RegAllocFast: Improve hinting heuristic

[arsenm]

Summary

• Trace through multiple COPYs when looking for a physreg source.
• Add hinting for vregs that will be copied into physregs (we only hinted for vregs getting copied to a physreg previously).
• Give hinted a register a bonus when deciding which value to spill.

This is part of my rewrite regallocfast series. In fact this one doesn't even have an effect unless you also flip the allocation to happen from back to front of a basic block. Nonetheless it helps to split this up to ease review of D52010

[arsenm]

lib/CodeGen/RegAllocFast.cpp

601–602

Isn't this the same as MI.isFullCopy()?

[arsenm]

I do see test changes with this (at least with D54366 reverted). At least one of them is a regression for this situation:

%11:sreg_32_xm0 = COPY %0.sub1
%12:sreg_32_xm0 = COPY %0.sub0

This inserts 2 64-bit restores of %0, and each one uses only one of the subregisters. Previously this did the right thing, and only restored once

[arsenm]

Use isFullCopy, add test changes

[arsenm]

Rebase, test updates

[qcolombet]

Looks good to me, modulo the order in which we check the hints.

lib/CodeGen/RegAllocFast.cpp

656

I would rather we try the hint that we've been given first. (I.e., Hint1 before Hint0).

[arsenm]

Swap hint order, which changed no tests

[arsenm]

Fix rename

[qcolombet]

LGTM

[arsenm]

r360887

lib/CodeGen/RegAllocFast.cpp

bool isRegUsedInInstr(MCPhysReg PhysReg) const {
         if (UsedInInstr.count(*Units))
           return true;
       return false;
     }
 
     enum : unsigned {
       spillClean = 50,
       spillDirty = 100,
+      spillPrefBonus = 20,
       spillImpossible = ~0u
     };
 
   public:
     StringRef getPassName() const override { return "Fast Register Allocator"; }
 
     void getAnalysisUsage(AnalysisUsage &AU) const override {
       AU.setPreservesCFG();
private:
     void allocVirtRegUndef(MachineOperand &MO);
     MCPhysReg defineVirtReg(MachineInstr &MI, unsigned OpNum, unsigned VirtReg,
                             unsigned Hint);
     LiveReg &reloadVirtReg(MachineInstr &MI, unsigned OpNum, unsigned VirtReg,
                            unsigned Hint);
     void spillAll(MachineBasicBlock::iterator MI, bool OnlyLiveOut);
     bool setPhysReg(MachineInstr &MI, MachineOperand &MO, MCPhysReg PhysReg);
 
+    unsigned traceCopies(unsigned VirtReg) const;
+    unsigned traceCopyChain(unsigned Reg) const;
+
     int getStackSpaceFor(unsigned VirtReg);
     void spill(MachineBasicBlock::iterator Before, unsigned VirtReg,
                MCPhysReg AssignedReg, bool Kill);
     void reload(MachineBasicBlock::iterator Before, unsigned VirtReg,
                 MCPhysReg PhysReg);
 
     bool mayLiveOut(unsigned VirtReg);
 
void RegAllocFast::assignVirtToPhysReg(LiveReg &LR, MCPhysReg PhysReg) {
   LLVM_DEBUG(dbgs() << "Assigning " << printReg(VirtReg, TRI) << " to "
                     << printReg(PhysReg, TRI) << '\n');
   assert(LR.PhysReg == 0 && "Already assigned a physreg");
   assert(PhysReg != 0 && "Trying to assign no register");
   LR.PhysReg = PhysReg;
   setPhysRegState(PhysReg, VirtReg);
 }
 
+static bool isCoalescable(const MachineInstr &MI) {
+  return MI.isFullCopy();
+}
+
+unsigned RegAllocFast::traceCopyChain(unsigned Reg) const {
+  static const unsigned ChainLengthLimit = 3;
+  unsigned C = 0;
+  do {
+    if (TargetRegisterInfo::isPhysicalRegister(Reg))
+      return Reg;
+    assert(TargetRegisterInfo::isVirtualRegister(Reg));
+
+    MachineInstr *VRegDef = MRI->getUniqueVRegDef(Reg);
+    if (!VRegDef || !isCoalescable(*VRegDef))
+      return 0;
+    Reg = VRegDef->getOperand(1).getReg();
+  } while (++C <= ChainLengthLimit);
+  return 0;
+}
+
+/// Check if any of \p VirtReg's definitions is a copy. If it is follow the
+/// chain of copies to check whether we reach a physical register we can
+/// coalesce with.
+unsigned RegAllocFast::traceCopies(unsigned VirtReg) const {
+  static const unsigned DefLimit = 3;
+  unsigned C = 0;
+  for (const MachineInstr &MI : MRI->def_instructions(VirtReg)) {
+    if (isCoalescable(MI)) {
+      unsigned Reg = MI.getOperand(1).getReg();
+      Reg = traceCopyChain(Reg);
+      if (Reg != 0)
+        return Reg;
+    }
+
+    if (++C >= DefLimit)
+      break;
+  }
+  return 0;
+}
+
 /// Allocates a physical register for VirtReg.
-void RegAllocFast::allocVirtReg(MachineInstr &MI, LiveReg &LR, unsigned Hint) {
+void RegAllocFast::allocVirtReg(MachineInstr &MI, LiveReg &LR, unsigned Hint1) {
   const unsigned VirtReg = LR.VirtReg;
 
   assert(TargetRegisterInfo::isVirtualRegister(VirtReg) &&
          "Can only allocate virtual registers");
 
   const TargetRegisterClass &RC = *MRI->getRegClass(VirtReg);
   LLVM_DEBUG(dbgs() << "Search register for " << printReg(VirtReg)
                     << " in class " << TRI->getRegClassName(&RC)
-                    << " with hint " << printReg(Hint, TRI) << '\n');
+                    << " with hint " << printReg(Hint1, TRI) << '\n');
 
   // Take hint when possible.
-  if (TargetRegisterInfo::isPhysicalRegister(Hint) &&
-      MRI->isAllocatable(Hint) && RC.contains(Hint)) {
+  unsigned Hint0 = traceCopies(VirtReg);
+  if (TargetRegisterInfo::isPhysicalRegister(Hint0) &&
+      MRI->isAllocatable(Hint0) && RC.contains(Hint0) &&
+      !isRegUsedInInstr(Hint0)) {
     // Ignore the hint if we would have to spill a dirty register.
-    unsigned Cost = calcSpillCost(Hint);
+    unsigned Cost = calcSpillCost(Hint0);
     if (Cost < spillDirty) {
+      LLVM_DEBUG(dbgs() << "\tPreferred Register 0: " << printReg(Hint0, TRI)
+                        << '\n');
       if (Cost)
-        definePhysReg(MI, Hint, regFree);
-      assignVirtToPhysReg(LR, Hint);
+        definePhysReg(MI, Hint0, regFree);
+      assignVirtToPhysReg(LR, Hint0);
       return;
+    } else {
+      LLVM_DEBUG(dbgs() << "\tPreferred Register 0: " << printReg(Hint0, TRI)
+                        << "occupied\n");
     }
+  } else {
+    Hint0 = 0;
+  }
+
+  // Try other hint.
+  if (TargetRegisterInfo::isPhysicalRegister(Hint1) &&
+      MRI->isAllocatable(Hint1) && RC.contains(Hint1)) {
+    // Ignore the hint if we would have to spill a dirty register.
+    unsigned Cost = calcSpillCost(Hint1);
+    if (Cost < spillDirty) {
+      LLVM_DEBUG(dbgs() << "\tPreferred Register 1: " << printReg(Hint1, TRI)
+                        << '\n');
+      if (Cost)
+        definePhysReg(MI, Hint1, regFree);
+      assignVirtToPhysReg(LR, Hint1);
+      return;
+    } else {
+      LLVM_DEBUG(dbgs() << "\tPreferred Register 1: " << printReg(Hint1, TRI)
+                        << "occupied\n");
+    }
+  } else {
+    Hint1 = 0;
   }
 
   MCPhysReg BestReg = 0;
   unsigned BestCost = spillImpossible;
   ArrayRef<MCPhysReg> AllocationOrder = RegClassInfo.getOrder(&RC);
   for (MCPhysReg PhysReg : AllocationOrder) {
     LLVM_DEBUG(dbgs() << "\tRegister: " << printReg(PhysReg, TRI) << ' ');
     unsigned Cost = calcSpillCost(PhysReg);
     LLVM_DEBUG(dbgs() << "Cost: " << Cost << " BestCost: " << BestCost << '\n');
     // Immediate take a register with cost 0.
     if (Cost == 0) {
       assignVirtToPhysReg(LR, PhysReg);
       return;
     }
+
+    if (PhysReg == Hint0 || PhysReg == Hint1)
+      Cost -= spillPrefBonus;
+
     if (Cost < BestCost) {
       BestReg = PhysReg;
       BestCost = Cost;
     }
   }
 
   if (!BestReg) {
     // Nothing we can do: Report an error and keep going with an invalid

test/CodeGen/AArch64/arm64-fast-isel-br.ll

if.end8:                                          ; preds = %if.then7, %if.end4
   ret i8 %3
 }
 
 declare void @foo1()
 
 ; rdar://15174028
 define i32 @trunc64(i64 %foo) nounwind {
 ; CHECK: trunc64
-; CHECK: and  [[REG1:x[0-9]+]], x0, #0x1
-; CHECK: mov  x[[REG2:[0-9]+]], [[REG1]]
-; CHECK: tbz w[[REG2]], #0, LBB5_2
+; CHECK: and x[[REG1:[0-9]+]], x0, #0x1
+; CHECK: tbz w[[REG1]], #0, LBB5_2
   %a = and i64 %foo, 1
   %b = trunc i64 %a to i1
   br i1 %b, label %if.then, label %if.else
 
 if.then:
   ret i32 1
 
 if.else:
   ret i32 0
 }

test/CodeGen/AArch64/arm64-fast-isel-conversion.ll

 ; RUN: llc -O0 -fast-isel -fast-isel-abort=1 -verify-machineinstrs -mtriple=arm64-apple-darwin -mcpu=cyclone < %s | FileCheck %s
 
 ;; Test various conversions.
 define zeroext i32 @trunc_(i8 zeroext %a, i16 zeroext %b, i32 %c, i64 %d) nounwind ssp {
 entry:
 ; CHECK-LABEL: trunc_
 ; CHECK: sub sp, sp, #16
 ; CHECK: strb w0, [sp, #15]
 ; CHECK: strh w1, [sp, #12]
 ; CHECK: str w2, [sp, #8]
 ; CHECK: str x3, [sp]
 ; CHECK: ldr x8, [sp]
-; CHECK: mov x9, x8
-; CHECK: str w9, [sp, #8]
-; CHECK: ldr w9, [sp, #8]
-; CHECK: strh w9, [sp, #12]
-; CHECK: ldrh w9, [sp, #12]
-; CHECK: strb w9, [sp, #15]
+; CHECK: ; kill: def $w8 killed $w8 killed $x8
+; CHECK: str w8, [sp, #8]
+; CHECK: ldr w8, [sp, #8]
+; CHECK: strh w8, [sp, #12]
+; CHECK: ldrh w8, [sp, #12]
+; CHECK: strb w8, [sp, #15]
 ; CHECK: ldrb w0, [sp, #15]
 ; CHECK: add sp, sp, #16
 ; CHECK: ret
   %a.addr = alloca i8, align 1
   %b.addr = alloca i16, align 2
   %c.addr = alloca i32, align 4
   %d.addr = alloca i64, align 8
   store i8 %a, i8* %a.addr, align 1

 ; CHECK: ucvtf d0, x0
   %conv = uitofp i64 %a to double
   ret double %conv
 }
 
 define i32 @i64_trunc_i32(i64 %a) nounwind ssp {
 entry:
 ; CHECK-LABEL: i64_trunc_i32
-; CHECK: mov x1, x0
+; CHECK-NOT: mov
+; CHECK: ret
   %conv = trunc i64 %a to i32
   ret i32 %conv
 }
 
 define zeroext i16 @i64_trunc_i16(i64 %a) nounwind ssp {
 entry:
 ; CHECK-LABEL: i64_trunc_i16
-; CHECK: mov x[[REG:[0-9]+]], x0
-; CHECK: and [[REG2:w[0-9]+]], w[[REG]], #0xffff
+; CHECK: and [[REG2:w[0-9]+]], w0, #0xffff
 ; CHECK: uxth w0, [[REG2]]
   %conv = trunc i64 %a to i16
   ret i16 %conv
 }
 
 define zeroext i8 @i64_trunc_i8(i64 %a) nounwind ssp {
 entry:
 ; CHECK-LABEL: i64_trunc_i8
-; CHECK: mov x[[REG:[0-9]+]], x0
-; CHECK: and [[REG2:w[0-9]+]], w[[REG]], #0xff
+; CHECK: and [[REG2:w[0-9]+]], w0, #0xff
 ; CHECK: uxtb w0, [[REG2]]
   %conv = trunc i64 %a to i8
   ret i8 %conv
 }
 
 define zeroext i1 @i64_trunc_i1(i64 %a) nounwind ssp {
 entry:
 ; CHECK-LABEL: i64_trunc_i1
-; CHECK: mov x[[REG:[0-9]+]], x0
-; CHECK: and [[REG2:w[0-9]+]], w[[REG]], #0x1
+; CHECK: and [[REG2:w[0-9]+]], w0, #0x1
 ; CHECK: and w0, [[REG2]], #0x1
   %conv = trunc i64 %a to i1
   ret i1 %conv
 }
 
 ; rdar://15101939
 define void @stack_trunc() nounwind {
 ; CHECK-LABEL: stack_trunc
 ; CHECK: sub  sp, sp, #16
-; CHECK: ldr  [[REG:x[0-9]+]], [sp]
-; CHECK: mov  x[[REG2:[0-9]+]], [[REG]]
-; CHECK: and  [[REG3:w[0-9]+]], w[[REG2]], #0xff
+; CHECK: ldr  x[[REG:[0-9]+]], [sp]
+; CHECK: and  [[REG3:w[0-9]+]], w[[REG]], #0xff
 ; CHECK: strb [[REG3]], [sp, #15]
 ; CHECK: add  sp, sp, #16
   %a = alloca i8, align 1
   %b = alloca i64, align 8
   %c = load i64, i64* %b, align 8
   %d = trunc i64 %c to i8
   store i8 %d, i8* %a, align 1
   ret void

test/CodeGen/AMDGPU/indirect-addressing-term.ll

 ; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
 ; RUN: llc -O0 -amdgpu-scalarize-global-loads=false -march=amdgcn -mcpu=gfx900 -mattr=-flat-for-global -verify-machineinstrs -stop-after=regallocfast < %s | FileCheck -check-prefixes=GCN %s
 
 ; Verify that we consider the xor at the end of the waterfall loop emitted for
 ; divergent indirect addressing as a terminator.
 
 declare i32 @llvm.amdgcn.workitem.id.x() #1
 
 ; There should be no spill code inserted between the xor and the real terminator
 define amdgpu_kernel void @extract_w_offset_vgpr(i32 addrspace(1)* %out) {
   ; GCN-LABEL: name: extract_w_offset_vgpr
   ; GCN: bb.0.entry:
   ; GCN:   successors: %bb.1(0x80000000)
   ; GCN:   liveins: $vgpr0, $sgpr0_sgpr1
   ; GCN:   renamable $sgpr0_sgpr1 = S_LOAD_DWORDX2_IMM killed renamable $sgpr0_sgpr1, 36, 0, 0 :: (dereferenceable invariant load 8 from %ir.out.kernarg.offset.cast, align 4, addrspace 4)
   ; GCN:   renamable $sgpr2 = COPY renamable $sgpr1
-  ; GCN:   renamable $sgpr4 = COPY renamable $sgpr0, implicit killed $sgpr0_sgpr1
-  ; GCN:   renamable $sgpr5 = S_MOV_B32 61440
-  ; GCN:   renamable $sgpr6 = S_MOV_B32 -1
-  ; GCN:   undef renamable $sgpr8 = COPY killed renamable $sgpr4, implicit-def $sgpr8_sgpr9_sgpr10_sgpr11
+  ; GCN:   renamable $sgpr0 = COPY renamable $sgpr0, implicit killed $sgpr0_sgpr1
+  ; GCN:   renamable $sgpr1 = S_MOV_B32 61440
+  ; GCN:   renamable $sgpr4 = S_MOV_B32 -1
+  ; GCN:   undef renamable $sgpr8 = COPY killed renamable $sgpr0, implicit-def $sgpr8_sgpr9_sgpr10_sgpr11
   ; GCN:   renamable $sgpr9 = COPY killed renamable $sgpr2
-  ; GCN:   renamable $sgpr10 = COPY killed renamable $sgpr6
-  ; GCN:   renamable $sgpr11 = COPY killed renamable $sgpr5
-  ; GCN:   renamable $sgpr2 = S_MOV_B32 16
-  ; GCN:   renamable $sgpr4 = S_MOV_B32 15
-  ; GCN:   renamable $sgpr5 = S_MOV_B32 14
-  ; GCN:   renamable $sgpr6 = S_MOV_B32 13
-  ; GCN:   renamable $sgpr7 = S_MOV_B32 12
-  ; GCN:   renamable $sgpr12 = S_MOV_B32 11
-  ; GCN:   renamable $sgpr13 = S_MOV_B32 10
-  ; GCN:   renamable $sgpr14 = S_MOV_B32 9
-  ; GCN:   renamable $sgpr15 = S_MOV_B32 8
-  ; GCN:   renamable $sgpr16 = S_MOV_B32 7
-  ; GCN:   renamable $sgpr17 = S_MOV_B32 6
-  ; GCN:   renamable $sgpr18 = S_MOV_B32 5
-  ; GCN:   renamable $sgpr19 = S_MOV_B32 3
-  ; GCN:   renamable $sgpr20 = S_MOV_B32 2
-  ; GCN:   renamable $sgpr21 = S_MOV_B32 1
-  ; GCN:   renamable $sgpr22 = S_MOV_B32 0
-  ; GCN:   renamable $vgpr1 = COPY killed renamable $sgpr22
-  ; GCN:   renamable $vgpr2 = COPY killed renamable $sgpr21
-  ; GCN:   renamable $vgpr3 = COPY killed renamable $sgpr20
-  ; GCN:   renamable $vgpr4 = COPY killed renamable $sgpr19
-  ; GCN:   renamable $vgpr5 = COPY killed renamable $sgpr18
-  ; GCN:   renamable $vgpr6 = COPY killed renamable $sgpr17
-  ; GCN:   renamable $vgpr7 = COPY killed renamable $sgpr16
-  ; GCN:   renamable $vgpr8 = COPY killed renamable $sgpr15
-  ; GCN:   renamable $vgpr9 = COPY killed renamable $sgpr14
-  ; GCN:   renamable $vgpr10 = COPY killed renamable $sgpr13
-  ; GCN:   renamable $vgpr11 = COPY killed renamable $sgpr12
-  ; GCN:   renamable $vgpr12 = COPY killed renamable $sgpr7
-  ; GCN:   renamable $vgpr13 = COPY killed renamable $sgpr6
-  ; GCN:   renamable $vgpr14 = COPY killed renamable $sgpr5
-  ; GCN:   renamable $vgpr15 = COPY killed renamable $sgpr4
-  ; GCN:   renamable $vgpr16 = COPY killed renamable $sgpr2
+  ; GCN:   renamable $sgpr10 = COPY killed renamable $sgpr4
+  ; GCN:   renamable $sgpr11 = COPY killed renamable $sgpr1
+  ; GCN:   renamable $sgpr0 = S_MOV_B32 16
+  ; GCN:   renamable $sgpr1 = S_MOV_B32 15
+  ; GCN:   renamable $sgpr2 = S_MOV_B32 14
+  ; GCN:   renamable $sgpr4 = S_MOV_B32 13
+  ; GCN:   renamable $sgpr5 = S_MOV_B32 12
+  ; GCN:   renamable $sgpr6 = S_MOV_B32 11
+  ; GCN:   renamable $sgpr7 = S_MOV_B32 10
+  ; GCN:   renamable $sgpr12 = S_MOV_B32 9
+  ; GCN:   renamable $sgpr13 = S_MOV_B32 8
+  ; GCN:   renamable $sgpr14 = S_MOV_B32 7
+  ; GCN:   renamable $sgpr15 = S_MOV_B32 6
+  ; GCN:   renamable $sgpr16 = S_MOV_B32 5
+  ; GCN:   renamable $sgpr17 = S_MOV_B32 3
+  ; GCN:   renamable $sgpr18 = S_MOV_B32 2
+  ; GCN:   renamable $sgpr19 = S_MOV_B32 1
+  ; GCN:   renamable $sgpr20 = S_MOV_B32 0
+  ; GCN:   renamable $vgpr1 = COPY killed renamable $sgpr20
+  ; GCN:   renamable $vgpr2 = COPY killed renamable $sgpr19
+  ; GCN:   renamable $vgpr3 = COPY killed renamable $sgpr18
+  ; GCN:   renamable $vgpr4 = COPY killed renamable $sgpr17
+  ; GCN:   renamable $vgpr5 = COPY killed renamable $sgpr16
+  ; GCN:   renamable $vgpr6 = COPY killed renamable $sgpr15
+  ; GCN:   renamable $vgpr7 = COPY killed renamable $sgpr14
+  ; GCN:   renamable $vgpr8 = COPY killed renamable $sgpr13
+  ; GCN:   renamable $vgpr9 = COPY killed renamable $sgpr12
+  ; GCN:   renamable $vgpr10 = COPY killed renamable $sgpr7
+  ; GCN:   renamable $vgpr11 = COPY killed renamable $sgpr6
+  ; GCN:   renamable $vgpr12 = COPY killed renamable $sgpr5
+  ; GCN:   renamable $vgpr13 = COPY killed renamable $sgpr4
+  ; GCN:   renamable $vgpr14 = COPY killed renamable $sgpr2
+  ; GCN:   renamable $vgpr15 = COPY killed renamable $sgpr1
+  ; GCN:   renamable $vgpr16 = COPY killed renamable $sgpr0
   ; GCN:   undef renamable $vgpr17 = COPY killed renamable $vgpr1, implicit-def $vgpr17_vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31_vgpr32
   ; GCN:   renamable $vgpr18 = COPY killed renamable $vgpr2
   ; GCN:   renamable $vgpr19 = COPY killed renamable $vgpr3
   ; GCN:   renamable $vgpr20 = COPY killed renamable $vgpr4
   ; GCN:   renamable $vgpr21 = COPY killed renamable $vgpr5
   ; GCN:   renamable $vgpr22 = COPY killed renamable $vgpr6
   ; GCN:   renamable $vgpr23 = COPY killed renamable $vgpr7
   ; GCN:   renamable $vgpr24 = COPY killed renamable $vgpr8
   ; GCN:   renamable $vgpr25 = COPY killed renamable $vgpr9
   ; GCN:   renamable $vgpr26 = COPY killed renamable $vgpr10
   ; GCN:   renamable $vgpr27 = COPY killed renamable $vgpr11
   ; GCN:   renamable $vgpr28 = COPY killed renamable $vgpr12
   ; GCN:   renamable $vgpr29 = COPY killed renamable $vgpr13
   ; GCN:   renamable $vgpr30 = COPY killed renamable $vgpr14
   ; GCN:   renamable $vgpr31 = COPY killed renamable $vgpr15
   ; GCN:   renamable $vgpr32 = COPY killed renamable $vgpr16
-  ; GCN:   renamable $sgpr0_sgpr1 = S_MOV_B64 $exec
+  ; GCN:   renamable $sgpr22_sgpr23 = S_MOV_B64 $exec
   ; GCN:   renamable $vgpr1 = IMPLICIT_DEF
   ; GCN:   renamable $sgpr24_sgpr25 = IMPLICIT_DEF
   ; GCN:   SI_SPILL_V32_SAVE killed $vgpr0, %stack.0, $sgpr96_sgpr97_sgpr98_sgpr99, $sgpr3, 0, implicit $exec :: (store 4 into %stack.0, addrspace 5)
   ; GCN:   SI_SPILL_S128_SAVE killed $sgpr8_sgpr9_sgpr10_sgpr11, %stack.1, implicit $exec, implicit $sgpr96_sgpr97_sgpr98_sgpr99, implicit $sgpr3, implicit-def dead $m0 :: (store 16 into %stack.1, align 4, addrspace 5)
   ; GCN:   SI_SPILL_V512_SAVE killed $vgpr17_vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31_vgpr32, %stack.2, $sgpr96_sgpr97_sgpr98_sgpr99, $sgpr3, 0, implicit $exec :: (store 64 into %stack.2, align 4, addrspace 5)
-  ; GCN:   SI_SPILL_S64_SAVE killed $sgpr0_sgpr1, %stack.3, implicit $exec, implicit $sgpr96_sgpr97_sgpr98_sgpr99, implicit $sgpr3, implicit-def dead $m0 :: (store 8 into %stack.3, align 4, addrspace 5)
+  ; GCN:   SI_SPILL_S64_SAVE killed $sgpr22_sgpr23, %stack.3, implicit $exec, implicit $sgpr96_sgpr97_sgpr98_sgpr99, implicit $sgpr3, implicit-def dead $m0 :: (store 8 into %stack.3, align 4, addrspace 5)
   ; GCN:   SI_SPILL_V32_SAVE killed $vgpr1, %stack.4, $sgpr96_sgpr97_sgpr98_sgpr99, $sgpr3, 0, implicit $exec :: (store 4 into %stack.4, addrspace 5)
   ; GCN:   SI_SPILL_S64_SAVE killed $sgpr24_sgpr25, %stack.5, implicit $exec, implicit $sgpr96_sgpr97_sgpr98_sgpr99, implicit $sgpr3, implicit-def dead $m0 :: (store 8 into %stack.5, align 4, addrspace 5)
   ; GCN: bb.1:
   ; GCN:   successors: %bb.1(0x40000000), %bb.2(0x40000000)
   ; GCN:   $sgpr0_sgpr1 = SI_SPILL_S64_RESTORE %stack.5, implicit $exec, implicit $sgpr96_sgpr97_sgpr98_sgpr99, implicit $sgpr3, implicit-def dead $m0 :: (load 8 from %stack.5, align 4, addrspace 5)
   ; GCN:   $vgpr0 = SI_SPILL_V32_RESTORE %stack.4, $sgpr96_sgpr97_sgpr98_sgpr99, $sgpr3, 0, implicit $exec :: (load 4 from %stack.4, addrspace 5)
   ; GCN:   $vgpr1 = SI_SPILL_V32_RESTORE %stack.0, $sgpr96_sgpr97_sgpr98_sgpr99, $sgpr3, 0, implicit $exec :: (load 4 from %stack.0, addrspace 5)
   ; GCN:   renamable $sgpr2 = V_READFIRSTLANE_B32 $vgpr1, implicit $exec

test/CodeGen/Mips/atomic.ll

 ; MIPS64R6-NEXT:  # %bb.2: # %entry
 ; MIPS64R6-NEXT:    jrc $ra
 ;
 ; MIPS64R6O0-LABEL: AtomicLoadAdd32:
 ; MIPS64R6O0:       # %bb.0: # %entry
 ; MIPS64R6O0-NEXT:    lui $1, %hi(%neg(%gp_rel(AtomicLoadAdd32)))
 ; MIPS64R6O0-NEXT:    daddu $1, $1, $25
 ; MIPS64R6O0-NEXT:    daddiu $1, $1, %lo(%neg(%gp_rel(AtomicLoadAdd32)))
-; MIPS64R6O0-NEXT:    move $2, $4
+; MIPS64R6O0-NEXT:    # kill: def $a0 killed $a0 killed $a0_64
 ; MIPS64R6O0-NEXT:    ld $1, %got_disp(x)($1)
 ; MIPS64R6O0-NEXT:  .LBB0_1: # %entry
 ; MIPS64R6O0-NEXT:    # =>This Inner Loop Header: Depth=1
-; MIPS64R6O0-NEXT:    ll $3, 0($1)
-; MIPS64R6O0-NEXT:    addu $5, $3, $2
-; MIPS64R6O0-NEXT:    sc $5, 0($1)
-; MIPS64R6O0-NEXT:    beqzc $5, .LBB0_1
+; MIPS64R6O0-NEXT:    ll $2, 0($1)
+; MIPS64R6O0-NEXT:    addu $3, $2, $4
+; MIPS64R6O0-NEXT:    sc $3, 0($1)
+; MIPS64R6O0-NEXT:    beqzc $3, .LBB0_1
+; MIPS64R6O0-NEXT:    nop
 ; MIPS64R6O0-NEXT:  # %bb.2: # %entry
-; MIPS64R6O0-NEXT:    move $2, $3
 ; MIPS64R6O0-NEXT:    jrc $ra
 ;
 ; MM32-LABEL: AtomicLoadAdd32:
 ; MM32:       # %bb.0: # %entry
 ; MM32-NEXT:    lui $2, %hi(_gp_disp)
 ; MM32-NEXT:    addiu $2, $2, %lo(_gp_disp)
 ; MM32-NEXT:    addu $2, $2, $25
 ; MM32-NEXT:    lw $1, %got(x)($2)

