:    v_writelane_b32 v40, s48, 13
 ; GISEL-NEXT:    v_writelane_b32 v40, s49, 14
+; GISEL-NEXT:    v_writelane_b32 v40, s30, 15
+; GISEL-NEXT:    v_writelane_b32 v40, s31, 16
 ; GISEL-NEXT:    s_mov_b32 s42, s14
 ; GISEL-NEXT:    s_mov_b32 s43, s13
 ; GISEL-NEXT:    s_mov_b32 s44, s12
 ; GISEL-NEXT:    s_mov_b64 s[34:35], s[10:11]
 ; GISEL-NEXT:    s_mov_b64 s[36:37], s[8:9]
 ; GISEL-NEXT:    s_mov_b64 s[38:39], s[6:7]
 ; GISEL-NEXT:    s_mov_b64 s[40:41], s[4:5]
-; GISEL-NEXT:    v_writelane_b32 v40, s30, 15
-; GISEL-NEXT:    v_writelane_b32 v40, s31, 16
 ; GISEL-NEXT:    s_mov_b64 s[46:47], exec
 ; GISEL-NEXT:  .LBB3_1: ; =>This Inner Loop Header: Depth=1
 ; GISEL-NEXT:    v_readfirstlane_b32 s16, v0
 ; GISEL-NEXT:    v_readfirstlane_b32 s17, v1
 ; GISEL-NEXT:    v_cmp_eq_u64_e32 vcc, s[16:17], v[0:1]
 ; GISEL-NEXT:    s_and_saveexec_b64 s[48:49], vcc
 ; GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
 ; GISEL-NEXT:    s_mov_b64 s[4:5], s[40:41]
 ; GISEL-NEXT:    s_mov_b64 s[6:7], s[38:39]
 ; GISEL-NEXT:    s_mov_b64 s[8:9], s[36:37]
 ; GISEL-NEXT:    s_mov_b64 s[10:11], s[34:35]
 ; GISEL-NEXT:    s_mov_b32 s12, s44
 ; GISEL-NEXT:    s_mov_b32 s13, s43
 ; GISEL-NEXT:    s_mov_b32 s14, s42
 ; GISEL-NEXT:    s_swappc_b64 s[30:31], s[16:17]
 ; GISEL-NEXT:    ; implicit-def: $vgpr0_vgpr1
 ; GISEL-NEXT:    ; implicit-def: $vgpr31
 ; GISEL-NEXT:    s_xor_b64 exec, exec, s[48:49]
 ; GISEL-NEXT:    s_cbranch_execnz .LBB3_1
 ; GISEL-NEXT:  ; %bb.2:
 ; GISEL-NEXT:    s_mov_b64 exec, s[46:47]
-; GISEL-NEXT:    v_readlane_b32 s4, v40, 15
-; GISEL-NEXT:    v_readlane_b32 s5, v40, 16
+; GISEL-NEXT:    v_readlane_b32 s30, v40, 15
+; GISEL-NEXT:    v_readlane_b32 s31, v40, 16
 ; GISEL-NEXT:    v_readlane_b32 s49, v40, 14
 ; GISEL-NEXT:    v_readlane_b32 s48, v40, 13
 ; GISEL-NEXT:    v_readlane_b32 s47, v40, 12
 ; GISEL-NEXT:    v_readlane_b32 s46, v40, 11
 ; GISEL-NEXT:    v_readlane_b32 s44, v40, 10
 ; GISEL-NEXT:    v_readlane_b32 s43, v40, 9
 ; GISEL-NEXT:    v_readlane_b32 s42, v40, 8
 ; GISEL-NEXT:    v_readlane_b32 s41, v40, 7
 ; GISEL-NEXT:    v_readlane_b32 s40, v40, 6
 ; GISEL-NEXT:    v_readlane_b32 s39, v40, 5
 ; GISEL-NEXT:    v_readlane_b32 s38, v40, 4
 ; GISEL-NEXT:    v_readlane_b32 s37, v40, 3
 ; GISEL-NEXT:    v_readlane_b32 s36, v40, 2
 ; GISEL-NEXT:    v_readlane_b32 s35, v40, 1
 ; GISEL-NEXT:    v_readlane_b32 s34, v40, 0
 ; GISEL-NEXT:    s_addk_i32 s32, 0xfc00
 ; GISEL-NEXT:    v_readlane_b32 s33, v40, 17
-; GISEL-NEXT:    s_or_saveexec_b64 s[6:7], -1
+; GISEL-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GISEL-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
-; GISEL-NEXT:    s_mov_b64 exec, s[6:7]
+; GISEL-NEXT:    s_mov_b64 exec, s[4:5]
 ; GISEL-NEXT:    s_waitcnt vmcnt(0)
-; GISEL-NEXT:    s_setpc_b64 s[4:5]
+; GISEL-NEXT:    s_setpc_b64 s[30:31]
   call void %fptr(i32 123)
   ret void
 }
 
 define i32 @test_indirect_call_vgpr_ptr_ret(i32()* %fptr) {
 ; GCN-LABEL: test_indirect_call_vgpr_ptr_ret:
 ; GCN:       ; %bb.0:
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)

 ; GCN-NEXT:    v_mov_b32_e32 v2, v0
 ; GCN-NEXT:    ; implicit-def: $vgpr0_vgpr1
 ; GCN-NEXT:    ; implicit-def: $vgpr31
 ; GCN-NEXT:    s_xor_b64 exec, exec, s[48:49]
 ; GCN-NEXT:    s_cbranch_execnz .LBB4_1
 ; GCN-NEXT:  ; %bb.2:
 ; GCN-NEXT:    s_mov_b64 exec, s[46:47]
 ; GCN-NEXT:    v_add_i32_e32 v0, vcc, 1, v2
-; GCN-NEXT:    v_readlane_b32 s4, v40, 15
-; GCN-NEXT:    v_readlane_b32 s5, v40, 16
+; GCN-NEXT:    v_readlane_b32 s30, v40, 15
+; GCN-NEXT:    v_readlane_b32 s31, v40, 16
 ; GCN-NEXT:    v_readlane_b32 s49, v40, 14
 ; GCN-NEXT:    v_readlane_b32 s48, v40, 13
 ; GCN-NEXT:    v_readlane_b32 s47, v40, 12
 ; GCN-NEXT:    v_readlane_b32 s46, v40, 11
 ; GCN-NEXT:    v_readlane_b32 s44, v40, 10
 ; GCN-NEXT:    v_readlane_b32 s43, v40, 9
 ; GCN-NEXT:    v_readlane_b32 s42, v40, 8
 ; GCN-NEXT:    v_readlane_b32 s41, v40, 7
 ; GCN-NEXT:    v_readlane_b32 s40, v40, 6
 ; GCN-NEXT:    v_readlane_b32 s39, v40, 5
 ; GCN-NEXT:    v_readlane_b32 s38, v40, 4
 ; GCN-NEXT:    v_readlane_b32 s37, v40, 3
 ; GCN-NEXT:    v_readlane_b32 s36, v40, 2
 ; GCN-NEXT:    v_readlane_b32 s35, v40, 1
 ; GCN-NEXT:    v_readlane_b32 s34, v40, 0
 ; GCN-NEXT:    s_addk_i32 s32, 0xfc00
 ; GCN-NEXT:    v_readlane_b32 s33, v40, 17
-; GCN-NEXT:    s_or_saveexec_b64 s[6:7], -1
+; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
-; GCN-NEXT:    s_mov_b64 exec, s[6:7]
+; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
-; GCN-NEXT:    s_setpc_b64 s[4:5]
+; GCN-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GISEL-LABEL: test_indirect_call_vgpr_ptr_ret:
 ; GISEL:       ; %bb.0:
 ; GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GISEL-NEXT:    s_or_saveexec_b64 s[16:17], -1
 ; GISEL-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GISEL-NEXT:    s_mov_b64 exec, s[16:17]
 ; GISEL-NEXT:    v_writelane_b32 v40, s33, 17

 ; GISEL-NEXT:    v_writelane_b32 v40, s41, 7
 ; GISEL-NEXT:    v_writelane_b32 v40, s42, 8
 ; GISEL-NEXT:    v_writelane_b32 v40, s43, 9
 ; GISEL-NEXT:    v_writelane_b32 v40, s44, 10
 ; GISEL-NEXT:    v_writelane_b32 v40, s46, 11
 ; GISEL-NEXT:    v_writelane_b32 v40, s47, 12
 ; GISEL-NEXT:    v_writelane_b32 v40, s48, 13
 ; GISEL-NEXT:    v_writelane_b32 v40, s49, 14
+; GISEL-NEXT:    v_writelane_b32 v40, s30, 15
+; GISEL-NEXT:    v_writelane_b32 v40, s31, 16
 ; GISEL-NEXT:    s_mov_b32 s42, s14
 ; GISEL-NEXT:    s_mov_b32 s43, s13
 ; GISEL-NEXT:    s_mov_b32 s44, s12
 ; GISEL-NEXT:    s_mov_b64 s[34:35], s[10:11]
 ; GISEL-NEXT:    s_mov_b64 s[36:37], s[8:9]
 ; GISEL-NEXT:    s_mov_b64 s[38:39], s[6:7]
 ; GISEL-NEXT:    s_mov_b64 s[40:41], s[4:5]
-; GISEL-NEXT:    v_writelane_b32 v40, s30, 15
-; GISEL-NEXT:    v_writelane_b32 v40, s31, 16
 ; GISEL-NEXT:    s_mov_b64 s[46:47], exec
 ; GISEL-NEXT:  .LBB4_1: ; =>This Inner Loop Header: Depth=1
 ; GISEL-NEXT:    v_readfirstlane_b32 s16, v0
 ; GISEL-NEXT:    v_readfirstlane_b32 s17, v1
 ; GISEL-NEXT:    v_cmp_eq_u64_e32 vcc, s[16:17], v[0:1]
 ; GISEL-NEXT:    s_and_saveexec_b64 s[48:49], vcc
 ; GISEL-NEXT:    s_mov_b64 s[4:5], s[40:41]
 ; GISEL-NEXT:    s_mov_b64 s[6:7], s[38:39]
 ; GISEL-NEXT:    s_mov_b64 s[8:9], s[36:37]
 ; GISEL-NEXT:    s_mov_b64 s[10:11], s[34:35]
 ; GISEL-NEXT:    s_mov_b32 s12, s44
 ; GISEL-NEXT:    s_mov_b32 s13, s43
 ; GISEL-NEXT:    s_mov_b32 s14, s42
 ; GISEL-NEXT:    s_swappc_b64 s[30:31], s[16:17]
 ; GISEL-NEXT:    v_mov_b32_e32 v2, v0
 ; GISEL-NEXT:    ; implicit-def: $vgpr0_vgpr1
 ; GISEL-NEXT:    ; implicit-def: $vgpr31
 ; GISEL-NEXT:    s_xor_b64 exec, exec, s[48:49]
 ; GISEL-NEXT:    s_cbranch_execnz .LBB4_1
 ; GISEL-NEXT:  ; %bb.2:
 ; GISEL-NEXT:    s_mov_b64 exec, s[46:47]
 ; GISEL-NEXT:    v_add_i32_e32 v0, vcc, 1, v2
-; GISEL-NEXT:    v_readlane_b32 s4, v40, 15
-; GISEL-NEXT:    v_readlane_b32 s5, v40, 16
+; GISEL-NEXT:    v_readlane_b32 s30, v40, 15
+; GISEL-NEXT:    v_readlane_b32 s31, v40, 16
 ; GISEL-NEXT:    v_readlane_b32 s49, v40, 14
 ; GISEL-NEXT:    v_readlane_b32 s48, v40, 13
 ; GISEL-NEXT:    v_readlane_b32 s47, v40, 12
 ; GISEL-NEXT:    v_readlane_b32 s46, v40, 11
 ; GISEL-NEXT:    v_readlane_b32 s44, v40, 10
 ; GISEL-NEXT:    v_readlane_b32 s43, v40, 9
 ; GISEL-NEXT:    v_readlane_b32 s42, v40, 8
 ; GISEL-NEXT:    v_readlane_b32 s41, v40, 7
 ; GISEL-NEXT:    v_readlane_b32 s40, v40, 6
 ; GISEL-NEXT:    v_readlane_b32 s39, v40, 5
 ; GISEL-NEXT:    v_readlane_b32 s38, v40, 4
 ; GISEL-NEXT:    v_readlane_b32 s37, v40, 3
 ; GISEL-NEXT:    v_readlane_b32 s36, v40, 2
 ; GISEL-NEXT:    v_readlane_b32 s35, v40, 1
 ; GISEL-NEXT:    v_readlane_b32 s34, v40, 0
 ; GISEL-NEXT:    s_addk_i32 s32, 0xfc00
 ; GISEL-NEXT:    v_readlane_b32 s33, v40, 17
-; GISEL-NEXT:    s_or_saveexec_b64 s[6:7], -1
+; GISEL-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GISEL-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
-; GISEL-NEXT:    s_mov_b64 exec, s[6:7]
+; GISEL-NEXT:    s_mov_b64 exec, s[4:5]
 ; GISEL-NEXT:    s_waitcnt vmcnt(0)
-; GISEL-NEXT:    s_setpc_b64 s[4:5]
+; GISEL-NEXT:    s_setpc_b64 s[30:31]
   %a = call i32 %fptr()
   %b = add i32 %a, 1
   ret i32 %b
 }
 
 define void @test_indirect_call_vgpr_ptr_in_branch(void()* %fptr, i1 %cond) {
 ; GCN-LABEL: test_indirect_call_vgpr_ptr_in_branch:
 ; GCN:       ; %bb.0: ; %bb0

 ; GCN-NEXT:    v_writelane_b32 v40, s43, 9
 ; GCN-NEXT:    v_writelane_b32 v40, s44, 10
 ; GCN-NEXT:    v_writelane_b32 v40, s46, 11
 ; GCN-NEXT:    v_writelane_b32 v40, s47, 12
 ; GCN-NEXT:    v_writelane_b32 v40, s48, 13
 ; GCN-NEXT:    v_writelane_b32 v40, s49, 14
 ; GCN-NEXT:    v_writelane_b32 v40, s50, 15
 ; GCN-NEXT:    v_writelane_b32 v40, s51, 16
+; GCN-NEXT:    v_writelane_b32 v40, s30, 17
+; GCN-NEXT:    v_writelane_b32 v40, s31, 18
 ; GCN-NEXT:    s_mov_b32 s42, s14
 ; GCN-NEXT:    s_mov_b32 s43, s13
 ; GCN-NEXT:    s_mov_b32 s44, s12
 ; GCN-NEXT:    s_mov_b64 s[34:35], s[10:11]
 ; GCN-NEXT:    s_mov_b64 s[36:37], s[8:9]
 ; GCN-NEXT:    s_mov_b64 s[38:39], s[6:7]
 ; GCN-NEXT:    s_mov_b64 s[40:41], s[4:5]
 ; GCN-NEXT:    v_and_b32_e32 v2, 1, v2
 ; GCN-NEXT:    v_cmp_eq_u32_e32 vcc, 1, v2
 ; GCN-NEXT:    s_and_saveexec_b64 s[46:47], vcc
 ; GCN-NEXT:    s_cbranch_execz .LBB5_4
 ; GCN-NEXT:  ; %bb.1: ; %bb1
-; GCN-NEXT:    v_writelane_b32 v40, s30, 17
-; GCN-NEXT:    v_writelane_b32 v40, s31, 18
 ; GCN-NEXT:    s_mov_b64 s[48:49], exec
 ; GCN-NEXT:  .LBB5_2: ; =>This Inner Loop Header: Depth=1
 ; GCN-NEXT:    v_readfirstlane_b32 s16, v0
 ; GCN-NEXT:    v_readfirstlane_b32 s17, v1
 ; GCN-NEXT:    v_cmp_eq_u64_e32 vcc, s[16:17], v[0:1]
 ; GCN-NEXT:    s_and_saveexec_b64 s[50:51], vcc
 ; GCN-NEXT:    s_mov_b64 s[4:5], s[40:41]
 ; GCN-NEXT:    s_mov_b64 s[6:7], s[38:39]
 ; GCN-NEXT:    s_mov_b64 s[8:9], s[36:37]
 ; GCN-NEXT:    s_mov_b64 s[10:11], s[34:35]
 ; GCN-NEXT:    s_mov_b32 s12, s44
 ; GCN-NEXT:    s_mov_b32 s13, s43
 ; GCN-NEXT:    s_mov_b32 s14, s42
 ; GCN-NEXT:    s_swappc_b64 s[30:31], s[16:17]
 ; GCN-NEXT:    ; implicit-def: $vgpr0_vgpr1
 ; GCN-NEXT:    ; implicit-def: $vgpr31
 ; GCN-NEXT:    s_xor_b64 exec, exec, s[50:51]
 ; GCN-NEXT:    s_cbranch_execnz .LBB5_2
 ; GCN-NEXT:  ; %bb.3:
 ; GCN-NEXT:    s_mov_b64 exec, s[48:49]
-; GCN-NEXT:    v_readlane_b32 s30, v40, 17
-; GCN-NEXT:    v_readlane_b32 s31, v40, 18
 ; GCN-NEXT:  .LBB5_4: ; %bb2
 ; GCN-NEXT:    s_or_b64 exec, exec, s[46:47]
+; GCN-NEXT:    v_readlane_b32 s30, v40, 17
+; GCN-NEXT:    v_readlane_b32 s31, v40, 18
 ; GCN-NEXT:    v_readlane_b32 s51, v40, 16
 ; GCN-NEXT:    v_readlane_b32 s50, v40, 15
 ; GCN-NEXT:    v_readlane_b32 s49, v40, 14
 ; GCN-NEXT:    v_readlane_b32 s48, v40, 13
 ; GCN-NEXT:    v_readlane_b32 s47, v40, 12
 ; GCN-NEXT:    v_readlane_b32 s46, v40, 11
 ; GCN-NEXT:    v_readlane_b32 s44, v40, 10
 ; GCN-NEXT:    v_readlane_b32 s43, v40, 9

 ; GISEL-NEXT:    v_writelane_b32 v40, s43, 9
 ; GISEL-NEXT:    v_writelane_b32 v40, s44, 10
 ; GISEL-NEXT:    v_writelane_b32 v40, s46, 11
 ; GISEL-NEXT:    v_writelane_b32 v40, s47, 12
 ; GISEL-NEXT:    v_writelane_b32 v40, s48, 13
 ; GISEL-NEXT:    v_writelane_b32 v40, s49, 14
 ; GISEL-NEXT:    v_writelane_b32 v40, s50, 15
 ; GISEL-NEXT:    v_writelane_b32 v40, s51, 16
+; GISEL-NEXT:    v_writelane_b32 v40, s30, 17
+; GISEL-NEXT:    v_writelane_b32 v40, s31, 18
 ; GISEL-NEXT:    s_mov_b32 s42, s14
 ; GISEL-NEXT:    s_mov_b32 s43, s13
 ; GISEL-NEXT:    s_mov_b32 s44, s12
 ; GISEL-NEXT:    s_mov_b64 s[34:35], s[10:11]
 ; GISEL-NEXT:    s_mov_b64 s[36:37], s[8:9]
 ; GISEL-NEXT:    s_mov_b64 s[38:39], s[6:7]
 ; GISEL-NEXT:    s_mov_b64 s[40:41], s[4:5]
 ; GISEL-NEXT:    v_and_b32_e32 v2, 1, v2
 ; GISEL-NEXT:    v_cmp_ne_u32_e32 vcc, 0, v2
 ; GISEL-NEXT:    s_and_saveexec_b64 s[46:47], vcc
 ; GISEL-NEXT:    s_cbranch_execz .LBB5_4
 ; GISEL-NEXT:  ; %bb.1: ; %bb1
-; GISEL-NEXT:    v_writelane_b32 v40, s30, 17
-; GISEL-NEXT:    v_writelane_b32 v40, s31, 18
 ; GISEL-NEXT:    s_mov_b64 s[48:49], exec
 ; GISEL-NEXT:  .LBB5_2: ; =>This Inner Loop Header: Depth=1
 ; GISEL-NEXT:    v_readfirstlane_b32 s16, v0
 ; GISEL-NEXT:    v_readfirstlane_b32 s17, v1
 ; GISEL-NEXT:    v_cmp_eq_u64_e32 vcc, s[16:17], v[0:1]
 ; GISEL-NEXT:    s_and_saveexec_b64 s[50:51], vcc
 ; GISEL-NEXT:    s_mov_b64 s[4:5], s[40:41]
 ; GISEL-NEXT:    s_mov_b64 s[6:7], s[38:39]
 ; GISEL-NEXT:    s_mov_b64 s[8:9], s[36:37]
 ; GISEL-NEXT:    s_mov_b64 s[10:11], s[34:35]
 ; GISEL-NEXT:    s_mov_b32 s12, s44
 ; GISEL-NEXT:    s_mov_b32 s13, s43
 ; GISEL-NEXT:    s_mov_b32 s14, s42
 ; GISEL-NEXT:    s_swappc_b64 s[30:31], s[16:17]
 ; GISEL-NEXT:    ; implicit-def: $vgpr0_vgpr1
 ; GISEL-NEXT:    ; implicit-def: $vgpr31
 ; GISEL-NEXT:    s_xor_b64 exec, exec, s[50:51]
 ; GISEL-NEXT:    s_cbranch_execnz .LBB5_2
 ; GISEL-NEXT:  ; %bb.3:
 ; GISEL-NEXT:    s_mov_b64 exec, s[48:49]
-; GISEL-NEXT:    v_readlane_b32 s30, v40, 17
-; GISEL-NEXT:    v_readlane_b32 s31, v40, 18
 ; GISEL-NEXT:  .LBB5_4: ; %bb2
 ; GISEL-NEXT:    s_or_b64 exec, exec, s[46:47]
+; GISEL-NEXT:    v_readlane_b32 s30, v40, 17
+; GISEL-NEXT:    v_readlane_b32 s31, v40, 18
 ; GISEL-NEXT:    v_readlane_b32 s51, v40, 16
 ; GISEL-NEXT:    v_readlane_b32 s50, v40, 15
 ; GISEL-NEXT:    v_readlane_b32 s49, v40, 14
 ; GISEL-NEXT:    v_readlane_b32 s48, v40, 13
 ; GISEL-NEXT:    v_readlane_b32 s47, v40, 12
 ; GISEL-NEXT:    v_readlane_b32 s46, v40, 11
 ; GISEL-NEXT:    v_readlane_b32 s44, v40, 10
 ; GISEL-NEXT:    v_readlane_b32 s43, v40, 9

 
 define void @test_indirect_call_vgpr_ptr_inreg_arg(void(i32)* %fptr) {
 ; GCN-LABEL: test_indirect_call_vgpr_ptr_inreg_arg:
 ; GCN:       ; %bb.0:
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
-; GCN-NEXT:    v_writelane_b32 v40, s33, 30
+; GCN-NEXT:    v_writelane_b32 v40, s33, 32
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
 ; GCN-NEXT:    v_writelane_b32 v40, s34, 0
 ; GCN-NEXT:    v_writelane_b32 v40, s35, 1
 ; GCN-NEXT:    v_writelane_b32 v40, s36, 2
 ; GCN-NEXT:    v_writelane_b32 v40, s37, 3
 ; GCN-NEXT:    v_writelane_b32 v40, s38, 4
 ; GCN-NEXT:    v_writelane_b32 v40, s39, 5

