[RISC-V] Add proposed mapping for Ztso

[patrick-rivos]

Summary

Currently LLVM emits Ztso code for fences, loads, and stores (behind an experimental flag) [1]. This patch updates the mapping and implements support for LR/SC and AMO ops. This updated mapping is compatible with the RVWMO ABI present in the psABI. Additional context can be found in the psABI pull request [2].

[1] https://reviews.llvm.org/D143076
[2] https://github.com/riscv-non-isa/riscv-elf-psabi-doc/pull/391

[asb]

From a first pass through, this appears to match the proposal. I've left one style comment that should be addressed.

I'm not convinced by the new helper in RISCVInstrInfo - might it be cleaner to just pass the RISCVSubtarget to the helpers? I think conventionally we'd put the subtarget argument last for this kind of helper

llvm/lib/Target/RISCV/RISCVExpandAtomicPseudoInsts.cpp

160

LLVM style is to not use an else after return. https://llvm.org/docs/CodingStandards.html#don-t-use-else-after-a-return

Same applies elsewhere in this patch

[reames]

Meta comment - mostly just chiming is because I'm the person who did the prior implementation, and I want it to be clear I support the direction here.

Given what looks to be a defacto change in the WMO memory model lowering which is going to be widely deployed, we should change the TSO mapping to be compatible with the planned mapping. The goal of the original change was to allow WMO and TSO objects to be compatible; this change continues in the same spirit, it's just that the assumed WMO lowering is changing.

This case - which we did not know about at the time of the original patch - is exactly why we left ourselves the flexibility of only supporting this experimentally. We absolutely should not, under any circumstances, promote this out of experimental until the psABI changes for WMO and TSO have finalized.

[patrick-rivos]

Revised to address @asb review comments.

I had to unprotect the subtarget attribute in order to pass it to the helper functions. If there's a better way to pass the subtarget please let me know!

[asb]

@patrick-rivos: I've just posted D155840 to introduce a field for RISCVSubtarget in those pseudo expansion passes. I think that would probably be a good approach here, but you might want to wait for any feedback on that review before changing again in case others disagree with me!

LGTM once we resolve that (I'm expecting a quick review on that linked patch - if it's slow for some reason let's loop back and make sure this patch isn't blocked on it unreasonably). Thanks for such a rapid update, and of course for updating us to the most recent proposed mappings.

[asb]

Ok, D155840 was reviewed and committed so I think following that same approach here (and leaving the STI field of RISCVInstrInfo as protected) probably makes sense. Then this patch LGTM. Thanks for your patience on these minor implementation tweaks.

[patrick-rivos]

Revised to keep the STI field of RISCVInstrInfo protected and get STI in runOnMachineFunction (Same approach as D155840).
Also removed a stray else I had missed on the earlier revision.

[asb]

LGTM, thanks!

[patrick-rivos]

The psABI Ztso proposal was merged today: https://github.com/riscv-non-isa/riscv-elf-psabi-doc/pull/391

[asb]

In D155517#4571330, @patrick-rivos wrote:

The psABI Ztso proposal was merged today: https://github.com/riscv-non-isa/riscv-elf-psabi-doc/pull/391

Brilliant. My LGTM still stands so I think this is good to merge once it's rebased (I think you basically just need to regenerate the test changes with update_llc_test_checks.py). Do you have commit access?

[patrick-rivos]

Rebased to main and regenerated changes to the testsuite.

[patrick-rivos]

Do you have commit access?

AFAIK no I don't have commit access.

[asb]

I've committed for you. Thanks for working on this!

[patrick-rivos]

Great, thank you!

llvm/lib/Target/RISCV/RISCVExpandAtomicPseudoInsts.cpp

case RISCV::PseudoCmpXchg64:
     return expandAtomicCmpXchg(MBB, MBBI, false, 64, NextMBBI);
   case RISCV::PseudoMaskedCmpXchg32:
     return expandAtomicCmpXchg(MBB, MBBI, true, 32, NextMBBI);
   }
 
   return false;
 }
 
-static unsigned getLRForRMW32(AtomicOrdering Ordering) {
+static unsigned getLRForRMW32(bool HasStdExtZtso, AtomicOrdering Ordering) {
   switch (Ordering) {
   default:
     llvm_unreachable("Unexpected AtomicOrdering");
   case AtomicOrdering::Monotonic:
     return RISCV::LR_W;
   case AtomicOrdering::Acquire:
+    if (HasStdExtZtso)
+      return RISCV::LR_W;
+    else
       return RISCV::LR_W_AQ;
   case AtomicOrdering::Release:
     return RISCV::LR_W;
   case AtomicOrdering::AcquireRelease:
+    if (HasStdExtZtso)
+      return RISCV::LR_W;
+    else
       return RISCV::LR_W_AQ;
   case AtomicOrdering::SequentiallyConsistent:
     return RISCV::LR_W_AQ_RL;
   }
 }
 
-static unsigned getSCForRMW32(AtomicOrdering Ordering) {
+static unsigned getSCForRMW32(bool HasStdExtZtso, AtomicOrdering Ordering) {
   switch (Ordering) {
   default:
     llvm_unreachable("Unexpected AtomicOrdering");
   case AtomicOrdering::Monotonic:
     return RISCV::SC_W;
   case AtomicOrdering::Acquire:
     return RISCV::SC_W;
   case AtomicOrdering::Release:
+    if (HasStdExtZtso)
+      return RISCV::SC_W;
+    else
       return RISCV::SC_W_RL;
   case AtomicOrdering::AcquireRelease:
+    if (HasStdExtZtso)
+      return RISCV::SC_W;
+    else
       return RISCV::SC_W_RL;
   case AtomicOrdering::SequentiallyConsistent:
     return RISCV::SC_W_RL;
   }
 }
 
-static unsigned getLRForRMW64(AtomicOrdering Ordering) {
+static unsigned getLRForRMW64(bool HasStdExtZtso, AtomicOrdering Ordering) {
   switch (Ordering) {
   default:
     llvm_unreachable("Unexpected AtomicOrdering");
   case AtomicOrdering::Monotonic:
     return RISCV::LR_D;
   case AtomicOrdering::Acquire:
+    if (HasStdExtZtso)
+      return RISCV::LR_D;
+    else
       return RISCV::LR_D_AQ;
   case AtomicOrdering::Release:
     return RISCV::LR_D;
   case AtomicOrdering::AcquireRelease:
+    if (HasStdExtZtso)
+      return RISCV::LR_D;
+    else
       return RISCV::LR_D_AQ;
   case AtomicOrdering::SequentiallyConsistent:
     return RISCV::LR_D_AQ_RL;
   }
 }
 
-static unsigned getSCForRMW64(AtomicOrdering Ordering) {
+static unsigned getSCForRMW64(bool HasStdExtZtso, AtomicOrdering Ordering) {
   switch (Ordering) {
   default:
     llvm_unreachable("Unexpected AtomicOrdering");
   case AtomicOrdering::Monotonic:
     return RISCV::SC_D;
   case AtomicOrdering::Acquire:
     return RISCV::SC_D;
   case AtomicOrdering::Release:
+    if (HasStdExtZtso)
+      return RISCV::SC_D;
+    else
       return RISCV::SC_D_RL;
   case AtomicOrdering::AcquireRelease:
+    if (HasStdExtZtso)
+      return RISCV::SC_D;
+    else
       return RISCV::SC_D_RL;
   case AtomicOrdering::SequentiallyConsistent:
     return RISCV::SC_D_RL;
   }
 }
 
-static unsigned getLRForRMW(AtomicOrdering Ordering, int Width) {
+static unsigned getLRForRMW(bool HasStdExtZtso, AtomicOrdering Ordering,
+                            int Width) {
   if (Width == 32)
-    return getLRForRMW32(Ordering);
+    return getLRForRMW32(HasStdExtZtso, Ordering);
   if (Width == 64)
-    return getLRForRMW64(Ordering);
+    return getLRForRMW64(HasStdExtZtso, Ordering);
   llvm_unreachable("Unexpected LR width\n");
 }
 
-static unsigned getSCForRMW(AtomicOrdering Ordering, int Width) {
+static unsigned getSCForRMW(bool HasStdExtZtso, AtomicOrdering Ordering,
+                            int Width) {
   if (Width == 32)
-    return getSCForRMW32(Ordering);
+    return getSCForRMW32(HasStdExtZtso, Ordering);
   if (Width == 64)
-    return getSCForRMW64(Ordering);
+    return getSCForRMW64(HasStdExtZtso, Ordering);
   llvm_unreachable("Unexpected SC width\n");
 }
 
 static void doAtomicBinOpExpansion(const RISCVInstrInfo *TII, MachineInstr &MI,
                                    DebugLoc DL, MachineBasicBlock *ThisMBB,
                                    MachineBasicBlock *LoopMBB,
                                    MachineBasicBlock *DoneMBB,
                                    AtomicRMWInst::BinOp BinOp, int Width) {
   Register DestReg = MI.getOperand(0).getReg();
   Register ScratchReg = MI.getOperand(1).getReg();
   Register AddrReg = MI.getOperand(2).getReg();
   Register IncrReg = MI.getOperand(3).getReg();
   AtomicOrdering Ordering =
       static_cast<AtomicOrdering>(MI.getOperand(4).getImm());
 