 ; MIPS64R6-NEXT:  # %bb.2: # %entry
 ; MIPS64R6-NEXT:    jrc $ra
 ;
 ; MIPS64R6O0-LABEL: AtomicLoadSub32:
 ; MIPS64R6O0:       # %bb.0: # %entry
 ; MIPS64R6O0-NEXT:    lui $1, %hi(%neg(%gp_rel(AtomicLoadSub32)))
 ; MIPS64R6O0-NEXT:    daddu $1, $1, $25
 ; MIPS64R6O0-NEXT:    daddiu $1, $1, %lo(%neg(%gp_rel(AtomicLoadSub32)))
-; MIPS64R6O0-NEXT:    move $2, $4
+; MIPS64R6O0-NEXT:    # kill: def $a0 killed $a0 killed $a0_64
 ; MIPS64R6O0-NEXT:    ld $1, %got_disp(x)($1)
 ; MIPS64R6O0-NEXT:  .LBB1_1: # %entry
 ; MIPS64R6O0-NEXT:    # =>This Inner Loop Header: Depth=1
-; MIPS64R6O0-NEXT:    ll $3, 0($1)
-; MIPS64R6O0-NEXT:    subu $5, $3, $2
-; MIPS64R6O0-NEXT:    sc $5, 0($1)
-; MIPS64R6O0-NEXT:    beqzc $5, .LBB1_1
+; MIPS64R6O0-NEXT:    ll $2, 0($1)
+; MIPS64R6O0-NEXT:    subu $3, $2, $4
+; MIPS64R6O0-NEXT:    sc $3, 0($1)
+; MIPS64R6O0-NEXT:    beqzc $3, .LBB1_1
+; MIPS64R6O0-NEXT:    nop
 ; MIPS64R6O0-NEXT:  # %bb.2: # %entry
-; MIPS64R6O0-NEXT:    move $2, $3
 ; MIPS64R6O0-NEXT:    jrc $ra
 ;
 ; MM32-LABEL: AtomicLoadSub32:
 ; MM32:       # %bb.0: # %entry
 ; MM32-NEXT:    lui $2, %hi(_gp_disp)
 ; MM32-NEXT:    addiu $2, $2, %lo(_gp_disp)
 ; MM32-NEXT:    addu $2, $2, $25
 ; MM32-NEXT:    lw $1, %got(x)($2)

 ; MIPS64R6-NEXT:  # %bb.2: # %entry
 ; MIPS64R6-NEXT:    jrc $ra
 ;
 ; MIPS64R6O0-LABEL: AtomicLoadXor32:
 ; MIPS64R6O0:       # %bb.0: # %entry
 ; MIPS64R6O0-NEXT:    lui $1, %hi(%neg(%gp_rel(AtomicLoadXor32)))
 ; MIPS64R6O0-NEXT:    daddu $1, $1, $25
 ; MIPS64R6O0-NEXT:    daddiu $1, $1, %lo(%neg(%gp_rel(AtomicLoadXor32)))
-; MIPS64R6O0-NEXT:    move $2, $4
+; MIPS64R6O0-NEXT:    # kill: def $a0 killed $a0 killed $a0_64
 ; MIPS64R6O0-NEXT:    ld $1, %got_disp(x)($1)
 ; MIPS64R6O0-NEXT:  .LBB2_1: # %entry
 ; MIPS64R6O0-NEXT:    # =>This Inner Loop Header: Depth=1
-; MIPS64R6O0-NEXT:    ll $3, 0($1)
-; MIPS64R6O0-NEXT:    xor $5, $3, $2
-; MIPS64R6O0-NEXT:    sc $5, 0($1)
-; MIPS64R6O0-NEXT:    beqzc $5, .LBB2_1
+; MIPS64R6O0-NEXT:    ll $2, 0($1)
+; MIPS64R6O0-NEXT:    xor $3, $2, $4
+; MIPS64R6O0-NEXT:    sc $3, 0($1)
+; MIPS64R6O0-NEXT:    beqzc $3, .LBB2_1
+; MIPS64R6O0-NEXT:    nop
 ; MIPS64R6O0-NEXT:  # %bb.2: # %entry
-; MIPS64R6O0-NEXT:    move $2, $3
 ; MIPS64R6O0-NEXT:    jrc $ra
 ;
 ; MM32-LABEL: AtomicLoadXor32:
 ; MM32:       # %bb.0: # %entry
 ; MM32-NEXT:    lui $2, %hi(_gp_disp)
 ; MM32-NEXT:    addiu $2, $2, %lo(_gp_disp)
 ; MM32-NEXT:    addu $2, $2, $25
 ; MM32-NEXT:    lw $1, %got(x)($2)

 ; MIPS64R6-NEXT:  # %bb.2: # %entry
 ; MIPS64R6-NEXT:    jrc $ra
 ;
 ; MIPS64R6O0-LABEL: AtomicLoadOr32:
 ; MIPS64R6O0:       # %bb.0: # %entry
 ; MIPS64R6O0-NEXT:    lui $1, %hi(%neg(%gp_rel(AtomicLoadOr32)))
 ; MIPS64R6O0-NEXT:    daddu $1, $1, $25
 ; MIPS64R6O0-NEXT:    daddiu $1, $1, %lo(%neg(%gp_rel(AtomicLoadOr32)))
-; MIPS64R6O0-NEXT:    move $2, $4
+; MIPS64R6O0-NEXT:    # kill: def $a0 killed $a0 killed $a0_64
 ; MIPS64R6O0-NEXT:    ld $1, %got_disp(x)($1)
 ; MIPS64R6O0-NEXT:  .LBB3_1: # %entry
 ; MIPS64R6O0-NEXT:    # =>This Inner Loop Header: Depth=1
-; MIPS64R6O0-NEXT:    ll $3, 0($1)
-; MIPS64R6O0-NEXT:    or $5, $3, $2
-; MIPS64R6O0-NEXT:    sc $5, 0($1)
-; MIPS64R6O0-NEXT:    beqzc $5, .LBB3_1
+; MIPS64R6O0-NEXT:    ll $2, 0($1)
+; MIPS64R6O0-NEXT:    or $3, $2, $4
+; MIPS64R6O0-NEXT:    sc $3, 0($1)
+; MIPS64R6O0-NEXT:    beqzc $3, .LBB3_1
+; MIPS64R6O0-NEXT:    nop
 ; MIPS64R6O0-NEXT:  # %bb.2: # %entry
-; MIPS64R6O0-NEXT:    move $2, $3
 ; MIPS64R6O0-NEXT:    jrc $ra
 ;
 ; MM32-LABEL: AtomicLoadOr32:
 ; MM32:       # %bb.0: # %entry
 ; MM32-NEXT:    lui $2, %hi(_gp_disp)
 ; MM32-NEXT:    addiu $2, $2, %lo(_gp_disp)
 ; MM32-NEXT:    addu $2, $2, $25
 ; MM32-NEXT:    lw $1, %got(x)($2)

 ; MIPS64R6-NEXT:  # %bb.2: # %entry
 ; MIPS64R6-NEXT:    jrc $ra
 ;
 ; MIPS64R6O0-LABEL: AtomicLoadAnd32:
 ; MIPS64R6O0:       # %bb.0: # %entry
 ; MIPS64R6O0-NEXT:    lui $1, %hi(%neg(%gp_rel(AtomicLoadAnd32)))
 ; MIPS64R6O0-NEXT:    daddu $1, $1, $25
 ; MIPS64R6O0-NEXT:    daddiu $1, $1, %lo(%neg(%gp_rel(AtomicLoadAnd32)))
-; MIPS64R6O0-NEXT:    move $2, $4
+; MIPS64R6O0-NEXT:    # kill: def $a0 killed $a0 killed $a0_64
 ; MIPS64R6O0-NEXT:    ld $1, %got_disp(x)($1)
 ; MIPS64R6O0-NEXT:  .LBB4_1: # %entry
 ; MIPS64R6O0-NEXT:    # =>This Inner Loop Header: Depth=1
-; MIPS64R6O0-NEXT:    ll $3, 0($1)
-; MIPS64R6O0-NEXT:    and $5, $3, $2
-; MIPS64R6O0-NEXT:    sc $5, 0($1)
-; MIPS64R6O0-NEXT:    beqzc $5, .LBB4_1
+; MIPS64R6O0-NEXT:    ll $2, 0($1)
+; MIPS64R6O0-NEXT:    and $3, $2, $4
+; MIPS64R6O0-NEXT:    sc $3, 0($1)
+; MIPS64R6O0-NEXT:    beqzc $3, .LBB4_1
+; MIPS64R6O0-NEXT:    nop
 ; MIPS64R6O0-NEXT:  # %bb.2: # %entry
-; MIPS64R6O0-NEXT:    move $2, $3
 ; MIPS64R6O0-NEXT:    jrc $ra
 ;
 ; MM32-LABEL: AtomicLoadAnd32:
 ; MM32:       # %bb.0: # %entry
 ; MM32-NEXT:    lui $2, %hi(_gp_disp)
 ; MM32-NEXT:    addiu $2, $2, %lo(_gp_disp)
 ; MM32-NEXT:    addu $2, $2, $25
 ; MM32-NEXT:    lw $1, %got(x)($2)

 ; MIPS64R6-NEXT:  # %bb.2: # %entry
 ; MIPS64R6-NEXT:    jrc $ra
 ;
 ; MIPS64R6O0-LABEL: AtomicLoadNand32:
 ; MIPS64R6O0:       # %bb.0: # %entry
 ; MIPS64R6O0-NEXT:    lui $1, %hi(%neg(%gp_rel(AtomicLoadNand32)))
 ; MIPS64R6O0-NEXT:    daddu $1, $1, $25
 ; MIPS64R6O0-NEXT:    daddiu $1, $1, %lo(%neg(%gp_rel(AtomicLoadNand32)))
-; MIPS64R6O0-NEXT:    move $2, $4
+; MIPS64R6O0-NEXT:    # kill: def $a0 killed $a0 killed $a0_64
 ; MIPS64R6O0-NEXT:    ld $1, %got_disp(x)($1)
 ; MIPS64R6O0-NEXT:  .LBB5_1: # %entry
 ; MIPS64R6O0-NEXT:    # =>This Inner Loop Header: Depth=1
-; MIPS64R6O0-NEXT:    ll $3, 0($1)
-; MIPS64R6O0-NEXT:    and $5, $3, $2
-; MIPS64R6O0-NEXT:    nor $5, $zero, $5
-; MIPS64R6O0-NEXT:    sc $5, 0($1)
-; MIPS64R6O0-NEXT:    beqzc $5, .LBB5_1
+; MIPS64R6O0-NEXT:    ll $2, 0($1)
+; MIPS64R6O0-NEXT:    and $3, $2, $4
+; MIPS64R6O0-NEXT:    nor $3, $zero, $3
+; MIPS64R6O0-NEXT:    sc $3, 0($1)
+; MIPS64R6O0-NEXT:    beqzc $3, .LBB5_1
+; MIPS64R6O0-NEXT:    nop
 ; MIPS64R6O0-NEXT:  # %bb.2: # %entry
-; MIPS64R6O0-NEXT:    move $2, $3
 ; MIPS64R6O0-NEXT:    jrc $ra
 ;
 ; MM32-LABEL: AtomicLoadNand32:
 ; MM32:       # %bb.0: # %entry
 ; MM32-NEXT:    lui $2, %hi(_gp_disp)
 ; MM32-NEXT:    addiu $2, $2, %lo(_gp_disp)
 ; MM32-NEXT:    addu $2, $2, $25
 ; MM32-NEXT:    lw $1, %got(x)($2)

 ; MIPS64R6-NEXT:    daddiu $sp, $sp, 16
 ;
 ; MIPS64R6O0-LABEL: AtomicSwap32:
 ; MIPS64R6O0:       # %bb.0: # %entry
 ; MIPS64R6O0-NEXT:    daddiu $sp, $sp, -16
 ; MIPS64R6O0-NEXT:    lui $1, %hi(%neg(%gp_rel(AtomicSwap32)))
 ; MIPS64R6O0-NEXT:    daddu $1, $1, $25
 ; MIPS64R6O0-NEXT:    daddiu $1, $1, %lo(%neg(%gp_rel(AtomicSwap32)))
-; MIPS64R6O0-NEXT:    move $2, $4
-; MIPS64R6O0-NEXT:    sw $2, 12($sp)
+; MIPS64R6O0-NEXT:    # kill: def $a0 killed $a0 killed $a0_64
+; MIPS64R6O0-NEXT:    sw $4, 12($sp)
 ; MIPS64R6O0-NEXT:    lw $2, 12($sp)
 ; MIPS64R6O0-NEXT:    ld $1, %got_disp(x)($1)
 ; MIPS64R6O0-NEXT:  .LBB6_1: # %entry
 ; MIPS64R6O0-NEXT:    # =>This Inner Loop Header: Depth=1
 ; MIPS64R6O0-NEXT:    ll $3, 0($1)
-; MIPS64R6O0-NEXT:    move $5, $2
-; MIPS64R6O0-NEXT:    sc $5, 0($1)
-; MIPS64R6O0-NEXT:    beqzc $5, .LBB6_1
+; MIPS64R6O0-NEXT:    move $4, $2
+; MIPS64R6O0-NEXT:    sc $4, 0($1)
+; MIPS64R6O0-NEXT:    beqzc $4, .LBB6_1
 ; MIPS64R6O0-NEXT:  # %bb.2: # %entry
 ; MIPS64R6O0-NEXT:    move $2, $3
 ; MIPS64R6O0-NEXT:    daddiu $sp, $sp, 16
 ; MIPS64R6O0-NEXT:    jrc $ra
 ;
 ; MM32-LABEL: AtomicSwap32:
 ; MM32:       # %bb.0: # %entry
 ; MM32-NEXT:    lui $2, %hi(_gp_disp)

 ; MIPS64R6-NEXT:    daddiu $sp, $sp, 16
 ;
 ; MIPS64R6O0-LABEL: AtomicCmpSwap32:
 ; MIPS64R6O0:       # %bb.0: # %entry
 ; MIPS64R6O0-NEXT:    daddiu $sp, $sp, -16
 ; MIPS64R6O0-NEXT:    lui $1, %hi(%neg(%gp_rel(AtomicCmpSwap32)))
 ; MIPS64R6O0-NEXT:    daddu $1, $1, $25
 ; MIPS64R6O0-NEXT:    daddiu $1, $1, %lo(%neg(%gp_rel(AtomicCmpSwap32)))
-; MIPS64R6O0-NEXT:    move $2, $4
-; MIPS64R6O0-NEXT:    move $3, $5
-; MIPS64R6O0-NEXT:    sw $3, 12($sp)
-; MIPS64R6O0-NEXT:    lw $3, 12($sp)
+; MIPS64R6O0-NEXT:    # kill: def $a0 killed $a0 killed $a0_64
+; MIPS64R6O0-NEXT:    # kill: def $a1 killed $a1 killed $a1_64
+; MIPS64R6O0-NEXT:    sw $5, 12($sp)
+; MIPS64R6O0-NEXT:    lw $2, 12($sp)
 ; MIPS64R6O0-NEXT:    ld $1, %got_disp(x)($1)
 ; MIPS64R6O0-NEXT:  .LBB7_1: # %entry
 ; MIPS64R6O0-NEXT:    # =>This Inner Loop Header: Depth=1
-; MIPS64R6O0-NEXT:    ll $6, 0($1)
-; MIPS64R6O0-NEXT:    bnec $6, $2, .LBB7_3
+; MIPS64R6O0-NEXT:    ll $3, 0($1)
+; MIPS64R6O0-NEXT:    bnec $3, $4, .LBB7_3
 ; MIPS64R6O0-NEXT:  # %bb.2: # %entry
 ; MIPS64R6O0-NEXT:    # in Loop: Header=BB7_1 Depth=1
-; MIPS64R6O0-NEXT:    move $7, $3
-; MIPS64R6O0-NEXT:    sc $7, 0($1)
-; MIPS64R6O0-NEXT:    beqzc $7, .LBB7_1
+; MIPS64R6O0-NEXT:    move $5, $2
+; MIPS64R6O0-NEXT:    sc $5, 0($1)
+; MIPS64R6O0-NEXT:    beqzc $5, .LBB7_1
 ; MIPS64R6O0-NEXT:  .LBB7_3: # %entry
-; MIPS64R6O0-NEXT:    move $2, $6
+; MIPS64R6O0-NEXT:    move $2, $3
 ; MIPS64R6O0-NEXT:    daddiu $sp, $sp, 16
 ; MIPS64R6O0-NEXT:    jrc $ra
 ;
 ; MM32-LABEL: AtomicCmpSwap32:
 ; MM32:       # %bb.0: # %entry
 ; MM32-NEXT:    lui $2, %hi(_gp_disp)
 ; MM32-NEXT:    addiu $2, $2, %lo(_gp_disp)
 ; MM32-NEXT:    addiu $sp, $sp, -8

 ; MIPS64R6-NEXT:    seb $2, $1
 ;
 ; MIPS64R6O0-LABEL: AtomicLoadAdd8:
 ; MIPS64R6O0:       # %bb.0: # %entry
 ; MIPS64R6O0-NEXT:    daddiu $sp, $sp, -16
 ; MIPS64R6O0-NEXT:    lui $1, %hi(%neg(%gp_rel(AtomicLoadAdd8)))
 ; MIPS64R6O0-NEXT:    daddu $1, $1, $25
 ; MIPS64R6O0-NEXT:    daddiu $1, $1, %lo(%neg(%gp_rel(AtomicLoadAdd8)))
-; MIPS64R6O0-NEXT:    move $2, $4
+; MIPS64R6O0-NEXT:    # kill: def $a0 killed $a0 killed $a0_64
 ; MIPS64R6O0-NEXT:    ld $1, %got_disp(y)($1)
-; MIPS64R6O0-NEXT:    daddiu $3, $zero, -4
-; MIPS64R6O0-NEXT:    and $3, $1, $3
-; MIPS64R6O0-NEXT:    andi $5, $1, 3
-; MIPS64R6O0-NEXT:    xori $5, $5, 3
-; MIPS64R6O0-NEXT:    sll $5, $5, 3
-; MIPS64R6O0-NEXT:    ori $6, $zero, 255
-; MIPS64R6O0-NEXT:    sllv $6, $6, $5
-; MIPS64R6O0-NEXT:    nor $7, $zero, $6
-; MIPS64R6O0-NEXT:    sllv $2, $2, $5
+; MIPS64R6O0-NEXT:    daddiu $2, $zero, -4
+; MIPS64R6O0-NEXT:    and $2, $1, $2
+; MIPS64R6O0-NEXT:    andi $3, $1, 3
+; MIPS64R6O0-NEXT:    xori $3, $3, 3
+; MIPS64R6O0-NEXT:    sll $3, $3, 3
+; MIPS64R6O0-NEXT:    ori $5, $zero, 255
+; MIPS64R6O0-NEXT:    sllv $5, $5, $3
+; MIPS64R6O0-NEXT:    nor $6, $zero, $5
+; MIPS64R6O0-NEXT:    sllv $4, $4, $3
 ; MIPS64R6O0-NEXT:  .LBB8_1: # %entry
 ; MIPS64R6O0-NEXT:    # =>This Inner Loop Header: Depth=1
-; MIPS64R6O0-NEXT:    ll $9, 0($3)
-; MIPS64R6O0-NEXT:    addu $10, $9, $2
-; MIPS64R6O0-NEXT:    and $10, $10, $6
-; MIPS64R6O0-NEXT:    and $11, $9, $7
-; MIPS64R6O0-NEXT:    or $11, $11, $10
-; MIPS64R6O0-NEXT:    sc $11, 0($3)
-; MIPS64R6O0-NEXT:    beqzc $11, .LBB8_1
+; MIPS64R6O0-NEXT:    ll $8, 0($2)
+; MIPS64R6O0-NEXT:    addu $9, $8, $4
+; MIPS64R6O0-NEXT:    and $9, $9, $5
+; MIPS64R6O0-NEXT:    and $10, $8, $6
+; MIPS64R6O0-NEXT:    or $10, $10, $9
+; MIPS64R6O0-NEXT:    sc $10, 0($2)
+; MIPS64R6O0-NEXT:    beqzc $10, .LBB8_1
 ; MIPS64R6O0-NEXT:  # %bb.2: # %entry
-; MIPS64R6O0-NEXT:    and $8, $9, $6
-; MIPS64R6O0-NEXT:    srlv $8, $8, $5
-; MIPS64R6O0-NEXT:    seb $8, $8
+; MIPS64R6O0-NEXT:    and $7, $8, $5
+; MIPS64R6O0-NEXT:    srlv $7, $7, $3
+; MIPS64R6O0-NEXT:    seb $7, $7
 ; MIPS64R6O0-NEXT:  # %bb.3: # %entry
-; MIPS64R6O0-NEXT:    sw $8, 12($sp) # 4-byte Folded Spill
+; MIPS64R6O0-NEXT:    sw $7, 12($sp) # 4-byte Folded Spill
 ; MIPS64R6O0-NEXT:  # %bb.4: # %entry
 ; MIPS64R6O0-NEXT:    lw $1, 12($sp) # 4-byte Folded Reload
 ; MIPS64R6O0-NEXT:    seb $2, $1
 ; MIPS64R6O0-NEXT:    daddiu $sp, $sp, 16
 ; MIPS64R6O0-NEXT:    jrc $ra
 ;
 ; MM32-LABEL: AtomicLoadAdd8:
 ; MM32:       # %bb.0: # %entry