 ; GCN-NEXT:    v_writelane_b32 v40, s56, 22
 ; GCN-NEXT:    v_writelane_b32 v40, s57, 23
 ; GCN-NEXT:    v_writelane_b32 v40, s58, 24
 ; GCN-NEXT:    v_writelane_b32 v40, s59, 25
 ; GCN-NEXT:    v_writelane_b32 v40, s60, 26
 ; GCN-NEXT:    v_writelane_b32 v40, s61, 27
 ; GCN-NEXT:    v_writelane_b32 v40, s62, 28
 ; GCN-NEXT:    v_writelane_b32 v40, s63, 29
-; GCN-NEXT:    s_mov_b64 s[6:7], s[30:31]
-; GCN-NEXT:    s_mov_b64 s[8:9], exec
+; GCN-NEXT:    v_writelane_b32 v40, s30, 30
+; GCN-NEXT:    v_writelane_b32 v40, s31, 31
+; GCN-NEXT:    s_mov_b64 s[6:7], exec
 ; GCN-NEXT:    s_movk_i32 s4, 0x7b
 ; GCN-NEXT:  .LBB6_1: ; =>This Inner Loop Header: Depth=1
-; GCN-NEXT:    v_readfirstlane_b32 s12, v0
-; GCN-NEXT:    v_readfirstlane_b32 s13, v1
-; GCN-NEXT:    v_cmp_eq_u64_e32 vcc, s[12:13], v[0:1]
-; GCN-NEXT:    s_and_saveexec_b64 s[10:11], vcc
-; GCN-NEXT:    s_swappc_b64 s[30:31], s[12:13]
+; GCN-NEXT:    v_readfirstlane_b32 s10, v0
+; GCN-NEXT:    v_readfirstlane_b32 s11, v1
+; GCN-NEXT:    v_cmp_eq_u64_e32 vcc, s[10:11], v[0:1]
+; GCN-NEXT:    s_and_saveexec_b64 s[8:9], vcc
+; GCN-NEXT:    s_swappc_b64 s[30:31], s[10:11]
 ; GCN-NEXT:    ; implicit-def: $vgpr0_vgpr1
-; GCN-NEXT:    s_xor_b64 exec, exec, s[10:11]
+; GCN-NEXT:    s_xor_b64 exec, exec, s[8:9]
 ; GCN-NEXT:    s_cbranch_execnz .LBB6_1
 ; GCN-NEXT:  ; %bb.2:
-; GCN-NEXT:    s_mov_b64 exec, s[8:9]
+; GCN-NEXT:    s_mov_b64 exec, s[6:7]
+; GCN-NEXT:    v_readlane_b32 s30, v40, 30
+; GCN-NEXT:    v_readlane_b32 s31, v40, 31
 ; GCN-NEXT:    v_readlane_b32 s63, v40, 29
 ; GCN-NEXT:    v_readlane_b32 s62, v40, 28
 ; GCN-NEXT:    v_readlane_b32 s61, v40, 27
 ; GCN-NEXT:    v_readlane_b32 s60, v40, 26
 ; GCN-NEXT:    v_readlane_b32 s59, v40, 25
 ; GCN-NEXT:    v_readlane_b32 s58, v40, 24
 ; GCN-NEXT:    v_readlane_b32 s57, v40, 23
 ; GCN-NEXT:    v_readlane_b32 s56, v40, 22

 ; GCN-NEXT:    v_readlane_b32 s40, v40, 6
 ; GCN-NEXT:    v_readlane_b32 s39, v40, 5
 ; GCN-NEXT:    v_readlane_b32 s38, v40, 4
 ; GCN-NEXT:    v_readlane_b32 s37, v40, 3
 ; GCN-NEXT:    v_readlane_b32 s36, v40, 2
 ; GCN-NEXT:    v_readlane_b32 s35, v40, 1
 ; GCN-NEXT:    v_readlane_b32 s34, v40, 0
 ; GCN-NEXT:    s_addk_i32 s32, 0xfc00
-; GCN-NEXT:    v_readlane_b32 s33, v40, 30
+; GCN-NEXT:    v_readlane_b32 s33, v40, 32
 ; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_load_dword v40, 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[6:7]
+; GCN-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GISEL-LABEL: test_indirect_call_vgpr_ptr_inreg_arg:
 ; GISEL:       ; %bb.0:
 ; GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GISEL-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GISEL-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GISEL-NEXT:    s_mov_b64 exec, s[4:5]
-; GISEL-NEXT:    v_writelane_b32 v40, s33, 30
+; GISEL-NEXT:    v_writelane_b32 v40, s33, 32
 ; GISEL-NEXT:    s_mov_b32 s33, s32
 ; GISEL-NEXT:    s_addk_i32 s32, 0x400
 ; GISEL-NEXT:    v_writelane_b32 v40, s34, 0
 ; GISEL-NEXT:    v_writelane_b32 v40, s35, 1
 ; GISEL-NEXT:    v_writelane_b32 v40, s36, 2
 ; GISEL-NEXT:    v_writelane_b32 v40, s37, 3
 ; GISEL-NEXT:    v_writelane_b32 v40, s38, 4
 ; GISEL-NEXT:    v_writelane_b32 v40, s39, 5

 ; GISEL-NEXT:    v_writelane_b32 v40, s56, 22
 ; GISEL-NEXT:    v_writelane_b32 v40, s57, 23
 ; GISEL-NEXT:    v_writelane_b32 v40, s58, 24
 ; GISEL-NEXT:    v_writelane_b32 v40, s59, 25
 ; GISEL-NEXT:    v_writelane_b32 v40, s60, 26
 ; GISEL-NEXT:    v_writelane_b32 v40, s61, 27
 ; GISEL-NEXT:    v_writelane_b32 v40, s62, 28
 ; GISEL-NEXT:    v_writelane_b32 v40, s63, 29
-; GISEL-NEXT:    s_mov_b64 s[6:7], s[30:31]
+; GISEL-NEXT:    v_writelane_b32 v40, s30, 30
+; GISEL-NEXT:    v_writelane_b32 v40, s31, 31
 ; GISEL-NEXT:    s_movk_i32 s4, 0x7b
-; GISEL-NEXT:    s_mov_b64 s[8:9], exec
+; GISEL-NEXT:    s_mov_b64 s[6:7], exec
 ; GISEL-NEXT:  .LBB6_1: ; =>This Inner Loop Header: Depth=1
-; GISEL-NEXT:    v_readfirstlane_b32 s10, v0
-; GISEL-NEXT:    v_readfirstlane_b32 s11, v1
-; GISEL-NEXT:    v_cmp_eq_u64_e32 vcc, s[10:11], v[0:1]
-; GISEL-NEXT:    s_and_saveexec_b64 s[12:13], vcc
-; GISEL-NEXT:    s_swappc_b64 s[30:31], s[10:11]
+; GISEL-NEXT:    v_readfirstlane_b32 s8, v0
+; GISEL-NEXT:    v_readfirstlane_b32 s9, v1
+; GISEL-NEXT:    v_cmp_eq_u64_e32 vcc, s[8:9], v[0:1]
+; GISEL-NEXT:    s_and_saveexec_b64 s[10:11], vcc
+; GISEL-NEXT:    s_swappc_b64 s[30:31], s[8:9]
 ; GISEL-NEXT:    ; implicit-def: $vgpr0_vgpr1
-; GISEL-NEXT:    s_xor_b64 exec, exec, s[12:13]
+; GISEL-NEXT:    s_xor_b64 exec, exec, s[10:11]
 ; GISEL-NEXT:    s_cbranch_execnz .LBB6_1
 ; GISEL-NEXT:  ; %bb.2:
-; GISEL-NEXT:    s_mov_b64 exec, s[8:9]
+; GISEL-NEXT:    s_mov_b64 exec, s[6:7]
+; GISEL-NEXT:    v_readlane_b32 s30, v40, 30
+; GISEL-NEXT:    v_readlane_b32 s31, v40, 31
 ; GISEL-NEXT:    v_readlane_b32 s63, v40, 29
 ; GISEL-NEXT:    v_readlane_b32 s62, v40, 28
 ; GISEL-NEXT:    v_readlane_b32 s61, v40, 27
 ; GISEL-NEXT:    v_readlane_b32 s60, v40, 26
 ; GISEL-NEXT:    v_readlane_b32 s59, v40, 25
 ; GISEL-NEXT:    v_readlane_b32 s58, v40, 24
 ; GISEL-NEXT:    v_readlane_b32 s57, v40, 23
 ; GISEL-NEXT:    v_readlane_b32 s56, v40, 22

 ; GISEL-NEXT:    v_readlane_b32 s40, v40, 6
 ; GISEL-NEXT:    v_readlane_b32 s39, v40, 5
 ; GISEL-NEXT:    v_readlane_b32 s38, v40, 4
 ; GISEL-NEXT:    v_readlane_b32 s37, v40, 3
 ; GISEL-NEXT:    v_readlane_b32 s36, v40, 2
 ; GISEL-NEXT:    v_readlane_b32 s35, v40, 1
 ; GISEL-NEXT:    v_readlane_b32 s34, v40, 0
 ; GISEL-NEXT:    s_addk_i32 s32, 0xfc00
-; GISEL-NEXT:    v_readlane_b32 s33, v40, 30
+; GISEL-NEXT:    v_readlane_b32 s33, v40, 32
 ; GISEL-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GISEL-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GISEL-NEXT:    s_mov_b64 exec, s[4:5]
 ; GISEL-NEXT:    s_waitcnt vmcnt(0)
-; GISEL-NEXT:    s_setpc_b64 s[6:7]
+; GISEL-NEXT:    s_setpc_b64 s[30:31]
   call amdgpu_gfx void %fptr(i32 inreg 123)
   ret void
 }
 
 define i32 @test_indirect_call_vgpr_ptr_arg_and_reuse(i32 %i, void(i32)* %fptr) {
 ; GCN-LABEL: test_indirect_call_vgpr_ptr_arg_and_reuse:
 ; GCN:       ; %bb.0:
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
-; GCN-NEXT:    v_writelane_b32 v40, s33, 30
+; GCN-NEXT:    v_writelane_b32 v40, s33, 32
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
 ; GCN-NEXT:    buffer_store_dword v41, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GCN-NEXT:    v_writelane_b32 v40, s34, 0
 ; GCN-NEXT:    v_writelane_b32 v40, s35, 1
 ; GCN-NEXT:    v_writelane_b32 v40, s36, 2
 ; GCN-NEXT:    v_writelane_b32 v40, s37, 3
 ; GCN-NEXT:    v_writelane_b32 v40, s38, 4

 ; GCN-NEXT:    v_writelane_b32 v40, s56, 22
 ; GCN-NEXT:    v_writelane_b32 v40, s57, 23
 ; GCN-NEXT:    v_writelane_b32 v40, s58, 24
 ; GCN-NEXT:    v_writelane_b32 v40, s59, 25
 ; GCN-NEXT:    v_writelane_b32 v40, s60, 26
 ; GCN-NEXT:    v_writelane_b32 v40, s61, 27
 ; GCN-NEXT:    v_writelane_b32 v40, s62, 28
 ; GCN-NEXT:    v_writelane_b32 v40, s63, 29
-; GCN-NEXT:    s_mov_b64 s[4:5], s[30:31]
+; GCN-NEXT:    v_writelane_b32 v40, s30, 30
+; GCN-NEXT:    v_writelane_b32 v40, s31, 31
 ; GCN-NEXT:    v_mov_b32_e32 v41, v0
-; GCN-NEXT:    s_mov_b64 s[6:7], exec
+; GCN-NEXT:    s_mov_b64 s[4:5], exec
 ; GCN-NEXT:  .LBB7_1: ; =>This Inner Loop Header: Depth=1
-; GCN-NEXT:    v_readfirstlane_b32 s10, v1
-; GCN-NEXT:    v_readfirstlane_b32 s11, v2
-; GCN-NEXT:    v_cmp_eq_u64_e32 vcc, s[10:11], v[1:2]
-; GCN-NEXT:    s_and_saveexec_b64 s[8:9], vcc
+; GCN-NEXT:    v_readfirstlane_b32 s8, v1
+; GCN-NEXT:    v_readfirstlane_b32 s9, v2
+; GCN-NEXT:    v_cmp_eq_u64_e32 vcc, s[8:9], v[1:2]
+; GCN-NEXT:    s_and_saveexec_b64 s[6:7], vcc
 ; GCN-NEXT:    v_mov_b32_e32 v0, v41
-; GCN-NEXT:    s_swappc_b64 s[30:31], s[10:11]
+; GCN-NEXT:    s_swappc_b64 s[30:31], s[8:9]
 ; GCN-NEXT:    ; implicit-def: $vgpr1_vgpr2
-; GCN-NEXT:    s_xor_b64 exec, exec, s[8:9]
+; GCN-NEXT:    s_xor_b64 exec, exec, s[6:7]
 ; GCN-NEXT:    s_cbranch_execnz .LBB7_1
 ; GCN-NEXT:  ; %bb.2:
-; GCN-NEXT:    s_mov_b64 exec, s[6:7]
+; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    v_mov_b32_e32 v0, v41
+; GCN-NEXT:    v_readlane_b32 s30, v40, 30
+; GCN-NEXT:    v_readlane_b32 s31, v40, 31
 ; GCN-NEXT:    v_readlane_b32 s63, v40, 29
 ; GCN-NEXT:    v_readlane_b32 s62, v40, 28
 ; GCN-NEXT:    v_readlane_b32 s61, v40, 27
 ; GCN-NEXT:    v_readlane_b32 s60, v40, 26
 ; GCN-NEXT:    v_readlane_b32 s59, v40, 25
 ; GCN-NEXT:    v_readlane_b32 s58, v40, 24
 ; GCN-NEXT:    v_readlane_b32 s57, v40, 23
 ; GCN-NEXT:    v_readlane_b32 s56, v40, 22

 ; GCN-NEXT:    v_readlane_b32 s39, v40, 5
 ; GCN-NEXT:    v_readlane_b32 s38, v40, 4
 ; GCN-NEXT:    v_readlane_b32 s37, v40, 3
 ; GCN-NEXT:    v_readlane_b32 s36, v40, 2
 ; GCN-NEXT:    v_readlane_b32 s35, v40, 1
 ; GCN-NEXT:    v_readlane_b32 s34, v40, 0
 ; GCN-NEXT:    buffer_load_dword v41, off, s[0:3], s33 ; 4-byte Folded Reload
 ; GCN-NEXT:    s_addk_i32 s32, 0xfc00
-; GCN-NEXT:    v_readlane_b32 s33, v40, 30
-; GCN-NEXT:    s_or_saveexec_b64 s[6:7], -1
+; GCN-NEXT:    v_readlane_b32 s33, v40, 32
+; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
-; GCN-NEXT:    s_mov_b64 exec, s[6:7]
+; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
-; GCN-NEXT:    s_setpc_b64 s[4:5]
+; GCN-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GISEL-LABEL: test_indirect_call_vgpr_ptr_arg_and_reuse:
 ; GISEL:       ; %bb.0:
 ; GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GISEL-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GISEL-NEXT:    buffer_store_dword v40, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill
 ; GISEL-NEXT:    s_mov_b64 exec, s[4:5]
-; GISEL-NEXT:    v_writelane_b32 v40, s33, 30
+; GISEL-NEXT:    v_writelane_b32 v40, s33, 32
 ; GISEL-NEXT:    s_mov_b32 s33, s32
 ; GISEL-NEXT:    s_addk_i32 s32, 0x400
 ; GISEL-NEXT:    buffer_store_dword v41, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GISEL-NEXT:    v_writelane_b32 v40, s34, 0
 ; GISEL-NEXT:    v_writelane_b32 v40, s35, 1
 ; GISEL-NEXT:    v_writelane_b32 v40, s36, 2
 ; GISEL-NEXT:    v_writelane_b32 v40, s37, 3
 ; GISEL-NEXT:    v_writelane_b32 v40, s38, 4

 ; GISEL-NEXT:    v_writelane_b32 v40, s56, 22
 ; GISEL-NEXT:    v_writelane_b32 v40, s57, 23
 ; GISEL-NEXT:    v_writelane_b32 v40, s58, 24
 ; GISEL-NEXT:    v_writelane_b32 v40, s59, 25
 ; GISEL-NEXT:    v_writelane_b32 v40, s60, 26
 ; GISEL-NEXT:    v_writelane_b32 v40, s61, 27
 ; GISEL-NEXT:    v_writelane_b32 v40, s62, 28
 ; GISEL-NEXT:    v_writelane_b32 v40, s63, 29
+; GISEL-NEXT:    v_writelane_b32 v40, s30, 30
+; GISEL-NEXT:    v_writelane_b32 v40, s31, 31
 ; GISEL-NEXT:    v_mov_b32_e32 v41, v0
-; GISEL-NEXT:    s_mov_b64 s[4:5], s[30:31]
-; GISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GISEL-NEXT:    s_mov_b64 s[4:5], exec
 ; GISEL-NEXT:  .LBB7_1: ; =>This Inner Loop Header: Depth=1
-; GISEL-NEXT:    v_readfirstlane_b32 s8, v1
-; GISEL-NEXT:    v_readfirstlane_b32 s9, v2
-; GISEL-NEXT:    v_cmp_eq_u64_e32 vcc, s[8:9], v[1:2]
-; GISEL-NEXT:    s_and_saveexec_b64 s[10:11], vcc
+; GISEL-NEXT:    v_readfirstlane_b32 s6, v1
+; GISEL-NEXT:    v_readfirstlane_b32 s7, v2
+; GISEL-NEXT:    v_cmp_eq_u64_e32 vcc, s[6:7], v[1:2]
+; GISEL-NEXT:    s_and_saveexec_b64 s[8:9], vcc
 ; GISEL-NEXT:    v_mov_b32_e32 v0, v41
-; GISEL-NEXT:    s_swappc_b64 s[30:31], s[8:9]
+; GISEL-NEXT:    s_swappc_b64 s[30:31], s[6:7]
 ; GISEL-NEXT:    ; implicit-def: $vgpr1_vgpr2
-; GISEL-NEXT:    s_xor_b64 exec, exec, s[10:11]
+; GISEL-NEXT:    s_xor_b64 exec, exec, s[8:9]
 ; GISEL-NEXT:    s_cbranch_execnz .LBB7_1
 ; GISEL-NEXT:  ; %bb.2:
-; GISEL-NEXT:    s_mov_b64 exec, s[6:7]
+; GISEL-NEXT:    s_mov_b64 exec, s[4:5]
 ; GISEL-NEXT:    v_mov_b32_e32 v0, v41
+; GISEL-NEXT:    v_readlane_b32 s30, v40, 30
+; GISEL-NEXT:    v_readlane_b32 s31, v40, 31
 ; GISEL-NEXT:    v_readlane_b32 s63, v40, 29
 ; GISEL-NEXT:    v_readlane_b32 s62, v40, 28
 ; GISEL-NEXT:    v_readlane_b32 s61, v40, 27
 ; GISEL-NEXT:    v_readlane_b32 s60, v40, 26
 ; GISEL-NEXT:    v_readlane_b32 s59, v40, 25
 ; GISEL-NEXT:    v_readlane_b32 s58, v40, 24
 ; GISEL-NEXT:    v_readlane_b32 s57, v40, 23
 ; GISEL-NEXT:    v_readlane_b32 s56, v40, 22

 ; GISEL-NEXT:    v_readlane_b32 s39, v40, 5
 ; GISEL-NEXT:    v_readlane_b32 s38, v40, 4
 ; GISEL-NEXT:    v_readlane_b32 s37, v40, 3
 ; GISEL-NEXT:    v_readlane_b32 s36, v40, 2
 ; GISEL-NEXT:    v_readlane_b32 s35, v40, 1
 ; GISEL-NEXT:    v_readlane_b32 s34, v40, 0
 ; GISEL-NEXT:    buffer_load_dword v41, off, s[0:3], s33 ; 4-byte Folded Reload
 ; GISEL-NEXT:    s_addk_i32 s32, 0xfc00
-; GISEL-NEXT:    v_readlane_b32 s33, v40, 30
-; GISEL-NEXT:    s_or_saveexec_b64 s[6:7], -1
+; GISEL-NEXT:    v_readlane_b32 s33, v40, 32
+; GISEL-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GISEL-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
-; GISEL-NEXT:    s_mov_b64 exec, s[6:7]
+; GISEL-NEXT:    s_mov_b64 exec, s[4:5]
 ; GISEL-NEXT:    s_waitcnt vmcnt(0)
-; GISEL-NEXT:    s_setpc_b64 s[4:5]
+; GISEL-NEXT:    s_setpc_b64 s[30:31]
   call amdgpu_gfx void %fptr(i32 %i)
   ret i32 %i
 }
 
 ; Use a variable inside a waterfall loop and use the return variable after the loop.
 ; TODO The argument and return variable could be in the same physical register, but the register
 ; allocator is not able to do that because the return value clashes with the liverange of an
 ; IMPLICIT_DEF of the argument.
 define i32 @test_indirect_call_vgpr_ptr_arg_and_return(i32 %i, i32(i32)* %fptr) {
 ; GCN-LABEL: test_indirect_call_vgpr_ptr_arg_and_return:
 ; GCN:       ; %bb.0:
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
-; GCN-NEXT:    v_writelane_b32 v40, s33, 30
+; GCN-NEXT:    v_writelane_b32 v40, s33, 32
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
 ; GCN-NEXT:    v_writelane_b32 v40, s34, 0
 ; GCN-NEXT:    v_writelane_b32 v40, s35, 1
 ; GCN-NEXT:    v_writelane_b32 v40, s36, 2
 ; GCN-NEXT:    v_writelane_b32 v40, s37, 3
 ; GCN-NEXT:    v_writelane_b32 v40, s38, 4
 ; GCN-NEXT:    v_writelane_b32 v40, s39, 5

 ; GCN-NEXT:    v_writelane_b32 v40, s56, 22
 ; GCN-NEXT:    v_writelane_b32 v40, s57, 23
 ; GCN-NEXT:    v_writelane_b32 v40, s58, 24
 ; GCN-NEXT:    v_writelane_b32 v40, s59, 25
 ; GCN-NEXT:    v_writelane_b32 v40, s60, 26
 ; GCN-NEXT:    v_writelane_b32 v40, s61, 27
 ; GCN-NEXT:    v_writelane_b32 v40, s62, 28
 ; GCN-NEXT:    v_writelane_b32 v40, s63, 29
-; GCN-NEXT:    s_mov_b64 s[4:5], s[30:31]
-; GCN-NEXT:    s_mov_b64 s[6:7], exec
+; GCN-NEXT:    v_writelane_b32 v40, s30, 30
+; GCN-NEXT:    v_writelane_b32 v40, s31, 31
+; GCN-NEXT:    s_mov_b64 s[4:5], exec
 ; GCN-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
-; GCN-NEXT:    v_readfirstlane_b32 s10, v1
-; GCN-NEXT:    v_readfirstlane_b32 s11, v2
-; GCN-NEXT:    v_cmp_eq_u64_e32 vcc, s[10:11], v[1:2]
-; GCN-NEXT:    s_and_saveexec_b64 s[8:9], vcc
-; GCN-NEXT:    s_swappc_b64 s[30:31], s[10:11]
+; GCN-NEXT:    v_readfirstlane_b32 s8, v1
+; GCN-NEXT:    v_readfirstlane_b32 s9, v2
+; GCN-NEXT:    v_cmp_eq_u64_e32 vcc, s[8:9], v[1:2]
+; GCN-NEXT:    s_and_saveexec_b64 s[6:7], vcc
+; GCN-NEXT:    s_swappc_b64 s[30:31], s[8:9]
 ; GCN-NEXT:    v_mov_b32_e32 v3, v0
 ; GCN-NEXT:    ; implicit-def: $vgpr1_vgpr2
 ; GCN-NEXT:    ; implicit-def: $vgpr0
-; GCN-NEXT:    s_xor_b64 exec, exec, s[8:9]
+; GCN-NEXT:    s_xor_b64 exec, exec, s[6:7]
 ; GCN-NEXT:    s_cbranch_execnz .LBB8_1
 ; GCN-NEXT:  ; %bb.2:
-; GCN-NEXT:    s_mov_b64 exec, s[6:7]
+; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    v_mov_b32_e32 v0, v3
+; GCN-NEXT:    v_readlane_b32 s30, v40, 30
+; GCN-NEXT:    v_readlane_b32 s31, v40, 31
 ; GCN-NEXT:    v_readlane_b32 s63, v40, 29
 ; GCN-NEXT:    v_readlane_b32 s62, v40, 28
 ; GCN-NEXT:    v_readlane_b32 s61, v40, 27
 ; GCN-NEXT:    v_readlane_b32 s60, v40, 26
 ; GCN-NEXT:    v_readlane_b32 s59, v40, 25
 ; GCN-NEXT:    v_readlane_b32 s58, v40, 24
 ; GCN-NEXT:    v_readlane_b32 s57, v40, 23
 ; GCN-NEXT:    v_readlane_b32 s56, v40, 22