   // .loop:
   //   lr.[w|d] dest, (addr)
   //   binop scratch, dest, val
   //   sc.[w|d] scratch, scratch, (addr)
   //   bnez scratch, loop
-  BuildMI(LoopMBB, DL, TII->get(getLRForRMW(Ordering, Width)), DestReg)
+  BuildMI(LoopMBB, DL,
+          TII->get(getLRForRMW(TII->subtargetHasStdExtZtso(), Ordering, Width)),
+          DestReg)
       .addReg(AddrReg);
   switch (BinOp) {
   default:
     llvm_unreachable("Unexpected AtomicRMW BinOp");
   case AtomicRMWInst::Nand:
     BuildMI(LoopMBB, DL, TII->get(RISCV::AND), ScratchReg)
         .addReg(DestReg)
         .addReg(IncrReg);
     BuildMI(LoopMBB, DL, TII->get(RISCV::XORI), ScratchReg)
         .addReg(ScratchReg)
         .addImm(-1);
     break;
   }
-  BuildMI(LoopMBB, DL, TII->get(getSCForRMW(Ordering, Width)), ScratchReg)
+  BuildMI(LoopMBB, DL,
+          TII->get(getSCForRMW(TII->subtargetHasStdExtZtso(), Ordering, Width)),
+          ScratchReg)
       .addReg(AddrReg)
       .addReg(ScratchReg);
   BuildMI(LoopMBB, DL, TII->get(RISCV::BNE))
       .addReg(ScratchReg)
       .addReg(RISCV::X0)
       .addMBB(LoopMBB);
 }
 
static void doMaskedAtomicBinOpExpansion(
   // .loop:
   //   lr.w destreg, (alignedaddr)
   //   binop scratch, destreg, incr
   //   xor scratch, destreg, scratch
   //   and scratch, scratch, masktargetdata
   //   xor scratch, destreg, scratch
   //   sc.w scratch, scratch, (alignedaddr)
   //   bnez scratch, loop
-  BuildMI(LoopMBB, DL, TII->get(getLRForRMW32(Ordering)), DestReg)
+  BuildMI(LoopMBB, DL,
+          TII->get(getLRForRMW32(TII->subtargetHasStdExtZtso(), Ordering)),
+          DestReg)
       .addReg(AddrReg);
   switch (BinOp) {
   default:
     llvm_unreachable("Unexpected AtomicRMW BinOp");
   case AtomicRMWInst::Xchg:
     BuildMI(LoopMBB, DL, TII->get(RISCV::ADDI), ScratchReg)
         .addReg(IncrReg)
         .addImm(0);
BuildMI(LoopMBB, DL, TII->get(RISCV::XORI), ScratchReg)
         .addReg(ScratchReg)
         .addImm(-1);
     break;
   }
 
   insertMaskedMerge(TII, DL, LoopMBB, ScratchReg, DestReg, ScratchReg, MaskReg,
                     ScratchReg);
 
-  BuildMI(LoopMBB, DL, TII->get(getSCForRMW32(Ordering)), ScratchReg)
+  BuildMI(LoopMBB, DL,
+          TII->get(getSCForRMW32(TII->subtargetHasStdExtZtso(), Ordering)),
+          ScratchReg)
       .addReg(AddrReg)
       .addReg(ScratchReg);
   BuildMI(LoopMBB, DL, TII->get(RISCV::BNE))
       .addReg(ScratchReg)
       .addReg(RISCV::X0)
       .addMBB(LoopMBB);
 }
 
bool RISCVExpandAtomicPseudo::expandAtomicMinMaxOp(
 
   //
   // .loophead:
   //   lr.w destreg, (alignedaddr)
   //   and scratch2, destreg, mask
   //   mv scratch1, destreg
   //   [sext scratch2 if signed min/max]
   //   ifnochangeneeded scratch2, incr, .looptail
-  BuildMI(LoopHeadMBB, DL, TII->get(getLRForRMW32(Ordering)), DestReg)
+  BuildMI(LoopHeadMBB, DL,
+          TII->get(getLRForRMW32(TII->subtargetHasStdExtZtso(), Ordering)),
+          DestReg)
       .addReg(AddrReg);
   BuildMI(LoopHeadMBB, DL, TII->get(RISCV::AND), Scratch2Reg)
       .addReg(DestReg)
       .addReg(MaskReg);
   BuildMI(LoopHeadMBB, DL, TII->get(RISCV::ADDI), Scratch1Reg)
       .addReg(DestReg)
       .addImm(0);
 
bool RISCVExpandAtomicPseudo::expandAtomicMinMaxOp(
   //   and scratch1, scratch1, mask
   //   xor scratch1, destreg, scratch1
   insertMaskedMerge(TII, DL, LoopIfBodyMBB, Scratch1Reg, DestReg, IncrReg,
                     MaskReg, Scratch1Reg);
 
   // .looptail:
   //   sc.w scratch1, scratch1, (addr)
   //   bnez scratch1, loop
-  BuildMI(LoopTailMBB, DL, TII->get(getSCForRMW32(Ordering)), Scratch1Reg)
+  BuildMI(LoopTailMBB, DL,
+          TII->get(getSCForRMW32(TII->subtargetHasStdExtZtso(), Ordering)),
+          Scratch1Reg)
       .addReg(AddrReg)
       .addReg(Scratch1Reg);
   BuildMI(LoopTailMBB, DL, TII->get(RISCV::BNE))
       .addReg(Scratch1Reg)
       .addReg(RISCV::X0)
       .addMBB(LoopHeadMBB);
 
   NextMBBI = MBB.end();
bool RISCVExpandAtomicPseudo::expandAtomicCmpXchg(
 
   AtomicOrdering Ordering =
       static_cast<AtomicOrdering>(MI.getOperand(IsMasked ? 6 : 5).getImm());
 
   if (!IsMasked) {
     // .loophead:
     //   lr.[w|d] dest, (addr)
     //   bne dest, cmpval, done
-    BuildMI(LoopHeadMBB, DL, TII->get(getLRForRMW(Ordering, Width)), DestReg)
+    BuildMI(
+        LoopHeadMBB, DL,
+        TII->get(getLRForRMW(TII->subtargetHasStdExtZtso(), Ordering, Width)),
+        DestReg)
         .addReg(AddrReg);
     BuildMI(LoopHeadMBB, DL, TII->get(RISCV::BNE))
         .addReg(DestReg)
         .addReg(CmpValReg)
         .addMBB(LoopHeadBNETarget);
     // .looptail:
     //   sc.[w|d] scratch, newval, (addr)
     //   bnez scratch, loophead
-    BuildMI(LoopTailMBB, DL, TII->get(getSCForRMW(Ordering, Width)), ScratchReg)
+    BuildMI(
+        LoopTailMBB, DL,
+        TII->get(getSCForRMW(TII->subtargetHasStdExtZtso(), Ordering, Width)),
+        ScratchReg)
         .addReg(AddrReg)
         .addReg(NewValReg);
     BuildMI(LoopTailMBB, DL, TII->get(RISCV::BNE))
         .addReg(ScratchReg)
         .addReg(RISCV::X0)
         .addMBB(LoopHeadMBB);
   } else {
     // .loophead:
     //   lr.w dest, (addr)
     //   and scratch, dest, mask
     //   bne scratch, cmpval, done
     Register MaskReg = MI.getOperand(5).getReg();
-    BuildMI(LoopHeadMBB, DL, TII->get(getLRForRMW(Ordering, Width)), DestReg)
+    BuildMI(
+        LoopHeadMBB, DL,
+        TII->get(getLRForRMW(TII->subtargetHasStdExtZtso(), Ordering, Width)),
+        DestReg)
         .addReg(AddrReg);
     BuildMI(LoopHeadMBB, DL, TII->get(RISCV::AND), ScratchReg)
         .addReg(DestReg)
         .addReg(MaskReg);
     BuildMI(LoopHeadMBB, DL, TII->get(RISCV::BNE))
         .addReg(ScratchReg)
         .addReg(CmpValReg)
         .addMBB(LoopHeadBNETarget);
 
     // .looptail:
     //   xor scratch, dest, newval
     //   and scratch, scratch, mask
     //   xor scratch, dest, scratch
     //   sc.w scratch, scratch, (adrr)
     //   bnez scratch, loophead
     insertMaskedMerge(TII, DL, LoopTailMBB, ScratchReg, DestReg, NewValReg,
                       MaskReg, ScratchReg);
-    BuildMI(LoopTailMBB, DL, TII->get(getSCForRMW(Ordering, Width)), ScratchReg)
+    BuildMI(
+        LoopTailMBB, DL,
+        TII->get(getSCForRMW(TII->subtargetHasStdExtZtso(), Ordering, Width)),
+        ScratchReg)
         .addReg(AddrReg)
         .addReg(ScratchReg);
     BuildMI(LoopTailMBB, DL, TII->get(RISCV::BNE))
         .addReg(ScratchReg)
         .addReg(RISCV::X0)
         .addMBB(LoopHeadMBB);
   }
 

llvm/lib/Target/RISCV/RISCVFeatures.td

def HasStdExtSvinval : Predicate<"Subtarget->hasStdExtSvinval()">,
                                 "'Svinval' (Fine-Grained Address-Translation Cache Invalidation)">;
 
 def FeatureStdExtZtso
     : SubtargetFeature<"experimental-ztso", "HasStdExtZtso", "true",
                        "'Ztso' (Memory Model - Total Store Order)">;
 def HasStdExtZtso : Predicate<"Subtarget->hasStdExtZTso()">,
                               AssemblerPredicate<(all_of FeatureStdExtZtso),
                               "'Ztso' (Memory Model - Total Store Order)">;
+def NotHasStdExtZtso : Predicate<"!Subtarget->hasStdExtZtso()">;
 
 def FeatureStdExtZawrs : SubtargetFeature<"zawrs", "HasStdExtZawrs", "true",
                                           "'Zawrs' (Wait on Reservation Set)">;
 def HasStdExtZawrs : Predicate<"Subtarget->hasStdExtZawrs()">,
                                AssemblerPredicate<(all_of FeatureStdExtZawrs),
                                "'Zawrs' (Wait on Reservation Set)">;
 
 def FeatureStdExtZvbb

llvm/lib/Target/RISCV/RISCVISelLowering.cpp

Instruction *RISCVTargetLowering::emitLeadingFence(IRBuilderBase &Builder,
   if (isa<StoreInst>(Inst) && isReleaseOrStronger(Ord))
     return Builder.CreateFence(AtomicOrdering::Release);
   return nullptr;
 }
 