 ; MIPS64R6-NEXT:    seb $2, $1
 ;
 ; MIPS64R6O0-LABEL: AtomicLoadSub8:
 ; MIPS64R6O0:       # %bb.0: # %entry
 ; MIPS64R6O0-NEXT:    daddiu $sp, $sp, -16
 ; MIPS64R6O0-NEXT:    lui $1, %hi(%neg(%gp_rel(AtomicLoadSub8)))
 ; MIPS64R6O0-NEXT:    daddu $1, $1, $25
 ; MIPS64R6O0-NEXT:    daddiu $1, $1, %lo(%neg(%gp_rel(AtomicLoadSub8)))
-; MIPS64R6O0-NEXT:    move $2, $4
+; MIPS64R6O0-NEXT:    # kill: def $a0 killed $a0 killed $a0_64
 ; MIPS64R6O0-NEXT:    ld $1, %got_disp(y)($1)
-; MIPS64R6O0-NEXT:    daddiu $3, $zero, -4
-; MIPS64R6O0-NEXT:    and $3, $1, $3
-; MIPS64R6O0-NEXT:    andi $5, $1, 3
-; MIPS64R6O0-NEXT:    xori $5, $5, 3
-; MIPS64R6O0-NEXT:    sll $5, $5, 3
-; MIPS64R6O0-NEXT:    ori $6, $zero, 255
-; MIPS64R6O0-NEXT:    sllv $6, $6, $5
-; MIPS64R6O0-NEXT:    nor $7, $zero, $6
-; MIPS64R6O0-NEXT:    sllv $2, $2, $5
+; MIPS64R6O0-NEXT:    daddiu $2, $zero, -4
+; MIPS64R6O0-NEXT:    and $2, $1, $2
+; MIPS64R6O0-NEXT:    andi $3, $1, 3
+; MIPS64R6O0-NEXT:    xori $3, $3, 3
+; MIPS64R6O0-NEXT:    sll $3, $3, 3
+; MIPS64R6O0-NEXT:    ori $5, $zero, 255
+; MIPS64R6O0-NEXT:    sllv $5, $5, $3
+; MIPS64R6O0-NEXT:    nor $6, $zero, $5
+; MIPS64R6O0-NEXT:    sllv $4, $4, $3
 ; MIPS64R6O0-NEXT:  .LBB9_1: # %entry
 ; MIPS64R6O0-NEXT:    # =>This Inner Loop Header: Depth=1
-; MIPS64R6O0-NEXT:    ll $9, 0($3)
-; MIPS64R6O0-NEXT:    subu $10, $9, $2
-; MIPS64R6O0-NEXT:    and $10, $10, $6
-; MIPS64R6O0-NEXT:    and $11, $9, $7
-; MIPS64R6O0-NEXT:    or $11, $11, $10
-; MIPS64R6O0-NEXT:    sc $11, 0($3)
-; MIPS64R6O0-NEXT:    beqzc $11, .LBB9_1
+; MIPS64R6O0-NEXT:    ll $8, 0($2)
+; MIPS64R6O0-NEXT:    subu $9, $8, $4
+; MIPS64R6O0-NEXT:    and $9, $9, $5
+; MIPS64R6O0-NEXT:    and $10, $8, $6
+; MIPS64R6O0-NEXT:    or $10, $10, $9
+; MIPS64R6O0-NEXT:    sc $10, 0($2)
+; MIPS64R6O0-NEXT:    beqzc $10, .LBB9_1
 ; MIPS64R6O0-NEXT:  # %bb.2: # %entry
-; MIPS64R6O0-NEXT:    and $8, $9, $6
-; MIPS64R6O0-NEXT:    srlv $8, $8, $5
-; MIPS64R6O0-NEXT:    seb $8, $8
+; MIPS64R6O0-NEXT:    and $7, $8, $5
+; MIPS64R6O0-NEXT:    srlv $7, $7, $3
+; MIPS64R6O0-NEXT:    seb $7, $7
 ; MIPS64R6O0-NEXT:  # %bb.3: # %entry
-; MIPS64R6O0-NEXT:    sw $8, 12($sp) # 4-byte Folded Spill
+; MIPS64R6O0-NEXT:    sw $7, 12($sp) # 4-byte Folded Spill
 ; MIPS64R6O0-NEXT:  # %bb.4: # %entry
 ; MIPS64R6O0-NEXT:    lw $1, 12($sp) # 4-byte Folded Reload
 ; MIPS64R6O0-NEXT:    seb $2, $1
 ; MIPS64R6O0-NEXT:    daddiu $sp, $sp, 16
 ; MIPS64R6O0-NEXT:    jrc $ra
 ;
 ; MM32-LABEL: AtomicLoadSub8:
 ; MM32:       # %bb.0: # %entry

 ; MIPS64R6-NEXT:    seb $2, $1
 ;
 ; MIPS64R6O0-LABEL: AtomicLoadNand8:
 ; MIPS64R6O0:       # %bb.0: # %entry
 ; MIPS64R6O0-NEXT:    daddiu $sp, $sp, -16
 ; MIPS64R6O0-NEXT:    lui $1, %hi(%neg(%gp_rel(AtomicLoadNand8)))
 ; MIPS64R6O0-NEXT:    daddu $1, $1, $25
 ; MIPS64R6O0-NEXT:    daddiu $1, $1, %lo(%neg(%gp_rel(AtomicLoadNand8)))
-; MIPS64R6O0-NEXT:    move $2, $4
+; MIPS64R6O0-NEXT:    # kill: def $a0 killed $a0 killed $a0_64
 ; MIPS64R6O0-NEXT:    ld $1, %got_disp(y)($1)
-; MIPS64R6O0-NEXT:    daddiu $3, $zero, -4
-; MIPS64R6O0-NEXT:    and $3, $1, $3
-; MIPS64R6O0-NEXT:    andi $5, $1, 3
-; MIPS64R6O0-NEXT:    xori $5, $5, 3
-; MIPS64R6O0-NEXT:    sll $5, $5, 3
-; MIPS64R6O0-NEXT:    ori $6, $zero, 255
-; MIPS64R6O0-NEXT:    sllv $6, $6, $5
-; MIPS64R6O0-NEXT:    nor $7, $zero, $6
-; MIPS64R6O0-NEXT:    sllv $2, $2, $5
+; MIPS64R6O0-NEXT:    daddiu $2, $zero, -4
+; MIPS64R6O0-NEXT:    and $2, $1, $2
+; MIPS64R6O0-NEXT:    andi $3, $1, 3
+; MIPS64R6O0-NEXT:    xori $3, $3, 3
+; MIPS64R6O0-NEXT:    sll $3, $3, 3
+; MIPS64R6O0-NEXT:    ori $5, $zero, 255
+; MIPS64R6O0-NEXT:    sllv $5, $5, $3
+; MIPS64R6O0-NEXT:    nor $6, $zero, $5
+; MIPS64R6O0-NEXT:    sllv $4, $4, $3
 ; MIPS64R6O0-NEXT:  .LBB10_1: # %entry
 ; MIPS64R6O0-NEXT:    # =>This Inner Loop Header: Depth=1
-; MIPS64R6O0-NEXT:    ll $9, 0($3)
-; MIPS64R6O0-NEXT:    and $10, $9, $2
-; MIPS64R6O0-NEXT:    nor $10, $zero, $10
-; MIPS64R6O0-NEXT:    and $10, $10, $6
-; MIPS64R6O0-NEXT:    and $11, $9, $7
-; MIPS64R6O0-NEXT:    or $11, $11, $10
-; MIPS64R6O0-NEXT:    sc $11, 0($3)
-; MIPS64R6O0-NEXT:    beqzc $11, .LBB10_1
+; MIPS64R6O0-NEXT:    ll $8, 0($2)
+; MIPS64R6O0-NEXT:    and $9, $8, $4
+; MIPS64R6O0-NEXT:    nor $9, $zero, $9
+; MIPS64R6O0-NEXT:    and $9, $9, $5
+; MIPS64R6O0-NEXT:    and $10, $8, $6
+; MIPS64R6O0-NEXT:    or $10, $10, $9
+; MIPS64R6O0-NEXT:    sc $10, 0($2)
+; MIPS64R6O0-NEXT:    beqzc $10, .LBB10_1
 ; MIPS64R6O0-NEXT:  # %bb.2: # %entry
-; MIPS64R6O0-NEXT:    and $8, $9, $6
-; MIPS64R6O0-NEXT:    srlv $8, $8, $5
-; MIPS64R6O0-NEXT:    seb $8, $8
+; MIPS64R6O0-NEXT:    and $7, $8, $5
+; MIPS64R6O0-NEXT:    srlv $7, $7, $3
+; MIPS64R6O0-NEXT:    seb $7, $7
 ; MIPS64R6O0-NEXT:  # %bb.3: # %entry
-; MIPS64R6O0-NEXT:    sw $8, 12($sp) # 4-byte Folded Spill
+; MIPS64R6O0-NEXT:    sw $7, 12($sp) # 4-byte Folded Spill
 ; MIPS64R6O0-NEXT:  # %bb.4: # %entry
 ; MIPS64R6O0-NEXT:    lw $1, 12($sp) # 4-byte Folded Reload
 ; MIPS64R6O0-NEXT:    seb $2, $1
 ; MIPS64R6O0-NEXT:    daddiu $sp, $sp, 16
 ; MIPS64R6O0-NEXT:    jrc $ra
 ;
 ; MM32-LABEL: AtomicLoadNand8:
 ; MM32:       # %bb.0: # %entry

 ; MIPS64R6-NEXT:    seb $2, $1
 ;
 ; MIPS64R6O0-LABEL: AtomicSwap8:
 ; MIPS64R6O0:       # %bb.0: # %entry
 ; MIPS64R6O0-NEXT:    daddiu $sp, $sp, -16
 ; MIPS64R6O0-NEXT:    lui $1, %hi(%neg(%gp_rel(AtomicSwap8)))
 ; MIPS64R6O0-NEXT:    daddu $1, $1, $25
 ; MIPS64R6O0-NEXT:    daddiu $1, $1, %lo(%neg(%gp_rel(AtomicSwap8)))
-; MIPS64R6O0-NEXT:    move $2, $4
+; MIPS64R6O0-NEXT:    # kill: def $a0 killed $a0 killed $a0_64
 ; MIPS64R6O0-NEXT:    ld $1, %got_disp(y)($1)
-; MIPS64R6O0-NEXT:    daddiu $3, $zero, -4
-; MIPS64R6O0-NEXT:    and $3, $1, $3
-; MIPS64R6O0-NEXT:    andi $5, $1, 3
-; MIPS64R6O0-NEXT:    xori $5, $5, 3
-; MIPS64R6O0-NEXT:    sll $5, $5, 3
-; MIPS64R6O0-NEXT:    ori $6, $zero, 255
-; MIPS64R6O0-NEXT:    sllv $6, $6, $5
-; MIPS64R6O0-NEXT:    nor $7, $zero, $6
-; MIPS64R6O0-NEXT:    sllv $2, $2, $5
+; MIPS64R6O0-NEXT:    daddiu $2, $zero, -4
+; MIPS64R6O0-NEXT:    and $2, $1, $2
+; MIPS64R6O0-NEXT:    andi $3, $1, 3
+; MIPS64R6O0-NEXT:    xori $3, $3, 3
+; MIPS64R6O0-NEXT:    sll $3, $3, 3
+; MIPS64R6O0-NEXT:    ori $5, $zero, 255
+; MIPS64R6O0-NEXT:    sllv $5, $5, $3
+; MIPS64R6O0-NEXT:    nor $6, $zero, $5
+; MIPS64R6O0-NEXT:    sllv $4, $4, $3
 ; MIPS64R6O0-NEXT:  .LBB11_1: # %entry
 ; MIPS64R6O0-NEXT:    # =>This Inner Loop Header: Depth=1
-; MIPS64R6O0-NEXT:    ll $9, 0($3)
-; MIPS64R6O0-NEXT:    and $10, $2, $6
-; MIPS64R6O0-NEXT:    and $11, $9, $7
-; MIPS64R6O0-NEXT:    or $11, $11, $10
-; MIPS64R6O0-NEXT:    sc $11, 0($3)
-; MIPS64R6O0-NEXT:    beqzc $11, .LBB11_1
+; MIPS64R6O0-NEXT:    ll $8, 0($2)
+; MIPS64R6O0-NEXT:    and $9, $4, $5
+; MIPS64R6O0-NEXT:    and $10, $8, $6
+; MIPS64R6O0-NEXT:    or $10, $10, $9
+; MIPS64R6O0-NEXT:    sc $10, 0($2)
+; MIPS64R6O0-NEXT:    beqzc $10, .LBB11_1
 ; MIPS64R6O0-NEXT:  # %bb.2: # %entry
-; MIPS64R6O0-NEXT:    and $8, $9, $6
-; MIPS64R6O0-NEXT:    srlv $8, $8, $5
-; MIPS64R6O0-NEXT:    seb $8, $8
+; MIPS64R6O0-NEXT:    and $7, $8, $5
+; MIPS64R6O0-NEXT:    srlv $7, $7, $3
+; MIPS64R6O0-NEXT:    seb $7, $7
 ; MIPS64R6O0-NEXT:  # %bb.3: # %entry
-; MIPS64R6O0-NEXT:    sw $8, 12($sp) # 4-byte Folded Spill
+; MIPS64R6O0-NEXT:    sw $7, 12($sp) # 4-byte Folded Spill
 ; MIPS64R6O0-NEXT:  # %bb.4: # %entry
 ; MIPS64R6O0-NEXT:    lw $1, 12($sp) # 4-byte Folded Reload
 ; MIPS64R6O0-NEXT:    seb $2, $1
 ; MIPS64R6O0-NEXT:    daddiu $sp, $sp, 16
 ; MIPS64R6O0-NEXT:    jrc $ra
 ;
 ; MM32-LABEL: AtomicSwap8:
 ; MM32:       # %bb.0: # %entry

 ; MIPS64R6-NEXT:    jrc $ra
 ;
 ; MIPS64R6O0-LABEL: AtomicCmpSwap8:
 ; MIPS64R6O0:       # %bb.0: # %entry
 ; MIPS64R6O0-NEXT:    daddiu $sp, $sp, -16
 ; MIPS64R6O0-NEXT:    lui $1, %hi(%neg(%gp_rel(AtomicCmpSwap8)))
 ; MIPS64R6O0-NEXT:    daddu $1, $1, $25
 ; MIPS64R6O0-NEXT:    daddiu $1, $1, %lo(%neg(%gp_rel(AtomicCmpSwap8)))
-; MIPS64R6O0-NEXT:    move $2, $5
-; MIPS64R6O0-NEXT:    move $3, $4
+; MIPS64R6O0-NEXT:    # kill: def $a1 killed $a1 killed $a1_64
+; MIPS64R6O0-NEXT:    # kill: def $a0 killed $a0 killed $a0_64
 ; MIPS64R6O0-NEXT:    ld $1, %got_disp(y)($1)
-; MIPS64R6O0-NEXT:    daddiu $4, $zero, -4
-; MIPS64R6O0-NEXT:    and $4, $1, $4
-; MIPS64R6O0-NEXT:    andi $6, $1, 3
-; MIPS64R6O0-NEXT:    xori $6, $6, 3
-; MIPS64R6O0-NEXT:    sll $6, $6, 3
-; MIPS64R6O0-NEXT:    ori $7, $zero, 255
-; MIPS64R6O0-NEXT:    sllv $7, $7, $6
-; MIPS64R6O0-NEXT:    nor $8, $zero, $7
-; MIPS64R6O0-NEXT:    andi $3, $3, 255
-; MIPS64R6O0-NEXT:    sllv $3, $3, $6
-; MIPS64R6O0-NEXT:    andi $2, $2, 255
-; MIPS64R6O0-NEXT:    sllv $2, $2, $6
+; MIPS64R6O0-NEXT:    daddiu $2, $zero, -4
+; MIPS64R6O0-NEXT:    and $2, $1, $2
+; MIPS64R6O0-NEXT:    andi $3, $1, 3
+; MIPS64R6O0-NEXT:    xori $3, $3, 3
+; MIPS64R6O0-NEXT:    sll $3, $3, 3
+; MIPS64R6O0-NEXT:    ori $6, $zero, 255
+; MIPS64R6O0-NEXT:    sllv $6, $6, $3
+; MIPS64R6O0-NEXT:    nor $7, $zero, $6
+; MIPS64R6O0-NEXT:    andi $4, $4, 255
+; MIPS64R6O0-NEXT:    sllv $4, $4, $3
+; MIPS64R6O0-NEXT:    andi $5, $5, 255
+; MIPS64R6O0-NEXT:    sllv $5, $5, $3
 ; MIPS64R6O0-NEXT:  .LBB12_1: # %entry
 ; MIPS64R6O0-NEXT:    # =>This Inner Loop Header: Depth=1
-; MIPS64R6O0-NEXT:    ll $10, 0($4)
-; MIPS64R6O0-NEXT:    and $11, $10, $7
-; MIPS64R6O0-NEXT:    bnec $11, $3, .LBB12_3
+; MIPS64R6O0-NEXT:    ll $9, 0($2)
+; MIPS64R6O0-NEXT:    and $10, $9, $6
+; MIPS64R6O0-NEXT:    bnec $10, $4, .LBB12_3
 ; MIPS64R6O0-NEXT:  # %bb.2: # %entry
 ; MIPS64R6O0-NEXT:    # in Loop: Header=BB12_1 Depth=1
-; MIPS64R6O0-NEXT:    and $10, $10, $8
-; MIPS64R6O0-NEXT:    or $10, $10, $2
-; MIPS64R6O0-NEXT:    sc $10, 0($4)
-; MIPS64R6O0-NEXT:    beqzc $10, .LBB12_1
+; MIPS64R6O0-NEXT:    and $9, $9, $7
+; MIPS64R6O0-NEXT:    or $9, $9, $5
+; MIPS64R6O0-NEXT:    sc $9, 0($2)
+; MIPS64R6O0-NEXT:    beqzc $9, .LBB12_1
 ; MIPS64R6O0-NEXT:  .LBB12_3: # %entry
-; MIPS64R6O0-NEXT:    srlv $9, $11, $6
-; MIPS64R6O0-NEXT:    seb $9, $9
+; MIPS64R6O0-NEXT:    srlv $8, $10, $3
+; MIPS64R6O0-NEXT:    seb $8, $8
 ; MIPS64R6O0-NEXT:  # %bb.4: # %entry
-; MIPS64R6O0-NEXT:    sw $9, 12($sp) # 4-byte Folded Spill
+; MIPS64R6O0-NEXT:    sw $8, 12($sp) # 4-byte Folded Spill
 ; MIPS64R6O0-NEXT:  # %bb.5: # %entry
 ; MIPS64R6O0-NEXT:    lw $2, 12($sp) # 4-byte Folded Reload
 ; MIPS64R6O0-NEXT:    daddiu $sp, $sp, 16
 ; MIPS64R6O0-NEXT:    jrc $ra
 ;
 ; MM32-LABEL: AtomicCmpSwap8:
 ; MM32:       # %bb.0: # %entry
 ; MM32-NEXT:    lui $2, %hi(_gp_disp)

 ; MIPS64R6-NEXT:  # %bb.4: # %entry
 ; MIPS64R6-NEXT:    xor $1, $1, $5
 ; MIPS64R6-NEXT:    jr $ra
 ; MIPS64R6-NEXT:    sltiu $2, $1, 1
 ;
 ; MIPS64R6O0-LABEL: AtomicCmpSwapRes8:
 ; MIPS64R6O0:       # %bb.0: # %entry
 ; MIPS64R6O0-NEXT:    daddiu $sp, $sp, -16
-; MIPS64R6O0-NEXT:    move $1, $6
-; MIPS64R6O0-NEXT:    move $2, $5
-; MIPS64R6O0-NEXT:    move $3, $4
-; MIPS64R6O0-NEXT:    daddiu $5, $zero, -4
-; MIPS64R6O0-NEXT:    and $5, $4, $5
-; MIPS64R6O0-NEXT:    andi $7, $4, 3
-; MIPS64R6O0-NEXT:    xori $7, $7, 3
-; MIPS64R6O0-NEXT:    sll $7, $7, 3
-; MIPS64R6O0-NEXT:    ori $8, $zero, 255
-; MIPS64R6O0-NEXT:    sllv $8, $8, $7
-; MIPS64R6O0-NEXT:    nor $9, $zero, $8
-; MIPS64R6O0-NEXT:    andi $10, $2, 255
-; MIPS64R6O0-NEXT:    sllv $10, $10, $7
-; MIPS64R6O0-NEXT:    andi $1, $1, 255
-; MIPS64R6O0-NEXT:    sllv $1, $1, $7
+; MIPS64R6O0-NEXT:    # kill: def $a2 killed $a2 killed $a2_64
+; MIPS64R6O0-NEXT:    # kill: def $a1 killed $a1 killed $a1_64
+; MIPS64R6O0-NEXT:    move $1, $4
+; MIPS64R6O0-NEXT:    daddiu $2, $zero, -4
+; MIPS64R6O0-NEXT:    and $2, $4, $2
+; MIPS64R6O0-NEXT:    andi $3, $4, 3
+; MIPS64R6O0-NEXT:    xori $3, $3, 3
+; MIPS64R6O0-NEXT:    sll $3, $3, 3
+; MIPS64R6O0-NEXT:    ori $7, $zero, 255
+; MIPS64R6O0-NEXT:    sllv $7, $7, $3
+; MIPS64R6O0-NEXT:    nor $8, $zero, $7
+; MIPS64R6O0-NEXT:    andi $9, $5, 255
+; MIPS64R6O0-NEXT:    sllv $9, $9, $3
+; MIPS64R6O0-NEXT:    andi $6, $6, 255
+; MIPS64R6O0-NEXT:    sllv $6, $6, $3
 ; MIPS64R6O0-NEXT:  .LBB13_1: # %entry
 ; MIPS64R6O0-NEXT:    # =>This Inner Loop Header: Depth=1
-; MIPS64R6O0-NEXT:    ll $12, 0($5)
-; MIPS64R6O0-NEXT:    and $13, $12, $8
-; MIPS64R6O0-NEXT:    bnec $13, $10, .LBB13_3
+; MIPS64R6O0-NEXT:    ll $11, 0($2)
+; MIPS64R6O0-NEXT:    and $12, $11, $7
+; MIPS64R6O0-NEXT:    bnec $12, $9, .LBB13_3
 ; MIPS64R6O0-NEXT:  # %bb.2: # %entry
 ; MIPS64R6O0-NEXT:    # in Loop: Header=BB13_1 Depth=1
-; MIPS64R6O0-NEXT:    and $12, $12, $9
-; MIPS64R6O0-NEXT:    or $12, $12, $1
-; MIPS64R6O0-NEXT:    sc $12, 0($5)
-; MIPS64R6O0-NEXT:    beqzc $12, .LBB13_1
+; MIPS64R6O0-NEXT:    and $11, $11, $8
+; MIPS64R6O0-NEXT:    or $11, $11, $6
+; MIPS64R6O0-NEXT:    sc $11, 0($2)
+; MIPS64R6O0-NEXT:    beqzc $11, .LBB13_1
 ; MIPS64R6O0-NEXT:  .LBB13_3: # %entry
-; MIPS64R6O0-NEXT:    srlv $11, $13, $7
-; MIPS64R6O0-NEXT:    seb $11, $11
+; MIPS64R6O0-NEXT:    srlv $10, $12, $3
+; MIPS64R6O0-NEXT:    seb $10, $10
 ; MIPS64R6O0-NEXT:  # %bb.4: # %entry
-; MIPS64R6O0-NEXT:    sw $2, 12($sp) # 4-byte Folded Spill
-; MIPS64R6O0-NEXT:    sw $11, 8($sp) # 4-byte Folded Spill
+; MIPS64R6O0-NEXT:    sw $5, 12($sp) # 4-byte Folded Spill
+; MIPS64R6O0-NEXT:    sw $10, 8($sp) # 4-byte Folded Spill
 ; MIPS64R6O0-NEXT:  # %bb.5: # %entry
 ; MIPS64R6O0-NEXT:    lw $1, 8($sp) # 4-byte Folded Reload
 ; MIPS64R6O0-NEXT:    lw $2, 12($sp) # 4-byte Folded Reload
 ; MIPS64R6O0-NEXT:    xor $1, $1, $2
 ; MIPS64R6O0-NEXT:    sltiu $2, $1, 1
 ; MIPS64R6O0-NEXT:    daddiu $sp, $sp, 16
 ; MIPS64R6O0-NEXT:    jrc $ra
 ;

 ; MIPS64R6-NEXT:    seh $2, $1
 ;
 ; MIPS64R6O0-LABEL: AtomicLoadAdd16:
 ; MIPS64R6O0:       # %bb.0: # %entry
 ; MIPS64R6O0-NEXT:    daddiu $sp, $sp, -16
 ; MIPS64R6O0-NEXT:    lui $1, %hi(%neg(%gp_rel(AtomicLoadAdd16)))
 ; MIPS64R6O0-NEXT:    daddu $1, $1, $25
 ; MIPS64R6O0-NEXT:    daddiu $1, $1, %lo(%neg(%gp_rel(AtomicLoadAdd16)))
-; MIPS64R6O0-NEXT:    move $2, $4
+; MIPS64R6O0-NEXT:    # kill: def $a0 killed $a0 killed $a0_64
 ; MIPS64R6O0-NEXT:    ld $1, %got_disp(z)($1)
-; MIPS64R6O0-NEXT:    daddiu $3, $zero, -4
-; MIPS64R6O0-NEXT:    and $3, $1, $3
-; MIPS64R6O0-NEXT:    andi $5, $1, 3
-; MIPS64R6O0-NEXT:    xori $5, $5, 2
-; MIPS64R6O0-NEXT:    sll $5, $5, 3
-; MIPS64R6O0-NEXT:    ori $6, $zero, 65535
-; MIPS64R6O0-NEXT:    sllv $6, $6, $5
-; MIPS64R6O0-NEXT:    nor $7, $zero, $6
-; MIPS64R6O0-NEXT:    sllv $2, $2, $5
+; MIPS64R6O0-NEXT:    daddiu $2, $zero, -4
+; MIPS64R6O0-NEXT:    and $2, $1, $2
+; MIPS64R6O0-NEXT:    andi $3, $1, 3
+; MIPS64R6O0-NEXT:    xori $3, $3, 2
+; MIPS64R6O0-NEXT:    sll $3, $3, 3
+; MIPS64R6O0-NEXT:    ori $5, $zero, 65535
+; MIPS64R6O0-NEXT:    sllv $5, $5, $3
+; MIPS64R6O0-NEXT:    nor $6, $zero, $5
+; MIPS64R6O0-NEXT:    sllv $4, $4, $3
 ; MIPS64R6O0-NEXT:  .LBB14_1: # %entry
 ; MIPS64R6O0-NEXT:    # =>This Inner Loop Header: Depth=1
-; MIPS64R6O0-NEXT:    ll $9, 0($3)
-; MIPS64R6O0-NEXT:    addu $10, $9, $2
-; MIPS64R6O0-NEXT:    and $10, $10, $6
-; MIPS64R6O0-NEXT:    and $11, $9, $7
-; MIPS64R6O0-NEXT:    or $11, $11, $10
-; MIPS64R6O0-NEXT:    sc $11, 0($3)
-; MIPS64R6O0-NEXT:    beqzc $11, .LBB14_1
+; MIPS64R6O0-NEXT:    ll $8, 0($2)
+; MIPS64R6O0-NEXT:    addu $9, $8, $4
+; MIPS64R6O0-NEXT:    and $9, $9, $5
+; MIPS64R6O0-NEXT:    and $10, $8, $6
+; MIPS64R6O0-NEXT:    or $10, $10, $9
+; MIPS64R6O0-NEXT:    sc $10, 0($2)
+; MIPS64R6O0-NEXT:    beqzc $10, .LBB14_1
 ; MIPS64R6O0-NEXT:  # %bb.2: # %entry
-; MIPS64R6O0-NEXT:    and $8, $9, $6
-; MIPS64R6O0-NEXT:    srlv $8, $8, $5
-; MIPS64R6O0-NEXT:    seh $8, $8
+; MIPS64R6O0-NEXT:    and $7, $8, $5
+; MIPS64R6O0-NEXT:    srlv $7, $7, $3
+; MIPS64R6O0-NEXT:    seh $7, $7
 ; MIPS64R6O0-NEXT:  # %bb.3: # %entry
-; MIPS64R6O0-NEXT:    sw $8, 12($sp) # 4-byte Folded Spill
+; MIPS64R6O0-NEXT:    sw $7, 12($sp) # 4-byte Folded Spill
 ; MIPS64R6O0-NEXT:  # %bb.4: # %entry
 ; MIPS64R6O0-NEXT:    lw $1, 12($sp) # 4-byte Folded Reload
 ; MIPS64R6O0-NEXT:    seh $2, $1
 ; MIPS64R6O0-NEXT:    daddiu $sp, $sp, 16
 ; MIPS64R6O0-NEXT:    jrc $ra
 ;
 ; MM32-LABEL: AtomicLoadAdd16:
 ; MM32:       # %bb.0: # %entry