 ; GCN-NEXT:    v_readlane_b32 s40, v40, 6
 ; GCN-NEXT:    v_readlane_b32 s39, v40, 5
 ; GCN-NEXT:    v_readlane_b32 s38, v40, 4
 ; GCN-NEXT:    v_readlane_b32 s37, v40, 3
 ; GCN-NEXT:    v_readlane_b32 s36, v40, 2
 ; GCN-NEXT:    v_readlane_b32 s35, v40, 1
 ; GCN-NEXT:    v_readlane_b32 s34, v40, 0
 ; GCN-NEXT:    s_addk_i32 s32, 0xfc00
-; GCN-NEXT:    v_readlane_b32 s33, v40, 30
-; GCN-NEXT:    s_or_saveexec_b64 s[6:7], -1
+; GCN-NEXT:    v_readlane_b32 s33, v40, 32
+; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
-; GCN-NEXT:    s_mov_b64 exec, s[6:7]
+; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
-; GCN-NEXT:    s_setpc_b64 s[4:5]
+; GCN-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GISEL-LABEL: test_indirect_call_vgpr_ptr_arg_and_return:
 ; GISEL:       ; %bb.0:
 ; GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GISEL-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GISEL-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GISEL-NEXT:    s_mov_b64 exec, s[4:5]
-; GISEL-NEXT:    v_writelane_b32 v40, s33, 30
+; GISEL-NEXT:    v_writelane_b32 v40, s33, 32
 ; GISEL-NEXT:    s_mov_b32 s33, s32
 ; GISEL-NEXT:    s_addk_i32 s32, 0x400
 ; GISEL-NEXT:    v_writelane_b32 v40, s34, 0
 ; GISEL-NEXT:    v_writelane_b32 v40, s35, 1
 ; GISEL-NEXT:    v_writelane_b32 v40, s36, 2
 ; GISEL-NEXT:    v_writelane_b32 v40, s37, 3
 ; GISEL-NEXT:    v_writelane_b32 v40, s38, 4
 ; GISEL-NEXT:    v_writelane_b32 v40, s39, 5

 ; GISEL-NEXT:    v_writelane_b32 v40, s56, 22
 ; GISEL-NEXT:    v_writelane_b32 v40, s57, 23
 ; GISEL-NEXT:    v_writelane_b32 v40, s58, 24
 ; GISEL-NEXT:    v_writelane_b32 v40, s59, 25
 ; GISEL-NEXT:    v_writelane_b32 v40, s60, 26
 ; GISEL-NEXT:    v_writelane_b32 v40, s61, 27
 ; GISEL-NEXT:    v_writelane_b32 v40, s62, 28
 ; GISEL-NEXT:    v_writelane_b32 v40, s63, 29
-; GISEL-NEXT:    s_mov_b64 s[4:5], s[30:31]
-; GISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GISEL-NEXT:    v_writelane_b32 v40, s30, 30
+; GISEL-NEXT:    v_writelane_b32 v40, s31, 31
+; GISEL-NEXT:    s_mov_b64 s[4:5], exec
 ; GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
-; GISEL-NEXT:    v_readfirstlane_b32 s8, v1
-; GISEL-NEXT:    v_readfirstlane_b32 s9, v2
-; GISEL-NEXT:    v_cmp_eq_u64_e32 vcc, s[8:9], v[1:2]
-; GISEL-NEXT:    s_and_saveexec_b64 s[10:11], vcc
-; GISEL-NEXT:    s_swappc_b64 s[30:31], s[8:9]
+; GISEL-NEXT:    v_readfirstlane_b32 s6, v1
+; GISEL-NEXT:    v_readfirstlane_b32 s7, v2
+; GISEL-NEXT:    v_cmp_eq_u64_e32 vcc, s[6:7], v[1:2]
+; GISEL-NEXT:    s_and_saveexec_b64 s[8:9], vcc
+; GISEL-NEXT:    s_swappc_b64 s[30:31], s[6:7]
 ; GISEL-NEXT:    v_mov_b32_e32 v3, v0
 ; GISEL-NEXT:    ; implicit-def: $vgpr1_vgpr2
 ; GISEL-NEXT:    ; implicit-def: $vgpr0
-; GISEL-NEXT:    s_xor_b64 exec, exec, s[10:11]
+; GISEL-NEXT:    s_xor_b64 exec, exec, s[8:9]
 ; GISEL-NEXT:    s_cbranch_execnz .LBB8_1
 ; GISEL-NEXT:  ; %bb.2:
-; GISEL-NEXT:    s_mov_b64 exec, s[6:7]
+; GISEL-NEXT:    s_mov_b64 exec, s[4:5]
 ; GISEL-NEXT:    v_mov_b32_e32 v0, v3
+; GISEL-NEXT:    v_readlane_b32 s30, v40, 30
+; GISEL-NEXT:    v_readlane_b32 s31, v40, 31
 ; GISEL-NEXT:    v_readlane_b32 s63, v40, 29
 ; GISEL-NEXT:    v_readlane_b32 s62, v40, 28
 ; GISEL-NEXT:    v_readlane_b32 s61, v40, 27
 ; GISEL-NEXT:    v_readlane_b32 s60, v40, 26
 ; GISEL-NEXT:    v_readlane_b32 s59, v40, 25
 ; GISEL-NEXT:    v_readlane_b32 s58, v40, 24
 ; GISEL-NEXT:    v_readlane_b32 s57, v40, 23
 ; GISEL-NEXT:    v_readlane_b32 s56, v40, 22

 ; GISEL-NEXT:    v_readlane_b32 s40, v40, 6
 ; GISEL-NEXT:    v_readlane_b32 s39, v40, 5
 ; GISEL-NEXT:    v_readlane_b32 s38, v40, 4
 ; GISEL-NEXT:    v_readlane_b32 s37, v40, 3
 ; GISEL-NEXT:    v_readlane_b32 s36, v40, 2
 ; GISEL-NEXT:    v_readlane_b32 s35, v40, 1
 ; GISEL-NEXT:    v_readlane_b32 s34, v40, 0
 ; GISEL-NEXT:    s_addk_i32 s32, 0xfc00
-; GISEL-NEXT:    v_readlane_b32 s33, v40, 30
-; GISEL-NEXT:    s_or_saveexec_b64 s[6:7], -1
+; GISEL-NEXT:    v_readlane_b32 s33, v40, 32
+; GISEL-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GISEL-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
-; GISEL-NEXT:    s_mov_b64 exec, s[6:7]
+; GISEL-NEXT:    s_mov_b64 exec, s[4:5]
 ; GISEL-NEXT:    s_waitcnt vmcnt(0)
-; GISEL-NEXT:    s_setpc_b64 s[4:5]
+; GISEL-NEXT:    s_setpc_b64 s[30:31]
   %ret = call amdgpu_gfx i32 %fptr(i32 %i)
   ret i32 %ret
 }
 
 ; Calling a vgpr can never be a tail call.
 define void @test_indirect_tail_call_vgpr_ptr(void()* %fptr) {
 ; GCN-LABEL: test_indirect_tail_call_vgpr_ptr:
 ; GCN:       ; %bb.0:
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
-; GCN-NEXT:    v_writelane_b32 v40, s33, 30
+; GCN-NEXT:    v_writelane_b32 v40, s33, 32
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
 ; GCN-NEXT:    v_writelane_b32 v40, s34, 0
 ; GCN-NEXT:    v_writelane_b32 v40, s35, 1
 ; GCN-NEXT:    v_writelane_b32 v40, s36, 2
 ; GCN-NEXT:    v_writelane_b32 v40, s37, 3
 ; GCN-NEXT:    v_writelane_b32 v40, s38, 4
 ; GCN-NEXT:    v_writelane_b32 v40, s39, 5

 ; GCN-NEXT:    v_writelane_b32 v40, s56, 22
 ; GCN-NEXT:    v_writelane_b32 v40, s57, 23
 ; GCN-NEXT:    v_writelane_b32 v40, s58, 24
 ; GCN-NEXT:    v_writelane_b32 v40, s59, 25
 ; GCN-NEXT:    v_writelane_b32 v40, s60, 26
 ; GCN-NEXT:    v_writelane_b32 v40, s61, 27
 ; GCN-NEXT:    v_writelane_b32 v40, s62, 28
 ; GCN-NEXT:    v_writelane_b32 v40, s63, 29
-; GCN-NEXT:    s_mov_b64 s[4:5], s[30:31]
-; GCN-NEXT:    s_mov_b64 s[6:7], exec
+; GCN-NEXT:    v_writelane_b32 v40, s30, 30
+; GCN-NEXT:    v_writelane_b32 v40, s31, 31
+; GCN-NEXT:    s_mov_b64 s[4:5], exec
 ; GCN-NEXT:  .LBB9_1: ; =>This Inner Loop Header: Depth=1
-; GCN-NEXT:    v_readfirstlane_b32 s10, v0
-; GCN-NEXT:    v_readfirstlane_b32 s11, v1
-; GCN-NEXT:    v_cmp_eq_u64_e32 vcc, s[10:11], v[0:1]
-; GCN-NEXT:    s_and_saveexec_b64 s[8:9], vcc
-; GCN-NEXT:    s_swappc_b64 s[30:31], s[10:11]
+; GCN-NEXT:    v_readfirstlane_b32 s8, v0
+; GCN-NEXT:    v_readfirstlane_b32 s9, v1
+; GCN-NEXT:    v_cmp_eq_u64_e32 vcc, s[8:9], v[0:1]
+; GCN-NEXT:    s_and_saveexec_b64 s[6:7], vcc
+; GCN-NEXT:    s_swappc_b64 s[30:31], s[8:9]
 ; GCN-NEXT:    ; implicit-def: $vgpr0_vgpr1
-; GCN-NEXT:    s_xor_b64 exec, exec, s[8:9]
+; GCN-NEXT:    s_xor_b64 exec, exec, s[6:7]
 ; GCN-NEXT:    s_cbranch_execnz .LBB9_1
 ; GCN-NEXT:  ; %bb.2:
-; GCN-NEXT:    s_mov_b64 exec, s[6:7]
+; GCN-NEXT:    s_mov_b64 exec, s[4:5]
+; GCN-NEXT:    v_readlane_b32 s30, v40, 30
+; GCN-NEXT:    v_readlane_b32 s31, v40, 31
 ; GCN-NEXT:    v_readlane_b32 s63, v40, 29
 ; GCN-NEXT:    v_readlane_b32 s62, v40, 28
 ; GCN-NEXT:    v_readlane_b32 s61, v40, 27
 ; GCN-NEXT:    v_readlane_b32 s60, v40, 26
 ; GCN-NEXT:    v_readlane_b32 s59, v40, 25
 ; GCN-NEXT:    v_readlane_b32 s58, v40, 24
 ; GCN-NEXT:    v_readlane_b32 s57, v40, 23
 ; GCN-NEXT:    v_readlane_b32 s56, v40, 22

 ; GCN-NEXT:    v_readlane_b32 s40, v40, 6
 ; GCN-NEXT:    v_readlane_b32 s39, v40, 5
 ; GCN-NEXT:    v_readlane_b32 s38, v40, 4
 ; GCN-NEXT:    v_readlane_b32 s37, v40, 3
 ; GCN-NEXT:    v_readlane_b32 s36, v40, 2
 ; GCN-NEXT:    v_readlane_b32 s35, v40, 1
 ; GCN-NEXT:    v_readlane_b32 s34, v40, 0
 ; GCN-NEXT:    s_addk_i32 s32, 0xfc00
-; GCN-NEXT:    v_readlane_b32 s33, v40, 30
-; GCN-NEXT:    s_or_saveexec_b64 s[6:7], -1
+; GCN-NEXT:    v_readlane_b32 s33, v40, 32
+; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
-; GCN-NEXT:    s_mov_b64 exec, s[6:7]
+; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
-; GCN-NEXT:    s_setpc_b64 s[4:5]
+; GCN-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GISEL-LABEL: test_indirect_tail_call_vgpr_ptr:
 ; GISEL:       ; %bb.0:
 ; GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GISEL-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GISEL-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GISEL-NEXT:    s_mov_b64 exec, s[4:5]
-; GISEL-NEXT:    v_writelane_b32 v40, s33, 30
+; GISEL-NEXT:    v_writelane_b32 v40, s33, 32
 ; GISEL-NEXT:    s_mov_b32 s33, s32
 ; GISEL-NEXT:    s_addk_i32 s32, 0x400
 ; GISEL-NEXT:    v_writelane_b32 v40, s34, 0
 ; GISEL-NEXT:    v_writelane_b32 v40, s35, 1
 ; GISEL-NEXT:    v_writelane_b32 v40, s36, 2
 ; GISEL-NEXT:    v_writelane_b32 v40, s37, 3
 ; GISEL-NEXT:    v_writelane_b32 v40, s38, 4
 ; GISEL-NEXT:    v_writelane_b32 v40, s39, 5

 ; GISEL-NEXT:    v_writelane_b32 v40, s56, 22
 ; GISEL-NEXT:    v_writelane_b32 v40, s57, 23
 ; GISEL-NEXT:    v_writelane_b32 v40, s58, 24
 ; GISEL-NEXT:    v_writelane_b32 v40, s59, 25
 ; GISEL-NEXT:    v_writelane_b32 v40, s60, 26
 ; GISEL-NEXT:    v_writelane_b32 v40, s61, 27
 ; GISEL-NEXT:    v_writelane_b32 v40, s62, 28
 ; GISEL-NEXT:    v_writelane_b32 v40, s63, 29
-; GISEL-NEXT:    s_mov_b64 s[4:5], s[30:31]
-; GISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GISEL-NEXT:    v_writelane_b32 v40, s30, 30
+; GISEL-NEXT:    v_writelane_b32 v40, s31, 31
+; GISEL-NEXT:    s_mov_b64 s[4:5], exec
 ; GISEL-NEXT:  .LBB9_1: ; =>This Inner Loop Header: Depth=1
-; GISEL-NEXT:    v_readfirstlane_b32 s8, v0
-; GISEL-NEXT:    v_readfirstlane_b32 s9, v1
-; GISEL-NEXT:    v_cmp_eq_u64_e32 vcc, s[8:9], v[0:1]
-; GISEL-NEXT:    s_and_saveexec_b64 s[10:11], vcc
-; GISEL-NEXT:    s_swappc_b64 s[30:31], s[8:9]
+; GISEL-NEXT:    v_readfirstlane_b32 s6, v0
+; GISEL-NEXT:    v_readfirstlane_b32 s7, v1
+; GISEL-NEXT:    v_cmp_eq_u64_e32 vcc, s[6:7], v[0:1]
+; GISEL-NEXT:    s_and_saveexec_b64 s[8:9], vcc
+; GISEL-NEXT:    s_swappc_b64 s[30:31], s[6:7]
 ; GISEL-NEXT:    ; implicit-def: $vgpr0_vgpr1
-; GISEL-NEXT:    s_xor_b64 exec, exec, s[10:11]
+; GISEL-NEXT:    s_xor_b64 exec, exec, s[8:9]
 ; GISEL-NEXT:    s_cbranch_execnz .LBB9_1
 ; GISEL-NEXT:  ; %bb.2:
-; GISEL-NEXT:    s_mov_b64 exec, s[6:7]
+; GISEL-NEXT:    s_mov_b64 exec, s[4:5]
+; GISEL-NEXT:    v_readlane_b32 s30, v40, 30
+; GISEL-NEXT:    v_readlane_b32 s31, v40, 31
 ; GISEL-NEXT:    v_readlane_b32 s63, v40, 29
 ; GISEL-NEXT:    v_readlane_b32 s62, v40, 28
 ; GISEL-NEXT:    v_readlane_b32 s61, v40, 27
 ; GISEL-NEXT:    v_readlane_b32 s60, v40, 26
 ; GISEL-NEXT:    v_readlane_b32 s59, v40, 25
 ; GISEL-NEXT:    v_readlane_b32 s58, v40, 24
 ; GISEL-NEXT:    v_readlane_b32 s57, v40, 23
 ; GISEL-NEXT:    v_readlane_b32 s56, v40, 22

 ; GISEL-NEXT:    v_readlane_b32 s40, v40, 6
 ; GISEL-NEXT:    v_readlane_b32 s39, v40, 5
 ; GISEL-NEXT:    v_readlane_b32 s38, v40, 4
 ; GISEL-NEXT:    v_readlane_b32 s37, v40, 3
 ; GISEL-NEXT:    v_readlane_b32 s36, v40, 2
 ; GISEL-NEXT:    v_readlane_b32 s35, v40, 1
 ; GISEL-NEXT:    v_readlane_b32 s34, v40, 0
 ; GISEL-NEXT:    s_addk_i32 s32, 0xfc00
-; GISEL-NEXT:    v_readlane_b32 s33, v40, 30
-; GISEL-NEXT:    s_or_saveexec_b64 s[6:7], -1
+; GISEL-NEXT:    v_readlane_b32 s33, v40, 32
+; GISEL-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GISEL-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
-; GISEL-NEXT:    s_mov_b64 exec, s[6:7]
+; GISEL-NEXT:    s_mov_b64 exec, s[4:5]
 ; GISEL-NEXT:    s_waitcnt vmcnt(0)
-; GISEL-NEXT:    s_setpc_b64 s[4:5]
+; GISEL-NEXT:    s_setpc_b64 s[30:31]
   tail call amdgpu_gfx void %fptr()
   ret void
 }

llvm/test/CodeGen/AMDGPU/inline-asm.i128.ll

 ; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
 ; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx908 -stop-after=finalize-isel -o - %s | FileCheck -check-prefix=GFX908 %s
 ; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx90a -stop-after=finalize-isel -o - %s | FileCheck -check-prefix=GFX90A %s
 
 ; Make sure we only use one 128-bit register instead of 2 for i128 asm
 ; constraints
 
 define amdgpu_kernel void @s_input_output_i128() {
   ; GFX908-LABEL: name: s_input_output_i128
   ; GFX908: bb.0 (%ir-block.0):
-  ; GFX908-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 5767178 /* regdef:SGPR_128 */, def %4
+  ; GFX908-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 5242890 /* regdef:SGPR_128 */, def %4
   ; GFX908-NEXT:   [[COPY:%[0-9]+]]:sgpr_128 = COPY %4
-  ; GFX908-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 5767177 /* reguse:SGPR_128 */, [[COPY]]
+  ; GFX908-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 5242889 /* reguse:SGPR_128 */, [[COPY]]
   ; GFX908-NEXT:   S_ENDPGM 0
   ; GFX90A-LABEL: name: s_input_output_i128
   ; GFX90A: bb.0 (%ir-block.0):
-  ; GFX90A-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 5767178 /* regdef:SGPR_128 */, def %4
+  ; GFX90A-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 5242890 /* regdef:SGPR_128 */, def %4
   ; GFX90A-NEXT:   [[COPY:%[0-9]+]]:sgpr_128 = COPY %4
-  ; GFX90A-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 5767177 /* reguse:SGPR_128 */, [[COPY]]
+  ; GFX90A-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 5242889 /* reguse:SGPR_128 */, [[COPY]]
   ; GFX90A-NEXT:   S_ENDPGM 0
   %val = tail call i128 asm sideeffect "; def $0", "=s"()
   call void asm sideeffect "; use $0", "s"(i128 %val)
   ret void
 }
 
 define amdgpu_kernel void @v_input_output_i128() {
   ; GFX908-LABEL: name: v_input_output_i128
   ; GFX908: bb.0 (%ir-block.0):
-  ; GFX908-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 5308426 /* regdef:VReg_128 */, def %4
+  ; GFX908-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 4784138 /* regdef:VReg_128 */, def %4
   ; GFX908-NEXT:   [[COPY:%[0-9]+]]:vreg_128 = COPY %4
-  ; GFX908-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 5308425 /* reguse:VReg_128 */, [[COPY]]
+  ; GFX908-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 4784137 /* reguse:VReg_128 */, [[COPY]]
   ; GFX908-NEXT:   S_ENDPGM 0
   ; GFX90A-LABEL: name: v_input_output_i128
   ; GFX90A: bb.0 (%ir-block.0):
-  ; GFX90A-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 5505034 /* regdef:VReg_128_Align2 */, def %4
+  ; GFX90A-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 4980746 /* regdef:VReg_128_Align2 */, def %4
   ; GFX90A-NEXT:   [[COPY:%[0-9]+]]:vreg_128_align2 = COPY %4
-  ; GFX90A-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 5505033 /* reguse:VReg_128_Align2 */, [[COPY]]
+  ; GFX90A-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 4980745 /* reguse:VReg_128_Align2 */, [[COPY]]
   ; GFX90A-NEXT:   S_ENDPGM 0
   %val = tail call i128 asm sideeffect "; def $0", "=v"()
   call void asm sideeffect "; use $0", "v"(i128 %val)
   ret void
 }
 
 define amdgpu_kernel void @a_input_output_i128() {
   ; GFX908-LABEL: name: a_input_output_i128
   ; GFX908: bb.0 (%ir-block.0):
-  ; GFX908-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 5242890 /* regdef:AReg_128 */, def %4
+  ; GFX908-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 4718602 /* regdef:AReg_128 */, def %4
   ; GFX908-NEXT:   [[COPY:%[0-9]+]]:areg_128 = COPY %4
-  ; GFX908-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 5242889 /* reguse:AReg_128 */, [[COPY]]
+  ; GFX908-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 4718601 /* reguse:AReg_128 */, [[COPY]]
   ; GFX908-NEXT:   S_ENDPGM 0
   ; GFX90A-LABEL: name: a_input_output_i128
   ; GFX90A: bb.0 (%ir-block.0):
-  ; GFX90A-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 5439498 /* regdef:AReg_128_Align2 */, def %4
+  ; GFX90A-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 4915210 /* regdef:AReg_128_Align2 */, def %4
   ; GFX90A-NEXT:   [[COPY:%[0-9]+]]:areg_128_align2 = COPY %4
-  ; GFX90A-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 5439497 /* reguse:AReg_128_Align2 */, [[COPY]]
+  ; GFX90A-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 4915209 /* reguse:AReg_128_Align2 */, [[COPY]]
   ; GFX90A-NEXT:   S_ENDPGM 0
   %val = call i128 asm sideeffect "; def $0", "=a"()
   call void asm sideeffect "; use $0", "a"(i128 %val)
   ret void
 }

llvm/test/CodeGen/AMDGPU/ipra.ll

 define amdgpu_kernel void @kernel_call() #0 {
   %vgpr = load volatile i32, i32 addrspace(1)* undef
   tail call void @func()
   store volatile i32 %vgpr, i32 addrspace(1)* undef
   ret void
 }
 
 ; GCN-LABEL: {{^}}func_regular_call:
-; GCN-NOT: buffer_store
 ; GCN-NOT: buffer_load
 ; GCN-NOT: readlane
-; GCN-NOT: writelane
-; GCN: flat_load_dword v8
+; GCN: flat_load_dword v9
 ; GCN: s_swappc_b64
-; GCN-NOT: buffer_store
 ; GCN-NOT: buffer_load
 ; GCN-NOT: readlane
-; GCN-NOT: writelane
-; GCN: flat_store_dword v{{\[[0-9]+:[0-9]+\]}}, v8
+; GCN: flat_store_dword v{{\[[0-9]+:[0-9]+\]}}, v9
 