 Instruction *RISCVTargetLowering::emitTrailingFence(IRBuilderBase &Builder,
                                                     Instruction *Inst,
                                                     AtomicOrdering Ord) const {
-  if (Subtarget.hasStdExtZtso())
+  if (Subtarget.hasStdExtZtso()) {
+    if (isa<StoreInst>(Inst) && Ord == AtomicOrdering::SequentiallyConsistent)
+      return Builder.CreateFence(Ord);
     return nullptr;
+  }
 
   if (isa<LoadInst>(Inst) && isAcquireOrStronger(Ord))
     return Builder.CreateFence(AtomicOrdering::Acquire);
   if (Subtarget.enableSeqCstTrailingFence() && isa<StoreInst>(Inst) &&
       Ord == AtomicOrdering::SequentiallyConsistent)
     return Builder.CreateFence(AtomicOrdering::SequentiallyConsistent);
   return nullptr;
 }

llvm/lib/Target/RISCV/RISCVInstrInfo.h

public:
   bool isAssociativeAndCommutative(const MachineInstr &Inst,
                                    bool Invert) const override;
 
   std::optional<unsigned> getInverseOpcode(unsigned Opcode) const override;
 
   ArrayRef<std::pair<MachineMemOperand::Flags, const char *>>
   getSerializableMachineMemOperandTargetFlags() const override;
 
+  bool subtargetHasStdExtZtso() const;
+
 protected:
   const RISCVSubtarget &STI;
 
 private:
   unsigned getInstBundleLength(const MachineInstr &MI) const;
 };
 
 namespace RISCV {

llvm/lib/Target/RISCV/RISCVInstrInfo.cpp

 ArrayRef<std::pair<MachineMemOperand::Flags, const char *>>
 RISCVInstrInfo::getSerializableMachineMemOperandTargetFlags() const {
   static const std::pair<MachineMemOperand::Flags, const char *> TargetFlags[] =
       {{MONontemporalBit0, "riscv-nontemporal-domain-bit-0"},
        {MONontemporalBit1, "riscv-nontemporal-domain-bit-1"}};
   return ArrayRef(TargetFlags);
 }
 
+bool RISCVInstrInfo::subtargetHasStdExtZtso() const {
+  return STI.hasStdExtZtso();
+}
+
 // Returns true if this is the sext.w pattern, addiw rd, rs1, 0.
 bool RISCV::isSEXT_W(const MachineInstr &MI) {
   return MI.getOpcode() == RISCV::ADDIW && MI.getOperand(1).isReg() &&
          MI.getOperand(2).isImm() && MI.getOperand(2).getImm() == 0;
 }
 
 // Returns true if this is the zext.w pattern, adduw rd, rs1, x0.
 bool RISCV::isZEXT_W(const MachineInstr &MI) {

llvm/lib/Target/RISCV/RISCVInstrInfoA.td

let Predicates = [HasAtomicLdSt] in {
   def : AtomicStPat<atomic_store_32, SW, GPR>;
 }
 
 let Predicates = [HasAtomicLdSt, IsRV64] in {
   def : LdPat<atomic_load_64, LD, i64>;
   def : AtomicStPat<atomic_store_64, SD, GPR, i64>;
 }
 
-let Predicates = [HasStdExtA] in {
-
 /// AMOs
 
 multiclass AMOPat<string AtomicOp, string BaseInst, ValueType vt = XLenVT> {
+let Predicates = [HasStdExtA, NotHasStdExtZtso] in {
   def : PatGprGpr<!cast<PatFrag>(AtomicOp#"_monotonic"),
                   !cast<RVInst>(BaseInst), vt>;
   def : PatGprGpr<!cast<PatFrag>(AtomicOp#"_acquire"),
                   !cast<RVInst>(BaseInst#"_AQ"), vt>;
   def : PatGprGpr<!cast<PatFrag>(AtomicOp#"_release"),
                   !cast<RVInst>(BaseInst#"_RL"), vt>;
   def : PatGprGpr<!cast<PatFrag>(AtomicOp#"_acq_rel"),
                   !cast<RVInst>(BaseInst#"_AQ_RL"), vt>;
   def : PatGprGpr<!cast<PatFrag>(AtomicOp#"_seq_cst"),
                   !cast<RVInst>(BaseInst#"_AQ_RL"), vt>;
 }
+let Predicates = [HasStdExtA, HasStdExtZtso] in {
+  def : PatGprGpr<!cast<PatFrag>(AtomicOp#"_monotonic"),
+                  !cast<RVInst>(BaseInst), vt>;
+  def : PatGprGpr<!cast<PatFrag>(AtomicOp#"_acquire"),
+                  !cast<RVInst>(BaseInst), vt>;
+  def : PatGprGpr<!cast<PatFrag>(AtomicOp#"_release"),
+                  !cast<RVInst>(BaseInst), vt>;
+  def : PatGprGpr<!cast<PatFrag>(AtomicOp#"_acq_rel"),
+                  !cast<RVInst>(BaseInst), vt>;
+  def : PatGprGpr<!cast<PatFrag>(AtomicOp#"_seq_cst"),
+                  !cast<RVInst>(BaseInst), vt>;
+}
+}
+
+let Predicates = [HasStdExtA] in {
 
 defm : AMOPat<"atomic_swap_32", "AMOSWAP_W">;
 defm : AMOPat<"atomic_load_add_32", "AMOADD_W">;
 defm : AMOPat<"atomic_load_and_32", "AMOAND_W">;
 defm : AMOPat<"atomic_load_or_32", "AMOOR_W">;
 defm : AMOPat<"atomic_load_xor_32", "AMOXOR_W">;
 defm : AMOPat<"atomic_load_max_32", "AMOMAX_W">;
 defm : AMOPat<"atomic_load_min_32", "AMOMIN_W">;

llvm/test/CodeGen/RISCV/atomic-cmpxchg.ll

 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc -mtriple=riscv32 -verify-machineinstrs < %s \
 ; RUN:   | FileCheck -check-prefix=RV32I %s
 ; RUN: llc -mtriple=riscv32 -mattr=+a -verify-machineinstrs < %s \
-; RUN:   | FileCheck -check-prefix=RV32IA %s
+; RUN:   | FileCheck -check-prefixes=RV32IA,RV32IA-WMO %s
 ; RUN: llc -mtriple=riscv32 -mattr=+a,+experimental-ztso -verify-machineinstrs < %s \
-; RUN:   | FileCheck -check-prefix=RV32IA %s
+; RUN:   | FileCheck -check-prefixes=RV32IA,RV32IA-TSO %s
 ; RUN: llc -mtriple=riscv64 -verify-machineinstrs < %s \
 ; RUN:   | FileCheck -check-prefix=RV64I %s
 ; RUN: llc -mtriple=riscv64 -mattr=+a -verify-machineinstrs < %s \
-; RUN:   | FileCheck -check-prefix=RV64IA %s
+; RUN:   | FileCheck -check-prefixes=RV64IA,RV64IA-WMO %s
 ; RUN: llc -mtriple=riscv64 -mattr=+a,+experimental-ztso -verify-machineinstrs < %s \
-; RUN:   | FileCheck -check-prefix=RV64IA %s
+; RUN:   | FileCheck -check-prefixes=RV64IA,RV64IA-TSO %s
 
 define void @cmpxchg_i8_monotonic_monotonic(ptr %ptr, i8 %cmp, i8 %val) nounwind {
 ; RV32I-LABEL: cmpxchg_i8_monotonic_monotonic:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    sb a1, 11(sp)
 ; RV32I-NEXT:    addi a1, sp, 11