 ; MIPS64R6-NEXT:    sltiu $3, $1, 1
 ; MIPS64R6-NEXT:    sync
 ; MIPS64R6-NEXT:    jrc $ra
 ;
 ; MIPS64R6O0-LABEL: foo:
 ; MIPS64R6O0:       # %bb.0:
 ; MIPS64R6O0-NEXT:    daddiu $sp, $sp, -16
 ; MIPS64R6O0-NEXT:    .cfi_def_cfa_offset 16
-; MIPS64R6O0-NEXT:    move $1, $7
-; MIPS64R6O0-NEXT:    sll $1, $1, 0
-; MIPS64R6O0-NEXT:    move $2, $6
-; MIPS64R6O0-NEXT:    sll $2, $2, 0
-; MIPS64R6O0-NEXT:    move $3, $5
-; MIPS64R6O0-NEXT:    sll $3, $3, 0
-; MIPS64R6O0-NEXT:    move $5, $4
+; MIPS64R6O0-NEXT:    # kill: def $a3 killed $a3 killed $a3_64
+; MIPS64R6O0-NEXT:    sll $1, $7, 0
+; MIPS64R6O0-NEXT:    # kill: def $a2 killed $a2 killed $a2_64
+; MIPS64R6O0-NEXT:    sll $2, $6, 0
+; MIPS64R6O0-NEXT:    # kill: def $a1 killed $a1 killed $a1_64
+; MIPS64R6O0-NEXT:    sll $3, $5, 0
+; MIPS64R6O0-NEXT:    move $8, $4
 ; MIPS64R6O0-NEXT:    addu $2, $3, $2
 ; MIPS64R6O0-NEXT:    sync
-; MIPS64R6O0-NEXT:    daddiu $6, $zero, -4
-; MIPS64R6O0-NEXT:    and $6, $4, $6
+; MIPS64R6O0-NEXT:    daddiu $9, $zero, -4
+; MIPS64R6O0-NEXT:    and $9, $4, $9
 ; MIPS64R6O0-NEXT:    andi $3, $4, 3
 ; MIPS64R6O0-NEXT:    xori $3, $3, 2
 ; MIPS64R6O0-NEXT:    sll $3, $3, 3
-; MIPS64R6O0-NEXT:    ori $8, $zero, 65535
-; MIPS64R6O0-NEXT:    sllv $8, $8, $3
-; MIPS64R6O0-NEXT:    nor $9, $zero, $8
-; MIPS64R6O0-NEXT:    andi $10, $2, 65535
-; MIPS64R6O0-NEXT:    sllv $10, $10, $3
+; MIPS64R6O0-NEXT:    ori $5, $zero, 65535
+; MIPS64R6O0-NEXT:    sllv $5, $5, $3
+; MIPS64R6O0-NEXT:    nor $6, $zero, $5
+; MIPS64R6O0-NEXT:    andi $7, $2, 65535
+; MIPS64R6O0-NEXT:    sllv $7, $7, $3
 ; MIPS64R6O0-NEXT:    andi $1, $1, 65535
 ; MIPS64R6O0-NEXT:    sllv $1, $1, $3
 ; MIPS64R6O0-NEXT:  .LBB15_1: # =>This Inner Loop Header: Depth=1
-; MIPS64R6O0-NEXT:    ll $12, 0($6)
-; MIPS64R6O0-NEXT:    and $13, $12, $8
-; MIPS64R6O0-NEXT:    bnec $13, $10, .LBB15_3
+; MIPS64R6O0-NEXT:    ll $11, 0($9)
+; MIPS64R6O0-NEXT:    and $12, $11, $5
+; MIPS64R6O0-NEXT:    bnec $12, $7, .LBB15_3
 ; MIPS64R6O0-NEXT:  # %bb.2: # in Loop: Header=BB15_1 Depth=1
-; MIPS64R6O0-NEXT:    and $12, $12, $9
-; MIPS64R6O0-NEXT:    or $12, $12, $1
-; MIPS64R6O0-NEXT:    sc $12, 0($6)
-; MIPS64R6O0-NEXT:    beqzc $12, .LBB15_1
+; MIPS64R6O0-NEXT:    and $11, $11, $6
+; MIPS64R6O0-NEXT:    or $11, $11, $1
+; MIPS64R6O0-NEXT:    sc $11, 0($9)
+; MIPS64R6O0-NEXT:    beqzc $11, .LBB15_1
 ; MIPS64R6O0-NEXT:  .LBB15_3:
-; MIPS64R6O0-NEXT:    srlv $11, $13, $3
-; MIPS64R6O0-NEXT:    seh $11, $11
+; MIPS64R6O0-NEXT:    srlv $10, $12, $3
+; MIPS64R6O0-NEXT:    seh $10, $10
 ; MIPS64R6O0-NEXT:  # %bb.4:
 ; MIPS64R6O0-NEXT:    sw $2, 12($sp) # 4-byte Folded Spill
-; MIPS64R6O0-NEXT:    sw $11, 8($sp) # 4-byte Folded Spill
+; MIPS64R6O0-NEXT:    sw $10, 8($sp) # 4-byte Folded Spill
 ; MIPS64R6O0-NEXT:  # %bb.5:
 ; MIPS64R6O0-NEXT:    lw $1, 12($sp) # 4-byte Folded Reload
 ; MIPS64R6O0-NEXT:    seh $2, $1
 ; MIPS64R6O0-NEXT:    lw $3, 8($sp) # 4-byte Folded Reload
 ; MIPS64R6O0-NEXT:    xor $2, $3, $2
 ; MIPS64R6O0-NEXT:    sltiu $3, $2, 1
 ; MIPS64R6O0-NEXT:    sync
 ; MIPS64R6O0-NEXT:    lw $2, 8($sp) # 4-byte Folded Reload

 ; MIPS64R6-NEXT:    sync
 ; MIPS64R6-NEXT:    jrc $ra
 ;
 ; MIPS64R6O0-LABEL: CheckSync:
 ; MIPS64R6O0:       # %bb.0: # %entry
 ; MIPS64R6O0-NEXT:    lui $1, %hi(%neg(%gp_rel(CheckSync)))
 ; MIPS64R6O0-NEXT:    daddu $1, $1, $25
 ; MIPS64R6O0-NEXT:    daddiu $1, $1, %lo(%neg(%gp_rel(CheckSync)))
-; MIPS64R6O0-NEXT:    move $2, $4
+; MIPS64R6O0-NEXT:    # kill: def $a0 killed $a0 killed $a0_64
 ; MIPS64R6O0-NEXT:    sync
 ; MIPS64R6O0-NEXT:    ld $1, %got_disp(countsint)($1)
 ; MIPS64R6O0-NEXT:  .LBB16_1: # %entry
 ; MIPS64R6O0-NEXT:    # =>This Inner Loop Header: Depth=1
-; MIPS64R6O0-NEXT:    ll $3, 0($1)
-; MIPS64R6O0-NEXT:    addu $5, $3, $2
-; MIPS64R6O0-NEXT:    sc $5, 0($1)
-; MIPS64R6O0-NEXT:    beqzc $5, .LBB16_1
+; MIPS64R6O0-NEXT:    ll $2, 0($1)
+; MIPS64R6O0-NEXT:    addu $3, $2, $4
+; MIPS64R6O0-NEXT:    sc $3, 0($1)
+; MIPS64R6O0-NEXT:    beqzc $3, .LBB16_1
 ; MIPS64R6O0-NEXT:  # %bb.2: # %entry
 ; MIPS64R6O0-NEXT:    sync
-; MIPS64R6O0-NEXT:    move $2, $3
 ; MIPS64R6O0-NEXT:    jrc $ra
 ;
 ; MM32-LABEL: CheckSync:
 ; MM32:       # %bb.0: # %entry
 ; MM32-NEXT:    lui $2, %hi(_gp_disp)
 ; MM32-NEXT:    addiu $2, $2, %lo(_gp_disp)
 ; MM32-NEXT:    addu $2, $2, $25
 ; MM32-NEXT:    sync

 ; MIPS64R6-NEXT:  # %bb.2: # %entry
 ; MIPS64R6-NEXT:    jrc $ra
 ;
 ; MIPS64R6O0-LABEL: AtomicLoadAdd32_OffGt9Bit:
 ; MIPS64R6O0:       # %bb.0: # %entry
 ; MIPS64R6O0-NEXT:    lui $1, %hi(%neg(%gp_rel(AtomicLoadAdd32_OffGt9Bit)))
 ; MIPS64R6O0-NEXT:    daddu $1, $1, $25
 ; MIPS64R6O0-NEXT:    daddiu $1, $1, %lo(%neg(%gp_rel(AtomicLoadAdd32_OffGt9Bit)))
-; MIPS64R6O0-NEXT:    move $2, $4
+; MIPS64R6O0-NEXT:    # kill: def $a0 killed $a0 killed $a0_64
 ; MIPS64R6O0-NEXT:    ld $1, %got_disp(x)($1)
 ; MIPS64R6O0-NEXT:    daddiu $1, $1, 1024
 ; MIPS64R6O0-NEXT:  .LBB18_1: # %entry
 ; MIPS64R6O0-NEXT:    # =>This Inner Loop Header: Depth=1
-; MIPS64R6O0-NEXT:    ll $3, 0($1)
-; MIPS64R6O0-NEXT:    addu $5, $3, $2
-; MIPS64R6O0-NEXT:    sc $5, 0($1)
-; MIPS64R6O0-NEXT:    beqzc $5, .LBB18_1
+; MIPS64R6O0-NEXT:    ll $2, 0($1)
+; MIPS64R6O0-NEXT:    addu $3, $2, $4
+; MIPS64R6O0-NEXT:    sc $3, 0($1)
+; MIPS64R6O0-NEXT:    beqzc $3, .LBB18_1
+; MIPS64R6O0-NEXT:    nop
 ; MIPS64R6O0-NEXT:  # %bb.2: # %entry
-; MIPS64R6O0-NEXT:    move $2, $3
 ; MIPS64R6O0-NEXT:    jrc $ra
 ;
 ; MM32-LABEL: AtomicLoadAdd32_OffGt9Bit:
 ; MM32:       # %bb.0: # %entry
 ; MM32-NEXT:    lui $2, %hi(_gp_disp)
 ; MM32-NEXT:    addiu $2, $2, %lo(_gp_disp)
 ; MM32-NEXT:    addu $2, $2, $25
 ; MM32-NEXT:    lw $1, %got(x)($2)

test/CodeGen/Mips/atomicCmpSwapPW.ll

 ; O32-NEXT:    nop
 ; O32-NEXT:  $BB0_3: # %entry
 ; O32-NEXT:    sync
 ; O32-NEXT:    jr $ra
 ; O32-NEXT:    nop
 ;
 ; N32-LABEL: foo:
 ; N32:       # %bb.0: # %entry
-; N32-NEXT:    move $1, $5
-; N32-NEXT:    sll $1, $1, 0
-; N32-NEXT:    move $2, $4
-; N32-NEXT:    sll $2, $2, 0
+; N32-NEXT:    # kill: def $a1 killed $a1 killed $a1_64
+; N32-NEXT:    sll $1, $5, 0
+; N32-NEXT:    # kill: def $a0 killed $a0 killed $a0_64
+; N32-NEXT:    sll $2, $4, 0
 ; N32-NEXT:    lui $3, %hi(sym)
 ; N32-NEXT:    lw $3, %lo(sym)($3)
 ; N32-NEXT:    sync
 ; N32-NEXT:  .LBB0_1: # %entry
 ; N32-NEXT:    # =>This Inner Loop Header: Depth=1
-; N32-NEXT:    ll $6, 0($3)
-; N32-NEXT:    bne $6, $2, .LBB0_3
+; N32-NEXT:    ll $4, 0($3)
+; N32-NEXT:    bne $4, $2, .LBB0_3
 ; N32-NEXT:    nop
 ; N32-NEXT:  # %bb.2: # %entry
 ; N32-NEXT:    # in Loop: Header=BB0_1 Depth=1
-; N32-NEXT:    move $7, $1
-; N32-NEXT:    sc $7, 0($3)
-; N32-NEXT:    beqz $7, .LBB0_1
+; N32-NEXT:    move $5, $1
+; N32-NEXT:    sc $5, 0($3)
+; N32-NEXT:    beqz $5, .LBB0_1
 ; N32-NEXT:    nop
 ; N32-NEXT:  .LBB0_3: # %entry
 ; N32-NEXT:    sync
 ; N32-NEXT:    jr $ra
 ; N32-NEXT:    nop
 ;
 ; N64-LABEL: foo:
 ; N64:       # %bb.0: # %entry
-; N64-NEXT:    move $1, $5
-; N64-NEXT:    sll $1, $1, 0
-; N64-NEXT:    move $2, $4
-; N64-NEXT:    sll $2, $2, 0
+; N64-NEXT:    # kill: def $a1 killed $a1 killed $a1_64
+; N64-NEXT:    sll $1, $5, 0
+; N64-NEXT:    # kill: def $a0 killed $a0 killed $a0_64
+; N64-NEXT:    sll $2, $4, 0
 ; N64-NEXT:    lui $3, %highest(sym)
 ; N64-NEXT:    daddiu $3, $3, %higher(sym)
 ; N64-NEXT:    dsll $3, $3, 16
 ; N64-NEXT:    daddiu $3, $3, %hi(sym)
 ; N64-NEXT:    dsll $3, $3, 16
 ; N64-NEXT:    ld $3, %lo(sym)($3)
 ; N64-NEXT:    sync
 ; N64-NEXT:  .LBB0_1: # %entry
 ; N64-NEXT:    # =>This Inner Loop Header: Depth=1
-; N64-NEXT:    ll $6, 0($3)
-; N64-NEXT:    bne $6, $2, .LBB0_3
+; N64-NEXT:    ll $4, 0($3)
+; N64-NEXT:    bne $4, $2, .LBB0_3
 ; N64-NEXT:    nop
 ; N64-NEXT:  # %bb.2: # %entry
 ; N64-NEXT:    # in Loop: Header=BB0_1 Depth=1
-; N64-NEXT:    move $7, $1
-; N64-NEXT:    sc $7, 0($3)
-; N64-NEXT:    beqz $7, .LBB0_1
+; N64-NEXT:    move $5, $1
+; N64-NEXT:    sc $5, 0($3)
+; N64-NEXT:    beqz $5, .LBB0_1
 ; N64-NEXT:    nop
 ; N64-NEXT:  .LBB0_3: # %entry
 ; N64-NEXT:    sync
 ; N64-NEXT:    jr $ra
 ; N64-NEXT:    nop
 entry:
   %0 = load i32 *, i32 ** @sym
   cmpxchg i32 * %0, i32 %new, i32 %old seq_cst seq_cst
   ret void
 }
 

test/CodeGen/PowerPC/vsx.ll

 ;
 ; CHECK-REG-LABEL: test63:
 ; CHECK-REG:       # %bb.0:
 ; CHECK-REG-NEXT:    xxlor f1, v2, v2
 ; CHECK-REG-NEXT:    blr
 ;
 ; CHECK-FISL-LABEL: test63:
 ; CHECK-FISL:       # %bb.0:
-; CHECK-FISL-NEXT:    xxlor f0, v2, v2
-; CHECK-FISL-NEXT:    fmr f1, f0
+; CHECK-FISL-NEXT:    # kill: def $vf2 killed $vf2 killed $v2
+; CHECK-FISL-NEXT:    xxlor f1, v2, v2
 ; CHECK-FISL-NEXT:    blr
 ;
 ; CHECK-LE-LABEL: test63:
 ; CHECK-LE:       # %bb.0:
 ; CHECK-LE-NEXT:    xxswapd vs1, v2
 ; CHECK-LE-NEXT:    # kill: def $f1 killed $f1 killed $vsl1
 ; CHECK-LE-NEXT:    blr
   %v = extractelement <2 x double> %a, i32 0

 ; CHECK-REG:       # %bb.0:
 ; CHECK-REG-NEXT:    xxswapd vs1, v2
 ; CHECK-REG-NEXT:    # kill: def $f1 killed $f1 killed $vsl1
 ; CHECK-REG-NEXT:    blr
 ;
 ; CHECK-FISL-LABEL: test64:
 ; CHECK-FISL:       # %bb.0:
 ; CHECK-FISL-NEXT:    xxswapd vs0, v2
+; CHECK-FISL-NEXT:    # kill: def $f0 killed $f0 killed $vsl0
 ; CHECK-FISL-NEXT:    fmr f1, f0
 ; CHECK-FISL-NEXT:    blr
 ;
 ; CHECK-LE-LABEL: test64:
 ; CHECK-LE:       # %bb.0:
 ; CHECK-LE-NEXT:    xxlor f1, v2, v2
 ; CHECK-LE-NEXT:    blr
   %v = extractelement <2 x double> %a, i32 1

 ; CHECK-REG-NEXT:    addi r3, r3, .LCPI65_0@toc@l
 ; CHECK-REG-NEXT:    lxvw4x v3, 0, r3
 ; CHECK-REG-NEXT:    xxspltw v2, vs0, 0
 ; CHECK-REG-NEXT:    vadduwm v2, v2, v3
 ; CHECK-REG-NEXT:    blr
 ;
 ; CHECK-FISL-LABEL: test80:
 ; CHECK-FISL:       # %bb.0:
-; CHECK-FISL-NEXT:    mr r4, r3
-; CHECK-FISL-NEXT:    stw r4, -16(r1)
-; CHECK-FISL-NEXT:    addi r3, r1, -16
-; CHECK-FISL-NEXT:    lxvw4x vs0, 0, r3
+; CHECK-FISL-NEXT:    # kill: def $r3 killed $r3 killed $x3
+; CHECK-FISL-NEXT:    stw r3, -16(r1)
+; CHECK-FISL-NEXT:    addi r4, r1, -16
+; CHECK-FISL-NEXT:    lxvw4x vs0, 0, r4
 ; CHECK-FISL-NEXT:    xxspltw v2, vs0, 0
-; CHECK-FISL-NEXT:    addis r3, r2, .LCPI65_0@toc@ha
-; CHECK-FISL-NEXT:    addi r3, r3, .LCPI65_0@toc@l
-; CHECK-FISL-NEXT:    lxvw4x v3, 0, r3
+; CHECK-FISL-NEXT:    addis r4, r2, .LCPI65_0@toc@ha
+; CHECK-FISL-NEXT:    addi r4, r4, .LCPI65_0@toc@l
+; CHECK-FISL-NEXT:    lxvw4x v3, 0, r4
 ; CHECK-FISL-NEXT:    vadduwm v2, v2, v3
 ; CHECK-FISL-NEXT:    blr
 ;
 ; CHECK-LE-LABEL: test80:
 ; CHECK-LE:       # %bb.0:
 ; CHECK-LE-NEXT:    mtvsrd f0, r3
 ; CHECK-LE-NEXT:    addis r4, r2, .LCPI65_0@toc@ha
 ; CHECK-LE-NEXT:    addi r3, r4, .LCPI65_0@toc@l

test/CodeGen/X86/atomic-monotonic.ll

 ; CHECK-O3-NEXT:    retq
   %v = load atomic i8, i8* %ptr monotonic, align 1
   ret i8 %v
 }
 
 define void @store_i8(i8* %ptr, i8 %v) {
 ; CHECK-O0-LABEL: store_i8:
 ; CHECK-O0:       # %bb.0:
-; CHECK-O0-NEXT:    movb %sil, %al
-; CHECK-O0-NEXT:    movb %al, (%rdi)
+; CHECK-O0-NEXT:    # kill: def $sil killed $sil killed $esi
+; CHECK-O0-NEXT:    movb %sil, (%rdi)
 ; CHECK-O0-NEXT:    retq
 ;
 ; CHECK-O3-LABEL: store_i8:
 ; CHECK-O3:       # %bb.0:
 ; CHECK-O3-NEXT:    movb %sil, (%rdi)
 ; CHECK-O3-NEXT:    retq
   store atomic i8 %v, i8* %ptr monotonic, align 1
   ret void
; CHECK-O3-NEXT:    retq
   %v = load atomic i16, i16* %ptr monotonic, align 2
   ret i16 %v
 }
 
 
 define void @store_i16(i16* %ptr, i16 %v) {
 ; CHECK-O0-LABEL: store_i16:
 ; CHECK-O0:       # %bb.0:
-; CHECK-O0-NEXT:    movw %si, %ax
-; CHECK-O0-NEXT:    movw %ax, (%rdi)
+; CHECK-O0-NEXT:    # kill: def $si killed $si killed $esi
+; CHECK-O0-NEXT:    movw %si, (%rdi)
 ; CHECK-O0-NEXT:    retq
 ;
 ; CHECK-O3-LABEL: store_i16:
 ; CHECK-O3:       # %bb.0:
 ; CHECK-O3-NEXT:    movw %si, (%rdi)
 ; CHECK-O3-NEXT:    retq
   store atomic i16 %v, i16* %ptr monotonic, align 2
   ret void

test/CodeGen/X86/atomic-unordered.ll

 ; CHECK-O3-NEXT:    retq
   %v = load atomic i8, i8* %ptr unordered, align 1
   ret i8 %v
 }
 
 define void @store_i8(i8* %ptr, i8 %v) {
 ; CHECK-O0-LABEL: store_i8:
 ; CHECK-O0:       # %bb.0:
-; CHECK-O0-NEXT:    movb %sil, %al
-; CHECK-O0-NEXT:    movb %al, (%rdi)
+; CHECK-O0-NEXT:    # kill: def $sil killed $sil killed $esi
+; CHECK-O0-NEXT:    movb %sil, (%rdi)
 ; CHECK-O0-NEXT:    retq
 ;
 ; CHECK-O3-LABEL: store_i8:
 ; CHECK-O3:       # %bb.0:
 ; CHECK-O3-NEXT:    movb %sil, (%rdi)
 ; CHECK-O3-NEXT:    retq
   store atomic i8 %v, i8* %ptr unordered, align 1
   ret void
; CHECK-O3-NEXT:    retq
   %v = load atomic i16, i16* %ptr unordered, align 2
   ret i16 %v
 }
 
 
 define void @store_i16(i16* %ptr, i16 %v) {
 ; CHECK-O0-LABEL: store_i16:
 ; CHECK-O0:       # %bb.0:
-; CHECK-O0-NEXT:    movw %si, %ax
-; CHECK-O0-NEXT:    movw %ax, (%rdi)
+; CHECK-O0-NEXT:    # kill: def $si killed $si killed $esi
+; CHECK-O0-NEXT:    movw %si, (%rdi)
 ; CHECK-O0-NEXT:    retq
 ;
 ; CHECK-O3-LABEL: store_i16:
 ; CHECK-O3:       # %bb.0:
 ; CHECK-O3-NEXT:    movw %si, (%rdi)
 ; CHECK-O3-NEXT:    retq
   store atomic i16 %v, i16* %ptr unordered, align 2
   ret void
; CHECK-O3-NEXT:    retq
   ret void
 }
 
 ; Must use a full width op, not a byte op
 define void @narrow_writeback_and(i64* %ptr) {
 ; CHECK-O0-LABEL: narrow_writeback_and:
 ; CHECK-O0:       # %bb.0:
 ; CHECK-O0-NEXT:    movq (%rdi), %rax
+; CHECK-O0-NEXT:    # kill: def $eax killed $eax killed $rax
+; CHECK-O0-NEXT:    andl $-256, %eax
 ; CHECK-O0-NEXT:    movl %eax, %ecx
-; CHECK-O0-NEXT:    andl $-256, %ecx
-; CHECK-O0-NEXT:    movl %ecx, %eax
-; CHECK-O0-NEXT:    movq %rax, (%rdi)
+; CHECK-O0-NEXT:    movq %rcx, (%rdi)
 ; CHECK-O0-NEXT:    retq
 ;
 ; CHECK-O3-LABEL: narrow_writeback_and:
 ; CHECK-O3:       # %bb.0:
 ; CHECK-O3-NEXT:    movl $4294967040, %eax # imm = 0xFFFFFF00
 ; CHECK-O3-NEXT:    andq %rax, (%rdi)
 ; CHECK-O3-NEXT:    retq
   %v = load atomic i64, i64* %ptr unordered, align 8

 ; CHECK-O0-NEXT:    movq %rdx, {{[-0-9]+}}(%r{{[sb]}}p) # 8-byte Spill
 ; CHECK-O0-NEXT:    movq %rsi, {{[-0-9]+}}(%r{{[sb]}}p) # 8-byte Spill
 ; CHECK-O0-NEXT:    movq %rdi, {{[-0-9]+}}(%r{{[sb]}}p) # 8-byte Spill
 ; CHECK-O0-NEXT:    movq %rax, {{[-0-9]+}}(%r{{[sb]}}p) # 8-byte Spill
 ; CHECK-O0-NEXT:    movq %rcx, {{[-0-9]+}}(%r{{[sb]}}p) # 8-byte Spill
 ; CHECK-O0-NEXT:    jmp .LBB16_1
 ; CHECK-O0-NEXT:  .LBB16_1: # %atomicrmw.start
 ; CHECK-O0-NEXT:    # =>This Inner Loop Header: Depth=1
-; CHECK-O0-NEXT:    movq {{[-0-9]+}}(%r{{[sb]}}p), %rax # 8-byte Reload
-; CHECK-O0-NEXT:    movq {{[-0-9]+}}(%r{{[sb]}}p), %rcx # 8-byte Reload
-; CHECK-O0-NEXT:    movq %rax, {{[-0-9]+}}(%r{{[sb]}}p) # 8-byte Spill
-; CHECK-O0-NEXT:    movq %rcx, %rax
 ; CHECK-O0-NEXT:    movq {{[-0-9]+}}(%r{{[sb]}}p), %rdx # 8-byte Reload
+; CHECK-O0-NEXT:    movq {{[-0-9]+}}(%r{{[sb]}}p), %rax # 8-byte Reload
 ; CHECK-O0-NEXT:    movq {{[-0-9]+}}(%r{{[sb]}}p), %rcx # 8-byte Reload
 ; CHECK-O0-NEXT:    movq {{[-0-9]+}}(%r{{[sb]}}p), %rbx # 8-byte Reload
 ; CHECK-O0-NEXT:    movq {{[-0-9]+}}(%r{{[sb]}}p), %rsi # 8-byte Reload
 ; CHECK-O0-NEXT:    lock cmpxchg16b (%rsi)
 ; CHECK-O0-NEXT:    movq %rdx, {{[-0-9]+}}(%r{{[sb]}}p) # 8-byte Spill
 ; CHECK-O0-NEXT:    movq %rax, {{[-0-9]+}}(%r{{[sb]}}p) # 8-byte Spill
 ; CHECK-O0-NEXT:    jne .LBB16_1
 ; CHECK-O0-NEXT:    jmp .LBB16_2
; CHECK-O3-NEXT:    retq
   ret void
 }
 