-; GCN: ; NumSgprs: 32
-; GCN: ; NumVgprs: 9
+; GCN: ; NumSgprs: 34
+; GCN: ; NumVgprs: 10
 define void @func_regular_call() #1 {
   %vgpr = load volatile i32, i32 addrspace(1)* undef
   tail call void @func()
   store volatile i32 %vgpr, i32 addrspace(1)* undef
   ret void
 }
 
 ; GCN-LABEL: {{^}}func_tail_call:
 ; GCN: s_waitcnt
 ; GCN-NEXT: s_getpc_b64 s[4:5]
 ; GCN-NEXT: s_add_u32 s4,
 ; GCN-NEXT: s_addc_u32 s5,
 ; GCN-NEXT: s_setpc_b64 s[4:5]
 
 ; GCN: ; NumSgprs: 32
 ; GCN: ; NumVgprs: 8
 define void @func_tail_call() #1 {
   tail call void @func()
   ret void
 }
 
 ; GCN-LABEL: {{^}}func_call_tail_call:
-; GCN: flat_load_dword v8
+; GCN: flat_load_dword v9
 ; GCN: s_swappc_b64
-; GCN: flat_store_dword v{{\[[0-9]+:[0-9]+\]}}, v8
+; GCN: flat_store_dword v{{\[[0-9]+:[0-9]+\]}}, v9
 ; GCN: s_setpc_b64
 
-; GCN: ; NumSgprs: 32
-; GCN: ; NumVgprs: 9
+; GCN: ; NumSgprs: 34
+; GCN: ; NumVgprs: 10
 define void @func_call_tail_call() #1 {
   %vgpr = load volatile i32, i32 addrspace(1)* undef
   tail call void @func()
   store volatile i32 %vgpr, i32 addrspace(1)* undef
   tail call void @func()
   ret void
 }
 
 define void @void_func_void() noinline {
   ret void
 }
 
 ; Make sure we don't get save/restore of FP between calls.
 ; GCN-LABEL: {{^}}test_funcx2:
-; GCN-NOT: s5
+; GCN: s_getpc_b64
 ; GCN-NOT: s32
+; GCN: s_swappc_b64
+; GCN-NOT: s32
+; GCN: s_swappc_b64
 define void @test_funcx2() #0 {
   call void @void_func_void()
   call void @void_func_void()
   ret void
 }
 
 attributes #0 = { nounwind }
 attributes #1 = { nounwind noinline }

llvm/test/CodeGen/AMDGPU/llvm.amdgcn.implicitarg.ptr.ll

 define amdgpu_kernel void @opencl_kernel_call_implicitarg_ptr_func([112 x i8]) #1 {
   call void @func_implicitarg_ptr()
   ret void
 }
 
 ; GCN-LABEL: {{^}}func_call_implicitarg_ptr_func:
 ; GCN-NOT: s4
 ; GCN-NOT: s5
-; GCN-NOT: s[4:5]
+; GCN: s_swappc_b64
+; GCN: s_setpc_b64 s[30:31]
 define void @func_call_implicitarg_ptr_func() #0 {
   call void @func_implicitarg_ptr()
   ret void
 }
 
 ; GCN-LABEL: {{^}}opencl_func_call_implicitarg_ptr_func:
 ; GCN-NOT: s4
 ; GCN-NOT: s5
-; GCN-NOT: s[4:5]
+; GCN: s_swappc_b64
+; GCN: s_setpc_b64 s[30:31]
 define void @opencl_func_call_implicitarg_ptr_func() #0 {
   call void @func_implicitarg_ptr()
   ret void
 }
 
 ; GCN-LABEL: {{^}}func_kernarg_implicitarg_ptr:
 ; GCN: s_waitcnt
 ; GCN-DAG: s_mov_b64 [[NULL:s\[[0-9]+:[0-9]+\]]], 0

llvm/test/CodeGen/AMDGPU/llvm.powi.ll

 }
 
 define float @v_powi_neg1_f32(float %l) {
 ; GFX7-LABEL: v_powi_neg1_f32:
 ; GFX7:       ; %bb.0:
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX7-NEXT:    v_div_scale_f32 v1, s[4:5], v0, v0, 1.0
 ; GFX7-NEXT:    v_rcp_f32_e32 v2, v1
-; GFX7-NEXT:    v_fma_f32 v3, -v1, v2, 1.0
-; GFX7-NEXT:    v_fma_f32 v2, v3, v2, v2
 ; GFX7-NEXT:    v_div_scale_f32 v3, vcc, 1.0, v0, 1.0
+; GFX7-NEXT:    v_fma_f32 v4, -v1, v2, 1.0
+; GFX7-NEXT:    v_fma_f32 v2, v4, v2, v2
 ; GFX7-NEXT:    v_mul_f32_e32 v4, v3, v2
 ; GFX7-NEXT:    v_fma_f32 v5, -v1, v4, v3
 ; GFX7-NEXT:    v_fma_f32 v4, v5, v2, v4
 ; GFX7-NEXT:    v_fma_f32 v1, -v1, v4, v3
 ; GFX7-NEXT:    v_div_fmas_f32 v1, v1, v2, v4
 ; GFX7-NEXT:    v_div_fixup_f32 v0, v1, v0, 1.0
 ; GFX7-NEXT:    s_setpc_b64 s[30:31]
 ;

 
 define float @v_powi_neg2_f32(float %l) {
 ; GFX7-LABEL: v_powi_neg2_f32:
 ; GFX7:       ; %bb.0:
 ; GFX7-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX7-NEXT:    v_mul_f32_e32 v0, v0, v0
 ; GFX7-NEXT:    v_div_scale_f32 v1, s[4:5], v0, v0, 1.0
 ; GFX7-NEXT:    v_rcp_f32_e32 v2, v1
-; GFX7-NEXT:    v_fma_f32 v3, -v1, v2, 1.0
-; GFX7-NEXT:    v_fma_f32 v2, v3, v2, v2
 ; GFX7-NEXT:    v_div_scale_f32 v3, vcc, 1.0, v0, 1.0
+; GFX7-NEXT:    v_fma_f32 v4, -v1, v2, 1.0
+; GFX7-NEXT:    v_fma_f32 v2, v4, v2, v2
 ; GFX7-NEXT:    v_mul_f32_e32 v4, v3, v2
 ; GFX7-NEXT:    v_fma_f32 v5, -v1, v4, v3
 ; GFX7-NEXT:    v_fma_f32 v4, v5, v2, v4
 ; GFX7-NEXT:    v_fma_f32 v1, -v1, v4, v3
 ; GFX7-NEXT:    v_div_fmas_f32 v1, v1, v2, v4
 ; GFX7-NEXT:    v_div_fixup_f32 v0, v1, v0, 1.0
 ; GFX7-NEXT:    s_setpc_b64 s[30:31]
 ;

 ; GFX7-NEXT:    v_mul_f32_e32 v0, v0, v0
 ; GFX7-NEXT:    v_mul_f32_e32 v0, v0, v0
 ; GFX7-NEXT:    v_mul_f32_e32 v0, v0, v0
 ; GFX7-NEXT:    v_mul_f32_e32 v0, v0, v0
 ; GFX7-NEXT:    v_mul_f32_e32 v0, v0, v0
 ; GFX7-NEXT:    v_mul_f32_e32 v0, v0, v0
 ; GFX7-NEXT:    v_div_scale_f32 v1, s[4:5], v0, v0, 1.0
 ; GFX7-NEXT:    v_rcp_f32_e32 v2, v1
-; GFX7-NEXT:    v_fma_f32 v3, -v1, v2, 1.0
-; GFX7-NEXT:    v_fma_f32 v2, v3, v2, v2
 ; GFX7-NEXT:    v_div_scale_f32 v3, vcc, 1.0, v0, 1.0
+; GFX7-NEXT:    v_fma_f32 v4, -v1, v2, 1.0
+; GFX7-NEXT:    v_fma_f32 v2, v4, v2, v2
 ; GFX7-NEXT:    v_mul_f32_e32 v4, v3, v2
 ; GFX7-NEXT:    v_fma_f32 v5, -v1, v4, v3
 ; GFX7-NEXT:    v_fma_f32 v4, v5, v2, v4
 ; GFX7-NEXT:    v_fma_f32 v1, -v1, v4, v3
 ; GFX7-NEXT:    v_div_fmas_f32 v1, v1, v2, v4
 ; GFX7-NEXT:    v_div_fixup_f32 v0, v1, v0, 1.0
 ; GFX7-NEXT:    s_setpc_b64 s[30:31]
 ;

llvm/test/CodeGen/AMDGPU/mul24-pass-ordering.ll

 ; GFX9-NEXT:    v_mad_u32_u24 v41, v42, v41, v43
 ; GFX9-NEXT:    v_mov_b32_e32 v0, v41
 ; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    v_add_u32_e32 v0, v41, v43
 ; GFX9-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-NEXT:    buffer_load_dword v43, off, s[0:3], s33 ; 4-byte Folded Reload
 ; GFX9-NEXT:    buffer_load_dword v42, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
 ; GFX9-NEXT:    buffer_load_dword v41, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload
-; GFX9-NEXT:    v_readlane_b32 s4, v40, 2
-; GFX9-NEXT:    v_readlane_b32 s5, v40, 3
+; GFX9-NEXT:    v_readlane_b32 s30, v40, 2
+; GFX9-NEXT:    v_readlane_b32 s31, v40, 3
 ; GFX9-NEXT:    v_readlane_b32 s35, v40, 1
 ; GFX9-NEXT:    v_readlane_b32 s34, v40, 0
 ; GFX9-NEXT:    s_addk_i32 s32, 0xf800
 ; GFX9-NEXT:    v_readlane_b32 s33, v40, 4
-; GFX9-NEXT:    s_or_saveexec_b64 s[6:7], -1
+; GFX9-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GFX9-NEXT:    buffer_load_dword v40, off, s[0:3], s32 offset:12 ; 4-byte Folded Reload
-; GFX9-NEXT:    s_mov_b64 exec, s[6:7]
+; GFX9-NEXT:    s_mov_b64 exec, s[4:5]
 ; GFX9-NEXT:    s_waitcnt vmcnt(0)
-; GFX9-NEXT:    s_setpc_b64 s[4:5]
+; GFX9-NEXT:    s_setpc_b64 s[30:31]
   %b = and i32 %b.arg, 16777215
   %s = and i32 %s.arg, 16777215
 
 ; CHECK-LABEL: @slsr1(
   ; foo(b * s);
   %mul0 = mul i32 %b, %s
 ; CHECK: mul i32
 ; CHECK-NOT: mul i32

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

 ; 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:    v_readlane_b32 s30, v1, 0
-; CHECK-NEXT:    v_readlane_b32 s31, v1, 1
 ; CHECK-NEXT:    ;;#ASMSTART
 ; CHECK-NEXT:    ; clobber csr v40
 ; CHECK-NEXT:    ;;#ASMEND
+; CHECK-NEXT:    v_readlane_b32 s30, v1, 0
+; CHECK-NEXT:    v_readlane_b32 s31, v1, 1
 ; 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:    v_readlane_b32 s33, v1, 2
 ; CHECK-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; CHECK-NEXT:    buffer_load_dword v1, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
 ; CHECK-NEXT:    s_mov_b64 exec, s[4:5]
 ; CHECK-NEXT:    s_waitcnt vmcnt(0)
 ; CHECK-NEXT:    s_setpc_b64 s[30:31]

llvm/test/CodeGen/AMDGPU/nested-calls.ll

 ; GCN-DAG: v_writelane_b32 v40, s33, 2
 ; GCN-DAG: s_mov_b32 s33, s32
 ; GCN-DAG: s_addk_i32 s32, 0x400
 ; GCN-DAG: v_writelane_b32 v40, s30, 0
 ; GCN-DAG: v_writelane_b32 v40, s31, 1
 
 ; GCN: s_swappc_b64
 
-; GCN: v_readlane_b32 s4, v40, 0
-; GCN: v_readlane_b32 s5, v40, 1
+; GCN: v_readlane_b32 s30, v40, 0
+; GCN: v_readlane_b32 s31, v40, 1
 
 ; GCN-NEXT: s_addk_i32 s32, 0xfc00
 ; GCN-NEXT: v_readlane_b32 s33, v40, 2
 ; GCN: s_or_saveexec_b64 [[COPY_EXEC1:s\[[0-9]+:[0-9]+\]]], -1{{$}}
 ; GCN-NEXT: buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GCN-NEXT: s_mov_b64 exec, [[COPY_EXEC1]]
 ; GCN-NEXT: s_waitcnt vmcnt(0)
-; GCN-NEXT: s_setpc_b64 s[4:5]
+; GCN-NEXT: s_setpc_b64 s[30:31]
 define void @test_func_call_external_void_func_i32_imm() #0 {
   call void @external_void_func_i32(i32 42)
   ret void
 }
 
 ; GCN-LABEL: {{^}}test_func_call_external_void_func_i32_imm_stack_use:
 ; GCN: s_waitcnt
 ; GCN: s_mov_b32 s33, s32

llvm/test/CodeGen/AMDGPU/no-remat-indirect-mov.mir

liveins:
   - { reg: '$vgpr14', virtual-reg: '%14' }
   - { reg: '$vgpr15', virtual-reg: '%15' }
   - { reg: '$vgpr16', virtual-reg: '%16' }
   - { reg: '$sgpr30_sgpr31', virtual-reg: '%17' }
 body:             |
   ; GFX9-LABEL: name: index_vgpr_waterfall_loop
   ; GFX9: bb.0:
   ; GFX9:   successors: %bb.1(0x80000000)
-  ; GFX9:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7, $vgpr8, $vgpr9, $vgpr10, $vgpr11, $vgpr12, $vgpr13, $vgpr14, $vgpr15, $vgpr16, $sgpr30_sgpr31
-  ; GFX9:   [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr30_sgpr31
+  ; GFX9:   liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7, $vgpr8, $vgpr9, $vgpr10, $vgpr11, $vgpr12, $vgpr13, $vgpr14, $vgpr15, $vgpr16
   ; GFX9:   [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr16
   ; GFX9:   undef %18.sub15:vreg_512 = COPY $vgpr15
   ; GFX9:   %18.sub14:vreg_512 = COPY $vgpr14
   ; GFX9:   %18.sub13:vreg_512 = COPY $vgpr13
   ; GFX9:   %18.sub12:vreg_512 = COPY $vgpr12
   ; GFX9:   %18.sub11:vreg_512 = COPY $vgpr11
   ; GFX9:   %18.sub10:vreg_512 = COPY $vgpr10
   ; GFX9:   %18.sub9:vreg_512 = COPY $vgpr9
body:             |
   ; GFX9:   [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64 = S_AND_SAVEEXEC_B64 [[V_CMP_EQ_U32_e64_]], implicit-def $exec, implicit-def dead $scc, implicit $exec
   ; GFX9:   S_SET_GPR_IDX_ON [[V_READFIRSTLANE_B32_]], 1, implicit-def $m0, implicit-def $mode, implicit undef $m0, implicit $mode
   ; GFX9:   [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 undef %18.sub0, implicit $exec, implicit %18, implicit $m0
   ; GFX9:   S_SET_GPR_IDX_OFF implicit-def $mode, implicit $mode
   ; GFX9:   $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def dead $scc
   ; GFX9:   S_CBRANCH_EXECNZ %bb.1, implicit $exec
   ; GFX9: bb.2:
   ; GFX9:   $exec = S_MOV_B64 [[S_MOV_B64_]]
-  ; GFX9:   $sgpr30_sgpr31 = COPY [[COPY]]
   ; GFX9:   $vgpr0 = COPY [[V_MOV_B32_e32_]]
-  ; GFX9:   S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0, implicit undef $vgpr1, implicit undef $vgpr2, implicit undef $vgpr3
+  ; GFX9:   S_SETPC_B64_return killed $sgpr30_sgpr31, implicit $vgpr0, implicit undef $vgpr1, implicit undef $vgpr2, implicit undef $vgpr3
   bb.0:
     successors: %bb.1
     liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7, $vgpr8, $vgpr9, $vgpr10, $vgpr11, $vgpr12, $vgpr13, $vgpr14, $vgpr15, $vgpr16, $sgpr30_sgpr31
 
     %17:sreg_64 = COPY killed $sgpr30_sgpr31
     %16:vgpr_32 = COPY killed $vgpr16
     %15:vgpr_32 = COPY killed $vgpr15
     %14:vgpr_32 = COPY killed $vgpr14

llvm/test/CodeGen/AMDGPU/no-source-locations-in-prologue.ll

 ; CHECK-NEXT:    s_load_dwordx2 s[4:5], s[4:5], 0x0
 ; 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_waitcnt lgkmcnt(0)
 ; CHECK-NEXT:    s_swappc_b64 s[30:31], s[4:5]
 ; CHECK-NEXT:  .Ltmp1:
+; CHECK-NEXT:    .loc 0 32 1 ; lane-info.cpp:32:1
 ; CHECK-NEXT:    v_readlane_b32 s30, v40, 0
 ; CHECK-NEXT:    v_readlane_b32 s31, v40, 1
 ; CHECK-NEXT:    s_add_i32 s32, s32, 0xfffffc00
 ; CHECK-NEXT:    v_readlane_b32 s33, v40, 2
 ; CHECK-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; CHECK-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
 ; CHECK-NEXT:    s_mov_b64 exec, s[4:5]
-; CHECK-NEXT:    .loc 0 32 1 ; lane-info.cpp:32:1
 ; CHECK-NEXT:    s_waitcnt vmcnt(0)
 ; CHECK-NEXT:    s_setpc_b64 s[30:31]
 ; CHECK-NEXT:  .Ltmp2:
 entry:
   call void @_ZL13sleep_foreverv(), !dbg !1646
   ret void, !dbg !1647
 }
 

llvm/test/CodeGen/AMDGPU/reserve-vgpr-for-sgpr-spill.ll

define void @reserve_lowest_available_vgpr() #0 {
   ,~{v250},~{v251},~{v252},~{v253}" () #0
   call void @child_function()
   ret void
 }
 
 ; GCN-LABEL: {{^}}reserve_vgpr_with_sgpr_spills:
 ; GCN-NOT:  buffer_store_dword v255, off, s[0:3], s32
 ; GCN: ; def s4
-; GCN: v_writelane_b32 v254, s4, 2
-; GCN: v_readlane_b32 s4, v254, 2
+; GCN: v_writelane_b32 v254, s4, 0
+; GCN: v_readlane_b32 s4, v254, 0
 ; GCN: ; use s4
 
 define void @reserve_vgpr_with_sgpr_spills() #0 {
   %alloca = alloca i32, align 4, addrspace(5)
   store volatile i32 0, i32 addrspace(5)* %alloca
 
   call void asm sideeffect "",
   "~{v0},~{v1},~{v2},~{v3},~{v4},~{v5},~{v6},~{v7},~{v8},~{v9}

llvm/test/CodeGen/AMDGPU/save-fp.ll

 ; RUN: llc -march=amdgcn -mcpu=gfx908 -verify-machineinstrs < %s | FileCheck -check-prefixes=GCN,GFX908 %s
 ; RUN: llc -march=amdgcn -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck -check-prefixes=GCN,GFX900 %s
 
 define void @foo() {
 bb:
   ret void
 }
 
 ; FIXME: We spill v40 into AGPR, but still save and restore FP
 ; which is not needed in this case.
 
 ; GCN-LABEL: {{^}}caller:
 
-; GFX900:     s_mov_b32 [[SAVED_FP:s[0-9]+]], s33
-; GFX900:     s_mov_b32 s33, s32
-; GFX908-NOT: s_mov_b32 s33, s32
+; GCN:     v_writelane_b32 v2, s33, 2
+; GCN:     s_mov_b32 s33, s32
 ; GFX900:     buffer_store_dword
-; GFX908-DAG: s_mov_b32 [[SAVED_FP:s[0-9]+]], s33
 ; GFX908-DAG: v_accvgpr_write_b32
 ; GCN:        s_swappc_b64
 ; GFX900:     buffer_load_dword
 ; GFX908:     v_accvgpr_read_b32
-; GCN:        s_mov_b32 s33, [[SAVED_FP]]
+; GCN:        v_readlane_b32 s33, v2, 2
 define i64 @caller() {
 bb:
   call void asm sideeffect "", "~{v40}" ()
   tail call void @foo()
   ret i64 0
 }

llvm/test/CodeGen/AMDGPU/sibling-call.ll

entry:
   ret i32 %ret
 }
 
 ; Have another non-tail in the function
 ; GCN-LABEL: {{^}}sibling_call_i32_fastcc_i32_i32_other_call:
 ; GCN: s_or_saveexec_b64 s{{\[[0-9]+:[0-9]+\]}}, -1
 ; GCN-NEXT: buffer_store_dword [[CSRV:v[0-9]+]], off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
 ; GCN-NEXT: s_mov_b64 exec
-; GCN: s_mov_b32 s33, s32
+; GCN: v_writelane_b32 [[CSRV]], s33, 2
 ; GCN-DAG: s_addk_i32 s32, 0x400
 
-; GCN-DAG: buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
-; GCN-DAG: buffer_store_dword v42, off, s[0:3], s33 ; 4-byte Folded Spill
-; GCN-DAG: v_writelane_b32 [[CSRV]], s34, 0
-; GCN-DAG: v_writelane_b32 [[CSRV]], s35, 1
-
 ; GCN-DAG: s_getpc_b64 s[4:5]
 ; GCN-DAG: s_add_u32 s4, s4, i32_fastcc_i32_i32@gotpcrel32@lo+4
 ; GCN-DAG: s_addc_u32 s5, s5, i32_fastcc_i32_i32@gotpcrel32@hi+12
 
+; GCN-DAG: v_writelane_b32 [[CSRV]], s30, 0
+; GCN-DAG: buffer_store_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
+; GCN-DAG: buffer_store_dword v42, off, s[0:3], s33 ; 4-byte Folded Spill
+; GCN-DAG: v_writelane_b32 [[CSRV]], s31, 1
+
 
 ; GCN: s_swappc_b64
 
 ; GCN-DAG: buffer_load_dword v42, off, s[0:3], s33 ; 4-byte Folded Reload
 ; GCN-DAG: buffer_load_dword v41, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
 
 ; GCN: s_getpc_b64 s[4:5]
 ; GCN-NEXT: s_add_u32 s4, s4, sibling_call_i32_fastcc_i32_i32@rel32@lo+4
 ; GCN-NEXT: s_addc_u32 s5, s5, sibling_call_i32_fastcc_i32_i32@rel32@hi+12
 
-; GCN-DAG: v_readlane_b32 s34, [[CSRV]], 0
-; GCN-DAG: v_readlane_b32 s35, [[CSRV]], 1
+; GCN-DAG: v_readlane_b32 s30, [[CSRV]], 0
+; GCN-DAG: v_readlane_b32 s31, [[CSRV]], 1
 
 ; GCN: s_addk_i32 s32, 0xfc00
 ; GCN-NEXT: v_readlane_b32 s33,
 ; GCN-NEXT: s_or_saveexec_b64 s[6:7], -1
 ; GCN-NEXT: buffer_load_dword [[CSRV]], off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
 ; GCN-NEXT: s_mov_b64 exec, s[6:7]
 ; GCN-NEXT: s_setpc_b64 s[4:5]
 define fastcc i32 @sibling_call_i32_fastcc_i32_i32_other_call(i32 %a, i32 %b, i32 %c) #1 {

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

 ; GCN: v_mov_b32_e32 v32, 0
 ; GCN: buffer_store_dword v32, off, s[0:3], s33 offset:1024
 ; GCN-NEXT: s_waitcnt vmcnt(0)
 ; GCN-NEXT: buffer_load_dword v{{[0-9]+}}, off, s[0:3], s34
 ; GCN-DAG: s_add_i32 s32, s32, 0x30000
 ; GCN: buffer_store_dword v{{[0-9]+}}, off, s[0:3], s32
 ; GCN: s_swappc_b64 s[30:31], s[4:5]
 