 ; RV32I-NEXT:    addi a1, sp, 11
 ; RV32I-NEXT:    li a3, 2
 ; RV32I-NEXT:    li a4, 0
 ; RV32I-NEXT:    call __atomic_compare_exchange_1@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: cmpxchg_i8_acquire_monotonic:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a3, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    li a4, 255
-; RV32IA-NEXT:    sll a4, a4, a0
-; RV32IA-NEXT:    andi a1, a1, 255
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:    andi a2, a2, 255
-; RV32IA-NEXT:    sll a0, a2, a0
-; RV32IA-NEXT:  .LBB1_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a2, (a3)
-; RV32IA-NEXT:    and a5, a2, a4
-; RV32IA-NEXT:    bne a5, a1, .LBB1_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB1_1 Depth=1
-; RV32IA-NEXT:    xor a5, a2, a0
-; RV32IA-NEXT:    and a5, a5, a4
-; RV32IA-NEXT:    xor a5, a2, a5
-; RV32IA-NEXT:    sc.w a5, a5, (a3)
-; RV32IA-NEXT:    bnez a5, .LBB1_1
-; RV32IA-NEXT:  .LBB1_3:
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: cmpxchg_i8_acquire_monotonic:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a3, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    li a4, 255
+; RV32IA-WMO-NEXT:    sll a4, a4, a0
+; RV32IA-WMO-NEXT:    andi a1, a1, 255
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:    andi a2, a2, 255
+; RV32IA-WMO-NEXT:    sll a0, a2, a0
+; RV32IA-WMO-NEXT:  .LBB1_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a2, (a3)
+; RV32IA-WMO-NEXT:    and a5, a2, a4
+; RV32IA-WMO-NEXT:    bne a5, a1, .LBB1_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB1_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a5, a2, a0
+; RV32IA-WMO-NEXT:    and a5, a5, a4
+; RV32IA-WMO-NEXT:    xor a5, a2, a5
+; RV32IA-WMO-NEXT:    sc.w a5, a5, (a3)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB1_1
+; RV32IA-WMO-NEXT:  .LBB1_3:
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: cmpxchg_i8_acquire_monotonic:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a3, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    li a4, 255
+; RV32IA-TSO-NEXT:    sll a4, a4, a0
+; RV32IA-TSO-NEXT:    andi a1, a1, 255
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:    andi a2, a2, 255
+; RV32IA-TSO-NEXT:    sll a0, a2, a0
+; RV32IA-TSO-NEXT:  .LBB1_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a2, (a3)
+; RV32IA-TSO-NEXT:    and a5, a2, a4
+; RV32IA-TSO-NEXT:    bne a5, a1, .LBB1_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB1_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a5, a2, a0
+; RV32IA-TSO-NEXT:    and a5, a5, a4
+; RV32IA-TSO-NEXT:    xor a5, a2, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a3)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB1_1
+; RV32IA-TSO-NEXT:  .LBB1_3:
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: cmpxchg_i8_acquire_monotonic:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sb a1, 7(sp)
 ; RV64I-NEXT:    addi a1, sp, 7
 ; RV64I-NEXT:    li a3, 2
 ; RV64I-NEXT:    li a4, 0
 ; RV64I-NEXT:    call __atomic_compare_exchange_1@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: cmpxchg_i8_acquire_monotonic:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a3, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    li a4, 255
-; RV64IA-NEXT:    sllw a4, a4, a0
-; RV64IA-NEXT:    andi a1, a1, 255
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:    andi a2, a2, 255
-; RV64IA-NEXT:    sllw a0, a2, a0
-; RV64IA-NEXT:  .LBB1_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a2, (a3)
-; RV64IA-NEXT:    and a5, a2, a4
-; RV64IA-NEXT:    bne a5, a1, .LBB1_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB1_1 Depth=1
-; RV64IA-NEXT:    xor a5, a2, a0
-; RV64IA-NEXT:    and a5, a5, a4
-; RV64IA-NEXT:    xor a5, a2, a5
-; RV64IA-NEXT:    sc.w a5, a5, (a3)
-; RV64IA-NEXT:    bnez a5, .LBB1_1
-; RV64IA-NEXT:  .LBB1_3:
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: cmpxchg_i8_acquire_monotonic:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a3, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    li a4, 255
+; RV64IA-WMO-NEXT:    sllw a4, a4, a0
+; RV64IA-WMO-NEXT:    andi a1, a1, 255
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:    andi a2, a2, 255
+; RV64IA-WMO-NEXT:    sllw a0, a2, a0
+; RV64IA-WMO-NEXT:  .LBB1_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a2, (a3)
+; RV64IA-WMO-NEXT:    and a5, a2, a4
+; RV64IA-WMO-NEXT:    bne a5, a1, .LBB1_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB1_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a5, a2, a0
+; RV64IA-WMO-NEXT:    and a5, a5, a4
+; RV64IA-WMO-NEXT:    xor a5, a2, a5
+; RV64IA-WMO-NEXT:    sc.w a5, a5, (a3)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB1_1
+; RV64IA-WMO-NEXT:  .LBB1_3:
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: cmpxchg_i8_acquire_monotonic:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a3, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    li a4, 255
+; RV64IA-TSO-NEXT:    sllw a4, a4, a0
+; RV64IA-TSO-NEXT:    andi a1, a1, 255
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:    andi a2, a2, 255
+; RV64IA-TSO-NEXT:    sllw a0, a2, a0
+; RV64IA-TSO-NEXT:  .LBB1_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a2, (a3)
+; RV64IA-TSO-NEXT:    and a5, a2, a4
+; RV64IA-TSO-NEXT:    bne a5, a1, .LBB1_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB1_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a5, a2, a0
+; RV64IA-TSO-NEXT:    and a5, a5, a4
+; RV64IA-TSO-NEXT:    xor a5, a2, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a3)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB1_1
+; RV64IA-TSO-NEXT:  .LBB1_3:
+; RV64IA-TSO-NEXT:    ret
   %res = cmpxchg ptr %ptr, i8 %cmp, i8 %val acquire monotonic
   ret void
 }
 
 define void @cmpxchg_i8_acquire_acquire(ptr %ptr, i8 %cmp, i8 %val) nounwind {
 ; RV32I-LABEL: cmpxchg_i8_acquire_acquire:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    sb a1, 11(sp)
 ; RV32I-NEXT:    addi a1, sp, 11
 ; RV32I-NEXT:    li a3, 2
 ; RV32I-NEXT:    li a4, 2
 ; RV32I-NEXT:    call __atomic_compare_exchange_1@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: cmpxchg_i8_acquire_acquire:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a3, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    li a4, 255
-; RV32IA-NEXT:    sll a4, a4, a0
-; RV32IA-NEXT:    andi a1, a1, 255
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:    andi a2, a2, 255
-; RV32IA-NEXT:    sll a0, a2, a0
-; RV32IA-NEXT:  .LBB2_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a2, (a3)
-; RV32IA-NEXT:    and a5, a2, a4
-; RV32IA-NEXT:    bne a5, a1, .LBB2_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB2_1 Depth=1
-; RV32IA-NEXT:    xor a5, a2, a0
-; RV32IA-NEXT:    and a5, a5, a4
-; RV32IA-NEXT:    xor a5, a2, a5
-; RV32IA-NEXT:    sc.w a5, a5, (a3)
-; RV32IA-NEXT:    bnez a5, .LBB2_1
-; RV32IA-NEXT:  .LBB2_3:
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: cmpxchg_i8_acquire_acquire:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a3, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    li a4, 255
+; RV32IA-WMO-NEXT:    sll a4, a4, a0
+; RV32IA-WMO-NEXT:    andi a1, a1, 255
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:    andi a2, a2, 255
+; RV32IA-WMO-NEXT:    sll a0, a2, a0
+; RV32IA-WMO-NEXT:  .LBB2_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a2, (a3)
+; RV32IA-WMO-NEXT:    and a5, a2, a4
+; RV32IA-WMO-NEXT:    bne a5, a1, .LBB2_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB2_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a5, a2, a0
+; RV32IA-WMO-NEXT:    and a5, a5, a4
+; RV32IA-WMO-NEXT:    xor a5, a2, a5
+; RV32IA-WMO-NEXT:    sc.w a5, a5, (a3)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB2_1
+; RV32IA-WMO-NEXT:  .LBB2_3:
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: cmpxchg_i8_acquire_acquire:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a3, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    li a4, 255
+; RV32IA-TSO-NEXT:    sll a4, a4, a0
+; RV32IA-TSO-NEXT:    andi a1, a1, 255
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:    andi a2, a2, 255
+; RV32IA-TSO-NEXT:    sll a0, a2, a0
+; RV32IA-TSO-NEXT:  .LBB2_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a2, (a3)
+; RV32IA-TSO-NEXT:    and a5, a2, a4
+; RV32IA-TSO-NEXT:    bne a5, a1, .LBB2_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB2_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a5, a2, a0
+; RV32IA-TSO-NEXT:    and a5, a5, a4
+; RV32IA-TSO-NEXT:    xor a5, a2, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a3)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB2_1
+; RV32IA-TSO-NEXT:  .LBB2_3:
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: cmpxchg_i8_acquire_acquire:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sb a1, 7(sp)
 ; RV64I-NEXT:    addi a1, sp, 7
 ; RV64I-NEXT:    li a3, 2
 ; RV64I-NEXT:    li a4, 2
 ; RV64I-NEXT:    call __atomic_compare_exchange_1@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: cmpxchg_i8_acquire_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a3, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    li a4, 255
-; RV64IA-NEXT:    sllw a4, a4, a0
-; RV64IA-NEXT:    andi a1, a1, 255
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:    andi a2, a2, 255
-; RV64IA-NEXT:    sllw a0, a2, a0
-; RV64IA-NEXT:  .LBB2_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a2, (a3)
-; RV64IA-NEXT:    and a5, a2, a4
-; RV64IA-NEXT:    bne a5, a1, .LBB2_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB2_1 Depth=1
-; RV64IA-NEXT:    xor a5, a2, a0
-; RV64IA-NEXT:    and a5, a5, a4
-; RV64IA-NEXT:    xor a5, a2, a5
-; RV64IA-NEXT:    sc.w a5, a5, (a3)
-; RV64IA-NEXT:    bnez a5, .LBB2_1
-; RV64IA-NEXT:  .LBB2_3:
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: cmpxchg_i8_acquire_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a3, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    li a4, 255
+; RV64IA-WMO-NEXT:    sllw a4, a4, a0
+; RV64IA-WMO-NEXT:    andi a1, a1, 255
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:    andi a2, a2, 255
+; RV64IA-WMO-NEXT:    sllw a0, a2, a0
+; RV64IA-WMO-NEXT:  .LBB2_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a2, (a3)
+; RV64IA-WMO-NEXT:    and a5, a2, a4
+; RV64IA-WMO-NEXT:    bne a5, a1, .LBB2_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB2_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a5, a2, a0
+; RV64IA-WMO-NEXT:    and a5, a5, a4
+; RV64IA-WMO-NEXT:    xor a5, a2, a5
+; RV64IA-WMO-NEXT:    sc.w a5, a5, (a3)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB2_1
+; RV64IA-WMO-NEXT:  .LBB2_3:
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: cmpxchg_i8_acquire_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a3, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    li a4, 255
+; RV64IA-TSO-NEXT:    sllw a4, a4, a0
+; RV64IA-TSO-NEXT:    andi a1, a1, 255
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:    andi a2, a2, 255
+; RV64IA-TSO-NEXT:    sllw a0, a2, a0
+; RV64IA-TSO-NEXT:  .LBB2_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a2, (a3)
+; RV64IA-TSO-NEXT:    and a5, a2, a4
+; RV64IA-TSO-NEXT:    bne a5, a1, .LBB2_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB2_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a5, a2, a0
+; RV64IA-TSO-NEXT:    and a5, a5, a4
+; RV64IA-TSO-NEXT:    xor a5, a2, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a3)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB2_1
+; RV64IA-TSO-NEXT:  .LBB2_3:
+; RV64IA-TSO-NEXT:    ret
   %res = cmpxchg ptr %ptr, i8 %cmp, i8 %val acquire acquire
   ret void
 }
 