 ; Legal to fold (TODO)
 define void @rmw_fold_shl2(i64* %p, i64 %v) {
 ; CHECK-O0-LABEL: rmw_fold_shl2:
 ; CHECK-O0:       # %bb.0:
 ; CHECK-O0-NEXT:    movq (%rdi), %rax
+; CHECK-O0-NEXT:    # kill: def $sil killed $sil killed $rsi
+; CHECK-O0-NEXT:    # implicit-def: $rcx
 ; CHECK-O0-NEXT:    movb %sil, %cl
-; CHECK-O0-NEXT:    # implicit-def: $rdx
-; CHECK-O0-NEXT:    movb %cl, %dl
-; CHECK-O0-NEXT:    shlxq %rdx, %rax, %rax
+; CHECK-O0-NEXT:    shlxq %rcx, %rax, %rax
 ; CHECK-O0-NEXT:    movq %rax, (%rdi)
 ; CHECK-O0-NEXT:    retq
 ;
 ; CHECK-O3-LABEL: rmw_fold_shl2:
 ; CHECK-O3:       # %bb.0:
 ; CHECK-O3-NEXT:    shlxq %rsi, (%rdi), %rax
 ; CHECK-O3-NEXT:    movq %rax, (%rdi)
 ; CHECK-O3-NEXT:    retq
; CHECK-O3-NEXT:    retq
   ret void
 }
 
 ; Legal to fold (TODO)
 define void @rmw_fold_lshr2(i64* %p, i64 %v) {
 ; CHECK-O0-LABEL: rmw_fold_lshr2:
 ; CHECK-O0:       # %bb.0:
 ; CHECK-O0-NEXT:    movq (%rdi), %rax
+; CHECK-O0-NEXT:    # kill: def $sil killed $sil killed $rsi
+; CHECK-O0-NEXT:    # implicit-def: $rcx
 ; CHECK-O0-NEXT:    movb %sil, %cl
-; CHECK-O0-NEXT:    # implicit-def: $rdx
-; CHECK-O0-NEXT:    movb %cl, %dl
-; CHECK-O0-NEXT:    shrxq %rdx, %rax, %rax
+; CHECK-O0-NEXT:    shrxq %rcx, %rax, %rax
 ; CHECK-O0-NEXT:    movq %rax, (%rdi)
 ; CHECK-O0-NEXT:    retq
 ;
 ; CHECK-O3-LABEL: rmw_fold_lshr2:
 ; CHECK-O3:       # %bb.0:
 ; CHECK-O3-NEXT:    shrxq %rsi, (%rdi), %rax
 ; CHECK-O3-NEXT:    movq %rax, (%rdi)
 ; CHECK-O3-NEXT:    retq
; CHECK-O3-NEXT:    retq
   ret void
 }
 
 ; Legal to fold (TODO)
 define void @rmw_fold_ashr2(i64* %p, i64 %v) {
 ; CHECK-O0-LABEL: rmw_fold_ashr2:
 ; CHECK-O0:       # %bb.0:
 ; CHECK-O0-NEXT:    movq (%rdi), %rax
+; CHECK-O0-NEXT:    # kill: def $sil killed $sil killed $rsi
+; CHECK-O0-NEXT:    # implicit-def: $rcx
 ; CHECK-O0-NEXT:    movb %sil, %cl
-; CHECK-O0-NEXT:    # implicit-def: $rdx
-; CHECK-O0-NEXT:    movb %cl, %dl
-; CHECK-O0-NEXT:    sarxq %rdx, %rax, %rax
+; CHECK-O0-NEXT:    sarxq %rcx, %rax, %rax
 ; CHECK-O0-NEXT:    movq %rax, (%rdi)
 ; CHECK-O0-NEXT:    retq
 ;
 ; CHECK-O3-LABEL: rmw_fold_ashr2:
 ; CHECK-O3:       # %bb.0:
 ; CHECK-O3-NEXT:    sarxq %rsi, (%rdi), %rax
 ; CHECK-O3-NEXT:    movq %rax, (%rdi)
 ; CHECK-O3-NEXT:    retq
   %prev = load atomic i64, i64* %p unordered, align 8
   %val = ashr i64 %prev, %v
   store atomic i64 %val, i64* %p unordered, align 8
   ret void
 }
 
 ; Legal, as expected
 define void @rmw_fold_and1(i64* %p, i64 %v) {
 ; CHECK-O0-LABEL: rmw_fold_and1:
 ; CHECK-O0:       # %bb.0:
 ; CHECK-O0-NEXT:    movq (%rdi), %rax
+; CHECK-O0-NEXT:    # kill: def $eax killed $eax killed $rax
+; CHECK-O0-NEXT:    andl $15, %eax
 ; CHECK-O0-NEXT:    movl %eax, %ecx
-; CHECK-O0-NEXT:    andl $15, %ecx
-; CHECK-O0-NEXT:    movl %ecx, %eax
-; CHECK-O0-NEXT:    movq %rax, (%rdi)
+; CHECK-O0-NEXT:    movq %rcx, (%rdi)
 ; CHECK-O0-NEXT:    retq
 ;
 ; CHECK-O3-LABEL: rmw_fold_and1:
 ; CHECK-O3:       # %bb.0:
 ; CHECK-O3-NEXT:    andq $15, (%rdi)
 ; CHECK-O3-NEXT:    retq
   %prev = load atomic i64, i64* %p unordered, align 8
   %val = and i64 %prev, 15

 ;; The next batch test truncations, in combination w/operations which could
 ;; be folded against the memory operation.
 
 ; Legal to reduce the load width (TODO)
 define i32 @fold_trunc(i64* %p) {
 ; CHECK-O0-LABEL: fold_trunc:
 ; CHECK-O0:       # %bb.0:
 ; CHECK-O0-NEXT:    movq (%rdi), %rax
-; CHECK-O0-NEXT:    movl %eax, %ecx
-; CHECK-O0-NEXT:    movl %ecx, %eax
+; CHECK-O0-NEXT:    # kill: def $eax killed $eax killed $rax
 ; CHECK-O0-NEXT:    retq
 ;
 ; CHECK-O3-LABEL: fold_trunc:
 ; CHECK-O3:       # %bb.0:
 ; CHECK-O3-NEXT:    movq (%rdi), %rax
 ; CHECK-O3-NEXT:    # kill: def $eax killed $eax killed $rax
 ; CHECK-O3-NEXT:    retq
   %v = load atomic i64, i64* %p unordered, align 8
   %ret = trunc i64 %v to i32
   ret i32 %ret
 }
 
 ; Legal to reduce the load width and fold the load (TODO)
 define i32 @fold_trunc_add(i64* %p, i32 %v2) {
 ; CHECK-O0-LABEL: fold_trunc_add:
 ; CHECK-O0:       # %bb.0:
 ; CHECK-O0-NEXT:    movq (%rdi), %rax
-; CHECK-O0-NEXT:    movl %eax, %ecx
-; CHECK-O0-NEXT:    addl %esi, %ecx
-; CHECK-O0-NEXT:    movl %ecx, %eax
+; CHECK-O0-NEXT:    # kill: def $eax killed $eax killed $rax
+; CHECK-O0-NEXT:    addl %esi, %eax
 ; CHECK-O0-NEXT:    retq
 ;
 ; CHECK-O3-LABEL: fold_trunc_add:
 ; CHECK-O3:       # %bb.0:
 ; CHECK-O3-NEXT:    movq (%rdi), %rax
 ; CHECK-O3-NEXT:    addl %esi, %eax
 ; CHECK-O3-NEXT:    # kill: def $eax killed $eax killed $rax
 ; CHECK-O3-NEXT:    retq
   %v = load atomic i64, i64* %p unordered, align 8
   %trunc = trunc i64 %v to i32
   %ret = add i32 %trunc, %v2
   ret i32 %ret
 }
 
 ; Legal to reduce the load width and fold the load (TODO)
 define i32 @fold_trunc_and(i64* %p, i32 %v2) {
 ; CHECK-O0-LABEL: fold_trunc_and:
 ; CHECK-O0:       # %bb.0:
 ; CHECK-O0-NEXT:    movq (%rdi), %rax
-; CHECK-O0-NEXT:    movl %eax, %ecx
-; CHECK-O0-NEXT:    andl %esi, %ecx
-; CHECK-O0-NEXT:    movl %ecx, %eax
+; CHECK-O0-NEXT:    # kill: def $eax killed $eax killed $rax
+; CHECK-O0-NEXT:    andl %esi, %eax
 ; CHECK-O0-NEXT:    retq
 ;
 ; CHECK-O3-LABEL: fold_trunc_and:
 ; CHECK-O3:       # %bb.0:
 ; CHECK-O3-NEXT:    movq (%rdi), %rax
 ; CHECK-O3-NEXT:    andl %esi, %eax
 ; CHECK-O3-NEXT:    # kill: def $eax killed $eax killed $rax
 ; CHECK-O3-NEXT:    retq
   %v = load atomic i64, i64* %p unordered, align 8
   %trunc = trunc i64 %v to i32
   %ret = and i32 %trunc, %v2
   ret i32 %ret
 }
 
 ; Legal to reduce the load width and fold the load (TODO)
 define i32 @fold_trunc_or(i64* %p, i32 %v2) {
 ; CHECK-O0-LABEL: fold_trunc_or:
 ; CHECK-O0:       # %bb.0:
 ; CHECK-O0-NEXT:    movq (%rdi), %rax
-; CHECK-O0-NEXT:    movl %eax, %ecx
-; CHECK-O0-NEXT:    orl %esi, %ecx
-; CHECK-O0-NEXT:    movl %ecx, %eax
+; CHECK-O0-NEXT:    # kill: def $eax killed $eax killed $rax
+; CHECK-O0-NEXT:    orl %esi, %eax
 ; CHECK-O0-NEXT:    retq
 ;
 ; CHECK-O3-LABEL: fold_trunc_or:
 ; CHECK-O3:       # %bb.0:
 ; CHECK-O3-NEXT:    movq (%rdi), %rax
 ; CHECK-O3-NEXT:    orl %esi, %eax
 ; CHECK-O3-NEXT:    # kill: def $eax killed $eax killed $rax
 ; CHECK-O3-NEXT:    retq
   %v = load atomic i64, i64* %p unordered, align 8
   %trunc = trunc i64 %v to i32
   %ret = or i32 %trunc, %v2
   ret i32 %ret
 }
 
 ; It's tempting to split the wide load into two smaller byte loads
 ; to reduce memory traffic, but this would be illegal for a atomic load
 define i32 @split_load(i64* %p) {
 ; CHECK-O0-LABEL: split_load:
 ; CHECK-O0:       # %bb.0:
 ; CHECK-O0-NEXT:    movq (%rdi), %rax
 ; CHECK-O0-NEXT:    movb %al, %cl
 ; CHECK-O0-NEXT:    shrq $32, %rax
-; CHECK-O0-NEXT:    movb %al, %dl
-; CHECK-O0-NEXT:    orb %dl, %cl
+; CHECK-O0-NEXT:    # kill: def $al killed $al killed $rax
+; CHECK-O0-NEXT:    orb %al, %cl
 ; CHECK-O0-NEXT:    movzbl %cl, %eax
 ; CHECK-O0-NEXT:    retq
 ;
 ; CHECK-O3-LABEL: split_load:
 ; CHECK-O3:       # %bb.0:
 ; CHECK-O3-NEXT:    movq (%rdi), %rax
 ; CHECK-O3-NEXT:    movq %rax, %rcx
 ; CHECK-O3-NEXT:    shrq $32, %rcx

test/CodeGen/X86/atomic6432.ll

 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc < %s -O0 -mtriple=i686-- -mcpu=corei7 -verify-machineinstrs | FileCheck %s --check-prefix X32
 
 @sc64 = external global i64
 
 define void @atomic_fetch_add64() nounwind {
 ; X32-LABEL: atomic_fetch_add64:
 ; X32:       # %bb.0: # %entry
 ; X32-NEXT:    pushl %ebx
 ; X32-NEXT:    pushl %edi
 ; X32-NEXT:    pushl %esi
-; X32-NEXT:    subl $72, %esp
+; X32-NEXT:    subl $56, %esp
 ; X32-NEXT:    movl sc64+4, %eax
 ; X32-NEXT:    movl sc64, %ecx
 ; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jmp .LBB0_1
 ; X32-NEXT:  .LBB0_1: # %atomicrmw.start
 ; X32-NEXT:    # =>This Inner Loop Header: Depth=1
+; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %ecx, %edx
-; X32-NEXT:    addl $1, %edx
-; X32-NEXT:    movl %eax, %esi
+; X32-NEXT:    movl %eax, %ecx
+; X32-NEXT:    addl $1, %ecx
+; X32-NEXT:    movl %edx, %esi
 ; X32-NEXT:    adcl $0, %esi
-; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %eax
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %edx
+; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %esi, %ecx
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload
 ; X32-NEXT:    lock cmpxchg8b sc64
 ; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jne .LBB0_1
 ; X32-NEXT:    jmp .LBB0_2
 ; X32-NEXT:  .LBB0_2: # %atomicrmw.end
 ; X32-NEXT:    movl sc64+4, %eax
 ; X32-NEXT:    movl sc64, %ecx
 ; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jmp .LBB0_3
 ; X32-NEXT:  .LBB0_3: # %atomicrmw.start2
 ; X32-NEXT:    # =>This Inner Loop Header: Depth=1
+; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %ecx, %edx
-; X32-NEXT:    addl $3, %edx
-; X32-NEXT:    movl %eax, %esi
+; X32-NEXT:    movl %eax, %ecx
+; X32-NEXT:    addl $3, %ecx
+; X32-NEXT:    movl %edx, %esi
 ; X32-NEXT:    adcl $0, %esi
-; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %eax
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %edx
+; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %esi, %ecx
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload
 ; X32-NEXT:    lock cmpxchg8b sc64
 ; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jne .LBB0_3
 ; X32-NEXT:    jmp .LBB0_4
 ; X32-NEXT:  .LBB0_4: # %atomicrmw.end1
 ; X32-NEXT:    movl sc64+4, %eax
 ; X32-NEXT:    movl sc64, %ecx
 ; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jmp .LBB0_5
 ; X32-NEXT:  .LBB0_5: # %atomicrmw.start8
 ; X32-NEXT:    # =>This Inner Loop Header: Depth=1
+; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %ecx, %edx
-; X32-NEXT:    addl $5, %edx
-; X32-NEXT:    movl %eax, %esi
+; X32-NEXT:    movl %eax, %ecx
+; X32-NEXT:    addl $5, %ecx
+; X32-NEXT:    movl %edx, %esi
 ; X32-NEXT:    adcl $0, %esi
-; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %eax
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %edx
+; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %esi, %ecx
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload
 ; X32-NEXT:    lock cmpxchg8b sc64
 ; X32-NEXT:    movl %eax, %ecx
 ; X32-NEXT:    movl %edx, %esi
 ; X32-NEXT:    movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jne .LBB0_5
 ; X32-NEXT:    jmp .LBB0_6
 ; X32-NEXT:  .LBB0_6: # %atomicrmw.end7
 ; X32-NEXT:    movl sc64+4, %eax
 ; X32-NEXT:    movl sc64, %ecx
 ; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jmp .LBB0_7
 ; X32-NEXT:  .LBB0_7: # %atomicrmw.start14
 ; X32-NEXT:    # =>This Inner Loop Header: Depth=1
+; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %ecx, %edx
+; X32-NEXT:    movl %eax, %ecx
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Reload
-; X32-NEXT:    addl %esi, %edx
-; X32-NEXT:    movl %eax, %edi
+; X32-NEXT:    addl %esi, %ecx
+; X32-NEXT:    movl %edx, %edi
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload
 ; X32-NEXT:    adcl %ebx, %edi
-; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %eax
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %edx, (%esp) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %edx
+; X32-NEXT:    movl %ecx, (%esp) # 4-byte Spill
 ; X32-NEXT:    movl %edi, %ecx
 ; X32-NEXT:    movl (%esp), %ebx # 4-byte Reload
 ; X32-NEXT:    lock cmpxchg8b sc64
 ; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jne .LBB0_7
 ; X32-NEXT:    jmp .LBB0_8
 ; X32-NEXT:  .LBB0_8: # %atomicrmw.end13
-; X32-NEXT:    addl $72, %esp
+; X32-NEXT:    addl $56, %esp
 ; X32-NEXT:    popl %esi
 ; X32-NEXT:    popl %edi
 ; X32-NEXT:    popl %ebx
 ; X32-NEXT:    retl
 entry:
   %t1 = atomicrmw add  i64* @sc64, i64 1 acquire
   %t2 = atomicrmw add  i64* @sc64, i64 3 acquire
   %t3 = atomicrmw add  i64* @sc64, i64 5 acquire
   %t4 = atomicrmw add  i64* @sc64, i64 %t3 acquire
   ret void
 }
 
 define void @atomic_fetch_sub64() nounwind {
 ; X32-LABEL: atomic_fetch_sub64:
 ; X32:       # %bb.0:
 ; X32-NEXT:    pushl %ebx
 ; X32-NEXT:    pushl %edi
 ; X32-NEXT:    pushl %esi
-; X32-NEXT:    subl $72, %esp
+; X32-NEXT:    subl $56, %esp
 ; X32-NEXT:    movl sc64+4, %eax
 ; X32-NEXT:    movl sc64, %ecx
 ; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jmp .LBB1_1
 ; X32-NEXT:  .LBB1_1: # %atomicrmw.start
 ; X32-NEXT:    # =>This Inner Loop Header: Depth=1
+; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %ecx, %edx
-; X32-NEXT:    addl $-1, %edx
-; X32-NEXT:    movl %eax, %esi
+; X32-NEXT:    movl %eax, %ecx
+; X32-NEXT:    addl $-1, %ecx
+; X32-NEXT:    movl %edx, %esi
 ; X32-NEXT:    adcl $-1, %esi
-; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %eax
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %edx
+; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %esi, %ecx
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload
 ; X32-NEXT:    lock cmpxchg8b sc64
 ; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jne .LBB1_1
 ; X32-NEXT:    jmp .LBB1_2
 ; X32-NEXT:  .LBB1_2: # %atomicrmw.end
 ; X32-NEXT:    movl sc64+4, %eax
 ; X32-NEXT:    movl sc64, %ecx
 ; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jmp .LBB1_3
 ; X32-NEXT:  .LBB1_3: # %atomicrmw.start2
 ; X32-NEXT:    # =>This Inner Loop Header: Depth=1
+; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %ecx, %edx
-; X32-NEXT:    addl $-3, %edx
-; X32-NEXT:    movl %eax, %esi
+; X32-NEXT:    movl %eax, %ecx
+; X32-NEXT:    addl $-3, %ecx
+; X32-NEXT:    movl %edx, %esi
 ; X32-NEXT:    adcl $-1, %esi
-; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %eax
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %edx
+; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %esi, %ecx
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload
 ; X32-NEXT:    lock cmpxchg8b sc64
 ; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jne .LBB1_3
 ; X32-NEXT:    jmp .LBB1_4
 ; X32-NEXT:  .LBB1_4: # %atomicrmw.end1
 ; X32-NEXT:    movl sc64+4, %eax
 ; X32-NEXT:    movl sc64, %ecx
 ; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jmp .LBB1_5
 ; X32-NEXT:  .LBB1_5: # %atomicrmw.start8
 ; X32-NEXT:    # =>This Inner Loop Header: Depth=1
+; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %ecx, %edx
-; X32-NEXT:    addl $-5, %edx
-; X32-NEXT:    movl %eax, %esi
+; X32-NEXT:    movl %eax, %ecx
+; X32-NEXT:    addl $-5, %ecx
+; X32-NEXT:    movl %edx, %esi
 ; X32-NEXT:    adcl $-1, %esi
-; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %eax
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %edx
+; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %esi, %ecx
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload
 ; X32-NEXT:    lock cmpxchg8b sc64
 ; X32-NEXT:    movl %eax, %ecx
 ; X32-NEXT:    movl %edx, %esi
 ; X32-NEXT:    movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jne .LBB1_5
 ; X32-NEXT:    jmp .LBB1_6
 ; X32-NEXT:  .LBB1_6: # %atomicrmw.end7
 ; X32-NEXT:    movl sc64+4, %eax
 ; X32-NEXT:    movl sc64, %ecx
 ; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jmp .LBB1_7
 ; X32-NEXT:  .LBB1_7: # %atomicrmw.start14
 ; X32-NEXT:    # =>This Inner Loop Header: Depth=1
+; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %ecx, %edx
+; X32-NEXT:    movl %eax, %ecx
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Reload
-; X32-NEXT:    subl %esi, %edx
-; X32-NEXT:    movl %eax, %edi
+; X32-NEXT:    subl %esi, %ecx
+; X32-NEXT:    movl %edx, %edi
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload
 ; X32-NEXT:    sbbl %ebx, %edi
-; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %eax
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %edx, (%esp) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %edx
+; X32-NEXT:    movl %ecx, (%esp) # 4-byte Spill
 ; X32-NEXT:    movl %edi, %ecx
 ; X32-NEXT:    movl (%esp), %ebx # 4-byte Reload
 ; X32-NEXT:    lock cmpxchg8b sc64
 ; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jne .LBB1_7
 ; X32-NEXT:    jmp .LBB1_8
 ; X32-NEXT:  .LBB1_8: # %atomicrmw.end13
-; X32-NEXT:    addl $72, %esp
+; X32-NEXT:    addl $56, %esp
 ; X32-NEXT:    popl %esi
 ; X32-NEXT:    popl %edi
 ; X32-NEXT:    popl %ebx
 ; X32-NEXT:    retl
   %t1 = atomicrmw sub  i64* @sc64, i64 1 acquire
   %t2 = atomicrmw sub  i64* @sc64, i64 3 acquire
   %t3 = atomicrmw sub  i64* @sc64, i64 5 acquire
   %t4 = atomicrmw sub  i64* @sc64, i64 %t3 acquire
   ret void
 }
 
 define void @atomic_fetch_and64() nounwind {
 ; X32-LABEL: atomic_fetch_and64:
 ; X32:       # %bb.0:
 ; X32-NEXT:    pushl %ebx
 ; X32-NEXT:    pushl %edi
 ; X32-NEXT:    pushl %esi
-; X32-NEXT:    subl $56, %esp
+; X32-NEXT:    subl $44, %esp
 ; X32-NEXT:    movl sc64+4, %eax
 ; X32-NEXT:    movl sc64, %ecx
 ; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jmp .LBB2_1
 ; X32-NEXT:  .LBB2_1: # %atomicrmw.start
 ; X32-NEXT:    # =>This Inner Loop Header: Depth=1
+; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %ecx, %edx
-; X32-NEXT:    andl $3, %edx
+; X32-NEXT:    movl %eax, %ecx
+; X32-NEXT:    andl $3, %ecx
 ; X32-NEXT:    xorl %esi, %esi
-; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %eax
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %edx
+; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %esi, %ecx
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload
 ; X32-NEXT:    lock cmpxchg8b sc64
 ; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jne .LBB2_1
 ; X32-NEXT:    jmp .LBB2_2
 ; X32-NEXT:  .LBB2_2: # %atomicrmw.end
 ; X32-NEXT:    movl sc64+4, %eax
 ; X32-NEXT:    movl sc64, %ecx
 ; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jmp .LBB2_3
 ; X32-NEXT:  .LBB2_3: # %atomicrmw.start2
 ; X32-NEXT:    # =>This Inner Loop Header: Depth=1
+; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %ecx, %edx
-; X32-NEXT:    andl $1, %edx
-; X32-NEXT:    movl %eax, %esi
+; X32-NEXT:    movl %eax, %ecx
+; X32-NEXT:    andl $1, %ecx
+; X32-NEXT:    movl %edx, %esi
 ; X32-NEXT:    andl $1, %esi
-; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %eax
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %edx
+; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %esi, %ecx
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload
 ; X32-NEXT:    lock cmpxchg8b sc64
 ; X32-NEXT:    movl %eax, %ecx
 ; X32-NEXT:    movl %edx, %esi
 ; X32-NEXT:    movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jne .LBB2_3
 ; X32-NEXT:    jmp .LBB2_4
 ; X32-NEXT:  .LBB2_4: # %atomicrmw.end1
 ; X32-NEXT:    movl sc64+4, %eax
 ; X32-NEXT:    movl sc64, %ecx
 ; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jmp .LBB2_5
 ; X32-NEXT:  .LBB2_5: # %atomicrmw.start8
 ; X32-NEXT:    # =>This Inner Loop Header: Depth=1
+; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %ecx, %edx
+; X32-NEXT:    movl %eax, %ecx
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Reload
-; X32-NEXT:    andl %esi, %edx
-; X32-NEXT:    movl %eax, %edi
+; X32-NEXT:    andl %esi, %ecx
+; X32-NEXT:    movl %edx, %edi
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload
 ; X32-NEXT:    andl %ebx, %edi
-; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %eax
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %edx, (%esp) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %edx
+; X32-NEXT:    movl %ecx, (%esp) # 4-byte Spill
 ; X32-NEXT:    movl %edi, %ecx
 ; X32-NEXT:    movl (%esp), %ebx # 4-byte Reload
 ; X32-NEXT:    lock cmpxchg8b sc64
 ; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jne .LBB2_5
 ; X32-NEXT:    jmp .LBB2_6
 ; X32-NEXT:  .LBB2_6: # %atomicrmw.end7
-; X32-NEXT:    addl $56, %esp
+; X32-NEXT:    addl $44, %esp
 ; X32-NEXT:    popl %esi
 ; X32-NEXT:    popl %edi
 ; X32-NEXT:    popl %ebx
 ; X32-NEXT:    retl
   %t1 = atomicrmw and  i64* @sc64, i64 3 acquire
   %t2 = atomicrmw and  i64* @sc64, i64 4294967297 acquire
   %t3 = atomicrmw and  i64* @sc64, i64 %t2 acquire
   ret void
 }
 