+; GCN: v_readlane_b32 s30, [[VGPR_REG]], 0
+; GCN-NEXT: v_readlane_b32 s31, [[VGPR_REG]], 1
 ; GCN: s_add_i32 s32, s32, 0xfffd0000
 ; GCN-NEXT: v_readlane_b32 s33, [[VGPR_REG]], 2
 ; GCN-NEXT: v_readlane_b32 s34, [[VGPR_REG]], 3
-; GCN-NEXT: s_or_saveexec_b64 s[6:7], -1
+; GCN-NEXT: s_or_saveexec_b64 s[4:5], -1
 ; GCN-NEXT: buffer_load_dword [[VGPR_REG]], off, s[0:3], s32 offset:1028 ; 4-byte Folded Reload
-; GCN-NEXT: s_mov_b64 exec, s[6:7]
+; GCN-NEXT: s_mov_b64 exec, s[4:5]
+; GCN: s_setpc_b64 s[30:31]
   %temp = alloca i32, align 1024, addrspace(5)
   store volatile i32 0, i32 addrspace(5)* %temp, align 1024
   call void @extern_func(<32 x i32> %a, i32 %b)
   ret void
 }
 
 %struct.Data = type { [9 x i32] }
 define i32 @needs_align1024_stack_args_used_inside_loop(%struct.Data addrspace(5)* nocapture readonly byval(%struct.Data) align 8 %arg) local_unnamed_addr #4 {

llvm/test/CodeGen/AMDGPU/strict_fadd.f16.ll

 ; GFX8-NEXT:    v_or_b32_e32 v0, v0, v5
 ; GFX8-NEXT:    v_or_b32_e32 v1, v1, v4
 ; GFX8-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: v_constained_fadd_v4f16_fpexcept_strict:
 ; GFX10:       ; %bb.0:
 ; GFX10-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
-; GFX10-NEXT:    v_add_f16_e32 v4, v0, v2
-; GFX10-NEXT:    v_mov_b32_e32 v5, 0xffff
+; GFX10-NEXT:    v_mov_b32_e32 v4, 0xffff
+; GFX10-NEXT:    v_add_f16_e32 v5, v0, v2
 ; GFX10-NEXT:    v_add_f16_e32 v6, v1, v3
 ; GFX10-NEXT:    v_add_f16_sdwa v1, v1, v3 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
 ; GFX10-NEXT:    v_add_f16_sdwa v0, v0, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
-; GFX10-NEXT:    v_and_b32_e32 v2, v5, v4
-; GFX10-NEXT:    v_and_b32_e32 v3, v5, v6
+; GFX10-NEXT:    v_and_b32_e32 v2, v4, v5
+; GFX10-NEXT:    v_and_b32_e32 v3, v4, v6
 ; GFX10-NEXT:    v_lshl_or_b32 v0, v0, 16, v2
 ; GFX10-NEXT:    v_lshl_or_b32 v1, v1, 16, v3
 ; GFX10-NEXT:    s_setpc_b64 s[30:31]
   %val = call <4 x half> @llvm.experimental.constrained.fadd.v4f16(<4 x half> %x, <4 x half> %y, metadata !"round.tonearest", metadata !"fpexcept.strict")
   ret <4 x half> %val
 }
 
 define amdgpu_ps half @s_constained_fadd_f16_fpexcept_strict(half inreg %x, half inreg %y) #0 {

llvm/test/CodeGen/AMDGPU/strict_fmul.f16.ll

 ; GFX8-NEXT:    v_or_b32_e32 v0, v0, v5
 ; GFX8-NEXT:    v_or_b32_e32 v1, v1, v4
 ; GFX8-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: v_constained_fmul_v4f16_fpexcept_strict:
 ; GFX10:       ; %bb.0:
 ; GFX10-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
-; GFX10-NEXT:    v_mul_f16_e32 v4, v0, v2
-; GFX10-NEXT:    v_mov_b32_e32 v5, 0xffff
+; GFX10-NEXT:    v_mov_b32_e32 v4, 0xffff
+; GFX10-NEXT:    v_mul_f16_e32 v5, v0, v2
 ; GFX10-NEXT:    v_mul_f16_e32 v6, v1, v3
 ; GFX10-NEXT:    v_mul_f16_sdwa v1, v1, v3 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
 ; GFX10-NEXT:    v_mul_f16_sdwa v0, v0, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
-; GFX10-NEXT:    v_and_b32_e32 v2, v5, v4
-; GFX10-NEXT:    v_and_b32_e32 v3, v5, v6
+; GFX10-NEXT:    v_and_b32_e32 v2, v4, v5
+; GFX10-NEXT:    v_and_b32_e32 v3, v4, v6
 ; GFX10-NEXT:    v_lshl_or_b32 v0, v0, 16, v2
 ; GFX10-NEXT:    v_lshl_or_b32 v1, v1, 16, v3
 ; GFX10-NEXT:    s_setpc_b64 s[30:31]
   %val = call <4 x half> @llvm.experimental.constrained.fmul.v4f16(<4 x half> %x, <4 x half> %y, metadata !"round.tonearest", metadata !"fpexcept.strict")
   ret <4 x half> %val
 }
 
 define amdgpu_ps half @s_constained_fmul_f16_fpexcept_strict(half inreg %x, half inreg %y) #0 {

llvm/test/CodeGen/AMDGPU/strict_fsub.f16.ll

 ; GFX8-NEXT:    v_or_b32_e32 v0, v0, v5
 ; GFX8-NEXT:    v_or_b32_e32 v1, v1, v4
 ; GFX8-NEXT:    s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: v_constained_fsub_v4f16_fpexcept_strict:
 ; GFX10:       ; %bb.0:
 ; GFX10-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
-; GFX10-NEXT:    v_sub_f16_e32 v4, v0, v2
-; GFX10-NEXT:    v_mov_b32_e32 v5, 0xffff
+; GFX10-NEXT:    v_mov_b32_e32 v4, 0xffff
+; GFX10-NEXT:    v_sub_f16_e32 v5, v0, v2
 ; GFX10-NEXT:    v_sub_f16_e32 v6, v1, v3
 ; GFX10-NEXT:    v_sub_f16_sdwa v1, v1, v3 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
 ; GFX10-NEXT:    v_sub_f16_sdwa v0, v0, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
-; GFX10-NEXT:    v_and_b32_e32 v2, v5, v4
-; GFX10-NEXT:    v_and_b32_e32 v3, v5, v6
+; GFX10-NEXT:    v_and_b32_e32 v2, v4, v5
+; GFX10-NEXT:    v_and_b32_e32 v3, v4, v6
 ; GFX10-NEXT:    v_lshl_or_b32 v0, v0, 16, v2
 ; GFX10-NEXT:    v_lshl_or_b32 v1, v1, 16, v3
 ; GFX10-NEXT:    s_setpc_b64 s[30:31]
   %val = call <4 x half> @llvm.experimental.constrained.fsub.v4f16(<4 x half> %x, <4 x half> %y, metadata !"round.tonearest", metadata !"fpexcept.strict")
   ret <4 x half> %val
 }
 
 define amdgpu_ps half @s_constained_fsub_f16_fpexcept_strict(half inreg %x, half inreg %y) #0 {

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

 
 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:    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, s33, 1
+; GCN-NEXT:    v_writelane_b32 v1, s33, 3
+; GCN-NEXT:    v_writelane_b32 v1, s4, 0
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
-; GCN-NEXT:    v_writelane_b32 v1, s4, 0
+; GCN-NEXT:    v_writelane_b32 v1, s30, 1
 ; GCN-NEXT:    v_add_f32_e32 v0, 1.0, v0
 ; GCN-NEXT:    s_mov_b32 s4, 2.0
-; GCN-NEXT:    s_mov_b64 s[36:37], s[30:31]
-; GCN-NEXT:    s_getpc_b64 s[30:31]
-; GCN-NEXT:    s_add_u32 s30, s30, callee@rel32@lo+4
-; GCN-NEXT:    s_addc_u32 s31, s31, callee@rel32@hi+12
-; GCN-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; 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 s30, v1, 1
+; GCN-NEXT:    v_readlane_b32 s31, v1, 2
 ; GCN-NEXT:    v_readlane_b32 s4, v1, 0
 ; GCN-NEXT:    s_addk_i32 s32, 0xfc00
-; GCN-NEXT:    v_readlane_b32 s33, v1, 1
-; GCN-NEXT:    s_or_saveexec_b64 s[30:31], -1
+; GCN-NEXT:    v_readlane_b32 s33, v1, 3
+; GCN-NEXT:    s_or_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[30:31]
+; GCN-NEXT:    s_mov_b64 exec, s[34:35]
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
-; GCN-NEXT:    s_setpc_b64 s[36:37]
+; 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/udiv.ll

 define i64 @v_test_udiv64_mulhi_fold(i64 %arg) {
 ; SI-LABEL: v_test_udiv64_mulhi_fold:
 ; SI:       ; %bb.0:
 ; SI-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; SI-NEXT:    v_mov_b32_e32 v2, 0x4f800000
 ; SI-NEXT:    v_madak_f32 v2, 0, v2, 0x47c35000
 ; SI-NEXT:    v_rcp_f32_e32 v2, v2
 ; SI-NEXT:    s_mov_b32 s4, 0xfffe7960
-; SI-NEXT:    v_mov_b32_e32 v9, 0
+; SI-NEXT:    v_mov_b32_e32 v8, 0
 ; SI-NEXT:    v_mul_f32_e32 v2, 0x5f7ffffc, v2
 ; SI-NEXT:    v_mul_f32_e32 v3, 0x2f800000, v2
 ; SI-NEXT:    v_trunc_f32_e32 v3, v3
 ; SI-NEXT:    v_mac_f32_e32 v2, 0xcf800000, v3
 ; SI-NEXT:    v_cvt_u32_f32_e32 v2, v2
 ; SI-NEXT:    v_cvt_u32_f32_e32 v3, v3
 ; SI-NEXT:    v_mul_hi_u32 v4, v2, s4
-; SI-NEXT:    v_mul_lo_u32 v5, v3, s4
-; SI-NEXT:    v_mul_lo_u32 v6, v2, s4
+; SI-NEXT:    v_mul_lo_u32 v6, v3, s4
+; SI-NEXT:    v_mul_lo_u32 v5, v2, s4
 ; SI-NEXT:    v_subrev_i32_e32 v4, vcc, v2, v4
-; SI-NEXT:    v_add_i32_e32 v4, vcc, v5, v4
-; SI-NEXT:    v_mul_lo_u32 v5, v2, v4
-; SI-NEXT:    v_mul_hi_u32 v7, v2, v6
-; SI-NEXT:    v_mul_hi_u32 v8, v2, v4
+; SI-NEXT:    v_add_i32_e32 v4, vcc, v6, v4
+; SI-NEXT:    v_mul_hi_u32 v7, v2, v5
+; SI-NEXT:    v_mul_lo_u32 v6, v2, v4
+; SI-NEXT:    v_mul_hi_u32 v9, v2, v4
 ; SI-NEXT:    v_mul_hi_u32 v10, v3, v4
 ; SI-NEXT:    v_mul_lo_u32 v4, v3, v4
-; SI-NEXT:    v_add_i32_e32 v5, vcc, v7, v5
-; SI-NEXT:    v_addc_u32_e32 v7, vcc, 0, v8, vcc
-; SI-NEXT:    v_mul_lo_u32 v8, v3, v6
-; SI-NEXT:    v_mul_hi_u32 v6, v3, v6
-; SI-NEXT:    v_add_i32_e32 v5, vcc, v5, v8
-; SI-NEXT:    v_addc_u32_e32 v5, vcc, v7, v6, vcc
-; SI-NEXT:    v_addc_u32_e32 v6, vcc, v10, v9, vcc
+; SI-NEXT:    v_add_i32_e32 v6, vcc, v7, v6
+; SI-NEXT:    v_addc_u32_e32 v7, vcc, 0, v9, vcc
+; SI-NEXT:    v_mul_lo_u32 v9, v3, v5
+; SI-NEXT:    v_mul_hi_u32 v5, v3, v5
+; SI-NEXT:    v_add_i32_e32 v6, vcc, v6, v9
+; SI-NEXT:    v_addc_u32_e32 v5, vcc, v7, v5, vcc
+; SI-NEXT:    v_addc_u32_e32 v6, vcc, v10, v8, vcc
 ; SI-NEXT:    v_add_i32_e32 v4, vcc, v5, v4
 ; SI-NEXT:    v_addc_u32_e32 v5, vcc, 0, v6, vcc
 ; SI-NEXT:    v_add_i32_e32 v2, vcc, v2, v4
 ; SI-NEXT:    v_addc_u32_e32 v3, vcc, v3, v5, vcc
 ; SI-NEXT:    v_mul_hi_u32 v4, v2, s4
 ; SI-NEXT:    v_mul_lo_u32 v5, v3, s4
 ; SI-NEXT:    v_mul_lo_u32 v6, v2, s4
 ; SI-NEXT:    s_mov_b32 s4, 0x186a0
 ; SI-NEXT:    v_subrev_i32_e32 v4, vcc, v2, v4
 ; SI-NEXT:    v_add_i32_e32 v4, vcc, v4, v5
 ; SI-NEXT:    v_mul_lo_u32 v5, v2, v4
 ; SI-NEXT:    v_mul_hi_u32 v7, v2, v6
-; SI-NEXT:    v_mul_hi_u32 v8, v2, v4
+; SI-NEXT:    v_mul_hi_u32 v9, v2, v4
 ; SI-NEXT:    v_mul_hi_u32 v10, v3, v4
 ; SI-NEXT:    v_mul_lo_u32 v4, v3, v4
 ; SI-NEXT:    v_add_i32_e32 v5, vcc, v7, v5
-; SI-NEXT:    v_addc_u32_e32 v7, vcc, 0, v8, vcc
-; SI-NEXT:    v_mul_lo_u32 v8, v3, v6
+; SI-NEXT:    v_addc_u32_e32 v7, vcc, 0, v9, vcc
+; SI-NEXT:    v_mul_lo_u32 v9, v3, v6
 ; SI-NEXT:    v_mul_hi_u32 v6, v3, v6
-; SI-NEXT:    v_add_i32_e32 v5, vcc, v5, v8
+; SI-NEXT:    v_add_i32_e32 v5, vcc, v5, v9
 ; SI-NEXT:    v_addc_u32_e32 v5, vcc, v7, v6, vcc
-; SI-NEXT:    v_addc_u32_e32 v6, vcc, v10, v9, vcc
+; SI-NEXT:    v_addc_u32_e32 v6, vcc, v10, v8, vcc
 ; SI-NEXT:    v_add_i32_e32 v4, vcc, v5, v4
 ; SI-NEXT:    v_addc_u32_e32 v5, vcc, 0, v6, vcc
 ; SI-NEXT:    v_add_i32_e32 v2, vcc, v2, v4
 ; SI-NEXT:    v_addc_u32_e32 v3, vcc, v3, v5, vcc
 ; SI-NEXT:    v_mul_lo_u32 v4, v0, v3
 ; SI-NEXT:    v_mul_hi_u32 v5, v0, v2
 ; SI-NEXT:    v_mul_hi_u32 v6, v0, v3
 ; SI-NEXT:    v_mul_hi_u32 v7, v1, v3
 ; SI-NEXT:    v_mul_lo_u32 v3, v1, v3
 ; SI-NEXT:    v_add_i32_e32 v4, vcc, v5, v4
 ; SI-NEXT:    v_addc_u32_e32 v5, vcc, 0, v6, vcc
 ; SI-NEXT:    v_mul_lo_u32 v6, v1, v2
 ; SI-NEXT:    v_mul_hi_u32 v2, v1, v2
 ; SI-NEXT:    v_add_i32_e32 v4, vcc, v4, v6
 ; SI-NEXT:    v_addc_u32_e32 v2, vcc, v5, v2, vcc
-; SI-NEXT:    v_addc_u32_e32 v4, vcc, v7, v9, vcc
+; SI-NEXT:    v_addc_u32_e32 v4, vcc, v7, v8, vcc
 ; SI-NEXT:    v_add_i32_e32 v2, vcc, v2, v3
 ; SI-NEXT:    v_addc_u32_e32 v3, vcc, 0, v4, vcc
 ; SI-NEXT:    v_mul_lo_u32 v4, v3, s4
 ; SI-NEXT:    v_mul_hi_u32 v5, v2, s4
 ; SI-NEXT:    v_mul_lo_u32 v6, v2, s4
 ; SI-NEXT:    v_add_i32_e32 v4, vcc, v5, v4
 ; SI-NEXT:    v_sub_i32_e32 v0, vcc, v0, v6
 ; SI-NEXT:    v_subb_u32_e32 v1, vcc, v1, v4, vcc

 ; VI-NEXT:    v_mul_f32_e32 v3, 0x2f800000, v2
 ; VI-NEXT:    v_trunc_f32_e32 v3, v3
 ; VI-NEXT:    v_mac_f32_e32 v2, 0xcf800000, v3
 ; VI-NEXT:    v_cvt_u32_f32_e32 v6, v2
 ; VI-NEXT:    v_cvt_u32_f32_e32 v7, v3
 ; VI-NEXT:    v_mad_u64_u32 v[2:3], s[4:5], v6, s6, 0
 ; VI-NEXT:    v_mul_lo_u32 v4, v7, s6
 ; VI-NEXT:    v_subrev_u32_e32 v3, vcc, v6, v3
-; VI-NEXT:    v_add_u32_e32 v5, vcc, v4, v3
-; VI-NEXT:    v_mad_u64_u32 v[3:4], s[4:5], v6, v5, 0
-; VI-NEXT:    v_mul_hi_u32 v8, v6, v2
-; VI-NEXT:    v_add_u32_e32 v8, vcc, v8, v3
+; VI-NEXT:    v_add_u32_e32 v8, vcc, v4, v3
+; VI-NEXT:    v_mul_hi_u32 v5, v6, v2
+; VI-NEXT:    v_mad_u64_u32 v[3:4], s[4:5], v6, v8, 0
+; VI-NEXT:    v_add_u32_e32 v10, vcc, v5, v3
 ; VI-NEXT:    v_mad_u64_u32 v[2:3], s[4:5], v7, v2, 0
-; VI-NEXT:    v_addc_u32_e32 v10, vcc, 0, v4, vcc
-; VI-NEXT:    v_mad_u64_u32 v[4:5], s[4:5], v7, v5, 0
-; VI-NEXT:    v_add_u32_e32 v2, vcc, v8, v2
-; VI-NEXT:    v_addc_u32_e32 v2, vcc, v10, v3, vcc
+; VI-NEXT:    v_addc_u32_e32 v11, vcc, 0, v4, vcc
+; VI-NEXT:    v_mad_u64_u32 v[4:5], s[4:5], v7, v8, 0
+; VI-NEXT:    v_add_u32_e32 v2, vcc, v10, v2
+; VI-NEXT:    v_addc_u32_e32 v2, vcc, v11, v3, vcc
 ; VI-NEXT:    v_addc_u32_e32 v3, vcc, v5, v9, vcc
 ; VI-NEXT:    v_add_u32_e32 v2, vcc, v2, v4
 ; VI-NEXT:    v_addc_u32_e32 v3, vcc, 0, v3, vcc
 ; VI-NEXT:    v_add_u32_e32 v6, vcc, v6, v2
 ; VI-NEXT:    v_addc_u32_e32 v7, vcc, v7, v3, vcc
 ; VI-NEXT:    v_mad_u64_u32 v[2:3], s[4:5], v6, s6, 0
 ; VI-NEXT:    v_mul_lo_u32 v4, v7, s6
 ; VI-NEXT:    s_mov_b32 s6, 0x186a0

 ; GCN-NEXT:    v_mul_f32_e32 v3, 0x2f800000, v2
 ; GCN-NEXT:    v_trunc_f32_e32 v3, v3
 ; GCN-NEXT:    v_mac_f32_e32 v2, 0xcf800000, v3
 ; GCN-NEXT:    v_cvt_u32_f32_e32 v6, v2
 ; GCN-NEXT:    v_cvt_u32_f32_e32 v7, v3
 ; GCN-NEXT:    v_mad_u64_u32 v[2:3], s[4:5], v6, s6, 0
 ; GCN-NEXT:    v_mul_lo_u32 v4, v7, s6
 ; GCN-NEXT:    v_subrev_u32_e32 v3, vcc, v6, v3
-; GCN-NEXT:    v_add_u32_e32 v5, vcc, v4, v3
-; GCN-NEXT:    v_mad_u64_u32 v[3:4], s[4:5], v6, v5, 0
-; GCN-NEXT:    v_mul_hi_u32 v8, v6, v2
-; GCN-NEXT:    v_add_u32_e32 v8, vcc, v8, v3
+; GCN-NEXT:    v_add_u32_e32 v8, vcc, v4, v3
+; GCN-NEXT:    v_mul_hi_u32 v5, v6, v2
+; GCN-NEXT:    v_mad_u64_u32 v[3:4], s[4:5], v6, v8, 0
+; GCN-NEXT:    v_add_u32_e32 v10, vcc, v5, v3
 ; GCN-NEXT:    v_mad_u64_u32 v[2:3], s[4:5], v7, v2, 0
-; GCN-NEXT:    v_addc_u32_e32 v10, vcc, 0, v4, vcc
-; GCN-NEXT:    v_mad_u64_u32 v[4:5], s[4:5], v7, v5, 0
-; GCN-NEXT:    v_add_u32_e32 v2, vcc, v8, v2
-; GCN-NEXT:    v_addc_u32_e32 v2, vcc, v10, v3, vcc
+; GCN-NEXT:    v_addc_u32_e32 v11, vcc, 0, v4, vcc
+; GCN-NEXT:    v_mad_u64_u32 v[4:5], s[4:5], v7, v8, 0
+; GCN-NEXT:    v_add_u32_e32 v2, vcc, v10, v2
+; GCN-NEXT:    v_addc_u32_e32 v2, vcc, v11, v3, vcc
 ; GCN-NEXT:    v_addc_u32_e32 v3, vcc, v5, v9, vcc
 ; GCN-NEXT:    v_add_u32_e32 v2, vcc, v2, v4
 ; GCN-NEXT:    v_addc_u32_e32 v3, vcc, 0, v3, vcc
 ; GCN-NEXT:    v_add_u32_e32 v6, vcc, v6, v2
 ; GCN-NEXT:    v_addc_u32_e32 v7, vcc, v7, v3, vcc
 ; GCN-NEXT:    v_mad_u64_u32 v[2:3], s[4:5], v6, s6, 0
 ; GCN-NEXT:    v_mul_lo_u32 v4, v7, s6
 ; GCN-NEXT:    s_mov_b32 s6, 0x186a0

llvm/test/CodeGen/AMDGPU/udiv64.ll

 ; GCN-NEXT:    v_rcp_f32_e32 v4, v4
 ; GCN-NEXT:    v_mov_b32_e32 v13, 0
 ; GCN-NEXT:    v_mul_f32_e32 v4, 0x5f7ffffc, v4
 ; GCN-NEXT:    v_mul_f32_e32 v5, 0x2f800000, v4
 ; GCN-NEXT:    v_trunc_f32_e32 v5, v5
 ; GCN-NEXT:    v_mac_f32_e32 v4, 0xcf800000, v5
 ; GCN-NEXT:    v_cvt_u32_f32_e32 v5, v5
 ; GCN-NEXT:    v_cvt_u32_f32_e32 v4, v4
-; GCN-NEXT:    v_mul_lo_u32 v9, v6, v5
-; GCN-NEXT:    v_mul_hi_u32 v8, v6, v4
+; GCN-NEXT:    v_mul_lo_u32 v8, v6, v5
+; GCN-NEXT:    v_mul_hi_u32 v9, v6, v4
 ; GCN-NEXT:    v_mul_lo_u32 v10, v7, v4
-; GCN-NEXT:    v_add_i32_e32 v8, vcc, v8, v9
+; GCN-NEXT:    v_add_i32_e32 v8, vcc, v9, v8
 ; GCN-NEXT:    v_mul_lo_u32 v9, v6, v4
 ; GCN-NEXT:    v_add_i32_e32 v8, vcc, v8, v10
 ; GCN-NEXT:    v_mul_lo_u32 v10, v4, v8
 ; GCN-NEXT:    v_mul_hi_u32 v11, v4, v9
 ; GCN-NEXT:    v_mul_hi_u32 v12, v4, v8
 ; GCN-NEXT:    v_mul_hi_u32 v14, v5, v8
 ; GCN-NEXT:    v_mul_lo_u32 v8, v5, v8
 ; GCN-NEXT:    v_add_i32_e32 v10, vcc, v11, v10