 define void @cmpxchg_i8_release_monotonic(ptr %ptr, i8 %cmp, i8 %val) nounwind {
 ; RV32I-LABEL: cmpxchg_i8_release_monotonic:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    sb a1, 11(sp)
 ; RV32I-NEXT:    addi a1, sp, 11
 ; RV32I-NEXT:    li a3, 3
 ; RV32I-NEXT:    li a4, 0
 ; RV32I-NEXT:    call __atomic_compare_exchange_1@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: cmpxchg_i8_release_monotonic:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a3, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    li a4, 255
-; RV32IA-NEXT:    sll a4, a4, a0
-; RV32IA-NEXT:    andi a1, a1, 255
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:    andi a2, a2, 255
-; RV32IA-NEXT:    sll a0, a2, a0
-; RV32IA-NEXT:  .LBB3_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w a2, (a3)
-; RV32IA-NEXT:    and a5, a2, a4
-; RV32IA-NEXT:    bne a5, a1, .LBB3_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB3_1 Depth=1
-; RV32IA-NEXT:    xor a5, a2, a0
-; RV32IA-NEXT:    and a5, a5, a4
-; RV32IA-NEXT:    xor a5, a2, a5
-; RV32IA-NEXT:    sc.w.rl a5, a5, (a3)
-; RV32IA-NEXT:    bnez a5, .LBB3_1
-; RV32IA-NEXT:  .LBB3_3:
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: cmpxchg_i8_release_monotonic:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a3, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    li a4, 255
+; RV32IA-WMO-NEXT:    sll a4, a4, a0
+; RV32IA-WMO-NEXT:    andi a1, a1, 255
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:    andi a2, a2, 255
+; RV32IA-WMO-NEXT:    sll a0, a2, a0
+; RV32IA-WMO-NEXT:  .LBB3_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w a2, (a3)
+; RV32IA-WMO-NEXT:    and a5, a2, a4
+; RV32IA-WMO-NEXT:    bne a5, a1, .LBB3_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB3_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a5, a2, a0
+; RV32IA-WMO-NEXT:    and a5, a5, a4
+; RV32IA-WMO-NEXT:    xor a5, a2, a5
+; RV32IA-WMO-NEXT:    sc.w.rl a5, a5, (a3)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB3_1
+; RV32IA-WMO-NEXT:  .LBB3_3:
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: cmpxchg_i8_release_monotonic:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a3, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    li a4, 255
+; RV32IA-TSO-NEXT:    sll a4, a4, a0
+; RV32IA-TSO-NEXT:    andi a1, a1, 255
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:    andi a2, a2, 255
+; RV32IA-TSO-NEXT:    sll a0, a2, a0
+; RV32IA-TSO-NEXT:  .LBB3_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a2, (a3)
+; RV32IA-TSO-NEXT:    and a5, a2, a4
+; RV32IA-TSO-NEXT:    bne a5, a1, .LBB3_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB3_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a5, a2, a0
+; RV32IA-TSO-NEXT:    and a5, a5, a4
+; RV32IA-TSO-NEXT:    xor a5, a2, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a3)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB3_1
+; RV32IA-TSO-NEXT:  .LBB3_3:
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: cmpxchg_i8_release_monotonic:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sb a1, 7(sp)
 ; RV64I-NEXT:    addi a1, sp, 7
 ; RV64I-NEXT:    li a3, 3
 ; RV64I-NEXT:    li a4, 0
 ; RV64I-NEXT:    call __atomic_compare_exchange_1@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: cmpxchg_i8_release_monotonic:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a3, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    li a4, 255
-; RV64IA-NEXT:    sllw a4, a4, a0
-; RV64IA-NEXT:    andi a1, a1, 255
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:    andi a2, a2, 255
-; RV64IA-NEXT:    sllw a0, a2, a0
-; RV64IA-NEXT:  .LBB3_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w a2, (a3)
-; RV64IA-NEXT:    and a5, a2, a4
-; RV64IA-NEXT:    bne a5, a1, .LBB3_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB3_1 Depth=1
-; RV64IA-NEXT:    xor a5, a2, a0
-; RV64IA-NEXT:    and a5, a5, a4
-; RV64IA-NEXT:    xor a5, a2, a5
-; RV64IA-NEXT:    sc.w.rl a5, a5, (a3)
-; RV64IA-NEXT:    bnez a5, .LBB3_1
-; RV64IA-NEXT:  .LBB3_3:
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: cmpxchg_i8_release_monotonic:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a3, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    li a4, 255
+; RV64IA-WMO-NEXT:    sllw a4, a4, a0
+; RV64IA-WMO-NEXT:    andi a1, a1, 255
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:    andi a2, a2, 255
+; RV64IA-WMO-NEXT:    sllw a0, a2, a0
+; RV64IA-WMO-NEXT:  .LBB3_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w a2, (a3)
+; RV64IA-WMO-NEXT:    and a5, a2, a4
+; RV64IA-WMO-NEXT:    bne a5, a1, .LBB3_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB3_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a5, a2, a0
+; RV64IA-WMO-NEXT:    and a5, a5, a4
+; RV64IA-WMO-NEXT:    xor a5, a2, a5
+; RV64IA-WMO-NEXT:    sc.w.rl a5, a5, (a3)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB3_1
+; RV64IA-WMO-NEXT:  .LBB3_3:
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: cmpxchg_i8_release_monotonic:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a3, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    li a4, 255
+; RV64IA-TSO-NEXT:    sllw a4, a4, a0
+; RV64IA-TSO-NEXT:    andi a1, a1, 255
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:    andi a2, a2, 255
+; RV64IA-TSO-NEXT:    sllw a0, a2, a0
+; RV64IA-TSO-NEXT:  .LBB3_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a2, (a3)
+; RV64IA-TSO-NEXT:    and a5, a2, a4
+; RV64IA-TSO-NEXT:    bne a5, a1, .LBB3_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB3_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a5, a2, a0
+; RV64IA-TSO-NEXT:    and a5, a5, a4
+; RV64IA-TSO-NEXT:    xor a5, a2, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a3)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB3_1
+; RV64IA-TSO-NEXT:  .LBB3_3:
+; RV64IA-TSO-NEXT:    ret
   %res = cmpxchg ptr %ptr, i8 %cmp, i8 %val release monotonic
   ret void
 }
 