 define void @atomic_fetch_or64() nounwind {
 ; X32-LABEL: atomic_fetch_or64:
 ; X32:       # %bb.0:
 ; X32-NEXT:    pushl %ebx
 ; X32-NEXT:    pushl %edi
 ; X32-NEXT:    pushl %esi
-; X32-NEXT:    subl $56, %esp
+; X32-NEXT:    subl $48, %esp
 ; X32-NEXT:    movl sc64+4, %eax
 ; X32-NEXT:    movl sc64, %ecx
 ; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jmp .LBB3_1
 ; X32-NEXT:  .LBB3_1: # %atomicrmw.start
 ; X32-NEXT:    # =>This Inner Loop Header: Depth=1
+; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %ecx, %edx
-; X32-NEXT:    orl $3, %edx
-; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %eax
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
+; X32-NEXT:    movl %eax, %ecx
+; X32-NEXT:    orl $3, %ecx
 ; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %edx
+; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Reload
+; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
+; X32-NEXT:    movl %esi, %ecx
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload
 ; X32-NEXT:    lock cmpxchg8b sc64
 ; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jne .LBB3_1
 ; X32-NEXT:    jmp .LBB3_2
 ; X32-NEXT:  .LBB3_2: # %atomicrmw.end
 ; X32-NEXT:    movl sc64+4, %eax
 ; X32-NEXT:    movl sc64, %ecx
 ; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jmp .LBB3_3
 ; X32-NEXT:  .LBB3_3: # %atomicrmw.start2
 ; X32-NEXT:    # =>This Inner Loop Header: Depth=1
+; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %ecx, %edx
-; X32-NEXT:    orl $1, %edx
-; X32-NEXT:    movl %eax, %esi
+; X32-NEXT:    movl %eax, %ecx
+; X32-NEXT:    orl $1, %ecx
+; X32-NEXT:    movl %edx, %esi
 ; X32-NEXT:    orl $1, %esi
-; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %eax
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %edx
+; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %esi, %ecx
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload
 ; X32-NEXT:    lock cmpxchg8b sc64
 ; X32-NEXT:    movl %eax, %ecx
 ; X32-NEXT:    movl %edx, %esi
 ; X32-NEXT:    movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jne .LBB3_3
 ; X32-NEXT:    jmp .LBB3_4
 ; X32-NEXT:  .LBB3_4: # %atomicrmw.end1
 ; X32-NEXT:    movl sc64+4, %eax
 ; X32-NEXT:    movl sc64, %ecx
 ; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jmp .LBB3_5
 ; X32-NEXT:  .LBB3_5: # %atomicrmw.start8
 ; X32-NEXT:    # =>This Inner Loop Header: Depth=1
+; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %ecx, %edx
+; X32-NEXT:    movl %eax, %ecx
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Reload
-; X32-NEXT:    orl %esi, %edx
-; X32-NEXT:    movl %eax, %edi
+; X32-NEXT:    orl %esi, %ecx
+; X32-NEXT:    movl %edx, %edi
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload
 ; X32-NEXT:    orl %ebx, %edi
-; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %eax
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %edx, (%esp) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %edx
+; X32-NEXT:    movl %ecx, (%esp) # 4-byte Spill
 ; X32-NEXT:    movl %edi, %ecx
 ; X32-NEXT:    movl (%esp), %ebx # 4-byte Reload
 ; X32-NEXT:    lock cmpxchg8b sc64
 ; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jne .LBB3_5
 ; X32-NEXT:    jmp .LBB3_6
 ; X32-NEXT:  .LBB3_6: # %atomicrmw.end7
-; X32-NEXT:    addl $56, %esp
+; X32-NEXT:    addl $48, %esp
 ; X32-NEXT:    popl %esi
 ; X32-NEXT:    popl %edi
 ; X32-NEXT:    popl %ebx
 ; X32-NEXT:    retl
   %t1 = atomicrmw or   i64* @sc64, i64 3 acquire
   %t2 = atomicrmw or   i64* @sc64, i64 4294967297 acquire
   %t3 = atomicrmw or   i64* @sc64, i64 %t2 acquire
   ret void
 }
 
 define void @atomic_fetch_xor64() nounwind {
 ; X32-LABEL: atomic_fetch_xor64:
 ; X32:       # %bb.0:
 ; X32-NEXT:    pushl %ebx
 ; X32-NEXT:    pushl %edi
 ; X32-NEXT:    pushl %esi
-; X32-NEXT:    subl $56, %esp
+; X32-NEXT:    subl $48, %esp
 ; X32-NEXT:    movl sc64+4, %eax
 ; X32-NEXT:    movl sc64, %ecx
 ; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jmp .LBB4_1
 ; X32-NEXT:  .LBB4_1: # %atomicrmw.start
 ; X32-NEXT:    # =>This Inner Loop Header: Depth=1
+; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %ecx, %edx
-; X32-NEXT:    xorl $3, %edx
-; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %eax
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
+; X32-NEXT:    movl %eax, %ecx
+; X32-NEXT:    xorl $3, %ecx
 ; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %edx
+; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Reload
+; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
+; X32-NEXT:    movl %esi, %ecx
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload
 ; X32-NEXT:    lock cmpxchg8b sc64
 ; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jne .LBB4_1
 ; X32-NEXT:    jmp .LBB4_2
 ; X32-NEXT:  .LBB4_2: # %atomicrmw.end
 ; X32-NEXT:    movl sc64+4, %eax
 ; X32-NEXT:    movl sc64, %ecx
 ; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jmp .LBB4_3
 ; X32-NEXT:  .LBB4_3: # %atomicrmw.start2
 ; X32-NEXT:    # =>This Inner Loop Header: Depth=1
+; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %ecx, %edx
-; X32-NEXT:    xorl $1, %edx
-; X32-NEXT:    movl %eax, %esi
+; X32-NEXT:    movl %eax, %ecx
+; X32-NEXT:    xorl $1, %ecx
+; X32-NEXT:    movl %edx, %esi
 ; X32-NEXT:    xorl $1, %esi
-; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %eax
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %edx
+; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %esi, %ecx
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload
 ; X32-NEXT:    lock cmpxchg8b sc64
 ; X32-NEXT:    movl %eax, %ecx
 ; X32-NEXT:    movl %edx, %esi
 ; X32-NEXT:    movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jne .LBB4_3
 ; X32-NEXT:    jmp .LBB4_4
 ; X32-NEXT:  .LBB4_4: # %atomicrmw.end1
 ; X32-NEXT:    movl sc64+4, %eax
 ; X32-NEXT:    movl sc64, %ecx
 ; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jmp .LBB4_5
 ; X32-NEXT:  .LBB4_5: # %atomicrmw.start8
 ; X32-NEXT:    # =>This Inner Loop Header: Depth=1
+; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %ecx, %edx
+; X32-NEXT:    movl %eax, %ecx
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Reload
-; X32-NEXT:    xorl %esi, %edx
-; X32-NEXT:    movl %eax, %edi
+; X32-NEXT:    xorl %esi, %ecx
+; X32-NEXT:    movl %edx, %edi
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload
 ; X32-NEXT:    xorl %ebx, %edi
-; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %eax
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %edx, (%esp) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %edx
+; X32-NEXT:    movl %ecx, (%esp) # 4-byte Spill
 ; X32-NEXT:    movl %edi, %ecx
 ; X32-NEXT:    movl (%esp), %ebx # 4-byte Reload
 ; X32-NEXT:    lock cmpxchg8b sc64
 ; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jne .LBB4_5
 ; X32-NEXT:    jmp .LBB4_6
 ; X32-NEXT:  .LBB4_6: # %atomicrmw.end7
-; X32-NEXT:    addl $56, %esp
+; X32-NEXT:    addl $48, %esp
 ; X32-NEXT:    popl %esi
 ; X32-NEXT:    popl %edi
 ; X32-NEXT:    popl %ebx
 ; X32-NEXT:    retl
   %t1 = atomicrmw xor  i64* @sc64, i64 3 acquire
   %t2 = atomicrmw xor  i64* @sc64, i64 4294967297 acquire
   %t3 = atomicrmw xor  i64* @sc64, i64 %t2 acquire
   ret void
 }
 
 define void @atomic_fetch_nand64(i64 %x) nounwind {
 ; X32-LABEL: atomic_fetch_nand64:
 ; X32:       # %bb.0:
 ; X32-NEXT:    pushl %ebx
 ; X32-NEXT:    pushl %edi
 ; X32-NEXT:    pushl %esi
-; X32-NEXT:    subl $24, %esp
+; X32-NEXT:    subl $16, %esp
 ; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
 ; X32-NEXT:    movl {{[0-9]+}}(%esp), %ecx
 ; X32-NEXT:    movl sc64+4, %edx
 ; X32-NEXT:    movl sc64, %esi
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
+; X32-NEXT:    movl %edx, (%esp) # 4-byte Spill
 ; X32-NEXT:    jmp .LBB5_1
 ; X32-NEXT:  .LBB5_1: # %atomicrmw.start
 ; X32-NEXT:    # =>This Inner Loop Header: Depth=1
+; X32-NEXT:    movl (%esp), %edx # 4-byte Reload
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %eax, %edx
+; X32-NEXT:    movl %edx, %ecx
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Reload
-; X32-NEXT:    andl %esi, %edx
-; X32-NEXT:    movl %ecx, %edi
+; X32-NEXT:    andl %esi, %ecx
+; X32-NEXT:    movl %eax, %edi
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload
 ; X32-NEXT:    andl %ebx, %edi
 ; X32-NEXT:    notl %edi
-; X32-NEXT:    notl %edx
-; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %eax
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %edx, (%esp) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %edx
-; X32-NEXT:    movl (%esp), %ecx # 4-byte Reload
+; X32-NEXT:    notl %ecx
 ; X32-NEXT:    movl %edi, %ebx
 ; X32-NEXT:    lock cmpxchg8b sc64
-; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
+; X32-NEXT:    movl %edx, (%esp) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jne .LBB5_1
 ; X32-NEXT:    jmp .LBB5_2
 ; X32-NEXT:  .LBB5_2: # %atomicrmw.end
-; X32-NEXT:    addl $24, %esp
+; X32-NEXT:    addl $16, %esp
 ; X32-NEXT:    popl %esi
 ; X32-NEXT:    popl %edi
 ; X32-NEXT:    popl %ebx
 ; X32-NEXT:    retl
   %t1 = atomicrmw nand i64* @sc64, i64 %x acquire
   ret void
 }
 
 define void @atomic_fetch_max64(i64 %x) nounwind {
 ; X32-LABEL: atomic_fetch_max64:
 ; X32:       # %bb.0:
 ; X32-NEXT:    pushl %ebx
 ; X32-NEXT:    pushl %edi
 ; X32-NEXT:    pushl %esi
-; X32-NEXT:    subl $28, %esp
+; X32-NEXT:    subl $24, %esp
 ; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
 ; X32-NEXT:    movl {{[0-9]+}}(%esp), %ecx
 ; X32-NEXT:    movl sc64+4, %edx
 ; X32-NEXT:    movl sc64, %esi
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jmp .LBB6_1
 ; X32-NEXT:  .LBB6_1: # %atomicrmw.start
 ; X32-NEXT:    # =>This Inner Loop Header: Depth=1
+; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %ecx, %edx
+; X32-NEXT:    movl %eax, %ecx
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Reload
-; X32-NEXT:    subl %esi, %edx
-; X32-NEXT:    movl %eax, %edi
+; X32-NEXT:    subl %esi, %ecx
+; X32-NEXT:    movl %edx, %edi
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload
 ; X32-NEXT:    sbbl %ebx, %edi
-; X32-NEXT:    cmovgel %eax, %ebx
-; X32-NEXT:    cmovgel %ecx, %esi
-; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %eax
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %edx
+; X32-NEXT:    cmovgel %edx, %ebx
+; X32-NEXT:    cmovgel %eax, %esi
+; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %ebx, %ecx
 ; X32-NEXT:    movl %esi, %ebx
 ; X32-NEXT:    lock cmpxchg8b sc64
 ; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %edi, (%esp) # 4-byte Spill
 ; X32-NEXT:    jne .LBB6_1
 ; X32-NEXT:    jmp .LBB6_2
 ; X32-NEXT:  .LBB6_2: # %atomicrmw.end
-; X32-NEXT:    addl $28, %esp
+; X32-NEXT:    addl $24, %esp
 ; X32-NEXT:    popl %esi
 ; X32-NEXT:    popl %edi
 ; X32-NEXT:    popl %ebx
 ; X32-NEXT:    retl
   %t1 = atomicrmw max  i64* @sc64, i64 %x acquire
   ret void
 }
 
 define void @atomic_fetch_min64(i64 %x) nounwind {
 ; X32-LABEL: atomic_fetch_min64:
 ; X32:       # %bb.0:
 ; X32-NEXT:    pushl %ebx
 ; X32-NEXT:    pushl %edi
 ; X32-NEXT:    pushl %esi
-; X32-NEXT:    subl $28, %esp
+; X32-NEXT:    subl $24, %esp
 ; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
 ; X32-NEXT:    movl {{[0-9]+}}(%esp), %ecx
 ; X32-NEXT:    movl sc64+4, %edx
 ; X32-NEXT:    movl sc64, %esi
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jmp .LBB7_1
 ; X32-NEXT:  .LBB7_1: # %atomicrmw.start
 ; X32-NEXT:    # =>This Inner Loop Header: Depth=1
+; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload
-; X32-NEXT:    subl %ecx, %edx
+; X32-NEXT:    subl %eax, %ecx
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Reload
-; X32-NEXT:    sbbl %eax, %esi
+; X32-NEXT:    sbbl %edx, %esi
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edi # 4-byte Reload
-; X32-NEXT:    cmovgel %eax, %edi
+; X32-NEXT:    cmovgel %edx, %edi
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload
-; X32-NEXT:    cmovgel %ecx, %ebx
-; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %eax
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %edx
+; X32-NEXT:    cmovgel %eax, %ebx
+; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %edi, %ecx
 ; X32-NEXT:    lock cmpxchg8b sc64
 ; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %esi, (%esp) # 4-byte Spill
 ; X32-NEXT:    jne .LBB7_1
 ; X32-NEXT:    jmp .LBB7_2
 ; X32-NEXT:  .LBB7_2: # %atomicrmw.end
-; X32-NEXT:    addl $28, %esp
+; X32-NEXT:    addl $24, %esp
 ; X32-NEXT:    popl %esi
 ; X32-NEXT:    popl %edi
 ; X32-NEXT:    popl %ebx
 ; X32-NEXT:    retl
   %t1 = atomicrmw min  i64* @sc64, i64 %x acquire
   ret void
 }
 
 define void @atomic_fetch_umax64(i64 %x) nounwind {
 ; X32-LABEL: atomic_fetch_umax64:
 ; X32:       # %bb.0:
 ; X32-NEXT:    pushl %ebx
 ; X32-NEXT:    pushl %edi
 ; X32-NEXT:    pushl %esi
-; X32-NEXT:    subl $28, %esp
+; X32-NEXT:    subl $24, %esp
 ; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
 ; X32-NEXT:    movl {{[0-9]+}}(%esp), %ecx
 ; X32-NEXT:    movl sc64+4, %edx
 ; X32-NEXT:    movl sc64, %esi
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jmp .LBB8_1
 ; X32-NEXT:  .LBB8_1: # %atomicrmw.start
 ; X32-NEXT:    # =>This Inner Loop Header: Depth=1
+; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload
-; X32-NEXT:    subl %ecx, %edx
+; X32-NEXT:    subl %eax, %ecx
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Reload
-; X32-NEXT:    sbbl %eax, %esi
+; X32-NEXT:    sbbl %edx, %esi
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edi # 4-byte Reload
-; X32-NEXT:    cmovbl %eax, %edi
+; X32-NEXT:    cmovbl %edx, %edi
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload
-; X32-NEXT:    cmovbl %ecx, %ebx
-; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %eax
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %edx
+; X32-NEXT:    cmovbl %eax, %ebx
+; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %edi, %ecx
 ; X32-NEXT:    lock cmpxchg8b sc64
 ; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %esi, (%esp) # 4-byte Spill
 ; X32-NEXT:    jne .LBB8_1
 ; X32-NEXT:    jmp .LBB8_2
 ; X32-NEXT:  .LBB8_2: # %atomicrmw.end
-; X32-NEXT:    addl $28, %esp
+; X32-NEXT:    addl $24, %esp
 ; X32-NEXT:    popl %esi
 ; X32-NEXT:    popl %edi
 ; X32-NEXT:    popl %ebx
 ; X32-NEXT:    retl
   %t1 = atomicrmw umax i64* @sc64, i64 %x acquire
   ret void
 }
 
 define void @atomic_fetch_umin64(i64 %x) nounwind {
 ; X32-LABEL: atomic_fetch_umin64:
 ; X32:       # %bb.0:
 ; X32-NEXT:    pushl %ebx
 ; X32-NEXT:    pushl %edi
 ; X32-NEXT:    pushl %esi
-; X32-NEXT:    subl $28, %esp
+; X32-NEXT:    subl $24, %esp
 ; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
 ; X32-NEXT:    movl {{[0-9]+}}(%esp), %ecx
 ; X32-NEXT:    movl sc64+4, %edx
 ; X32-NEXT:    movl sc64, %esi
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jmp .LBB9_1
 ; X32-NEXT:  .LBB9_1: # %atomicrmw.start
 ; X32-NEXT:    # =>This Inner Loop Header: Depth=1
+; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload
-; X32-NEXT:    subl %ecx, %edx
+; X32-NEXT:    subl %eax, %ecx
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Reload
-; X32-NEXT:    sbbl %eax, %esi
+; X32-NEXT:    sbbl %edx, %esi
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edi # 4-byte Reload
-; X32-NEXT:    cmovael %eax, %edi
+; X32-NEXT:    cmovael %edx, %edi
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload
-; X32-NEXT:    cmovael %ecx, %ebx
-; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %eax
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %edx
+; X32-NEXT:    cmovael %eax, %ebx
+; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %edi, %ecx
 ; X32-NEXT:    lock cmpxchg8b sc64
 ; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %esi, (%esp) # 4-byte Spill
 ; X32-NEXT:    jne .LBB9_1
 ; X32-NEXT:    jmp .LBB9_2
 ; X32-NEXT:  .LBB9_2: # %atomicrmw.end
-; X32-NEXT:    addl $28, %esp
+; X32-NEXT:    addl $24, %esp
 ; X32-NEXT:    popl %esi
 ; X32-NEXT:    popl %edi
 ; X32-NEXT:    popl %ebx
 ; X32-NEXT:    retl
   %t1 = atomicrmw umin i64* @sc64, i64 %x acquire
   ret void
 }
 
; X32-NEXT:    retl
   store atomic i64 %x, i64* @sc64 release, align 8
   ret void
 }
 
 define void @atomic_fetch_swap64(i64 %x) nounwind {
 ; X32-LABEL: atomic_fetch_swap64:
 ; X32:       # %bb.0:
 ; X32-NEXT:    pushl %ebx
-; X32-NEXT:    subl $20, %esp
+; X32-NEXT:    subl $16, %esp
 ; X32-NEXT:    movl {{[0-9]+}}(%esp), %ecx
 ; X32-NEXT:    movl {{[0-9]+}}(%esp), %ebx
 ; X32-NEXT:    movl sc64+4, %eax
 ; X32-NEXT:    movl sc64, %edx
 ; X32-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
+; X32-NEXT:    movl %eax, (%esp) # 4-byte Spill
 ; X32-NEXT:    jmp .LBB12_1
 ; X32-NEXT:  .LBB12_1: # %atomicrmw.start
 ; X32-NEXT:    # =>This Inner Loop Header: Depth=1
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
-; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
-; X32-NEXT:    movl %eax, (%esp) # 4-byte Spill
-; X32-NEXT:    movl %ecx, %eax
 ; X32-NEXT:    movl (%esp), %edx # 4-byte Reload
+; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
 ; X32-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload
 ; X32-NEXT:    lock cmpxchg8b sc64
-; X32-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
+; X32-NEXT:    movl %edx, (%esp) # 4-byte Spill
 ; X32-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; X32-NEXT:    jne .LBB12_1
 ; X32-NEXT:    jmp .LBB12_2
 ; X32-NEXT:  .LBB12_2: # %atomicrmw.end
-; X32-NEXT:    addl $20, %esp
+; X32-NEXT:    addl $16, %esp
 ; X32-NEXT:    popl %ebx
 ; X32-NEXT:    retl
   %t1 = atomicrmw xchg i64* @sc64, i64 %x acquire
   ret void
 }

test/CodeGen/X86/avx512-mask-zext-bugfix.ll

 ; CHECK-NEXT:    movl $8, %eax
 ; CHECK-NEXT:    movq %rdx, %rdi
 ; CHECK-NEXT:    movq %rdx, {{[-0-9]+}}(%r{{[sb]}}p) ## 8-byte Spill
 ; CHECK-NEXT:    movl %eax, %edx
 ; CHECK-NEXT:    movq %rcx, {{[-0-9]+}}(%r{{[sb]}}p) ## 8-byte Spill
 ; CHECK-NEXT:    vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
 ; CHECK-NEXT:    kmovw %k0, {{[-0-9]+}}(%r{{[sb]}}p) ## 2-byte Spill
 ; CHECK-NEXT:    callq _calc_expected_mask_val
-; CHECK-NEXT:    movl %eax, %edx
-; CHECK-NEXT:    movw %dx, %r8w
-; CHECK-NEXT:    movzwl %r8w, %esi
+; CHECK-NEXT:    ## kill: def $eax killed $eax killed $rax
+; CHECK-NEXT:    ## kill: def $ax killed $ax killed $eax
+; CHECK-NEXT:    movzwl %ax, %esi
 ; CHECK-NEXT:    kmovw {{[-0-9]+}}(%r{{[sb]}}p), %k0 ## 2-byte Reload
 ; CHECK-NEXT:    kmovb %k0, %edi
 ; CHECK-NEXT:    movq {{[-0-9]+}}(%r{{[sb]}}p), %rdx ## 8-byte Reload
 ; CHECK-NEXT:    movq {{[-0-9]+}}(%r{{[sb]}}p), %rcx ## 8-byte Reload
 ; CHECK-NEXT:    callq _check_mask16
 ; CHECK-NEXT:    vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
 ; CHECK-NEXT:    vpmovd2m %xmm0, %k0
 ; CHECK-NEXT:    kmovq %k0, %k1
 ; CHECK-NEXT:    kmovd %k0, %esi
-; CHECK-NEXT:    movb %sil, %r9b
-; CHECK-NEXT:    movzbl %r9b, %esi
-; CHECK-NEXT:    movw %si, %r8w
-; CHECK-NEXT:    movq {{[-0-9]+}}(%r{{[sb]}}p), %rdi ## 8-byte Reload
-; CHECK-NEXT:    movl $4, %esi
-; CHECK-NEXT:    movl %esi, {{[-0-9]+}}(%r{{[sb]}}p) ## 4-byte Spill
-; CHECK-NEXT:    movl {{[-0-9]+}}(%r{{[sb]}}p), %edx ## 4-byte Reload
+; CHECK-NEXT:    ## kill: def $sil killed $sil killed $esi
+; CHECK-NEXT:    movzbl %sil, %edi
+; CHECK-NEXT:    ## kill: def $di killed $di killed $edi
+; CHECK-NEXT:    movq {{[-0-9]+}}(%r{{[sb]}}p), %rcx ## 8-byte Reload
+; CHECK-NEXT:    movw %di, {{[-0-9]+}}(%r{{[sb]}}p) ## 2-byte Spill
+; CHECK-NEXT:    movq %rcx, %rdi
+; CHECK-NEXT:    movl $4, %r8d
+; CHECK-NEXT:    movl %r8d, %esi
+; CHECK-NEXT:    movl %r8d, %edx
 ; CHECK-NEXT:    movl %eax, {{[-0-9]+}}(%r{{[sb]}}p) ## 4-byte Spill
 ; CHECK-NEXT:    kmovw %k1, {{[-0-9]+}}(%r{{[sb]}}p) ## 2-byte Spill
-; CHECK-NEXT:    movw %r8w, (%rsp) ## 2-byte Spill
 ; CHECK-NEXT:    callq _calc_expected_mask_val
-; CHECK-NEXT:    movw %ax, %r8w
-; CHECK-NEXT:    movw (%rsp), %r10w ## 2-byte Reload
-; CHECK-NEXT:    movzwl %r10w, %edi
-; CHECK-NEXT:    movzwl %r8w, %esi
+; CHECK-NEXT:    ## kill: def $ax killed $ax killed $rax
+; CHECK-NEXT:    movw {{[-0-9]+}}(%r{{[sb]}}p), %r9w ## 2-byte Reload
+; CHECK-NEXT:    movzwl %r9w, %edi
+; CHECK-NEXT:    movzwl %ax, %esi
 ; CHECK-NEXT:    movq {{[-0-9]+}}(%r{{[sb]}}p), %rdx ## 8-byte Reload
 ; CHECK-NEXT:    movq {{[-0-9]+}}(%r{{[sb]}}p), %rcx ## 8-byte Reload
 ; CHECK-NEXT:    callq _check_mask16
 ; CHECK-NEXT:    addq $56, %rsp
 ; CHECK-NEXT:    retq
   %d2 = bitcast <2 x i64> %a to <8 x i16>
   %m2 = call i8 @llvm.x86.avx512.cvtw2mask.128(<8 x i16> %d2)
   %conv7 = zext i8 %m2 to i16

test/CodeGen/X86/crash-O0.ll

 ; aliased registers (AX and AL) - RegAllocFast does not like that.
 ; PR7312
 define i32 @div8() nounwind {
 ; CHECK-LABEL: div8:
 ; CHECK:       ## %bb.0: ## %entry
 ; CHECK-NEXT:    pushq %rbp
 ; CHECK-NEXT:    movq %rsp, %rbp
 ; CHECK-NEXT:    xorl %eax, %eax
-; CHECK-NEXT:    movb %al, %cl
-; CHECK-NEXT:    ## implicit-def: $rdx
-; CHECK-NEXT:    movb %dl, %sil
-; CHECK-NEXT:    movzbw %cl, %ax
-; CHECK-NEXT:    divb %sil
+; CHECK-NEXT:    ## kill: def $al killed $al killed $eax
+; CHECK-NEXT:    ## implicit-def: $rcx
+; CHECK-NEXT:    ## kill: def $cl killed $cl killed $rcx
 ; CHECK-NEXT:    movb %al, {{[-0-9]+}}(%r{{[sb]}}p) ## 1-byte Spill
-; CHECK-NEXT:    movzbw %cl, %ax
-; CHECK-NEXT:    divb %sil
+; CHECK-NEXT:    movzbw %al, %ax
+; CHECK-NEXT:    divb %cl
+; CHECK-NEXT:    movb {{[-0-9]+}}(%r{{[sb]}}p), %dl ## 1-byte Reload
+; CHECK-NEXT:    movb %al, {{[-0-9]+}}(%r{{[sb]}}p) ## 1-byte Spill
+; CHECK-NEXT:    movzbw %dl, %ax
+; CHECK-NEXT:    divb %cl
 ; CHECK-NEXT:    shrw $8, %ax
-; CHECK-NEXT:    movb %al, %cl
-; CHECK-NEXT:    cmpb %sil, %cl
+; CHECK-NEXT:    ## kill: def $al killed $al killed $ax
+; CHECK-NEXT:    cmpb %cl, %al
 ; CHECK-NEXT:    jae LBB0_2
 ; CHECK-NEXT:  ## %bb.1: ## %"39"
 ; CHECK-NEXT:    movb {{[-0-9]+}}(%r{{[sb]}}p), %al ## 1-byte Reload
 ; CHECK-NEXT:    movzbl %al, %ecx
 ; CHECK-NEXT:    ## implicit-def: $edx
 ; CHECK-NEXT:    imull %edx, %ecx
 ; CHECK-NEXT:    addl %edx, %ecx
 ; CHECK-NEXT:    cmpl %edx, %ecx

test/CodeGen/X86/fast-isel-ret-ext.ll

define zeroext i16 @test3(i32 %y) nounwind {
   ; CHECK-LABEL: test3:
   ; CHECK: movzwl {{.*}}, %eax
 }
 
 define signext i16 @test4(i32 %y) nounwind {
   %conv = trunc i32 %y to i16
   ret i16 %conv
   ; CHECK-LABEL: test4:
-  ; CHECK: {{(movswl.%.x, %eax|cwtl)}}
+  ; CHECK: {{(movswl.%di, %eax|cwtl)}}
 }
 
 define zeroext i1 @test5(i32 %y) nounwind {
   %conv = trunc i32 %y to i1
   ret i1 %conv
   ; CHECK-LABEL: test5:
   ; CHECK: andb $1
   ; CHECK: movzbl {{.*}}, %eax
 }

test/CodeGen/X86/fast-isel-select.ll

 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc < %s -mtriple=x86_64-apple-darwin -O0 | FileCheck %s
 