 define i64 @v_test_udiv_k_den_i64(i64 %x) {
 ; GCN-LABEL: v_test_udiv_k_den_i64:
 ; GCN:       ; %bb.0:
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GCN-NEXT:    v_mov_b32_e32 v2, 0x4f800000
 ; GCN-NEXT:    v_madak_f32 v2, 0, v2, 0x41c00000
 ; GCN-NEXT:    v_rcp_f32_e32 v2, v2
 ; GCN-NEXT:    s_movk_i32 s4, 0xffe8
-; GCN-NEXT:    v_mov_b32_e32 v9, 0
+; GCN-NEXT:    v_mov_b32_e32 v8, 0
 ; GCN-NEXT:    v_mul_f32_e32 v2, 0x5f7ffffc, v2
 ; GCN-NEXT:    v_mul_f32_e32 v3, 0x2f800000, v2
 ; GCN-NEXT:    v_trunc_f32_e32 v3, v3
 ; GCN-NEXT:    v_mac_f32_e32 v2, 0xcf800000, v3
 ; GCN-NEXT:    v_cvt_u32_f32_e32 v2, v2
 ; GCN-NEXT:    v_cvt_u32_f32_e32 v3, v3
 ; GCN-NEXT:    v_mul_hi_u32 v4, v2, s4
-; GCN-NEXT:    v_mul_lo_u32 v5, v3, s4
-; GCN-NEXT:    v_mul_lo_u32 v6, v2, s4
+; GCN-NEXT:    v_mul_lo_u32 v6, v3, s4
+; GCN-NEXT:    v_mul_lo_u32 v5, v2, s4
 ; GCN-NEXT:    v_subrev_i32_e32 v4, vcc, v2, v4
-; GCN-NEXT:    v_add_i32_e32 v4, vcc, v5, v4
-; GCN-NEXT:    v_mul_lo_u32 v5, v2, v4
-; GCN-NEXT:    v_mul_hi_u32 v7, v2, v6
-; GCN-NEXT:    v_mul_hi_u32 v8, v2, v4
+; GCN-NEXT:    v_add_i32_e32 v4, vcc, v6, v4
+; GCN-NEXT:    v_mul_hi_u32 v7, v2, v5
+; GCN-NEXT:    v_mul_lo_u32 v6, v2, v4
+; GCN-NEXT:    v_mul_hi_u32 v9, v2, v4
 ; GCN-NEXT:    v_mul_hi_u32 v10, v3, v4
 ; GCN-NEXT:    v_mul_lo_u32 v4, v3, v4
-; GCN-NEXT:    v_add_i32_e32 v5, vcc, v7, v5
-; GCN-NEXT:    v_addc_u32_e32 v7, vcc, 0, v8, vcc
-; GCN-NEXT:    v_mul_lo_u32 v8, v3, v6
-; GCN-NEXT:    v_mul_hi_u32 v6, v3, v6
-; GCN-NEXT:    v_add_i32_e32 v5, vcc, v5, v8
-; GCN-NEXT:    v_addc_u32_e32 v5, vcc, v7, v6, vcc
-; GCN-NEXT:    v_addc_u32_e32 v6, vcc, v10, v9, vcc
+; GCN-NEXT:    v_add_i32_e32 v6, vcc, v7, v6
+; GCN-NEXT:    v_addc_u32_e32 v7, vcc, 0, v9, vcc
+; GCN-NEXT:    v_mul_lo_u32 v9, v3, v5
+; GCN-NEXT:    v_mul_hi_u32 v5, v3, v5
+; GCN-NEXT:    v_add_i32_e32 v6, vcc, v6, v9
+; GCN-NEXT:    v_addc_u32_e32 v5, vcc, v7, v5, vcc
+; GCN-NEXT:    v_addc_u32_e32 v6, vcc, v10, v8, vcc
 ; GCN-NEXT:    v_add_i32_e32 v4, vcc, v5, v4
 ; GCN-NEXT:    v_addc_u32_e32 v5, vcc, 0, v6, vcc
 ; GCN-NEXT:    v_add_i32_e32 v2, vcc, v2, v4
 ; GCN-NEXT:    v_addc_u32_e32 v3, vcc, v3, v5, vcc
 ; GCN-NEXT:    v_mul_hi_u32 v4, v2, s4
 ; GCN-NEXT:    v_mul_lo_u32 v5, v3, s4
 ; GCN-NEXT:    v_mul_lo_u32 v6, v2, s4
 ; GCN-NEXT:    v_subrev_i32_e32 v4, vcc, v2, v4
 ; GCN-NEXT:    v_add_i32_e32 v4, vcc, v4, v5
 ; GCN-NEXT:    v_mul_lo_u32 v5, v2, v4
 ; GCN-NEXT:    v_mul_hi_u32 v7, v2, v6
-; GCN-NEXT:    v_mul_hi_u32 v8, v2, v4
+; GCN-NEXT:    v_mul_hi_u32 v9, v2, v4
 ; GCN-NEXT:    v_mul_hi_u32 v10, v3, v4
 ; GCN-NEXT:    v_mul_lo_u32 v4, v3, v4
 ; GCN-NEXT:    v_add_i32_e32 v5, vcc, v7, v5
-; GCN-NEXT:    v_addc_u32_e32 v7, vcc, 0, v8, vcc
-; GCN-NEXT:    v_mul_lo_u32 v8, v3, v6
+; GCN-NEXT:    v_addc_u32_e32 v7, vcc, 0, v9, vcc
+; GCN-NEXT:    v_mul_lo_u32 v9, v3, v6
 ; GCN-NEXT:    v_mul_hi_u32 v6, v3, v6
-; GCN-NEXT:    v_add_i32_e32 v5, vcc, v5, v8
+; GCN-NEXT:    v_add_i32_e32 v5, vcc, v5, v9
 ; GCN-NEXT:    v_addc_u32_e32 v5, vcc, v7, v6, vcc
-; GCN-NEXT:    v_addc_u32_e32 v6, vcc, v10, v9, vcc
+; GCN-NEXT:    v_addc_u32_e32 v6, vcc, v10, v8, vcc
 ; GCN-NEXT:    v_add_i32_e32 v4, vcc, v5, v4
 ; GCN-NEXT:    v_addc_u32_e32 v5, vcc, 0, v6, vcc
 ; GCN-NEXT:    v_add_i32_e32 v2, vcc, v2, v4
 ; GCN-NEXT:    v_addc_u32_e32 v3, vcc, v3, v5, vcc
 ; GCN-NEXT:    v_mul_lo_u32 v4, v0, v3
 ; GCN-NEXT:    v_mul_hi_u32 v5, v0, v2
 ; GCN-NEXT:    v_mul_hi_u32 v6, v0, v3
 ; GCN-NEXT:    v_mul_hi_u32 v7, v1, v3
 ; GCN-NEXT:    v_mul_lo_u32 v3, v1, v3
 ; GCN-NEXT:    v_add_i32_e32 v4, vcc, v5, v4
 ; GCN-NEXT:    v_addc_u32_e32 v5, vcc, 0, v6, vcc
 ; GCN-NEXT:    v_mul_lo_u32 v6, v1, v2
 ; GCN-NEXT:    v_mul_hi_u32 v2, v1, v2
 ; GCN-NEXT:    v_add_i32_e32 v4, vcc, v4, v6
 ; GCN-NEXT:    v_addc_u32_e32 v2, vcc, v5, v2, vcc
-; GCN-NEXT:    v_addc_u32_e32 v4, vcc, v7, v9, vcc
+; GCN-NEXT:    v_addc_u32_e32 v4, vcc, v7, v8, vcc
 ; GCN-NEXT:    v_add_i32_e32 v2, vcc, v2, v3
 ; GCN-NEXT:    v_addc_u32_e32 v3, vcc, 0, v4, vcc
 ; GCN-NEXT:    v_mul_lo_u32 v4, v3, 24
 ; GCN-NEXT:    v_mul_hi_u32 v5, v2, 24
 ; GCN-NEXT:    v_mul_lo_u32 v6, v2, 24
 ; GCN-NEXT:    v_add_i32_e32 v4, vcc, v5, v4
 ; GCN-NEXT:    v_sub_i32_e32 v0, vcc, v0, v6
 ; GCN-NEXT:    v_subb_u32_e32 v1, vcc, v1, v4, vcc

llvm/test/CodeGen/AMDGPU/unstructured-cfg-def-use-issue.ll

 ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
 ; RUN: llc -mtriple=amdgcn-amdhsa -verify-machineinstrs -simplifycfg-require-and-preserve-domtree=1 < %s | FileCheck -check-prefix=GCN %s
 ; RUN: opt -S -si-annotate-control-flow -mtriple=amdgcn-amdhsa -verify-machineinstrs -simplifycfg-require-and-preserve-domtree=1 < %s | FileCheck -check-prefix=SI-OPT %s
 
 define hidden void @widget() {
 ; GCN-LABEL: widget:
 ; GCN:       ; %bb.0: ; %bb
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    v_writelane_b32 v40, s33, 2
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; GCN-NEXT:    s_addk_i32 s32, 0x400
+; GCN-NEXT:    v_writelane_b32 v40, s30, 0
+; GCN-NEXT:    v_writelane_b32 v40, s31, 1
 ; GCN-NEXT:    v_mov_b32_e32 v0, 0
 ; GCN-NEXT:    v_mov_b32_e32 v1, 0
 ; GCN-NEXT:    flat_load_dword v0, v[0:1]
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
 ; GCN-NEXT:    v_cmp_gt_i32_e32 vcc, 21, v0
 ; GCN-NEXT:    s_and_b64 vcc, exec, vcc
-; GCN-NEXT:    v_writelane_b32 v40, s30, 0
-; GCN-NEXT:    v_writelane_b32 v40, s31, 1
 ; GCN-NEXT:    s_cbranch_vccz .LBB0_3
 ; GCN-NEXT:  ; %bb.1: ; %bb4
 ; GCN-NEXT:    v_cmp_ne_u32_e32 vcc, 9, v0
 ; GCN-NEXT:    s_and_b64 vcc, exec, vcc
 ; GCN-NEXT:    s_cbranch_vccnz .LBB0_4
 ; GCN-NEXT:  ; %bb.2: ; %bb7
 ; GCN-NEXT:    s_getpc_b64 s[4:5]
 ; GCN-NEXT:    s_add_u32 s4, s4, wibble@rel32@lo+4

 ; GCN-NEXT:  ; %bb.5: ; %bb9.bb12_crit_edge
 ; GCN-NEXT:    s_or_b64 exec, exec, s[4:5]
 ; GCN-NEXT:  .LBB0_6: ; %bb12
 ; GCN-NEXT:    v_mov_b32_e32 v2, 0
 ; GCN-NEXT:    v_mov_b32_e32 v0, 0
 ; GCN-NEXT:    v_mov_b32_e32 v1, 0
 ; GCN-NEXT:    flat_store_dword v[0:1], v2
 ; GCN-NEXT:  .LBB0_7: ; %UnifiedReturnBlock
-; GCN-NEXT:    v_readlane_b32 s4, v40, 0
-; GCN-NEXT:    v_readlane_b32 s5, v40, 1
+; GCN-NEXT:    v_readlane_b32 s30, v40, 0
+; GCN-NEXT:    v_readlane_b32 s31, v40, 1
 ; GCN-NEXT:    s_addk_i32 s32, 0xfc00
 ; GCN-NEXT:    v_readlane_b32 s33, v40, 2
-; GCN-NEXT:    s_or_saveexec_b64 s[6:7], -1
+; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload
-; GCN-NEXT:    s_mov_b64 exec, s[6:7]
+; GCN-NEXT:    s_mov_b64 exec, s[4:5]
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
-; GCN-NEXT:    s_setpc_b64 s[4:5]
+; GCN-NEXT:    s_setpc_b64 s[30:31]
 ; SI-OPT-LABEL: @widget(
 ; SI-OPT-NEXT:  bb:
 ; SI-OPT-NEXT:    [[TMP:%.*]] = load i32, i32 addrspace(1)* null, align 16
 ; SI-OPT-NEXT:    [[TMP1:%.*]] = icmp slt i32 [[TMP]], 21
 ; SI-OPT-NEXT:    br i1 [[TMP1]], label [[BB4:%.*]], label [[BB2:%.*]]
 ; SI-OPT:       bb2:
 ; SI-OPT-NEXT:    [[TMP3:%.*]] = icmp eq i32 [[TMP]], 21
 ; SI-OPT-NEXT:    br i1 [[TMP3]], label [[BB12:%.*]], label [[BB9:%.*]]

 ; SI-OPT-NEXT:    br label [[BB2]]
 ;
 ; GCN-LABEL: blam:
 ; GCN:       ; %bb.0: ; %bb
 ; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; GCN-NEXT:    s_or_saveexec_b64 s[4:5], -1
 ; GCN-NEXT:    buffer_store_dword v40, off, s[0:3], s32 offset:16 ; 4-byte Folded Spill
 ; GCN-NEXT:    s_mov_b64 exec, s[4:5]
-; GCN-NEXT:    v_writelane_b32 v40, s33, 4
+; GCN-NEXT:    v_writelane_b32 v40, s33, 6
 ; GCN-NEXT:    s_mov_b32 s33, s32
 ; GCN-NEXT:    s_addk_i32 s32, 0x800
 ; GCN-NEXT:    buffer_store_dword v41, off, s[0:3], s33 offset:12 ; 4-byte Folded Spill
 ; GCN-NEXT:    buffer_store_dword v42, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
 ; GCN-NEXT:    buffer_store_dword v43, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GCN-NEXT:    buffer_store_dword v44, off, s[0:3], s33 ; 4-byte Folded Spill
 ; GCN-NEXT:    v_writelane_b32 v40, s34, 0
 ; GCN-NEXT:    v_writelane_b32 v40, s35, 1
 ; GCN-NEXT:    v_writelane_b32 v40, s36, 2
 ; GCN-NEXT:    v_writelane_b32 v40, s37, 3
+; GCN-NEXT:    v_writelane_b32 v40, s30, 4
+; GCN-NEXT:    v_writelane_b32 v40, s31, 5
 ; GCN-NEXT:    s_mov_b64 s[4:5], 0
 ; GCN-NEXT:    v_mov_b32_e32 v1, 0
 ; GCN-NEXT:    v_mov_b32_e32 v2, 0
 ; GCN-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
 ; GCN-NEXT:    v_mov_b32_e32 v42, 0
 ; GCN-NEXT:    flat_load_dword v43, v[1:2]
 ; GCN-NEXT:    v_mov_b32_e32 v44, 0x7fc00000
 ; GCN-NEXT:    s_getpc_b64 s[36:37]

llvm/test/CodeGen/AMDGPU/urem-seteq-illegal-types.ll

 define <3 x i1> @test_urem_vec(<3 x i11> %X) nounwind {
 ; CHECK-LABEL: test_urem_vec:
 ; CHECK:       ; %bb.0:
 ; CHECK-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
 ; CHECK-NEXT:    s_movk_i32 s4, 0x7ff
 ; CHECK-NEXT:    s_mov_b32 s5, 0x8311eb33
 ; CHECK-NEXT:    s_mov_b32 s6, 0x20140c
 ; CHECK-NEXT:    s_mov_b32 s7, 0xb6db6db7
-; CHECK-NEXT:    s_mov_b32 s8, 0x49249249
-; CHECK-NEXT:    s_mov_b32 s9, 0x24924924
-; CHECK-NEXT:    s_mov_b32 s10, 0xaaaaaaab
-; CHECK-NEXT:    s_mov_b32 s11, 0x2aaaaaaa
+; CHECK-NEXT:    s_mov_b32 s8, 0x24924924
+; CHECK-NEXT:    s_mov_b32 s9, 0xaaaaaaab
+; CHECK-NEXT:    s_mov_b32 s10, 0x2aaaaaaa
 ; CHECK-NEXT:    v_and_b32_e32 v0, s4, v0
 ; CHECK-NEXT:    v_and_b32_e32 v1, s4, v1
 ; CHECK-NEXT:    v_and_b32_e32 v2, s4, v2
 ; CHECK-NEXT:    v_mul_lo_u32 v2, v2, s5
 ; CHECK-NEXT:    v_mul_lo_u32 v1, v1, s7
-; CHECK-NEXT:    v_mul_lo_u32 v0, v0, s10
+; CHECK-NEXT:    v_mul_lo_u32 v0, v0, s9
 ; CHECK-NEXT:    v_add_i32_e32 v2, vcc, 0xf9dc299a, v2
-; CHECK-NEXT:    v_add_i32_e32 v1, vcc, s8, v1
+; CHECK-NEXT:    v_add_i32_e32 v1, vcc, 0x49249249, v1
 ; CHECK-NEXT:    v_alignbit_b32 v0, v0, v0, 1
-; CHECK-NEXT:    v_cmp_lt_u32_e32 vcc, s11, v0
+; CHECK-NEXT:    v_cmp_lt_u32_e32 vcc, s10, v0
 ; CHECK-NEXT:    v_cndmask_b32_e64 v0, 0, 1, vcc
-; CHECK-NEXT:    v_cmp_lt_u32_e32 vcc, s9, v1
+; CHECK-NEXT:    v_cmp_lt_u32_e32 vcc, s8, v1
 ; CHECK-NEXT:    v_cndmask_b32_e64 v1, 0, 1, vcc
 ; CHECK-NEXT:    v_cmp_lt_u32_e32 vcc, s6, v2
 ; CHECK-NEXT:    v_cndmask_b32_e64 v2, 0, 1, vcc
 ; CHECK-NEXT:    s_setpc_b64 s[30:31]
   %urem = urem <3 x i11> %X, <i11 6, i11 7, i11 -5>
   %cmp = icmp ne <3 x i11> %urem, <i11 0, i11 1, i11 2>
   ret <3 x i1> %cmp
 }

llvm/test/CodeGen/AMDGPU/urem64.ll

 ; GCN-NEXT:    v_rcp_f32_e32 v4, v4
 ; GCN-NEXT:    v_mov_b32_e32 v13, 0
 ; GCN-NEXT:    v_mul_f32_e32 v4, 0x5f7ffffc, v4
 ; GCN-NEXT:    v_mul_f32_e32 v5, 0x2f800000, v4
 ; GCN-NEXT:    v_trunc_f32_e32 v5, v5
 ; GCN-NEXT:    v_mac_f32_e32 v4, 0xcf800000, v5
 ; GCN-NEXT:    v_cvt_u32_f32_e32 v5, v5
 ; GCN-NEXT:    v_cvt_u32_f32_e32 v4, v4
-; GCN-NEXT:    v_mul_lo_u32 v9, v6, v5
-; GCN-NEXT:    v_mul_hi_u32 v8, v6, v4
+; GCN-NEXT:    v_mul_lo_u32 v8, v6, v5
+; GCN-NEXT:    v_mul_hi_u32 v9, v6, v4
 ; GCN-NEXT:    v_mul_lo_u32 v10, v7, v4
-; GCN-NEXT:    v_add_i32_e32 v8, vcc, v8, v9
+; GCN-NEXT:    v_add_i32_e32 v8, vcc, v9, v8
 ; GCN-NEXT:    v_mul_lo_u32 v9, v6, v4
 ; GCN-NEXT:    v_add_i32_e32 v8, vcc, v8, v10
 ; GCN-NEXT:    v_mul_lo_u32 v10, v4, v8
 ; GCN-NEXT:    v_mul_hi_u32 v11, v4, v9
 ; GCN-NEXT:    v_mul_hi_u32 v12, v4, v8
 ; GCN-NEXT:    v_mul_hi_u32 v14, v5, v8
 ; GCN-NEXT:    v_mul_lo_u32 v8, v5, v8
 ; GCN-NEXT:    v_add_i32_e32 v10, vcc, v11, v10

llvm/test/CodeGen/AMDGPU/vgpr-tuple-allocation.ll

 ; GFX9: ;;#ASMSTART
 ; GFX9-NEXT: ;;#ASMEND
 ; GFX9: image_gather4_c_b_cl v[41:44], v[32:36], s[4:11], s[4:7] dmask:0x1
 ; GFX9-NEXT: s_addk_i32 s32, 0x800
 ; GFX9-NEXT: s_getpc_b64 s[4:5]
 ; GFX9-NEXT: s_add_u32 s4, s4, extern_func@gotpcrel32@lo+4
 ; GFX9-NEXT: s_addc_u32 s5, s5, extern_func@gotpcrel32@hi+12
 ; GFX9-NEXT: s_load_dwordx2 s[4:5], s[4:5], 0x0
-; GFX9-NEXT: v_writelane_b32 v40, s30, 0
+; GFX9: v_writelane_b32 v40, s30, 0
+; GFX9: v_writelane_b32 v40, s31, 1
 ; GFX9: s_waitcnt lgkmcnt(0)
 ; GFX9-NEXT: s_swappc_b64 s[30:31], s[4:5]
 
 ; GFX9: buffer_load_dword v44, off, s[0:3], s33 ; 4-byte Folded Reload
 ; GFX9-NEXT: buffer_load_dword v43, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
 ; GFX9-NEXT: buffer_load_dword v42, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload
 ; GFX9-NEXT: buffer_load_dword v41, off, s[0:3], s33 offset:12 ; 4-byte Folded Reload
+; GFX9: v_readlane_b32 s30, v40, 0
+; GFX9: v_readlane_b32 s31, v40, 1
 ; GFX9: buffer_load_dword v40, off, s[0:3], s32 offset:16 ; 4-byte Folded Reload
-; GFX9: s_setpc_b64 s[4:5]
+; GFX9: s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: non_preserved_vgpr_tuple8:
 ; GFX10: buffer_store_dword v40, off, s[0:3], s32 offset:16 ; 4-byte Folded Spill
 
 ; GFX10: v_mov_b32_e32 v36, v16
 ; GFX10-NEXT: v_mov_b32_e32 v35, v15
 ; GFX10-NEXT: v_mov_b32_e32 v34, v14
 ; GFX10-NEXT: v_mov_b32_e32 v33, v13
 ; GFX10-NEXT: v_mov_b32_e32 v32, v12
 
 ; GFX10: buffer_store_dword v41, off, s[0:3], s33 offset:12 ; 4-byte Folded Spill
 ; GFX10-NEXT: buffer_store_dword v42, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
 ; GFX10-NEXT: buffer_store_dword v43, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT: buffer_store_dword v44, off, s[0:3], s33 ; 4-byte Folded Spill
 
 ; GFX10: ;;#ASMSTART
 ; GFX10-NEXT: ;;#ASMEND
 
 ; GFX10: image_gather4_c_b_cl v[41:44], v[32:36], s[4:11], s[4:7] dmask:0x1 dim:SQ_RSRC_IMG_2D
 ; GFX10-NEXT: s_addk_i32 s32, 0x400
 ; GFX10-NEXT: s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT: s_getpc_b64 s[4:5]
 ; GFX10-NEXT: s_add_u32 s4, s4, extern_func@gotpcrel32@lo+4
 ; GFX10-NEXT: s_addc_u32 s5, s5, extern_func@gotpcrel32@hi+12
+; GFX10: v_writelane_b32 v40, s30, 0
 ; GFX10: s_load_dwordx2 s[4:5], s[4:5], 0x0
+; GFX10: v_writelane_b32 v40, s31, 1
 ; GFX10: s_waitcnt lgkmcnt(0)
 ; GFX10-NEXT: s_swappc_b64 s[30:31], s[4:5]
 