 define void @cmpxchg_i8_release_acquire(ptr %ptr, i8 %cmp, i8 %val) nounwind {
 ; RV32I-LABEL: cmpxchg_i8_release_acquire:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    sb a1, 11(sp)
 ; RV32I-NEXT:    addi a1, sp, 11
 ; RV32I-NEXT:    li a3, 3
 ; RV32I-NEXT:    li a4, 2
 ; RV32I-NEXT:    call __atomic_compare_exchange_1@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: cmpxchg_i8_release_acquire:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a3, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    li a4, 255
-; RV32IA-NEXT:    sll a4, a4, a0
-; RV32IA-NEXT:    andi a1, a1, 255
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:    andi a2, a2, 255
-; RV32IA-NEXT:    sll a0, a2, a0
-; RV32IA-NEXT:  .LBB4_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a2, (a3)
-; RV32IA-NEXT:    and a5, a2, a4
-; RV32IA-NEXT:    bne a5, a1, .LBB4_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB4_1 Depth=1
-; RV32IA-NEXT:    xor a5, a2, a0
-; RV32IA-NEXT:    and a5, a5, a4
-; RV32IA-NEXT:    xor a5, a2, a5
-; RV32IA-NEXT:    sc.w.rl a5, a5, (a3)
-; RV32IA-NEXT:    bnez a5, .LBB4_1
-; RV32IA-NEXT:  .LBB4_3:
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: cmpxchg_i8_release_acquire:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a3, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    li a4, 255
+; RV32IA-WMO-NEXT:    sll a4, a4, a0
+; RV32IA-WMO-NEXT:    andi a1, a1, 255
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:    andi a2, a2, 255
+; RV32IA-WMO-NEXT:    sll a0, a2, a0
+; RV32IA-WMO-NEXT:  .LBB4_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a2, (a3)
+; RV32IA-WMO-NEXT:    and a5, a2, a4
+; RV32IA-WMO-NEXT:    bne a5, a1, .LBB4_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB4_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a5, a2, a0
+; RV32IA-WMO-NEXT:    and a5, a5, a4
+; RV32IA-WMO-NEXT:    xor a5, a2, a5
+; RV32IA-WMO-NEXT:    sc.w.rl a5, a5, (a3)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB4_1
+; RV32IA-WMO-NEXT:  .LBB4_3:
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: cmpxchg_i8_release_acquire:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a3, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    li a4, 255
+; RV32IA-TSO-NEXT:    sll a4, a4, a0
+; RV32IA-TSO-NEXT:    andi a1, a1, 255
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:    andi a2, a2, 255
+; RV32IA-TSO-NEXT:    sll a0, a2, a0
+; RV32IA-TSO-NEXT:  .LBB4_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a2, (a3)
+; RV32IA-TSO-NEXT:    and a5, a2, a4
+; RV32IA-TSO-NEXT:    bne a5, a1, .LBB4_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB4_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a5, a2, a0
+; RV32IA-TSO-NEXT:    and a5, a5, a4
+; RV32IA-TSO-NEXT:    xor a5, a2, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a3)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB4_1
+; RV32IA-TSO-NEXT:  .LBB4_3:
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: cmpxchg_i8_release_acquire:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sb a1, 7(sp)
 ; RV64I-NEXT:    addi a1, sp, 7
 ; RV64I-NEXT:    li a3, 3
 ; RV64I-NEXT:    li a4, 2
 ; RV64I-NEXT:    call __atomic_compare_exchange_1@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: cmpxchg_i8_release_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a3, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    li a4, 255
-; RV64IA-NEXT:    sllw a4, a4, a0
-; RV64IA-NEXT:    andi a1, a1, 255
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:    andi a2, a2, 255
-; RV64IA-NEXT:    sllw a0, a2, a0
-; RV64IA-NEXT:  .LBB4_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a2, (a3)
-; RV64IA-NEXT:    and a5, a2, a4
-; RV64IA-NEXT:    bne a5, a1, .LBB4_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB4_1 Depth=1
-; RV64IA-NEXT:    xor a5, a2, a0
-; RV64IA-NEXT:    and a5, a5, a4
-; RV64IA-NEXT:    xor a5, a2, a5
-; RV64IA-NEXT:    sc.w.rl a5, a5, (a3)
-; RV64IA-NEXT:    bnez a5, .LBB4_1
-; RV64IA-NEXT:  .LBB4_3:
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: cmpxchg_i8_release_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a3, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    li a4, 255
+; RV64IA-WMO-NEXT:    sllw a4, a4, a0
+; RV64IA-WMO-NEXT:    andi a1, a1, 255
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:    andi a2, a2, 255
+; RV64IA-WMO-NEXT:    sllw a0, a2, a0
+; RV64IA-WMO-NEXT:  .LBB4_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a2, (a3)
+; RV64IA-WMO-NEXT:    and a5, a2, a4
+; RV64IA-WMO-NEXT:    bne a5, a1, .LBB4_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB4_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a5, a2, a0
+; RV64IA-WMO-NEXT:    and a5, a5, a4
+; RV64IA-WMO-NEXT:    xor a5, a2, a5
+; RV64IA-WMO-NEXT:    sc.w.rl a5, a5, (a3)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB4_1
+; RV64IA-WMO-NEXT:  .LBB4_3:
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: cmpxchg_i8_release_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a3, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    li a4, 255
+; RV64IA-TSO-NEXT:    sllw a4, a4, a0
+; RV64IA-TSO-NEXT:    andi a1, a1, 255
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:    andi a2, a2, 255
+; RV64IA-TSO-NEXT:    sllw a0, a2, a0
+; RV64IA-TSO-NEXT:  .LBB4_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a2, (a3)
+; RV64IA-TSO-NEXT:    and a5, a2, a4
+; RV64IA-TSO-NEXT:    bne a5, a1, .LBB4_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB4_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a5, a2, a0
+; RV64IA-TSO-NEXT:    and a5, a5, a4
+; RV64IA-TSO-NEXT:    xor a5, a2, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a3)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB4_1
+; RV64IA-TSO-NEXT:  .LBB4_3:
+; RV64IA-TSO-NEXT:    ret
   %res = cmpxchg ptr %ptr, i8 %cmp, i8 %val release acquire
   ret void
 }
 
 define void @cmpxchg_i8_acq_rel_monotonic(ptr %ptr, i8 %cmp, i8 %val) nounwind {
 ; RV32I-LABEL: cmpxchg_i8_acq_rel_monotonic:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    sb a1, 11(sp)
 ; RV32I-NEXT:    addi a1, sp, 11
 ; RV32I-NEXT:    li a3, 4
 ; RV32I-NEXT:    li a4, 0
 ; RV32I-NEXT:    call __atomic_compare_exchange_1@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: cmpxchg_i8_acq_rel_monotonic:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a3, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    li a4, 255
-; RV32IA-NEXT:    sll a4, a4, a0
-; RV32IA-NEXT:    andi a1, a1, 255
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:    andi a2, a2, 255
-; RV32IA-NEXT:    sll a0, a2, a0
-; RV32IA-NEXT:  .LBB5_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a2, (a3)
-; RV32IA-NEXT:    and a5, a2, a4
-; RV32IA-NEXT:    bne a5, a1, .LBB5_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB5_1 Depth=1
-; RV32IA-NEXT:    xor a5, a2, a0
-; RV32IA-NEXT:    and a5, a5, a4
-; RV32IA-NEXT:    xor a5, a2, a5
-; RV32IA-NEXT:    sc.w.rl a5, a5, (a3)
-; RV32IA-NEXT:    bnez a5, .LBB5_1
-; RV32IA-NEXT:  .LBB5_3:
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: cmpxchg_i8_acq_rel_monotonic:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a3, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    li a4, 255
+; RV32IA-WMO-NEXT:    sll a4, a4, a0
+; RV32IA-WMO-NEXT:    andi a1, a1, 255
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:    andi a2, a2, 255
+; RV32IA-WMO-NEXT:    sll a0, a2, a0
+; RV32IA-WMO-NEXT:  .LBB5_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a2, (a3)
+; RV32IA-WMO-NEXT:    and a5, a2, a4
+; RV32IA-WMO-NEXT:    bne a5, a1, .LBB5_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB5_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a5, a2, a0
+; RV32IA-WMO-NEXT:    and a5, a5, a4
+; RV32IA-WMO-NEXT:    xor a5, a2, a5
+; RV32IA-WMO-NEXT:    sc.w.rl a5, a5, (a3)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB5_1
+; RV32IA-WMO-NEXT:  .LBB5_3:
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: cmpxchg_i8_acq_rel_monotonic:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a3, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    li a4, 255
+; RV32IA-TSO-NEXT:    sll a4, a4, a0
+; RV32IA-TSO-NEXT:    andi a1, a1, 255
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:    andi a2, a2, 255
+; RV32IA-TSO-NEXT:    sll a0, a2, a0
+; RV32IA-TSO-NEXT:  .LBB5_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a2, (a3)
+; RV32IA-TSO-NEXT:    and a5, a2, a4
+; RV32IA-TSO-NEXT:    bne a5, a1, .LBB5_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB5_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a5, a2, a0
+; RV32IA-TSO-NEXT:    and a5, a5, a4
+; RV32IA-TSO-NEXT:    xor a5, a2, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a3)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB5_1
+; RV32IA-TSO-NEXT:  .LBB5_3:
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: cmpxchg_i8_acq_rel_monotonic:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sb a1, 7(sp)
 ; RV64I-NEXT:    addi a1, sp, 7
 ; RV64I-NEXT:    li a3, 4
 ; RV64I-NEXT:    li a4, 0
 ; RV64I-NEXT:    call __atomic_compare_exchange_1@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: cmpxchg_i8_acq_rel_monotonic:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a3, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    li a4, 255
-; RV64IA-NEXT:    sllw a4, a4, a0
-; RV64IA-NEXT:    andi a1, a1, 255
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:    andi a2, a2, 255
-; RV64IA-NEXT:    sllw a0, a2, a0
-; RV64IA-NEXT:  .LBB5_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a2, (a3)
-; RV64IA-NEXT:    and a5, a2, a4
-; RV64IA-NEXT:    bne a5, a1, .LBB5_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB5_1 Depth=1
-; RV64IA-NEXT:    xor a5, a2, a0
-; RV64IA-NEXT:    and a5, a5, a4
-; RV64IA-NEXT:    xor a5, a2, a5
-; RV64IA-NEXT:    sc.w.rl a5, a5, (a3)
-; RV64IA-NEXT:    bnez a5, .LBB5_1
-; RV64IA-NEXT:  .LBB5_3:
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: cmpxchg_i8_acq_rel_monotonic:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a3, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    li a4, 255
+; RV64IA-WMO-NEXT:    sllw a4, a4, a0
+; RV64IA-WMO-NEXT:    andi a1, a1, 255
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:    andi a2, a2, 255
+; RV64IA-WMO-NEXT:    sllw a0, a2, a0
+; RV64IA-WMO-NEXT:  .LBB5_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a2, (a3)
+; RV64IA-WMO-NEXT:    and a5, a2, a4
+; RV64IA-WMO-NEXT:    bne a5, a1, .LBB5_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB5_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a5, a2, a0
+; RV64IA-WMO-NEXT:    and a5, a5, a4
+; RV64IA-WMO-NEXT:    xor a5, a2, a5
+; RV64IA-WMO-NEXT:    sc.w.rl a5, a5, (a3)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB5_1
+; RV64IA-WMO-NEXT:  .LBB5_3:
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: cmpxchg_i8_acq_rel_monotonic:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a3, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    li a4, 255
+; RV64IA-TSO-NEXT:    sllw a4, a4, a0
+; RV64IA-TSO-NEXT:    andi a1, a1, 255
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:    andi a2, a2, 255
+; RV64IA-TSO-NEXT:    sllw a0, a2, a0
+; RV64IA-TSO-NEXT:  .LBB5_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a2, (a3)
+; RV64IA-TSO-NEXT:    and a5, a2, a4
+; RV64IA-TSO-NEXT:    bne a5, a1, .LBB5_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB5_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a5, a2, a0
+; RV64IA-TSO-NEXT:    and a5, a5, a4
+; RV64IA-TSO-NEXT:    xor a5, a2, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a3)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB5_1
+; RV64IA-TSO-NEXT:  .LBB5_3:
+; RV64IA-TSO-NEXT:    ret
   %res = cmpxchg ptr %ptr, i8 %cmp, i8 %val acq_rel monotonic
   ret void
 }
 