 ; Make sure we only use the less significant bit of the value that feeds the
 ; select. Otherwise, we may account for a non-zero value whereas the
 ; lsb is zero.
 ; <rdar://problem/15651765>
 
 define i32 @fastisel_select(i1 %exchSub2211_, i1 %trunc_8766) {
 ; CHECK-LABEL: fastisel_select:
 ; CHECK:       ## %bb.0:
-; CHECK-NEXT:    movb %sil, %al
-; CHECK-NEXT:    movb %dil, %cl
-; CHECK-NEXT:    xorl %edx, %edx
-; CHECK-NEXT:    subb %al, %cl
-; CHECK-NEXT:    testb $1, %cl
-; CHECK-NEXT:    movl $1204476887, %esi ## imm = 0x47CADBD7
-; CHECK-NEXT:    cmovnel %esi, %edx
-; CHECK-NEXT:    movl %edx, %eax
+; CHECK-NEXT:    ## kill: def $sil killed $sil killed $esi
+; CHECK-NEXT:    ## kill: def $dil killed $dil killed $edi
+; CHECK-NEXT:    xorl %eax, %eax
+; CHECK-NEXT:    subb %sil, %dil
+; CHECK-NEXT:    testb $1, %dil
+; CHECK-NEXT:    movl $1204476887, %ecx ## imm = 0x47CADBD7
+; CHECK-NEXT:    cmovnel %ecx, %eax
 ; CHECK-NEXT:    retq
   %shuffleInternal15257_8932 = sub i1 %exchSub2211_, %trunc_8766
   %counter_diff1345 = select i1 %shuffleInternal15257_8932, i32 1204476887, i32 0
   ret i32 %counter_diff1345
 }
 

test/CodeGen/X86/pr32241.ll

 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc -fast-isel-sink-local-values < %s -fast-isel -mtriple=i686-unknown-unknown -O0 -mcpu=skx | FileCheck %s
 
 define i32 @_Z3foov() {
 ; CHECK-LABEL: _Z3foov:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    subl $16, %esp
 ; CHECK-NEXT:    .cfi_def_cfa_offset 20
 ; CHECK-NEXT:    movw $10959, {{[0-9]+}}(%esp) # imm = 0x2ACF
 ; CHECK-NEXT:    movw $-15498, {{[0-9]+}}(%esp) # imm = 0xC376
 ; CHECK-NEXT:    movw $19417, {{[0-9]+}}(%esp) # imm = 0x4BD9
 ; CHECK-NEXT:    movzwl {{[0-9]+}}(%esp), %eax
 ; CHECK-NEXT:    cmpw $0, {{[0-9]+}}(%esp)
 ; CHECK-NEXT:    movb $1, %cl
-; CHECK-NEXT:    movl %eax, {{[0-9]+}}(%esp) # 4-byte Spill
-; CHECK-NEXT:    movb %cl, {{[0-9]+}}(%esp) # 1-byte Spill
+; CHECK-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
+; CHECK-NEXT:    movb %cl, {{[-0-9]+}}(%e{{[sb]}}p) # 1-byte Spill
 ; CHECK-NEXT:    jne .LBB0_2
 ; CHECK-NEXT:  # %bb.1: # %lor.rhs
 ; CHECK-NEXT:    xorl %eax, %eax
-; CHECK-NEXT:    movb %al, %cl
-; CHECK-NEXT:    movb %cl, {{[0-9]+}}(%esp) # 1-byte Spill
+; CHECK-NEXT:    # kill: def $al killed $al killed $eax
+; CHECK-NEXT:    movb %al, {{[-0-9]+}}(%e{{[sb]}}p) # 1-byte Spill
 ; CHECK-NEXT:    jmp .LBB0_2
 ; CHECK-NEXT:  .LBB0_2: # %lor.end
-; CHECK-NEXT:    movb {{[0-9]+}}(%esp), %al # 1-byte Reload
+; CHECK-NEXT:    movb {{[-0-9]+}}(%e{{[sb]}}p), %al # 1-byte Reload
 ; CHECK-NEXT:    andb $1, %al
 ; CHECK-NEXT:    movzbl %al, %ecx
-; CHECK-NEXT:    movl {{[0-9]+}}(%esp), %edx # 4-byte Reload
+; CHECK-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload
 ; CHECK-NEXT:    cmpl %ecx, %edx
 ; CHECK-NEXT:    setl %al
 ; CHECK-NEXT:    andb $1, %al
 ; CHECK-NEXT:    movzbl %al, %ecx
 ; CHECK-NEXT:    xorl $-1, %ecx
 ; CHECK-NEXT:    cmpl $0, %ecx
 ; CHECK-NEXT:    movb $1, %al
-; CHECK-NEXT:    movb %al, {{[0-9]+}}(%esp) # 1-byte Spill
+; CHECK-NEXT:    movb %al, {{[-0-9]+}}(%e{{[sb]}}p) # 1-byte Spill
 ; CHECK-NEXT:    jne .LBB0_4
 ; CHECK-NEXT:  # %bb.3: # %lor.rhs4
 ; CHECK-NEXT:    xorl %eax, %eax
-; CHECK-NEXT:    movb %al, %cl
-; CHECK-NEXT:    movb %cl, {{[0-9]+}}(%esp) # 1-byte Spill
+; CHECK-NEXT:    # kill: def $al killed $al killed $eax
+; CHECK-NEXT:    movb %al, {{[-0-9]+}}(%e{{[sb]}}p) # 1-byte Spill
 ; CHECK-NEXT:    jmp .LBB0_4
 ; CHECK-NEXT:  .LBB0_4: # %lor.end5
-; CHECK-NEXT:    movb {{[0-9]+}}(%esp), %al # 1-byte Reload
+; CHECK-NEXT:    movb {{[-0-9]+}}(%e{{[sb]}}p), %al # 1-byte Reload
 ; CHECK-NEXT:    andb $1, %al
 ; CHECK-NEXT:    movzbl %al, %ecx
-; CHECK-NEXT:    movw %cx, %dx
-; CHECK-NEXT:    movw %dx, {{[0-9]+}}(%esp)
+; CHECK-NEXT:    # kill: def $cx killed $cx killed $ecx
+; CHECK-NEXT:    movw %cx, {{[0-9]+}}(%esp)
 ; CHECK-NEXT:    movzwl {{[0-9]+}}(%esp), %eax
 ; CHECK-NEXT:    addl $16, %esp
 ; CHECK-NEXT:    .cfi_def_cfa_offset 4
 ; CHECK-NEXT:    retl
 entry:
   %aa = alloca i16, align 2
   %bb = alloca i16, align 2
   %cc = alloca i16, align 2

test/CodeGen/X86/pr32256.ll

 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc < %s -fast-isel -mtriple=i686-unknown-unknown -O0 -mcpu=skx | FileCheck %s
 
 @c = external global i8, align 1
 
 ; Function Attrs: noinline nounwind
 define void @_Z1av() {
 ; CHECK-LABEL: _Z1av:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    subl $2, %esp
 ; CHECK-NEXT:    .cfi_def_cfa_offset 6
 ; CHECK-NEXT:    xorl %eax, %eax
-; CHECK-NEXT:    movb %al, %cl
-; CHECK-NEXT:    movb c, %dl
-; CHECK-NEXT:    xorb $-1, %dl
-; CHECK-NEXT:    testb $1, %dl
-; CHECK-NEXT:    movb %cl, (%esp) # 1-byte Spill
+; CHECK-NEXT:    # kill: def $al killed $al killed $eax
+; CHECK-NEXT:    movb c, %cl
+; CHECK-NEXT:    xorb $-1, %cl
+; CHECK-NEXT:    testb $1, %cl
+; CHECK-NEXT:    movb %al, (%esp) # 1-byte Spill
 ; CHECK-NEXT:    jne .LBB0_1
 ; CHECK-NEXT:    jmp .LBB0_2
 ; CHECK-NEXT:  .LBB0_1: # %land.rhs
 ; CHECK-NEXT:    xorl %eax, %eax
-; CHECK-NEXT:    movb %al, %cl
-; CHECK-NEXT:    movb %cl, (%esp) # 1-byte Spill
+; CHECK-NEXT:    # kill: def $al killed $al killed $eax
+; CHECK-NEXT:    movb %al, (%esp) # 1-byte Spill
 ; CHECK-NEXT:    jmp .LBB0_2
 ; CHECK-NEXT:  .LBB0_2: # %land.end
 ; CHECK-NEXT:    movb (%esp), %al # 1-byte Reload
 ; CHECK-NEXT:    andb $1, %al
 ; CHECK-NEXT:    movb %al, {{[0-9]+}}(%esp)
 ; CHECK-NEXT:    addl $2, %esp
 ; CHECK-NEXT:    .cfi_def_cfa_offset 4
 ; CHECK-NEXT:    retl

test/CodeGen/X86/pr32284.ll

 ; X86-O0:       # %bb.0: # %entry
 ; X86-O0-NEXT:    xorl %eax, %eax
 ; X86-O0-NEXT:    movl %eax, %ecx
 ; X86-O0-NEXT:    xorl %eax, %eax
 ; X86-O0-NEXT:    movzbl c, %edx
 ; X86-O0-NEXT:    subl %edx, %eax
 ; X86-O0-NEXT:    movslq %eax, %rsi
 ; X86-O0-NEXT:    subq %rsi, %rcx
-; X86-O0-NEXT:    movb %cl, %dil
-; X86-O0-NEXT:    cmpb $0, %dil
-; X86-O0-NEXT:    setne %dil
-; X86-O0-NEXT:    andb $1, %dil
-; X86-O0-NEXT:    movb %dil, -{{[0-9]+}}(%rsp)
+; X86-O0-NEXT:    # kill: def $cl killed $cl killed $rcx
+; X86-O0-NEXT:    cmpb $0, %cl
+; X86-O0-NEXT:    setne %cl
+; X86-O0-NEXT:    andb $1, %cl
+; X86-O0-NEXT:    movb %cl, -{{[0-9]+}}(%rsp)
 ; X86-O0-NEXT:    cmpb $0, c
-; X86-O0-NEXT:    setne %dil
-; X86-O0-NEXT:    xorb $-1, %dil
-; X86-O0-NEXT:    xorb $-1, %dil
-; X86-O0-NEXT:    andb $1, %dil
-; X86-O0-NEXT:    movzbl %dil, %eax
+; X86-O0-NEXT:    setne %cl
+; X86-O0-NEXT:    xorb $-1, %cl
+; X86-O0-NEXT:    xorb $-1, %cl
+; X86-O0-NEXT:    andb $1, %cl
+; X86-O0-NEXT:    movzbl %cl, %eax
 ; X86-O0-NEXT:    movzbl c, %edx
 ; X86-O0-NEXT:    cmpl %edx, %eax
-; X86-O0-NEXT:    setle %dil
-; X86-O0-NEXT:    andb $1, %dil
-; X86-O0-NEXT:    movzbl %dil, %eax
+; X86-O0-NEXT:    setle %cl
+; X86-O0-NEXT:    andb $1, %cl
+; X86-O0-NEXT:    movzbl %cl, %eax
 ; X86-O0-NEXT:    movl %eax, -{{[0-9]+}}(%rsp)
 ; X86-O0-NEXT:    retq
 ;
 ; X64-LABEL: foo:
 ; X64:       # %bb.0: # %entry
 ; X64-NEXT:    movzbl {{.*}}(%rip), %eax
 ; X64-NEXT:    xorl %ecx, %ecx
 ; X64-NEXT:    testl %eax, %eax

 ; X86-O0:       # %bb.0: # %entry
 ; X86-O0-NEXT:    movzbl var_7, %eax
 ; X86-O0-NEXT:    cmpb $0, var_7
 ; X86-O0-NEXT:    setne %cl
 ; X86-O0-NEXT:    xorb $-1, %cl
 ; X86-O0-NEXT:    andb $1, %cl
 ; X86-O0-NEXT:    movzbl %cl, %edx
 ; X86-O0-NEXT:    xorl %edx, %eax
-; X86-O0-NEXT:    movw %ax, %si
-; X86-O0-NEXT:    movw %si, -{{[0-9]+}}(%rsp)
-; X86-O0-NEXT:    movzbl var_7, %eax
-; X86-O0-NEXT:    movw %ax, %si
-; X86-O0-NEXT:    cmpw $0, %si
+; X86-O0-NEXT:    # kill: def $ax killed $ax killed $eax
+; X86-O0-NEXT:    movw %ax, -{{[0-9]+}}(%rsp)
+; X86-O0-NEXT:    movzbl var_7, %edx
+; X86-O0-NEXT:    # kill: def $dx killed $dx killed $edx
+; X86-O0-NEXT:    cmpw $0, %dx
 ; X86-O0-NEXT:    setne %cl
 ; X86-O0-NEXT:    xorb $-1, %cl
 ; X86-O0-NEXT:    andb $1, %cl
-; X86-O0-NEXT:    movzbl %cl, %eax
-; X86-O0-NEXT:    movzbl var_7, %edx
-; X86-O0-NEXT:    cmpl %edx, %eax
+; X86-O0-NEXT:    movzbl %cl, %esi
+; X86-O0-NEXT:    movzbl var_7, %edi
+; X86-O0-NEXT:    cmpl %edi, %esi
 ; X86-O0-NEXT:    sete %cl
 ; X86-O0-NEXT:    andb $1, %cl
-; X86-O0-NEXT:    movzbl %cl, %eax
-; X86-O0-NEXT:    movw %ax, %si
-; X86-O0-NEXT:    # implicit-def: $rdi
-; X86-O0-NEXT:    movw %si, (%rdi)
+; X86-O0-NEXT:    movzbl %cl, %esi
+; X86-O0-NEXT:    # kill: def $si killed $si killed $esi
+; X86-O0-NEXT:    # implicit-def: $r8
+; X86-O0-NEXT:    movw %si, (%r8)
 ; X86-O0-NEXT:    retq
 ;
 ; X64-LABEL: f2:
 ; X64:       # %bb.0: # %entry
 ; X64-NEXT:    movzbl {{.*}}(%rip), %eax
 ; X64-NEXT:    xorl %ecx, %ecx
 ; X64-NEXT:    testl %eax, %eax
 ; X64-NEXT:    sete %cl
 ; X64-NEXT:    xorl %eax, %ecx
 ; X64-NEXT:    movw %cx, -{{[0-9]+}}(%rsp)
 ; X64-NEXT:    xorl %ecx, %ecx
 ; X64-NEXT:    testb %al, %al
 ; X64-NEXT:    sete %cl
 ; X64-NEXT:    xorl %edx, %edx
 ; X64-NEXT:    cmpl %eax, %ecx
 ; X64-NEXT:    sete %dl
 ; X64-NEXT:    movw %dx, (%rax)
 ; X64-NEXT:    retq
 ;
 ; 686-O0-LABEL: f2:
 ; 686-O0:       # %bb.0: # %entry
-; 686-O0-NEXT:    pushl %esi
+; 686-O0-NEXT:    pushl %edi
 ; 686-O0-NEXT:    .cfi_def_cfa_offset 8
+; 686-O0-NEXT:    pushl %esi
+; 686-O0-NEXT:    .cfi_def_cfa_offset 12
 ; 686-O0-NEXT:    subl $2, %esp
-; 686-O0-NEXT:    .cfi_def_cfa_offset 10
-; 686-O0-NEXT:    .cfi_offset %esi, -8
+; 686-O0-NEXT:    .cfi_def_cfa_offset 14
+; 686-O0-NEXT:    .cfi_offset %esi, -12
+; 686-O0-NEXT:    .cfi_offset %edi, -8
 ; 686-O0-NEXT:    movzbl var_7, %eax
 ; 686-O0-NEXT:    cmpb $0, var_7
 ; 686-O0-NEXT:    setne %cl
 ; 686-O0-NEXT:    xorb $-1, %cl
 ; 686-O0-NEXT:    andb $1, %cl
 ; 686-O0-NEXT:    movzbl %cl, %edx
 ; 686-O0-NEXT:    xorl %edx, %eax
-; 686-O0-NEXT:    movw %ax, %si
-; 686-O0-NEXT:    movw %si, (%esp)
-; 686-O0-NEXT:    movzbl var_7, %eax
-; 686-O0-NEXT:    movw %ax, %si
-; 686-O0-NEXT:    cmpw $0, %si
+; 686-O0-NEXT:    # kill: def $ax killed $ax killed $eax
+; 686-O0-NEXT:    movw %ax, (%esp)
+; 686-O0-NEXT:    movzbl var_7, %edx
+; 686-O0-NEXT:    # kill: def $dx killed $dx killed $edx
+; 686-O0-NEXT:    cmpw $0, %dx
 ; 686-O0-NEXT:    setne %cl
 ; 686-O0-NEXT:    xorb $-1, %cl
 ; 686-O0-NEXT:    andb $1, %cl
-; 686-O0-NEXT:    movzbl %cl, %eax
-; 686-O0-NEXT:    movzbl var_7, %edx
-; 686-O0-NEXT:    cmpl %edx, %eax
+; 686-O0-NEXT:    movzbl %cl, %esi
+; 686-O0-NEXT:    movzbl var_7, %edi
+; 686-O0-NEXT:    cmpl %edi, %esi
 ; 686-O0-NEXT:    sete %cl
 ; 686-O0-NEXT:    andb $1, %cl
-; 686-O0-NEXT:    movzbl %cl, %eax
-; 686-O0-NEXT:    movw %ax, %si
-; 686-O0-NEXT:    # implicit-def: $eax
-; 686-O0-NEXT:    movw %si, (%eax)
+; 686-O0-NEXT:    movzbl %cl, %esi
+; 686-O0-NEXT:    # kill: def $si killed $si killed $esi
+; 686-O0-NEXT:    # implicit-def: $edi
+; 686-O0-NEXT:    movw %si, (%edi)
 ; 686-O0-NEXT:    addl $2, %esp
-; 686-O0-NEXT:    .cfi_def_cfa_offset 8
+; 686-O0-NEXT:    .cfi_def_cfa_offset 12
 ; 686-O0-NEXT:    popl %esi
+; 686-O0-NEXT:    .cfi_def_cfa_offset 8
+; 686-O0-NEXT:    popl %edi
 ; 686-O0-NEXT:    .cfi_def_cfa_offset 4
 ; 686-O0-NEXT:    retl
 ;
 ; 686-LABEL: f2:
 ; 686:       # %bb.0: # %entry
 ; 686-NEXT:    subl $2, %esp
 ; 686-NEXT:    .cfi_def_cfa_offset 6
 ; 686-NEXT:    movzbl var_7, %eax

 ; X86-O0-NEXT:    cmpl $0, var_13
 ; X86-O0-NEXT:    setne %dl
 ; X86-O0-NEXT:    xorb $-1, %dl
 ; X86-O0-NEXT:    andb $1, %dl
 ; X86-O0-NEXT:    movzbl %dl, %eax
 ; X86-O0-NEXT:    movl %eax, %esi
 ; X86-O0-NEXT:    andq $0, %rsi
 ; X86-O0-NEXT:    orq %rsi, %rcx
-; X86-O0-NEXT:    movl %ecx, %eax
-; X86-O0-NEXT:    movl %eax, var_46
+; X86-O0-NEXT:    # kill: def $ecx killed $ecx killed $rcx
+; X86-O0-NEXT:    movl %ecx, var_46
 ; X86-O0-NEXT:    retq
 ;
 ; X64-LABEL: f3:
 ; X64:       # %bb.0: # %entry
 ; X64-NEXT:    movl {{.*}}(%rip), %eax
 ; X64-NEXT:    xorl %ecx, %ecx
 ; X64-NEXT:    testl %eax, %eax
 ; X64-NEXT:    notl %eax

test/CodeGen/X86/pr32340.ll

 ; X64-NEXT:    xorq %r9, %r8
 ; X64-NEXT:    movq %rdi, %r9
 ; X64-NEXT:    xorq %r8, %r9
 ; X64-NEXT:    xorq $-1, %r9
 ; X64-NEXT:    xorq %r9, %rdi
 ; X64-NEXT:    movq %rdi, %r8
 ; X64-NEXT:    orq var_57, %r8
 ; X64-NEXT:    orq %r8, %rdi
-; X64-NEXT:    movw %di, %r10w
-; X64-NEXT:    movw %r10w, var_900
+; X64-NEXT:    # kill: def $di killed $di killed $rdi
+; X64-NEXT:    movw %di, var_900
 ; X64-NEXT:    cmpq var_28, %rcx
-; X64-NEXT:    setne %r11b
-; X64-NEXT:    andb $1, %r11b
-; X64-NEXT:    movzbl %r11b, %eax
-; X64-NEXT:    movw %ax, %r10w
-; X64-NEXT:    movw %r10w, var_827
+; X64-NEXT:    setne %r10b
+; X64-NEXT:    andb $1, %r10b
+; X64-NEXT:    movzbl %r10b, %eax
+; X64-NEXT:    # kill: def $ax killed $ax killed $eax
+; X64-NEXT:    movw %ax, var_827
 ; X64-NEXT:    retq
 entry:
   store i16 0, i16* @var_825, align 2
   %v0 = load i16, i16* @var_32, align 2
   %conv = zext i16 %v0 to i32
   %v2 = load i16, i16* @var_901, align 2
   %conv2 = zext i16 %v2 to i32
   %xor = xor i32 %conv, %conv2