 ; GFX10: buffer_load_dword v44, off, s[0:3], s33
 ; GFX10-NEXT: buffer_load_dword v43, off, s[0:3], s33 offset:4
 ; GFX10-NEXT: buffer_load_dword v42, off, s[0:3], s33 offset:8
 ; GFX10-NEXT: buffer_load_dword v41, off, s[0:3], s33 offset:12
 
+; GFX10: v_readlane_b32 s30, v40, 0
+; GFX10: v_readlane_b32 s31, v40, 1
 ; GFX10: buffer_load_dword v40, off, s[0:3], s32 offset:16 ; 4-byte Folded Reload
-; GFX10: s_setpc_b64 s[4:5]
+; GFX10: s_setpc_b64 s[30:31]
 main_body:
   call void asm sideeffect "", "~{v0},~{v1},~{v2},~{v3},~{v4},~{v5},~{v6},~{v7}"() #0
   call void asm sideeffect "", "~{v8},~{v9},~{v10},~{v11},~{v12},~{v13},~{v14},~{v15}"() #0
   call void asm sideeffect "", "~{v16},~{v17},~{v18},~{v19},~{v20},~{v21},~{v22},~{v23}"() #0
   call void asm sideeffect "", "~{v24},~{v25},~{v26},~{v27},~{v28},~{v29},~{v30},~{v31}"() #0
   %v = call <4 x float> @llvm.amdgcn.image.gather4.c.b.cl.2d.v4f32.f32.f32(i32 1, float %bias, float %zcompare, float %s, float %t, float %clamp, <8 x i32> undef, <4 x i32> undef, i1 false, i32 0, i32 0)
   call void @extern_func()
   ret <4 x float> %v

 ; GFX9-NEXT: v_mov_b32_e32 v41, v12
 
 ; GFX9: image_gather4_c_b_cl v[0:3], v[41:45], s[36:43], s[4:7] dmask:0x1
 ; GFX9-NEXT: s_addk_i32 s32, 0x800
 ; GFX9-NEXT: s_getpc_b64 s[4:5]
 ; GFX9-NEXT: s_add_u32 s4, s4, extern_func@gotpcrel32@lo+4
 ; GFX9-NEXT: s_addc_u32 s5, s5, extern_func@gotpcrel32@hi+12
 ; GFX9-NEXT: s_load_dwordx2 s[4:5], s[4:5], 0x0
-; GFX9: s_waitcnt vmcnt(0)
+; GFX9-NEXT: v_writelane_b32 v40, s30, 8
+; GFX9-NEXT: v_writelane_b32 v40, s31, 9
+; GFX9-NEXT: s_waitcnt vmcnt(0)
 ; GFX9-NEXT: global_store_dwordx4 v[0:1], v[0:3], off
 ; GFX9-NEXT: s_waitcnt lgkmcnt(0)
 ; GFX9-NEXT: s_swappc_b64 s[30:31], s[4:5]
 ; GFX9-NEXT: image_gather4_c_b_cl v[0:3], v[41:45], s[36:43], s[4:7] dmask:0x1
 
 ; GFX9: buffer_load_dword v45, off, s[0:3], s33 ; 4-byte Folded Reload
 ; GFX9-NEXT: buffer_load_dword v44, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload
 ; GFX9-NEXT: buffer_load_dword v43, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload
 ; GFX9-NEXT: buffer_load_dword v42, off, s[0:3], s33 offset:12 ; 4-byte Folded Reload
 ; GFX9-NEXT: buffer_load_dword v41, off, s[0:3], s33 offset:16 ; 4-byte Folded Reload
 
+; GFX9: v_readlane_b32 s30, v40, 8
+; GFX9-NEXT: v_readlane_b32 s31, v40, 9
 ; GFX9: buffer_load_dword v40, off, s[0:3], s32 offset:20 ; 4-byte Folded Reload
-; GFX9: s_setpc_b64 s[4:5]
+; GFX9: s_setpc_b64 s[30:31]
 ;
 ; GFX10-LABEL: call_preserved_vgpr_tuple8:
 ; GFX10: buffer_store_dword v40, off, s[0:3], s32 offset:20 ; 4-byte Folded Spill
 ; GFX10: buffer_store_dword v41, off, s[0:3], s33 offset:16 ; 4-byte Folded Spill
 ; GFX10-NEXT: buffer_store_dword v42, off, s[0:3], s33 offset:12 ; 4-byte Folded Spill
 ; GFX10-NEXT: buffer_store_dword v43, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill
 ; GFX10-NEXT: buffer_store_dword v44, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill
 ; GFX10-NEXT: buffer_store_dword v45, off, s[0:3], s33 ; 4-byte Folded Spill
 
+; GFX10: v_mov_b32_e32 v41, v16
+; GFX10: v_mov_b32_e32 v42, v15
+; GFX10: v_mov_b32_e32 v43, v14
+; GFX10: v_mov_b32_e32 v44, v13
+; GFX10: v_mov_b32_e32 v45, v12
 
 ; GFX10:      image_gather4_c_b_cl v[0:3], v[12:16], s[36:43], s[4:7] dmask:0x1 dim:SQ_RSRC_IMG_2D
 ; GFX10-NEXT: s_waitcnt_depctr 0xffe3
 ; GFX10-NEXT: s_getpc_b64 s[4:5]
 ; GFX10-NEXT: s_add_u32 s4, s4, extern_func@gotpcrel32@lo+4
 ; GFX10-NEXT: s_addc_u32 s5, s5, extern_func@gotpcrel32@hi+12
 ; GFX10-NEXT: v_writelane_b32 v40, s30, 8
 ; GFX10-NEXT: s_load_dwordx2 s[4:5], s[4:5], 0x0
 ; GFX10-NEXT: v_writelane_b32 v40, s31, 9
 ; GFX10-NEXT: s_waitcnt vmcnt(0)
 ; GFX10-NEXT: global_store_dwordx4 v[0:1], v[0:3], off
 ; GFX10-NEXT: s_waitcnt lgkmcnt(0)
 ; GFX10-NEXT: s_swappc_b64 s[30:31], s[4:5]
 ; GFX10-NEXT: image_gather4_c_b_cl v[0:3], [v45, v44, v43, v42, v41], s[36:43], s[4:7] dmask:0x1 dim:SQ_RSRC_IMG_2D
 
 ; GFX10: buffer_load_dword v45, off, s[0:3], s33{{$}}
 ; GFX10-NEXT: buffer_load_dword v44, off, s[0:3], s33 offset:4
 ; GFX10-NEXT: buffer_load_dword v43, off, s[0:3], s33 offset:8
 ; GFX10-NEXT: buffer_load_dword v42, off, s[0:3], s33 offset:12
 ; GFX10-NEXT: buffer_load_dword v41, off, s[0:3], s33 offset:16
+; GFX10: v_readlane_b32 s30, v40, 8
+; GFX10-NEXT: v_readlane_b32 s31, v40, 9
 ; GFX10: buffer_load_dword v40, off, s[0:3], s32 offset:20
-; GFX10: s_setpc_b64 s[4:5]
+; GFX10: s_setpc_b64 s[30:31]
 main_body:
   %v = call <4 x float> @llvm.amdgcn.image.gather4.c.b.cl.2d.v4f32.f32.f32(i32 1, float %bias, float %zcompare, float %s, float %t, float %clamp, <8 x i32> undef, <4 x i32> undef, i1 false, i32 0, i32 0)
   store <4 x float> %v, <4 x float> addrspace(1)* undef
   call void @extern_func()
   %v1 = call <4 x float> @llvm.amdgcn.image.gather4.c.b.cl.2d.v4f32.f32.f32(i32 1, float %bias, float %zcompare, float %s, float %t, float %clamp, <8 x i32> undef, <4 x i32> undef, i1 false, i32 0, i32 0)
   ret <4 x float> %v1
 }
 
 declare <4 x float> @llvm.amdgcn.image.gather4.c.b.cl.2d.v4f32.f32.f32(i32 immarg, float, float, float, float, float, <8 x i32>, <4 x i32>, i1 immarg, i32 immarg, i32 immarg) #1
 
 attributes #0 = { nounwind writeonly }
 attributes #1 = { nounwind readonly }

llvm/test/CodeGen/AMDGPU/wave32.ll

 ; GCN: s_mov_b32 s33, s32
 ; GFX1064: s_addk_i32 s32, 0x400
 ; GFX1032: s_addk_i32 s32, 0x200
 
 
 ; GCN-DAG: v_writelane_b32 v40, s30, 0
 ; GCN-DAG: v_writelane_b32 v40, s31, 1
 ; GCN: s_swappc_b64
-; GCN-DAG: v_readlane_b32 s4, v40, 0
-; GCN-DAG: v_readlane_b32 s5, v40, 1
+; GCN-DAG: v_readlane_b32 s30, v40, 0
+; GCN-DAG: v_readlane_b32 s31, v40, 1
 
 
 ; GFX1064: s_addk_i32 s32, 0xfc00
 ; GFX1032: s_addk_i32 s32, 0xfe00
 ; GCN: v_readlane_b32 s33, v40, 2
 ; GFX1064: s_or_saveexec_b64 [[COPY_EXEC1:s\[[0-9]+:[0-9]+\]]], -1{{$}}
 ; GFX1032: s_or_saveexec_b32 [[COPY_EXEC1:s[0-9]]], -1{{$}}
 ; GCN-NEXT: buffer_load_dword v40, off, s[0:3], s32 ; 4-byte Folded Reload

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

 ; 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:    buffer_store_dword v5, off, s[0:3], s32 offset:24 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    buffer_store_dword v1, 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:    s_mov_b64 exec, s[34:35]
-; GFX9-O0-NEXT:    v_writelane_b32 v5, s30, 0
-; GFX9-O0-NEXT:    v_writelane_b32 v5, s31, 1
 ; GFX9-O0-NEXT:    s_mov_b32 s36, s4
 ; GFX9-O0-NEXT:    ; kill: def $sgpr36 killed $sgpr36 def $sgpr36_sgpr37_sgpr38_sgpr39
 ; GFX9-O0-NEXT:    s_mov_b32 s37, s5
 ; GFX9-O0-NEXT:    s_mov_b32 s38, s6
 ; GFX9-O0-NEXT:    s_mov_b32 s39, s7
 ; GFX9-O0-NEXT:    s_mov_b64 s[42:43], s[38:39]
 ; GFX9-O0-NEXT:    s_mov_b64 s[40:41], s[36:37]
-; GFX9-O0-NEXT:    v_writelane_b32 v5, s40, 2
-; GFX9-O0-NEXT:    v_writelane_b32 v5, s41, 3
-; GFX9-O0-NEXT:    v_writelane_b32 v5, s42, 4
-; GFX9-O0-NEXT:    v_writelane_b32 v5, s43, 5
-; GFX9-O0-NEXT:    s_mov_b32 s30, 0
-; GFX9-O0-NEXT:    buffer_load_dwordx2 v[3:4], off, s[36:39], s30
+; GFX9-O0-NEXT:    v_writelane_b32 v5, s40, 0
+; GFX9-O0-NEXT:    v_writelane_b32 v5, s41, 1
+; GFX9-O0-NEXT:    v_writelane_b32 v5, s42, 2
+; GFX9-O0-NEXT:    v_writelane_b32 v5, s43, 3
+; GFX9-O0-NEXT:    s_mov_b32 s34, 0
+; GFX9-O0-NEXT:    buffer_load_dwordx2 v[3:4], off, s[36:39], s34
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
 ; 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:    ; implicit-def: $sgpr34_sgpr35
+; GFX9-O0-NEXT:    ; implicit-def: $sgpr36_sgpr37
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v1, v3
 ; GFX9-O0-NEXT:    s_not_b64 exec, exec
-; GFX9-O0-NEXT:    v_mov_b32_e32 v1, s30
+; GFX9-O0-NEXT:    v_mov_b32_e32 v1, s34
 ; GFX9-O0-NEXT:    s_not_b64 exec, exec
-; GFX9-O0-NEXT:    s_or_saveexec_b64 s[34:35], -1
-; GFX9-O0-NEXT:    v_mov_b32_e32 v2, s30
+; GFX9-O0-NEXT:    s_or_saveexec_b64 s[36:37], -1
+; GFX9-O0-NEXT:    v_mov_b32_e32 v2, s34
 ; GFX9-O0-NEXT:    s_nop 1
 ; GFX9-O0-NEXT:    v_mov_b32_dpp v2, v1 row_bcast:31 row_mask:0xc bank_mask:0xf
 ; GFX9-O0-NEXT:    v_add_u32_e64 v1, v1, v2
-; GFX9-O0-NEXT:    s_mov_b64 exec, s[34:35]
+; GFX9-O0-NEXT:    s_mov_b64 exec, s[36:37]
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v3, v1
 ; GFX9-O0-NEXT:    buffer_store_dword v3, off, s[0:3], s32 offset:12 ; 4-byte Folded Spill
-; GFX9-O0-NEXT:    v_cmp_eq_u32_e64 s[34:35], v0, s30
-; GFX9-O0-NEXT:    v_mov_b32_e32 v0, s30
+; GFX9-O0-NEXT:    v_cmp_eq_u32_e64 s[36:37], v0, s34
+; GFX9-O0-NEXT:    v_mov_b32_e32 v0, s34
 ; GFX9-O0-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
-; GFX9-O0-NEXT:    s_mov_b64 s[30:31], exec
-; GFX9-O0-NEXT:    v_writelane_b32 v5, s30, 6
-; GFX9-O0-NEXT:    v_writelane_b32 v5, s31, 7
-; GFX9-O0-NEXT:    s_and_b64 s[30:31], s[30:31], s[34:35]
-; GFX9-O0-NEXT:    s_mov_b64 exec, s[30:31]
+; GFX9-O0-NEXT:    s_mov_b64 s[34:35], exec
+; GFX9-O0-NEXT:    v_writelane_b32 v5, s34, 4
+; GFX9-O0-NEXT:    v_writelane_b32 v5, s35, 5
+; GFX9-O0-NEXT:    s_and_b64 s[34:35], s[34:35], s[36:37]
+; GFX9-O0-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-O0-NEXT:    s_cbranch_execz .LBB1_2
 ; GFX9-O0-NEXT:  ; %bb.1: ; %if
 ; GFX9-O0-NEXT:    buffer_load_dword v3, off, s[0:3], s32 offset:16 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
 ; GFX9-O0-NEXT:    buffer_load_dword v4, off, s[0:3], s32 offset:20 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v0, v4
-; GFX9-O0-NEXT:    s_or_saveexec_b64 s[30:31], -1
+; GFX9-O0-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v1, 0
-; GFX9-O0-NEXT:    s_mov_b64 exec, s[30:31]
+; GFX9-O0-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX9-O0-NEXT:    s_not_b64 exec, exec
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v2, v1
 ; GFX9-O0-NEXT:    s_not_b64 exec, exec
-; GFX9-O0-NEXT:    s_or_saveexec_b64 s[30:31], -1
+; GFX9-O0-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-O0-NEXT:    v_mov_b32_dpp v1, v2 row_bcast:31 row_mask:0xc bank_mask:0xf
 ; GFX9-O0-NEXT:    v_add_u32_e64 v1, v2, v1
-; GFX9-O0-NEXT:    s_mov_b64 exec, s[30:31]
+; GFX9-O0-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v0, v1
 ; GFX9-O0-NEXT:    buffer_store_dword v0, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:  .LBB1_2: ; %merge
-; GFX9-O0-NEXT:    v_readlane_b32 s34, v5, 6
-; GFX9-O0-NEXT:    v_readlane_b32 s35, v5, 7
+; GFX9-O0-NEXT:    v_readlane_b32 s34, v5, 4
+; GFX9-O0-NEXT:    v_readlane_b32 s35, v5, 5
 ; GFX9-O0-NEXT:    s_or_b64 exec, exec, s[34:35]
-; GFX9-O0-NEXT:    v_readlane_b32 s30, v5, 0
-; GFX9-O0-NEXT:    v_readlane_b32 s31, v5, 1
-; GFX9-O0-NEXT:    v_readlane_b32 s36, v5, 2
-; GFX9-O0-NEXT:    v_readlane_b32 s37, v5, 3
-; GFX9-O0-NEXT:    v_readlane_b32 s38, v5, 4
-; GFX9-O0-NEXT:    v_readlane_b32 s39, v5, 5
+; GFX9-O0-NEXT:    v_readlane_b32 s36, v5, 0
+; GFX9-O0-NEXT:    v_readlane_b32 s37, v5, 1
+; GFX9-O0-NEXT:    v_readlane_b32 s38, v5, 2
+; GFX9-O0-NEXT:    v_readlane_b32 s39, v5, 3
 ; GFX9-O0-NEXT:    buffer_load_dword v0, off, s[0:3], s32 offset:12 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    buffer_load_dword v3, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-O0-NEXT:    v_cmp_eq_u32_e64 s[34:35], v0, v3
 ; GFX9-O0-NEXT:    v_cndmask_b32_e64 v0, 0, 1, s[34:35]
 ; GFX9-O0-NEXT:    s_mov_b32 s34, 1
 ; GFX9-O0-NEXT:    v_lshlrev_b32_e64 v0, s34, v0
 ; GFX9-O0-NEXT:    s_mov_b32 s34, 2

 ; GFX9-O0-NEXT:    ; kill: def $sgpr40_sgpr41_sgpr42_sgpr43 killed $sgpr36_sgpr37_sgpr38_sgpr39
 ; GFX9-O0-NEXT:    s_mov_b32 s34, 0
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v0, s8
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX9-O0-NEXT:    s_not_b64 exec, exec
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v2, s34
 ; GFX9-O0-NEXT:    s_not_b64 exec, exec
 ; GFX9-O0-NEXT:    s_or_saveexec_b64 s[40:41], -1
-; GFX9-O0-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-O0-NEXT:    s_add_u32 s30, s30, strict_wwm_called@rel32@lo+4
-; GFX9-O0-NEXT:    s_addc_u32 s31, s31, strict_wwm_called@rel32@hi+12
+; GFX9-O0-NEXT:    s_getpc_b64 s[42:43]
+; GFX9-O0-NEXT:    s_add_u32 s42, s42, strict_wwm_called@rel32@lo+4
+; GFX9-O0-NEXT:    s_addc_u32 s43, s43, strict_wwm_called@rel32@hi+12
 ; GFX9-O0-NEXT:    s_mov_b64 s[46:47], s[2:3]
 ; GFX9-O0-NEXT:    s_mov_b64 s[44:45], s[0:1]
 ; GFX9-O0-NEXT:    s_mov_b64 s[0:1], s[44:45]
 ; GFX9-O0-NEXT:    s_mov_b64 s[2:3], s[46:47]
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v0, v2
-; GFX9-O0-NEXT:    s_swappc_b64 s[30:31], s[30:31]
-; GFX9-O0-NEXT:    v_readlane_b32 s30, v3, 0
-; GFX9-O0-NEXT:    v_readlane_b32 s31, v3, 1
+; GFX9-O0-NEXT:    s_swappc_b64 s[30:31], s[42:43]
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v1, v0
 ; 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 s30, v3, 0
+; GFX9-O0-NEXT:    v_readlane_b32 s31, v3, 1
 ; GFX9-O0-NEXT:    s_add_i32 s32, s32, 0xfffffc00
 ; GFX9-O0-NEXT:    v_readlane_b32 s33, v3, 2
 ; GFX9-O0-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-O0-NEXT:    buffer_load_dword v3, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    buffer_load_dword v2, 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:    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:    buffer_store_dword v3, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    buffer_store_dword v2, 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:    s_mov_b64 exec, s[34:35]
-; GFX9-O3-NEXT:    s_mov_b32 s38, s33
+; GFX9-O3-NEXT:    v_writelane_b32 v3, s33, 2
+; GFX9-O3-NEXT:    v_writelane_b32 v3, s30, 0
 ; GFX9-O3-NEXT:    s_mov_b32 s33, s32
 ; GFX9-O3-NEXT:    s_addk_i32 s32, 0x400
-; GFX9-O3-NEXT:    s_mov_b64 s[36:37], s[30:31]
+; GFX9-O3-NEXT:    v_writelane_b32 v3, s31, 1
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v2, s8
 ; GFX9-O3-NEXT:    s_not_b64 exec, exec
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v2, 0
 ; GFX9-O3-NEXT:    s_not_b64 exec, exec
 ; GFX9-O3-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v0, v2
-; GFX9-O3-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-O3-NEXT:    s_add_u32 s30, s30, strict_wwm_called@rel32@lo+4
-; GFX9-O3-NEXT:    s_addc_u32 s31, s31, strict_wwm_called@rel32@hi+12
-; GFX9-O3-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-O3-NEXT:    s_getpc_b64 s[36:37]
+; GFX9-O3-NEXT:    s_add_u32 s36, s36, strict_wwm_called@rel32@lo+4
+; GFX9-O3-NEXT:    s_addc_u32 s37, s37, strict_wwm_called@rel32@hi+12
+; GFX9-O3-NEXT:    s_swappc_b64 s[30:31], s[36:37]
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v1, v0
 ; 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:    v_readlane_b32 s30, v3, 0
 ; 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:    s_addk_i32 s32, 0xfc00
-; GFX9-O3-NEXT:    s_mov_b32 s33, s38
-; GFX9-O3-NEXT:    s_or_saveexec_b64 s[30:31], -1
+; GFX9-O3-NEXT:    v_readlane_b32 s33, v3, 2
+; GFX9-O3-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-O3-NEXT:    buffer_load_dword v3, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload
 ; GFX9-O3-NEXT:    buffer_load_dword v2, 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:    s_mov_b64 exec, s[30:31]
+; GFX9-O3-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-O3-NEXT:    s_waitcnt vmcnt(0)
-; GFX9-O3-NEXT:    s_setpc_b64 s[36:37]
+; 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)
   %tmp136 = add i32 %tmp134, %tmp107
   %tmp137 = tail call i32 @llvm.amdgcn.strict.wwm.i32(i32 %tmp136)
   call void @llvm.amdgcn.raw.buffer.store.i32(i32 %tmp137, <4 x i32> %tmp14, i32 4, i32 0, i32 0)
   ret void
 }
 