 define void @cmpxchg_i8_acq_rel_acquire(ptr %ptr, i8 %cmp, i8 %val) nounwind {
 ; RV32I-LABEL: cmpxchg_i8_acq_rel_acquire:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    sb a1, 11(sp)
 ; RV32I-NEXT:    addi a1, sp, 11
 ; RV32I-NEXT:    li a3, 4
 ; RV32I-NEXT:    li a4, 2
 ; RV32I-NEXT:    call __atomic_compare_exchange_1@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: cmpxchg_i8_acq_rel_acquire:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a3, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    li a4, 255
-; RV32IA-NEXT:    sll a4, a4, a0
-; RV32IA-NEXT:    andi a1, a1, 255
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:    andi a2, a2, 255
-; RV32IA-NEXT:    sll a0, a2, a0
-; RV32IA-NEXT:  .LBB6_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a2, (a3)
-; RV32IA-NEXT:    and a5, a2, a4
-; RV32IA-NEXT:    bne a5, a1, .LBB6_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB6_1 Depth=1
-; RV32IA-NEXT:    xor a5, a2, a0
-; RV32IA-NEXT:    and a5, a5, a4
-; RV32IA-NEXT:    xor a5, a2, a5
-; RV32IA-NEXT:    sc.w.rl a5, a5, (a3)
-; RV32IA-NEXT:    bnez a5, .LBB6_1
-; RV32IA-NEXT:  .LBB6_3:
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: cmpxchg_i8_acq_rel_acquire:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a3, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    li a4, 255
+; RV32IA-WMO-NEXT:    sll a4, a4, a0
+; RV32IA-WMO-NEXT:    andi a1, a1, 255
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:    andi a2, a2, 255
+; RV32IA-WMO-NEXT:    sll a0, a2, a0
+; RV32IA-WMO-NEXT:  .LBB6_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a2, (a3)
+; RV32IA-WMO-NEXT:    and a5, a2, a4
+; RV32IA-WMO-NEXT:    bne a5, a1, .LBB6_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB6_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a5, a2, a0
+; RV32IA-WMO-NEXT:    and a5, a5, a4
+; RV32IA-WMO-NEXT:    xor a5, a2, a5
+; RV32IA-WMO-NEXT:    sc.w.rl a5, a5, (a3)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB6_1
+; RV32IA-WMO-NEXT:  .LBB6_3:
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: cmpxchg_i8_acq_rel_acquire:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a3, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    li a4, 255
+; RV32IA-TSO-NEXT:    sll a4, a4, a0
+; RV32IA-TSO-NEXT:    andi a1, a1, 255
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:    andi a2, a2, 255
+; RV32IA-TSO-NEXT:    sll a0, a2, a0
+; RV32IA-TSO-NEXT:  .LBB6_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a2, (a3)
+; RV32IA-TSO-NEXT:    and a5, a2, a4
+; RV32IA-TSO-NEXT:    bne a5, a1, .LBB6_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB6_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a5, a2, a0
+; RV32IA-TSO-NEXT:    and a5, a5, a4
+; RV32IA-TSO-NEXT:    xor a5, a2, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a3)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB6_1
+; RV32IA-TSO-NEXT:  .LBB6_3:
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: cmpxchg_i8_acq_rel_acquire:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sb a1, 7(sp)
 ; RV64I-NEXT:    addi a1, sp, 7
 ; RV64I-NEXT:    li a3, 4
 ; RV64I-NEXT:    li a4, 2
 ; RV64I-NEXT:    call __atomic_compare_exchange_1@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: cmpxchg_i8_acq_rel_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a3, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    li a4, 255
-; RV64IA-NEXT:    sllw a4, a4, a0
-; RV64IA-NEXT:    andi a1, a1, 255
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:    andi a2, a2, 255
-; RV64IA-NEXT:    sllw a0, a2, a0
-; RV64IA-NEXT:  .LBB6_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a2, (a3)
-; RV64IA-NEXT:    and a5, a2, a4
-; RV64IA-NEXT:    bne a5, a1, .LBB6_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB6_1 Depth=1
-; RV64IA-NEXT:    xor a5, a2, a0
-; RV64IA-NEXT:    and a5, a5, a4
-; RV64IA-NEXT:    xor a5, a2, a5
-; RV64IA-NEXT:    sc.w.rl a5, a5, (a3)
-; RV64IA-NEXT:    bnez a5, .LBB6_1
-; RV64IA-NEXT:  .LBB6_3:
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: cmpxchg_i8_acq_rel_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a3, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    li a4, 255
+; RV64IA-WMO-NEXT:    sllw a4, a4, a0
+; RV64IA-WMO-NEXT:    andi a1, a1, 255
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:    andi a2, a2, 255
+; RV64IA-WMO-NEXT:    sllw a0, a2, a0
+; RV64IA-WMO-NEXT:  .LBB6_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a2, (a3)
+; RV64IA-WMO-NEXT:    and a5, a2, a4
+; RV64IA-WMO-NEXT:    bne a5, a1, .LBB6_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB6_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a5, a2, a0
+; RV64IA-WMO-NEXT:    and a5, a5, a4
+; RV64IA-WMO-NEXT:    xor a5, a2, a5
+; RV64IA-WMO-NEXT:    sc.w.rl a5, a5, (a3)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB6_1
+; RV64IA-WMO-NEXT:  .LBB6_3:
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: cmpxchg_i8_acq_rel_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a3, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    li a4, 255
+; RV64IA-TSO-NEXT:    sllw a4, a4, a0
+; RV64IA-TSO-NEXT:    andi a1, a1, 255
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:    andi a2, a2, 255
+; RV64IA-TSO-NEXT:    sllw a0, a2, a0
+; RV64IA-TSO-NEXT:  .LBB6_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a2, (a3)
+; RV64IA-TSO-NEXT:    and a5, a2, a4
+; RV64IA-TSO-NEXT:    bne a5, a1, .LBB6_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB6_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a5, a2, a0
+; RV64IA-TSO-NEXT:    and a5, a5, a4
+; RV64IA-TSO-NEXT:    xor a5, a2, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a3)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB6_1
+; RV64IA-TSO-NEXT:  .LBB6_3:
+; RV64IA-TSO-NEXT:    ret
   %res = cmpxchg ptr %ptr, i8 %cmp, i8 %val acq_rel acquire
   ret void
 }
 
 define void @cmpxchg_i8_seq_cst_monotonic(ptr %ptr, i8 %cmp, i8 %val) nounwind {
 ; RV32I-LABEL: cmpxchg_i8_seq_cst_monotonic:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16