test/CodeGen/X86/pr32345.ll

 ; X640-NEXT:    xorl %ecx, %eax
 ; X640-NEXT:    movslq %eax, %rdx
 ; X640-NEXT:    movzwl var_27, %eax
 ; X640-NEXT:    subl $16610, %eax # imm = 0x40E2
 ; X640-NEXT:    movl %eax, %eax
 ; X640-NEXT:    movl %eax, %ecx
 ; X640-NEXT:    # kill: def $cl killed $rcx
 ; X640-NEXT:    sarq %cl, %rdx
-; X640-NEXT:    movb %dl, %cl
-; X640-NEXT:    # implicit-def: $rdx
-; X640-NEXT:    movb %cl, (%rdx)
+; X640-NEXT:    # kill: def $dl killed $dl killed $rdx
+; X640-NEXT:    # implicit-def: $rsi
+; X640-NEXT:    movb %dl, (%rsi)
 ; X640-NEXT:    retq
 ;
 ; 6860-LABEL: foo:
 ; 6860:       # %bb.0: # %bb
 ; 6860-NEXT:    pushl %ebp
 ; 6860-NEXT:    .cfi_def_cfa_offset 8
 ; 6860-NEXT:    .cfi_offset %ebp, -8
 ; 6860-NEXT:    movl %esp, %ebp
 ; 6860-NEXT:    .cfi_def_cfa_register %ebp
 ; 6860-NEXT:    pushl %ebx
 ; 6860-NEXT:    pushl %edi
 ; 6860-NEXT:    pushl %esi
 ; 6860-NEXT:    andl $-8, %esp
 ; 6860-NEXT:    subl $32, %esp
 ; 6860-NEXT:    .cfi_offset %esi, -20
 ; 6860-NEXT:    .cfi_offset %edi, -16
 ; 6860-NEXT:    .cfi_offset %ebx, -12
 ; 6860-NEXT:    movw var_22, %ax
 ; 6860-NEXT:    movzwl var_27, %ecx
 ; 6860-NEXT:    movw %cx, %dx
 ; 6860-NEXT:    xorw %dx, %ax
 ; 6860-NEXT:    # implicit-def: $esi
 ; 6860-NEXT:    movw %ax, %si
 ; 6860-NEXT:    xorl %ecx, %esi
-; 6860-NEXT:    movw %si, %ax
-; 6860-NEXT:    movzwl %ax, %ecx
+; 6860-NEXT:    # kill: def $si killed $si killed $esi
+; 6860-NEXT:    movzwl %si, %ecx
 ; 6860-NEXT:    movl %ecx, {{[0-9]+}}(%esp)
 ; 6860-NEXT:    movl $0, {{[0-9]+}}(%esp)
 ; 6860-NEXT:    movw var_22, %ax
 ; 6860-NEXT:    movzwl var_27, %ecx
 ; 6860-NEXT:    movw %cx, %dx
 ; 6860-NEXT:    xorw %dx, %ax
-; 6860-NEXT:    # implicit-def: $esi
-; 6860-NEXT:    movw %ax, %si
-; 6860-NEXT:    xorl %ecx, %esi
-; 6860-NEXT:    movw %si, %ax
-; 6860-NEXT:    movzwl %ax, %esi
-; 6860-NEXT:    movb %cl, %bl
-; 6860-NEXT:    addb $30, %bl
-; 6860-NEXT:    xorl %ecx, %ecx
-; 6860-NEXT:    movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; 6860-NEXT:    movb %bl, %cl
-; 6860-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %edi # 4-byte Reload
-; 6860-NEXT:    shrdl %cl, %edi, %esi
-; 6860-NEXT:    testb $32, %bl
-; 6860-NEXT:    movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
-; 6860-NEXT:    movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
+; 6860-NEXT:    # implicit-def: $edi
+; 6860-NEXT:    movw %ax, %di
+; 6860-NEXT:    xorl %ecx, %edi
+; 6860-NEXT:    # kill: def $di killed $di killed $edi
+; 6860-NEXT:    movzwl %di, %ebx
+; 6860-NEXT:    # kill: def $cl killed $cl killed $ecx
+; 6860-NEXT:    addb $30, %cl
+; 6860-NEXT:    xorl %eax, %eax
+; 6860-NEXT:    movb %cl, {{[-0-9]+}}(%e{{[sb]}}p) # 1-byte Spill
+; 6860-NEXT:    shrdl %cl, %eax, %ebx
+; 6860-NEXT:    movb {{[-0-9]+}}(%e{{[sb]}}p), %cl # 1-byte Reload
+; 6860-NEXT:    testb $32, %cl
+; 6860-NEXT:    movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
+; 6860-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; 6860-NEXT:    jne .LBB0_2
 ; 6860-NEXT:  # %bb.1: # %bb
 ; 6860-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
 ; 6860-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
 ; 6860-NEXT:  .LBB0_2: # %bb
 ; 6860-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
-; 6860-NEXT:    movb %al, %cl
-; 6860-NEXT:    # implicit-def: $eax
-; 6860-NEXT:    movb %cl, (%eax)
+; 6860-NEXT:    # kill: def $al killed $al killed $eax
+; 6860-NEXT:    # implicit-def: $ecx
+; 6860-NEXT:    movb %al, (%ecx)
 ; 6860-NEXT:    leal -12(%ebp), %esp
 ; 6860-NEXT:    popl %esi
 ; 6860-NEXT:    popl %edi
 ; 6860-NEXT:    popl %ebx
 ; 6860-NEXT:    popl %ebp
 ; 6860-NEXT:    .cfi_def_cfa %esp, 4
 ; 6860-NEXT:    retl
 ;

test/CodeGen/X86/pr34592.ll

 ; CHECK-NEXT:    vpblendd {{.*#+}} ymm2 = ymm6[0,1,2,3,4,5],ymm2[6,7]
 ; CHECK-NEXT:    vmovaps %xmm9, %xmm6
 ; CHECK-NEXT:    vmovaps %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
 ; CHECK-NEXT:    # implicit-def: $ymm0
 ; CHECK-NEXT:    vinserti128 $1, %xmm6, %ymm0, %ymm0
 ; CHECK-NEXT:    vpalignr {{.*#+}} ymm11 = ymm2[8,9,10,11,12,13,14,15],ymm11[0,1,2,3,4,5,6,7],ymm2[24,25,26,27,28,29,30,31],ymm11[16,17,18,19,20,21,22,23]
 ; CHECK-NEXT:    vpermq {{.*#+}} ymm11 = ymm11[2,3,2,0]
 ; CHECK-NEXT:    vpblendd {{.*#+}} ymm0 = ymm11[0,1,2,3],ymm0[4,5],ymm11[6,7]
-; CHECK-NEXT:    vmovaps %xmm2, %xmm6
-; CHECK-NEXT:    # implicit-def: $ymm2
-; CHECK-NEXT:    vinserti128 $1, %xmm6, %ymm2, %ymm2
-; CHECK-NEXT:    vextracti128 $1, %ymm7, %xmm6
-; CHECK-NEXT:    vmovq {{.*#+}} xmm6 = xmm6[0],zero
+; CHECK-NEXT:    # kill: def $xmm2 killed $xmm2 killed $ymm2
 ; CHECK-NEXT:    # implicit-def: $ymm11
-; CHECK-NEXT:    vmovaps %xmm6, %xmm11
-; CHECK-NEXT:    vpblendd {{.*#+}} ymm2 = ymm11[0,1,2,3],ymm2[4,5,6,7]
+; CHECK-NEXT:    vinserti128 $1, %xmm2, %ymm11, %ymm11
+; CHECK-NEXT:    vextracti128 $1, %ymm7, %xmm2
+; CHECK-NEXT:    vmovq {{.*#+}} xmm2 = xmm2[0],zero
+; CHECK-NEXT:    # implicit-def: $ymm6
+; CHECK-NEXT:    vmovaps %xmm2, %xmm6
+; CHECK-NEXT:    vpblendd {{.*#+}} ymm2 = ymm6[0,1,2,3],ymm11[4,5,6,7]
 ; CHECK-NEXT:    vmovaps %xmm7, %xmm6
 ; CHECK-NEXT:    vpslldq {{.*#+}} xmm6 = zero,zero,zero,zero,zero,zero,zero,zero,xmm6[0,1,2,3,4,5,6,7]
 ; CHECK-NEXT:    # implicit-def: $ymm11
 ; CHECK-NEXT:    vmovaps %xmm6, %xmm11
 ; CHECK-NEXT:    vpalignr {{.*#+}} ymm9 = ymm9[8,9,10,11,12,13,14,15],ymm5[0,1,2,3,4,5,6,7],ymm9[24,25,26,27,28,29,30,31],ymm5[16,17,18,19,20,21,22,23]
 ; CHECK-NEXT:    vpermq {{.*#+}} ymm9 = ymm9[0,1,0,3]
 ; CHECK-NEXT:    vpblendd {{.*#+}} ymm9 = ymm11[0,1,2,3],ymm9[4,5,6,7]
 ; CHECK-NEXT:    vpblendd {{.*#+}} ymm7 = ymm7[0,1],ymm8[2,3],ymm7[4,5,6,7]

test/CodeGen/X86/pr41678.ll

 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc < %s -O0 -mtriple=i386-pc-linux-gnu -mattr=avx512f | FileCheck %s
 
 define void @a() {
 ; CHECK-LABEL: a:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    subl $2, %esp
 ; CHECK-NEXT:    .cfi_def_cfa_offset 6
 ; CHECK-NEXT:    #APP
 ; CHECK-NEXT:    #NO_APP
 ; CHECK-NEXT:    kmovw %k6, %eax
-; CHECK-NEXT:    movw %ax, %cx
-; CHECK-NEXT:    movw %cx, (%esp)
+; CHECK-NEXT:    # kill: def $ax killed $ax killed $eax
+; CHECK-NEXT:    movw %ax, (%esp)
 ; CHECK-NEXT:    addl $2, %esp
 ; CHECK-NEXT:    .cfi_def_cfa_offset 4
 ; CHECK-NEXT:    retl
 entry:
   %b = alloca i16, align 2
   %0 = call i16 asm "", "={k6},~{dirflag},~{fpsr},~{flags}"() #1
   store i16 %0, i16* %b, align 2
   ret void
 }

test/CodeGen/X86/swift-return.ll

 ; CHECK-O0-NEXT:    pushq %rax
 ; CHECK-O0-NEXT:    .cfi_def_cfa_offset 16
 ; CHECK-O0-NEXT:    movl %edi, {{[0-9]+}}(%rsp)
 ; CHECK-O0-NEXT:    movl {{[0-9]+}}(%rsp), %edi
 ; CHECK-O0-NEXT:    callq gen
 ; CHECK-O0-NEXT:    movswl %ax, %ecx
 ; CHECK-O0-NEXT:    movsbl %dl, %esi
 ; CHECK-O0-NEXT:    addl %esi, %ecx
+; CHECK-O0-NEXT:    # kill: def $cx killed $cx killed $ecx
 ; CHECK-O0-NEXT:    movw %cx, %ax
 ; CHECK-O0-NEXT:    popq %rcx
 ; CHECK-O0-NEXT:    .cfi_def_cfa_offset 8
 ; CHECK-O0-NEXT:    retq
 entry:
   %key.addr = alloca i32, align 4
   store i32 %key, i32* %key.addr, align 4
   %0 = load i32, i32* %key.addr, align 4

 ; CHECK-NEXT:    movl %edi, %eax
 ; CHECK-NEXT:    movl %eax, %edx
 ; CHECK-NEXT:    movl %eax, %ecx
 ; CHECK-NEXT:    movl %eax, %r8d
 ; CHECK-NEXT:    retq
 ;
 ; CHECK-O0-LABEL: gen9:
 ; CHECK-O0:       # %bb.0:
+; CHECK-O0-NEXT:    # kill: def $dil killed $dil killed $edi
 ; CHECK-O0-NEXT:    movb %dil, %al
-; CHECK-O0-NEXT:    movb %al, {{[-0-9]+}}(%r{{[sb]}}p) # 1-byte Spill
-; CHECK-O0-NEXT:    movb {{[-0-9]+}}(%r{{[sb]}}p), %dl # 1-byte Reload
-; CHECK-O0-NEXT:    movb {{[-0-9]+}}(%r{{[sb]}}p), %cl # 1-byte Reload
-; CHECK-O0-NEXT:    movb {{[-0-9]+}}(%r{{[sb]}}p), %r8b # 1-byte Reload
+; CHECK-O0-NEXT:    movb %dil, %dl
+; CHECK-O0-NEXT:    movb %dil, %cl
+; CHECK-O0-NEXT:    movb %dil, %r8b
 ; CHECK-O0-NEXT:    retq
   %v0 = insertvalue { i8, i8, i8, i8 } undef, i8 %key, 0
   %v1 = insertvalue { i8, i8, i8, i8 } %v0, i8 %key, 1
   %v2 = insertvalue { i8, i8, i8, i8 } %v1, i8 %key, 2
   %v3 = insertvalue { i8, i8, i8, i8 } %v2, i8 %key, 3
   ret { i8, i8, i8, i8 } %v3
 }
 define swiftcc { double, double, double, double, i64, i64, i64, i64 } @gen10(double %keyd, i64 %keyi) {

test/DebugInfo/X86/convert-debugloc.ll

 ; RUN: llc -dwarf-version=5 -filetype=obj -O0 < %s | llvm-dwarfdump - | FileCheck %s --match-full-lines --check-prefix=DW5-CHECK
 ; RUN: llc -dwarf-version=4 -filetype=obj -O0 < %s | llvm-dwarfdump - | FileCheck %s --match-full-lines --check-prefix=DW4-CHECK
 
 ; DW5-CHECK: .debug_info contents:
-; DW5-CHECK-NEXT: 0x00000000: Compile Unit: length = 0x0000005c version = 0x0005 unit_type = DW_UT_compile abbr_offset = 0x0000 addr_size = 0x08 (next unit at 0x00000060)
+; DW5-CHECK-NEXT: 0x00000000: Compile Unit: length = 0x00000064 version = 0x0005 unit_type = DW_UT_compile abbr_offset = 0x0000 addr_size = 0x08 (next unit at 0x00000068)
 ; DW5-CHECK-EMPTY:
 ; DW5-CHECK-NEXT: 0x0000000c: DW_TAG_compile_unit
 ; DW5-CHECK-NEXT:               DW_AT_producer  ("clang version 9.0.0 (trunk 353791) (llvm/trunk 353801)")
 ; DW5-CHECK-NEXT:               DW_AT_language  (DW_LANG_C99)
 ; DW5-CHECK-NEXT:               DW_AT_name  ("dbg.c")
 ; DW5-CHECK-NEXT:               DW_AT_str_offsets_base  (0x00000008)
 ; DW5-CHECK-NEXT:               DW_AT_stmt_list (0x00000000)
 ; DW5-CHECK-NEXT:               DW_AT_comp_dir {{.*}}
 ; DW5-CHECK-NEXT:               DW_AT_addr_base (0x00000008)
 ; DW5-CHECK-NEXT:               DW_AT_low_pc  (0x0000000000000000)
-; DW5-CHECK-NEXT:               DW_AT_high_pc (0x0000000000000007)
-; DW5-CHECK-NEXT:               DW_AT_loclists_base (0x0000000c)
+; DW5-CHECK-NEXT:               DW_AT_high_pc (0x0000000000000005)
 ; DW5-CHECK-EMPTY:
-; DW5-CHECK-NEXT: 0x00000027:   DW_TAG_base_type
+; DW5-CHECK-NEXT: 0x00000023:   DW_TAG_base_type
 ; DW5-CHECK-NEXT:                 DW_AT_name  ("DW_ATE_signed_8")
 ; DW5-CHECK-NEXT:                 DW_AT_encoding  (DW_ATE_signed)
 ; DW5-CHECK-NEXT:                 DW_AT_byte_size (0x01)
 ; DW5-CHECK-EMPTY:
-; DW5-CHECK-NEXT: 0x0000002b:   DW_TAG_base_type
+; DW5-CHECK-NEXT: 0x00000027:   DW_TAG_base_type
 ; DW5-CHECK-NEXT:                 DW_AT_name  ("DW_ATE_signed_32")
 ; DW5-CHECK-NEXT:                 DW_AT_encoding  (DW_ATE_signed)
 ; DW5-CHECK-NEXT:                 DW_AT_byte_size (0x04)
 ; DW5-CHECK-EMPTY:
-; DW5-CHECK-NEXT: 0x0000002f:   DW_TAG_subprogram
+; DW5-CHECK-NEXT: 0x0000002b:   DW_TAG_subprogram
 ; DW5-CHECK-NEXT:                 DW_AT_low_pc  (0x0000000000000000)
-; DW5-CHECK-NEXT:                 DW_AT_high_pc (0x0000000000000007)
+; DW5-CHECK-NEXT:                 DW_AT_high_pc (0x0000000000000005)
 ; DW5-CHECK-NEXT:                 DW_AT_frame_base  (DW_OP_reg7 RSP)
 ; DW5-CHECK-NEXT:                 DW_AT_name  ("foo")
 ; DW5-CHECK-NEXT:                 DW_AT_decl_file {{.*}}
 ; DW5-CHECK-NEXT:                 DW_AT_decl_line (1)
 ; DW5-CHECK-NEXT:                 DW_AT_prototyped  (true)
-; DW5-CHECK-NEXT:                 DW_AT_type  (0x00000057 "signed char")
+; DW5-CHECK-NEXT:                 DW_AT_type  (0x0000005f "signed char")
 ; DW5-CHECK-NEXT:                 DW_AT_external  (true)
 ; DW5-CHECK-EMPTY:
-; DW5-CHECK-NEXT: 0x0000003e:     DW_TAG_formal_parameter
-; DW5-CHECK-NEXT:                   DW_AT_location  (0x0000000c
-; DW5-CHECK-NEXT:                      [0x0000000000000003, 0x0000000000000006): DW_OP_reg0 RAX)
+; DW5-CHECK-NEXT: 0x0000003a:     DW_TAG_formal_parameter
+; DW5-CHECK-NEXT:                   DW_AT_location  (DW_OP_reg5 RDI)
 ; DW5-CHECK-NEXT:                   DW_AT_name  ("x")
 ; DW5-CHECK-NEXT:                   DW_AT_decl_file {{.*}}
 ; DW5-CHECK-NEXT:                   DW_AT_decl_line (1)
-; DW5-CHECK-NEXT:                   DW_AT_type  (0x00000057 "signed char")
+; DW5-CHECK-NEXT:                   DW_AT_type  (0x0000005f "signed char")
 ; DW5-CHECK-EMPTY:
-; DW5-CHECK-NEXT: 0x0000004a:     DW_TAG_variable
-; DW5-CHECK-NEXT:                   DW_AT_location  (0x00000012
-; DW5-CHECK-NEXT:                      [0x0000000000000003, 0x0000000000000006): DW_OP_breg0 RAX+0, DW_OP_constu 0xff, DW_OP_and, DW_OP_convert (0x00000027) "DW_ATE_signed_8", DW_OP_convert (0x0000002b) "DW_ATE_signed_32", DW_OP_stack_value)
+; DW5-CHECK-NEXT: 0x00000044:     DW_TAG_variable
+; DW5-CHECK-NEXT:                   DW_AT_location  (DW_OP_breg5 RDI+0, DW_OP_constu 0xff, DW_OP_and, DW_OP_convert (0x00000023) "DW_ATE_signed_8", DW_OP_convert (0x00000027) "DW_ATE_signed_32", DW_OP_stack_value)
 ; DW5-CHECK-NEXT:                   DW_AT_name  ("y")
 ; DW5-CHECK-NEXT:                   DW_AT_decl_file {{.*}}
 ; DW5-CHECK-NEXT:                   DW_AT_decl_line (3)
-; DW5-CHECK-NEXT:                   DW_AT_type  (0x0000005b "int")
+; DW5-CHECK-NEXT:                   DW_AT_type  (0x00000063 "int")
 ; DW5-CHECK-EMPTY:
-; DW5-CHECK-NEXT: 0x00000056:     NULL
+; DW5-CHECK-NEXT: 0x0000005e:     NULL
 ; DW5-CHECK-EMPTY:
-; DW5-CHECK-NEXT: 0x00000057:   DW_TAG_base_type
+; DW5-CHECK-NEXT: 0x0000005f:   DW_TAG_base_type
 ; DW5-CHECK-NEXT:                 DW_AT_name  ("signed char")
 ; DW5-CHECK-NEXT:                 DW_AT_encoding  (DW_ATE_signed_char)
 ; DW5-CHECK-NEXT:                 DW_AT_byte_size (0x01)
 ; DW5-CHECK-EMPTY:
-; DW5-CHECK-NEXT: 0x0000005b:   DW_TAG_base_type
+; DW5-CHECK-NEXT: 0x00000063:   DW_TAG_base_type
 ; DW5-CHECK-NEXT:                 DW_AT_name  ("int")
 ; DW5-CHECK-NEXT:                 DW_AT_encoding  (DW_ATE_signed)
 ; DW5-CHECK-NEXT:                 DW_AT_byte_size (0x04)
 ; DW5-CHECK-EMPTY:
-; DW5-CHECK-NEXT: 0x0000005f:   NULL
+; DW5-CHECK-NEXT: 0x00000067:   NULL
 
 
 ; DW4-CHECK: .debug_info contents:
-; DW4-CHECK-NEXT: 0x00000000: Compile Unit: length = 0x0000006d version = 0x0004 abbr_offset = 0x0000 addr_size = 0x08 (next unit at 0x00000071)
+; DW4-CHECK-NEXT: 0x00000000: Compile Unit: length = 0x0000007a version = 0x0004 abbr_offset = 0x0000 addr_size = 0x08 (next unit at 0x0000007e)
 ; DW4-CHECK-EMPTY:
 ; DW4-CHECK-NEXT: 0x0000000b: DW_TAG_compile_unit
 ; DW4-CHECK-NEXT:               DW_AT_producer  ("clang version 9.0.0 (trunk 353791) (llvm/trunk 353801)")
 ; DW4-CHECK-NEXT:               DW_AT_language  (DW_LANG_C99)
 ; DW4-CHECK-NEXT:               DW_AT_name  ("dbg.c")
 ; DW4-CHECK-NEXT:               DW_AT_stmt_list (0x00000000)
 ; DW4-CHECK-NEXT:               DW_AT_comp_dir {{.*}}
 ; DW4-CHECK-NEXT:               DW_AT_low_pc  (0x0000000000000000)
-; DW4-CHECK-NEXT:               DW_AT_high_pc (0x0000000000000007)
+; DW4-CHECK-NEXT:               DW_AT_high_pc (0x0000000000000005)
 ; DW4-CHECK-EMPTY:
 ; DW4-CHECK-NEXT: 0x0000002a:   DW_TAG_subprogram
 ; DW4-CHECK-NEXT:                 DW_AT_low_pc  (0x0000000000000000)
-; DW4-CHECK-NEXT:                 DW_AT_high_pc (0x0000000000000007)
+; DW4-CHECK-NEXT:                 DW_AT_high_pc (0x0000000000000005)
 ; DW4-CHECK-NEXT:                 DW_AT_frame_base  (DW_OP_reg7 RSP)
 ; DW4-CHECK-NEXT:                 DW_AT_name  ("foo")
 ; DW4-CHECK-NEXT:                 DW_AT_decl_file {{.*}}
 ; DW4-CHECK-NEXT:                 DW_AT_decl_line (1)
 ; DW4-CHECK-NEXT:                 DW_AT_prototyped  (true)
-; DW4-CHECK-NEXT:                 DW_AT_type  (0x00000062 "signed char")
+; DW4-CHECK-NEXT:                 DW_AT_type  (0x0000006f "signed char")
 ; DW4-CHECK-NEXT:                 DW_AT_external  (true)
 ; DW4-CHECK-EMPTY:
 ; DW4-CHECK-NEXT: 0x00000043:     DW_TAG_formal_parameter
-; DW4-CHECK-NEXT:                   DW_AT_location  (0x00000000
-; DW4-CHECK-NEXT:                      [0x0000000000000003,  0x0000000000000006): DW_OP_reg0 RAX)
+; DW4-CHECK-NEXT:                   DW_AT_location  (DW_OP_reg5 RDI)
 ; DW4-CHECK-NEXT:                   DW_AT_name  ("x")
 ; DW4-CHECK-NEXT:                   DW_AT_decl_file {{.*}}
 ; DW4-CHECK-NEXT:                   DW_AT_decl_line (1)
-; DW4-CHECK-NEXT:                   DW_AT_type  (0x00000062 "signed char")
+; DW4-CHECK-NEXT:                   DW_AT_type  (0x0000006f "signed char")
 ; DW4-CHECK-EMPTY:
-; DW4-CHECK-NEXT: 0x00000052:     DW_TAG_variable
-; DW4-CHECK-NEXT:                   DW_AT_location  (0x00000023
-; DW4-CHECK-NEXT:                      [0x0000000000000003,  0x0000000000000006): DW_OP_breg0 RAX+0, DW_OP_constu 0xff, DW_OP_and, DW_OP_dup, DW_OP_constu 0x7, DW_OP_shr, DW_OP_lit0, DW_OP_not, DW_OP_mul, DW_OP_constu 0x8, DW_OP_shl, DW_OP_or, DW_OP_stack_value)
+; DW4-CHECK-NEXT: 0x00000050:     DW_TAG_variable
+; DW4-CHECK-NEXT:                   DW_AT_location  (DW_OP_breg5 RDI+0, DW_OP_constu 0xff, DW_OP_and, DW_OP_dup, DW_OP_constu 0x7, DW_OP_shr, DW_OP_lit0, DW_OP_not, DW_OP_mul, DW_OP_constu 0x8, DW_OP_shl, DW_OP_or, DW_OP_stack_value)
 ; DW4-CHECK-NEXT:                   DW_AT_name  ("y")
 ; DW4-CHECK-NEXT:                   DW_AT_decl_file {{.*}}
 ; DW4-CHECK-NEXT:                   DW_AT_decl_line (3)
-; DW4-CHECK-NEXT:                   DW_AT_type  (0x00000069 "int")
+; DW4-CHECK-NEXT:                   DW_AT_type  (0x00000076 "int")
 ; DW4-CHECK-EMPTY:
-; DW4-CHECK-NEXT: 0x00000061:     NULL
+; DW4-CHECK-NEXT: 0x0000006e:     NULL
 ; DW4-CHECK-EMPTY:
-; DW4-CHECK-NEXT: 0x00000062:   DW_TAG_base_type
+; DW4-CHECK-NEXT: 0x0000006f:   DW_TAG_base_type
 ; DW4-CHECK-NEXT:                 DW_AT_name  ("signed char")
 ; DW4-CHECK-NEXT:                 DW_AT_encoding  (DW_ATE_signed_char)
 ; DW4-CHECK-NEXT:                 DW_AT_byte_size (0x01)
 ; DW4-CHECK-EMPTY:
-; DW4-CHECK-NEXT: 0x00000069:   DW_TAG_base_type
+; DW4-CHECK-NEXT: 0x00000076:   DW_TAG_base_type
 ; DW4-CHECK-NEXT:                 DW_AT_name  ("int")
 ; DW4-CHECK-NEXT:                 DW_AT_encoding  (DW_ATE_signed)
 ; DW4-CHECK-NEXT:                 DW_AT_byte_size (0x04)
 ; DW4-CHECK-EMPTY:
-; DW4-CHECK-NEXT: 0x00000070:   NULL
+; DW4-CHECK-NEXT: 0x0000007d:   NULL
 
 
 ; ModuleID = 'dbg.ll'
 source_filename = "dbg.c"
 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
 target triple = "x86_64-unknown-linux-gnu"
 
 ; Function Attrs: noinline nounwind uwtable

test/DebugInfo/X86/subreg.ll

 ; RUN: llc %s -mtriple=x86_64-pc-linux-gnu -O0 -o - | FileCheck %s
 
 ; We are testing that a value in a 16 bit register gets reported as
 ; being in its superregister.
 
-; CHECK: .byte   80                      # super-register DW_OP_reg0
+; CHECK: .byte   85                      # super-register DW_OP_reg5
 ; No need to a piece at offset 0.
 ; CHECK-NOT: DW_OP_piece
 ; CHECK-NOT: DW_OP_bit_piece
 
 define i16 @f(i16 signext %zzz) nounwind !dbg !1 {
 entry:
   call void @llvm.dbg.value(metadata i16 %zzz, metadata !0, metadata !DIExpression()), !dbg !DILocation(scope: !1)
+  br label %exit
+
+exit:
   %conv = sext i16 %zzz to i32, !dbg !7
   %conv1 = trunc i32 %conv to i16
   ret i16 %conv1
 }
 
 declare void @llvm.dbg.value(metadata, metadata, metadata) nounwind readnone
 
 !llvm.dbg.cu = !{!3}