 ; GFX9-O0-NEXT:    buffer_store_dword v4, off, s[0:3], s32 offset:32 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    buffer_store_dword v5, off, s[0:3], s32 offset:36 ; 4-byte Folded Spill
 ; GFX9-O0-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-O0-NEXT:    v_writelane_b32 v10, s33, 8
 ; GFX9-O0-NEXT:    s_mov_b32 s33, s32
 ; GFX9-O0-NEXT:    s_add_i32 s32, s32, 0xc00
 ; GFX9-O0-NEXT:    v_writelane_b32 v10, s30, 0
 ; GFX9-O0-NEXT:    v_writelane_b32 v10, s31, 1
-; GFX9-O0-NEXT:    s_mov_b32 s34, s8
-; GFX9-O0-NEXT:    s_mov_b32 s36, s4
-; GFX9-O0-NEXT:    ; kill: def $sgpr36 killed $sgpr36 def $sgpr36_sgpr37_sgpr38_sgpr39
-; GFX9-O0-NEXT:    s_mov_b32 s37, s5
-; GFX9-O0-NEXT:    s_mov_b32 s38, s6
-; GFX9-O0-NEXT:    s_mov_b32 s39, s7
-; GFX9-O0-NEXT:    v_writelane_b32 v10, s36, 2
-; GFX9-O0-NEXT:    v_writelane_b32 v10, s37, 3
-; GFX9-O0-NEXT:    v_writelane_b32 v10, s38, 4
-; GFX9-O0-NEXT:    v_writelane_b32 v10, s39, 5
-; GFX9-O0-NEXT:    ; kill: def $sgpr34 killed $sgpr34 def $sgpr34_sgpr35
-; GFX9-O0-NEXT:    s_mov_b32 s35, s9
-; GFX9-O0-NEXT:    ; kill: def $sgpr30_sgpr31 killed $sgpr34_sgpr35
-; GFX9-O0-NEXT:    s_mov_b64 s[30:31], 0
-; GFX9-O0-NEXT:    v_mov_b32_e32 v0, s34
-; GFX9-O0-NEXT:    v_mov_b32_e32 v1, s35
+; GFX9-O0-NEXT:    s_mov_b32 s36, s8
+; GFX9-O0-NEXT:    s_mov_b32 s40, s4
+; GFX9-O0-NEXT:    ; kill: def $sgpr40 killed $sgpr40 def $sgpr40_sgpr41_sgpr42_sgpr43
+; GFX9-O0-NEXT:    s_mov_b32 s41, s5
+; GFX9-O0-NEXT:    s_mov_b32 s42, s6
+; GFX9-O0-NEXT:    s_mov_b32 s43, s7
+; GFX9-O0-NEXT:    v_writelane_b32 v10, s40, 2
+; GFX9-O0-NEXT:    v_writelane_b32 v10, s41, 3
+; GFX9-O0-NEXT:    v_writelane_b32 v10, s42, 4
+; GFX9-O0-NEXT:    v_writelane_b32 v10, s43, 5
+; GFX9-O0-NEXT:    ; kill: def $sgpr36 killed $sgpr36 def $sgpr36_sgpr37
+; GFX9-O0-NEXT:    s_mov_b32 s37, s9
+; GFX9-O0-NEXT:    ; kill: def $sgpr34_sgpr35 killed $sgpr36_sgpr37
+; GFX9-O0-NEXT:    s_mov_b64 s[34:35], 0
+; GFX9-O0-NEXT:    v_mov_b32_e32 v0, s36
+; GFX9-O0-NEXT:    v_mov_b32_e32 v1, s37
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v9, v1
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v8, v0
 ; GFX9-O0-NEXT:    s_not_b64 exec, exec
-; GFX9-O0-NEXT:    v_mov_b32_e32 v8, s30
-; GFX9-O0-NEXT:    v_mov_b32_e32 v9, s31
+; GFX9-O0-NEXT:    v_mov_b32_e32 v8, s34
+; GFX9-O0-NEXT:    v_mov_b32_e32 v9, s35
 ; GFX9-O0-NEXT:    s_not_b64 exec, exec
-; GFX9-O0-NEXT:    s_or_saveexec_b64 s[30:31], -1
-; GFX9-O0-NEXT:    v_writelane_b32 v10, s30, 6
-; GFX9-O0-NEXT:    v_writelane_b32 v10, s31, 7
+; GFX9-O0-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-O0-NEXT:    v_writelane_b32 v10, s34, 6
+; GFX9-O0-NEXT:    v_writelane_b32 v10, s35, 7
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v2, v8
-; GFX9-O0-NEXT:    s_mov_b32 s30, 32
-; GFX9-O0-NEXT:    ; implicit-def: $sgpr34_sgpr35
-; GFX9-O0-NEXT:    v_lshrrev_b64 v[3:4], s30, v[8:9]
-; GFX9-O0-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-O0-NEXT:    s_add_u32 s30, s30, strict_wwm_called_i64@gotpcrel32@lo+4
-; GFX9-O0-NEXT:    s_addc_u32 s31, s31, strict_wwm_called_i64@gotpcrel32@hi+12
-; GFX9-O0-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
+; GFX9-O0-NEXT:    s_mov_b32 s34, 32
+; GFX9-O0-NEXT:    ; implicit-def: $sgpr36_sgpr37
+; GFX9-O0-NEXT:    v_lshrrev_b64 v[3:4], s34, v[8:9]
+; GFX9-O0-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-O0-NEXT:    s_add_u32 s34, s34, strict_wwm_called_i64@gotpcrel32@lo+4
+; GFX9-O0-NEXT:    s_addc_u32 s35, s35, strict_wwm_called_i64@gotpcrel32@hi+12
+; GFX9-O0-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX9-O0-NEXT:    s_mov_b64 s[38:39], s[2:3]
 ; GFX9-O0-NEXT:    s_mov_b64 s[36:37], s[0:1]
 ; GFX9-O0-NEXT:    s_mov_b64 s[0:1], s[36:37]
 ; GFX9-O0-NEXT:    s_mov_b64 s[2:3], s[38:39]
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v0, v2
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v1, v3
 ; GFX9-O0-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX9-O0-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-O0-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-O0-NEXT:    v_readlane_b32 s34, v10, 6
 ; GFX9-O0-NEXT:    v_readlane_b32 s35, v10, 7
 ; GFX9-O0-NEXT:    v_readlane_b32 s36, v10, 2
 ; GFX9-O0-NEXT:    v_readlane_b32 s37, v10, 3
 ; GFX9-O0-NEXT:    v_readlane_b32 s38, v10, 4
 ; GFX9-O0-NEXT:    v_readlane_b32 s39, v10, 5
-; GFX9-O0-NEXT:    v_readlane_b32 s30, v10, 0
-; GFX9-O0-NEXT:    v_readlane_b32 s31, v10, 1
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v3, v1
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v4, v8
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v5, v9
 ; GFX9-O0-NEXT:    v_add_co_u32_e64 v2, s[40:41], v2, v4
 ; GFX9-O0-NEXT:    v_addc_co_u32_e64 v3, s[40:41], v3, v5, s[40:41]
 ; GFX9-O0-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v0, v2
 ; GFX9-O0-NEXT:    v_mov_b32_e32 v1, v3
 ; GFX9-O0-NEXT:    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 s30, v10, 0
+; GFX9-O0-NEXT:    v_readlane_b32 s31, v10, 1
 ; GFX9-O0-NEXT:    s_add_i32 s32, s32, 0xfffff400
 ; GFX9-O0-NEXT:    v_readlane_b32 s33, v10, 8
 ; GFX9-O0-NEXT:    s_or_saveexec_b64 s[34:35], -1
 ; GFX9-O0-NEXT:    buffer_load_dword v10, off, s[0:3], s32 offset:40 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
 ; GFX9-O0-NEXT:    buffer_load_dword v8, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-O0-NEXT:    s_nop 0
 ; GFX9-O0-NEXT:    buffer_load_dword v9, 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_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:    buffer_store_dword v8, off, s[0:3], s32 offset:24 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    buffer_store_dword v6, off, s[0:3], s32 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-O3-NEXT:    buffer_store_dword v7, 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:    buffer_store_dword v2, off, s[0:3], s32 offset:16 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    s_waitcnt vmcnt(0)
 ; GFX9-O3-NEXT:    buffer_store_dword v3, off, s[0:3], s32 offset:20 ; 4-byte Folded Spill
 ; GFX9-O3-NEXT:    s_mov_b64 exec, s[34:35]
-; GFX9-O3-NEXT:    s_mov_b32 s40, s33
+; GFX9-O3-NEXT:    v_writelane_b32 v8, s33, 2
+; GFX9-O3-NEXT:    v_writelane_b32 v8, s30, 0
 ; GFX9-O3-NEXT:    s_mov_b32 s33, s32
 ; GFX9-O3-NEXT:    s_addk_i32 s32, 0x800
-; GFX9-O3-NEXT:    s_mov_b64 s[36:37], s[30:31]
+; GFX9-O3-NEXT:    v_writelane_b32 v8, s31, 1
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v6, s8
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v7, s9
 ; GFX9-O3-NEXT:    s_not_b64 exec, exec
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v6, 0
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v7, 0
 ; GFX9-O3-NEXT:    s_not_b64 exec, exec
-; GFX9-O3-NEXT:    s_or_saveexec_b64 s[38:39], -1
-; GFX9-O3-NEXT:    s_getpc_b64 s[30:31]
-; GFX9-O3-NEXT:    s_add_u32 s30, s30, strict_wwm_called_i64@gotpcrel32@lo+4
-; GFX9-O3-NEXT:    s_addc_u32 s31, s31, strict_wwm_called_i64@gotpcrel32@hi+12
-; GFX9-O3-NEXT:    s_load_dwordx2 s[30:31], s[30:31], 0x0
+; GFX9-O3-NEXT:    s_or_saveexec_b64 s[36:37], -1
+; GFX9-O3-NEXT:    s_getpc_b64 s[34:35]
+; GFX9-O3-NEXT:    s_add_u32 s34, s34, strict_wwm_called_i64@gotpcrel32@lo+4
+; GFX9-O3-NEXT:    s_addc_u32 s35, s35, strict_wwm_called_i64@gotpcrel32@hi+12
+; GFX9-O3-NEXT:    s_load_dwordx2 s[34:35], s[34:35], 0x0
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v0, v6
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v1, v7
 ; GFX9-O3-NEXT:    s_waitcnt lgkmcnt(0)
-; GFX9-O3-NEXT:    s_swappc_b64 s[30:31], s[30:31]
+; GFX9-O3-NEXT:    s_swappc_b64 s[30:31], s[34:35]
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v2, v0
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v3, v1
 ; GFX9-O3-NEXT:    v_add_co_u32_e32 v2, vcc, v2, v6
 ; GFX9-O3-NEXT:    v_addc_co_u32_e32 v3, vcc, v3, v7, vcc
-; GFX9-O3-NEXT:    s_mov_b64 exec, s[38:39]
+; GFX9-O3-NEXT:    s_mov_b64 exec, s[36:37]
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v0, v2
 ; GFX9-O3-NEXT:    v_mov_b32_e32 v1, v3
+; GFX9-O3-NEXT:    v_readlane_b32 s30, v8, 0
 ; 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:    s_addk_i32 s32, 0xf800
-; GFX9-O3-NEXT:    s_mov_b32 s33, s40
-; GFX9-O3-NEXT:    s_or_saveexec_b64 s[30:31], -1
+; GFX9-O3-NEXT:    v_readlane_b32 s33, v8, 2
+; GFX9-O3-NEXT:    s_or_saveexec_b64 s[34:35], -1
+; GFX9-O3-NEXT:    buffer_load_dword v8, off, s[0:3], s32 offset:24 ; 4-byte Folded Reload
+; GFX9-O3-NEXT:    s_nop 0
 ; GFX9-O3-NEXT:    buffer_load_dword v6, off, s[0:3], s32 ; 4-byte Folded Reload
 ; GFX9-O3-NEXT:    s_nop 0
 ; GFX9-O3-NEXT:    buffer_load_dword v7, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload
 ; GFX9-O3-NEXT:    s_nop 0
 ; 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_nop 0
 ; GFX9-O3-NEXT:    buffer_load_dword v2, off, s[0:3], s32 offset:16 ; 4-byte Folded Reload
 ; GFX9-O3-NEXT:    s_nop 0
 ; GFX9-O3-NEXT:    buffer_load_dword v3, off, s[0:3], s32 offset:20 ; 4-byte Folded Reload
-; GFX9-O3-NEXT:    s_mov_b64 exec, s[30:31]
+; GFX9-O3-NEXT:    s_mov_b64 exec, s[34:35]
 ; GFX9-O3-NEXT:    s_waitcnt vmcnt(0)
-; GFX9-O3-NEXT:    s_setpc_b64 s[36:37]
+; GFX9-O3-NEXT:    s_setpc_b64 s[30:31]
   %tmp107 = tail call i64 @llvm.amdgcn.set.inactive.i64(i64 %arg, i64 0)
   %tmp134 = call amdgpu_gfx i64 @strict_wwm_called_i64(i64 %tmp107)
   %tmp136 = add i64 %tmp134, %tmp107
   %tmp137 = tail call i64 @llvm.amdgcn.strict.wwm.i64(i64 %tmp136)
   %tmp138 = bitcast i64 %tmp137 to <2 x i32>
   call void @llvm.amdgcn.raw.buffer.store.v2i32(<2 x i32> %tmp138, <4 x i32> %tmp14, i32 4, i32 0, i32 0)
   ret void
 }

llvm/test/CodeGen/MIR/AMDGPU/machine-metadata.mir

machineMetadataNodes:
   - '!6 = distinct !{!6, !7, !"Src"}'
   - '!11 = !{!4, !6}'
   - '!5 = !{!1, !6}'
   - '!8 = !{!4, !9}'
   - '!10 = !{!1, !9}'
   - '!7 = distinct !{!7, !"MemcpyLoweringDomain"}'
 body:             |
   bb.0 (%ir-block.0):
-    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $sgpr30_sgpr31
+    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
 
     ; CHECK-LABEL: name: test_memcpy
-    ; CHECK: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr30_sgpr31
     ; CHECK: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr3
     ; CHECK: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
     ; CHECK: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr1
     ; CHECK: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
     ; CHECK: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY2]], %subreg.sub0, [[COPY1]], %subreg.sub1
     ; CHECK: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY3]], %subreg.sub1
     ; CHECK: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]]
     ; CHECK: [[GLOBAL_LOAD_DWORDX4_:%[0-9]+]]:vreg_128 = GLOBAL_LOAD_DWORDX4 [[COPY5]], 16, 0, implicit $exec :: (load (s128) from %ir.p1, align 4, !alias.scope !5, !noalias !8, addrspace 1)
     ; CHECK: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]]
     ; CHECK: GLOBAL_STORE_DWORDX4 [[COPY6]], killed [[GLOBAL_LOAD_DWORDX4_]], 0, 0, implicit $exec :: (store (s128) into %ir.p0, align 4, !alias.scope !10, !noalias !11, addrspace 1)
     ; CHECK: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]]
     ; CHECK: [[GLOBAL_LOAD_DWORDX2_:%[0-9]+]]:vreg_64 = GLOBAL_LOAD_DWORDX2 killed [[COPY7]], 0, 0, implicit $exec :: (load (s64) from %ir.1, align 4, !alias.scope !3, !noalias !0, addrspace 1)
     ; CHECK: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[GLOBAL_LOAD_DWORDX2_]].sub0
     ; CHECK: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[GLOBAL_LOAD_DWORDX2_]].sub1
     ; CHECK: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 killed [[COPY8]], killed [[COPY9]], 0, implicit $exec
-    ; CHECK: [[COPY10:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
     ; CHECK: $vgpr0 = COPY [[V_ADD_U32_e64_]]
-    ; CHECK: [[COPY11:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY10]]
-    ; CHECK: S_SETPC_B64_return [[COPY11]], implicit $vgpr0
-    %4:sreg_64 = COPY $sgpr30_sgpr31
+    ; CHECK: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     %3:vgpr_32 = COPY $vgpr3
     %2:vgpr_32 = COPY $vgpr2
     %1:vgpr_32 = COPY $vgpr1
     %0:vgpr_32 = COPY $vgpr0
     %17:vreg_64 = REG_SEQUENCE %2, %subreg.sub0, %3, %subreg.sub1
     %18:vreg_64 = REG_SEQUENCE %0, %subreg.sub0, %1, %subreg.sub1
     %9:vreg_64 = COPY %18
     %8:vreg_128 = GLOBAL_LOAD_DWORDX4 %9, 16, 0, implicit $exec :: (load (s128) from %ir.p1, align 4, !alias.scope !5, !noalias !8, addrspace 1)
     %10:vreg_64 = COPY %18
     GLOBAL_STORE_DWORDX4 %10, killed %8, 0, 0, implicit $exec :: (store (s128) into %ir.p0, align 4, !alias.scope !10, !noalias !11, addrspace 1)
     %12:vreg_64 = COPY %17
     %11:vreg_64 = GLOBAL_LOAD_DWORDX2 killed %12, 0, 0, implicit $exec :: (load (s64) from %ir.1, align 4, !alias.scope !3, !noalias !0, addrspace 1)
     %13:vgpr_32 = COPY %11.sub0
     %14:vgpr_32 = COPY %11.sub1
     %15:vgpr_32 = V_ADD_U32_e64 killed %13, killed %14, 0, implicit $exec
-    %5:ccr_sgpr_64 = COPY %4
     $vgpr0 = COPY %15
-    %16:ccr_sgpr_64 = COPY %5
-    S_SETPC_B64_return %16, implicit $vgpr0
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
 
 ...
 ---
 name:            test_memcpy_inline
 machineMetadataNodes:
   - '!6 = distinct !{!6, !7, !"Src"}'
   - '!7 = distinct !{!7, !"MemcpyLoweringDomain"}'
   - '!9 = distinct !{!9, !7, !"Dst"}'
   - '!11 = !{!4, !6}'
   - '!5 = !{!1, !6}'
   - '!8 = !{!4, !9}'
   - '!10 = !{!1, !9}'
 body:             |
   bb.0 (%ir-block.0):
-    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $sgpr30_sgpr31
+    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
 
     ; CHECK-LABEL: name: test_memcpy_inline
-    ; CHECK: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr30_sgpr31
     ; CHECK: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr3
     ; CHECK: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
     ; CHECK: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr1
     ; CHECK: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
     ; CHECK: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY2]], %subreg.sub0, [[COPY1]], %subreg.sub1
     ; CHECK: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY3]], %subreg.sub1
     ; CHECK: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]]
     ; CHECK: [[GLOBAL_LOAD_DWORDX4_:%[0-9]+]]:vreg_128 = GLOBAL_LOAD_DWORDX4 [[COPY5]], 16, 0, implicit $exec :: (load (s128) from %ir.p1, align 4, !alias.scope !5, !noalias !8, addrspace 1)
     ; CHECK: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]]
     ; CHECK: GLOBAL_STORE_DWORDX4 [[COPY6]], killed [[GLOBAL_LOAD_DWORDX4_]], 0, 0, implicit $exec :: (store (s128) into %ir.p0, align 4, !alias.scope !10, !noalias !11, addrspace 1)
     ; CHECK: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]]
     ; CHECK: [[GLOBAL_LOAD_DWORDX2_:%[0-9]+]]:vreg_64 = GLOBAL_LOAD_DWORDX2 killed [[COPY7]], 0, 0, implicit $exec :: (load (s64) from %ir.1, align 4, !alias.scope !3, !noalias !0, addrspace 1)
     ; CHECK: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[GLOBAL_LOAD_DWORDX2_]].sub0
     ; CHECK: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[GLOBAL_LOAD_DWORDX2_]].sub1
     ; CHECK: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 killed [[COPY8]], killed [[COPY9]], 0, implicit $exec
-    ; CHECK: [[COPY10:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
     ; CHECK: $vgpr0 = COPY [[V_ADD_U32_e64_]]
-    ; CHECK: [[COPY11:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY10]]
-    ; CHECK: S_SETPC_B64_return [[COPY11]], implicit $vgpr0
-    %4:sreg_64 = COPY $sgpr30_sgpr31
+    ; CHECK: S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
     %3:vgpr_32 = COPY $vgpr3
     %2:vgpr_32 = COPY $vgpr2
     %1:vgpr_32 = COPY $vgpr1
     %0:vgpr_32 = COPY $vgpr0
     %17:vreg_64 = REG_SEQUENCE %2, %subreg.sub0, %3, %subreg.sub1
     %18:vreg_64 = REG_SEQUENCE %0, %subreg.sub0, %1, %subreg.sub1
     %9:vreg_64 = COPY %18
     %8:vreg_128 = GLOBAL_LOAD_DWORDX4 %9, 16, 0, implicit $exec :: (load (s128) from %ir.p1, align 4, !alias.scope !5, !noalias !8, addrspace 1)
     %10:vreg_64 = COPY %18
     GLOBAL_STORE_DWORDX4 %10, killed %8, 0, 0, implicit $exec :: (store (s128) into %ir.p0, align 4, !alias.scope !10, !noalias !11, addrspace 1)
     %12:vreg_64 = COPY %17
     %11:vreg_64 = GLOBAL_LOAD_DWORDX2 killed %12, 0, 0, implicit $exec :: (load (s64) from %ir.1, align 4, !alias.scope !3, !noalias !0, addrspace 1)
     %13:vgpr_32 = COPY %11.sub0
     %14:vgpr_32 = COPY %11.sub1
     %15:vgpr_32 = V_ADD_U32_e64 killed %13, killed %14, 0, implicit $exec
-    %5:ccr_sgpr_64 = COPY %4
     $vgpr0 = COPY %15
-    %16:ccr_sgpr_64 = COPY %5
-    S_SETPC_B64_return %16, implicit $vgpr0
+    S_SETPC_B64_return $sgpr30_sgpr31, implicit $vgpr0
 
 ...

llvm/test/CodeGen/MIR/AMDGPU/stack-id-assert.mir

 # CHECK-LABEL: name:            bar
 # CHECK: fixedStack:      []
 # CHECK-NEXT: {{^}}stack:
 # CHECK-NEXT:  - { id:
 
 
 ---
 name:            foo
-liveins:
-  - { reg: '$sgpr30_sgpr31', virtual-reg: '' }
 body:             |
   bb.0:
-    S_SETPC_B64_return killed renamable $sgpr30_sgpr31
+    SI_RETURN
 
 ...
 ---
 name:            bar
 tracksRegLiveness: true
 liveins:
   - { reg: '$vgpr0', virtual-reg: '' }
   - { reg: '$vgpr1', virtual-reg: '' }
   - { reg: '$vgpr2', virtual-reg: '' }
-  - { reg: '$sgpr30_sgpr31', virtual-reg: '' }
 stack:
   - { id: 0, name: '', type: spill-slot, offset: 0, size: 8, alignment: 4,
       stack-id: sgpr-spill, callee-saved-register: '', callee-saved-restored: true,
       debug-info-variable: '', debug-info-expression: '', debug-info-location: '' }
 machineFunctionInfo:
   scratchRSrcReg:  '$sgpr0_sgpr1_sgpr2_sgpr3'
   stackPtrOffsetReg: '$sgpr32'
 body:             |
   bb.0:
-    liveins: $vgpr0, $vgpr1, $vgpr2, $sgpr30_sgpr31
+    liveins: $vgpr0, $vgpr1, $vgpr2
 
     renamable $vgpr41 = COPY $vgpr2, implicit $exec
     renamable $vgpr40 = COPY $vgpr1, implicit $exec
     renamable $vcc = V_CMP_NE_U32_e64 42, killed $vgpr0, implicit $exec
     $sgpr4_sgpr5 = S_AND_SAVEEXEC_B64 $vcc, implicit-def $exec, implicit-def $scc, implicit $exec
     renamable $sgpr34_sgpr35 = S_XOR_B64 $exec, killed renamable $sgpr4_sgpr5, implicit-def dead $scc
 
     ADJCALLSTACKUP 0, 0, implicit-def dead $scc, implicit-def $sgpr32, implicit $sgpr32
     renamable $sgpr4_sgpr5 = SI_PC_ADD_REL_OFFSET target-flags(amdgpu-gotprel32-lo) @foo + 4, target-flags(amdgpu-gotprel32-hi) @foo + 12, implicit-def dead $scc
     renamable $sgpr4_sgpr5 = S_LOAD_DWORDX2_IMM killed renamable $sgpr4_sgpr5, 0, 0
-    SI_SPILL_S64_SAVE killed renamable $sgpr30_sgpr31, %stack.0, implicit $exec, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr32
     dead $sgpr30_sgpr31 = SI_CALL killed renamable $sgpr4_sgpr5, @foo, csr_amdgpu_highregs, implicit $sgpr0_sgpr1_sgpr2_sgpr3
-    renamable $sgpr30_sgpr31 = SI_SPILL_S64_RESTORE %stack.0, implicit $exec, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr32
     ADJCALLSTACKDOWN 0, 0, implicit-def dead $scc, implicit-def $sgpr32, implicit $sgpr32
 
-    S_SETPC_B64_return killed renamable $sgpr30_sgpr31
+    SI_RETURN
 
 ...