 ; RV32I-NEXT:    addi a1, sp, 10
 ; RV32I-NEXT:    li a3, 2
 ; RV32I-NEXT:    li a4, 0
 ; RV32I-NEXT:    call __atomic_compare_exchange_2@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: cmpxchg_i16_acquire_monotonic:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a3, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    lui a4, 16
-; RV32IA-NEXT:    addi a4, a4, -1
-; RV32IA-NEXT:    sll a5, a4, a0
-; RV32IA-NEXT:    and a1, a1, a4
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:    and a2, a2, a4
-; RV32IA-NEXT:    sll a0, a2, a0
-; RV32IA-NEXT:  .LBB11_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a2, (a3)
-; RV32IA-NEXT:    and a4, a2, a5
-; RV32IA-NEXT:    bne a4, a1, .LBB11_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB11_1 Depth=1
-; RV32IA-NEXT:    xor a4, a2, a0
-; RV32IA-NEXT:    and a4, a4, a5
-; RV32IA-NEXT:    xor a4, a2, a4
-; RV32IA-NEXT:    sc.w a4, a4, (a3)
-; RV32IA-NEXT:    bnez a4, .LBB11_1
-; RV32IA-NEXT:  .LBB11_3:
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: cmpxchg_i16_acquire_monotonic:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a3, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    lui a4, 16
+; RV32IA-WMO-NEXT:    addi a4, a4, -1
+; RV32IA-WMO-NEXT:    sll a5, a4, a0
+; RV32IA-WMO-NEXT:    and a1, a1, a4
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:    and a2, a2, a4
+; RV32IA-WMO-NEXT:    sll a0, a2, a0
+; RV32IA-WMO-NEXT:  .LBB11_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a2, (a3)
+; RV32IA-WMO-NEXT:    and a4, a2, a5
+; RV32IA-WMO-NEXT:    bne a4, a1, .LBB11_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB11_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a4, a2, a0
+; RV32IA-WMO-NEXT:    and a4, a4, a5
+; RV32IA-WMO-NEXT:    xor a4, a2, a4
+; RV32IA-WMO-NEXT:    sc.w a4, a4, (a3)
+; RV32IA-WMO-NEXT:    bnez a4, .LBB11_1
+; RV32IA-WMO-NEXT:  .LBB11_3:
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: cmpxchg_i16_acquire_monotonic:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a3, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    lui a4, 16
+; RV32IA-TSO-NEXT:    addi a4, a4, -1
+; RV32IA-TSO-NEXT:    sll a5, a4, a0
+; RV32IA-TSO-NEXT:    and a1, a1, a4
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:    and a2, a2, a4
+; RV32IA-TSO-NEXT:    sll a0, a2, a0
+; RV32IA-TSO-NEXT:  .LBB11_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a2, (a3)
+; RV32IA-TSO-NEXT:    and a4, a2, a5
+; RV32IA-TSO-NEXT:    bne a4, a1, .LBB11_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB11_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a4, a2, a0
+; RV32IA-TSO-NEXT:    and a4, a4, a5
+; RV32IA-TSO-NEXT:    xor a4, a2, a4
+; RV32IA-TSO-NEXT:    sc.w a4, a4, (a3)
+; RV32IA-TSO-NEXT:    bnez a4, .LBB11_1
+; RV32IA-TSO-NEXT:  .LBB11_3:
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: cmpxchg_i16_acquire_monotonic:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sh a1, 6(sp)
 ; RV64I-NEXT:    addi a1, sp, 6
 ; RV64I-NEXT:    li a3, 2
 ; RV64I-NEXT:    li a4, 0
 ; RV64I-NEXT:    call __atomic_compare_exchange_2@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: cmpxchg_i16_acquire_monotonic:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a3, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    lui a4, 16
-; RV64IA-NEXT:    addiw a4, a4, -1
-; RV64IA-NEXT:    sllw a5, a4, a0
-; RV64IA-NEXT:    and a1, a1, a4
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:    and a2, a2, a4
-; RV64IA-NEXT:    sllw a0, a2, a0
-; RV64IA-NEXT:  .LBB11_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a2, (a3)
-; RV64IA-NEXT:    and a4, a2, a5
-; RV64IA-NEXT:    bne a4, a1, .LBB11_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB11_1 Depth=1
-; RV64IA-NEXT:    xor a4, a2, a0
-; RV64IA-NEXT:    and a4, a4, a5
-; RV64IA-NEXT:    xor a4, a2, a4
-; RV64IA-NEXT:    sc.w a4, a4, (a3)
-; RV64IA-NEXT:    bnez a4, .LBB11_1
-; RV64IA-NEXT:  .LBB11_3:
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: cmpxchg_i16_acquire_monotonic:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a3, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    lui a4, 16
+; RV64IA-WMO-NEXT:    addiw a4, a4, -1
+; RV64IA-WMO-NEXT:    sllw a5, a4, a0
+; RV64IA-WMO-NEXT:    and a1, a1, a4
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:    and a2, a2, a4
+; RV64IA-WMO-NEXT:    sllw a0, a2, a0
+; RV64IA-WMO-NEXT:  .LBB11_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a2, (a3)
+; RV64IA-WMO-NEXT:    and a4, a2, a5
+; RV64IA-WMO-NEXT:    bne a4, a1, .LBB11_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB11_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a4, a2, a0
+; RV64IA-WMO-NEXT:    and a4, a4, a5
+; RV64IA-WMO-NEXT:    xor a4, a2, a4
+; RV64IA-WMO-NEXT:    sc.w a4, a4, (a3)
+; RV64IA-WMO-NEXT:    bnez a4, .LBB11_1
+; RV64IA-WMO-NEXT:  .LBB11_3:
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: cmpxchg_i16_acquire_monotonic:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a3, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    lui a4, 16
+; RV64IA-TSO-NEXT:    addiw a4, a4, -1
+; RV64IA-TSO-NEXT:    sllw a5, a4, a0
+; RV64IA-TSO-NEXT:    and a1, a1, a4
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:    and a2, a2, a4
+; RV64IA-TSO-NEXT:    sllw a0, a2, a0
+; RV64IA-TSO-NEXT:  .LBB11_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a2, (a3)
+; RV64IA-TSO-NEXT:    and a4, a2, a5
+; RV64IA-TSO-NEXT:    bne a4, a1, .LBB11_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB11_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a4, a2, a0
+; RV64IA-TSO-NEXT:    and a4, a4, a5
+; RV64IA-TSO-NEXT:    xor a4, a2, a4
+; RV64IA-TSO-NEXT:    sc.w a4, a4, (a3)
+; RV64IA-TSO-NEXT:    bnez a4, .LBB11_1
+; RV64IA-TSO-NEXT:  .LBB11_3:
+; RV64IA-TSO-NEXT:    ret
   %res = cmpxchg ptr %ptr, i16 %cmp, i16 %val acquire monotonic
   ret void
 }
 
 define void @cmpxchg_i16_acquire_acquire(ptr %ptr, i16 %cmp, i16 %val) nounwind {
 ; RV32I-LABEL: cmpxchg_i16_acquire_acquire:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    sh a1, 10(sp)
 ; RV32I-NEXT:    addi a1, sp, 10
 ; RV32I-NEXT:    li a3, 2
 ; RV32I-NEXT:    li a4, 2
 ; RV32I-NEXT:    call __atomic_compare_exchange_2@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: cmpxchg_i16_acquire_acquire:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a3, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    lui a4, 16
-; RV32IA-NEXT:    addi a4, a4, -1
-; RV32IA-NEXT:    sll a5, a4, a0
-; RV32IA-NEXT:    and a1, a1, a4
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:    and a2, a2, a4
-; RV32IA-NEXT:    sll a0, a2, a0
-; RV32IA-NEXT:  .LBB12_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a2, (a3)
-; RV32IA-NEXT:    and a4, a2, a5
-; RV32IA-NEXT:    bne a4, a1, .LBB12_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB12_1 Depth=1
-; RV32IA-NEXT:    xor a4, a2, a0
-; RV32IA-NEXT:    and a4, a4, a5
-; RV32IA-NEXT:    xor a4, a2, a4
-; RV32IA-NEXT:    sc.w a4, a4, (a3)
-; RV32IA-NEXT:    bnez a4, .LBB12_1
-; RV32IA-NEXT:  .LBB12_3:
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: cmpxchg_i16_acquire_acquire:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a3, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    lui a4, 16
+; RV32IA-WMO-NEXT:    addi a4, a4, -1
+; RV32IA-WMO-NEXT:    sll a5, a4, a0
+; RV32IA-WMO-NEXT:    and a1, a1, a4
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:    and a2, a2, a4
+; RV32IA-WMO-NEXT:    sll a0, a2, a0
+; RV32IA-WMO-NEXT:  .LBB12_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a2, (a3)
+; RV32IA-WMO-NEXT:    and a4, a2, a5
+; RV32IA-WMO-NEXT:    bne a4, a1, .LBB12_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB12_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a4, a2, a0
+; RV32IA-WMO-NEXT:    and a4, a4, a5
+; RV32IA-WMO-NEXT:    xor a4, a2, a4
+; RV32IA-WMO-NEXT:    sc.w a4, a4, (a3)
+; RV32IA-WMO-NEXT:    bnez a4, .LBB12_1
+; RV32IA-WMO-NEXT:  .LBB12_3:
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: cmpxchg_i16_acquire_acquire:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a3, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    lui a4, 16
+; RV32IA-TSO-NEXT:    addi a4, a4, -1
+; RV32IA-TSO-NEXT:    sll a5, a4, a0
+; RV32IA-TSO-NEXT:    and a1, a1, a4
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:    and a2, a2, a4
+; RV32IA-TSO-NEXT:    sll a0, a2, a0
+; RV32IA-TSO-NEXT:  .LBB12_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a2, (a3)
+; RV32IA-TSO-NEXT:    and a4, a2, a5
+; RV32IA-TSO-NEXT:    bne a4, a1, .LBB12_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB12_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a4, a2, a0
+; RV32IA-TSO-NEXT:    and a4, a4, a5
+; RV32IA-TSO-NEXT:    xor a4, a2, a4
+; RV32IA-TSO-NEXT:    sc.w a4, a4, (a3)
+; RV32IA-TSO-NEXT:    bnez a4, .LBB12_1
+; RV32IA-TSO-NEXT:  .LBB12_3:
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: cmpxchg_i16_acquire_acquire:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sh a1, 6(sp)
 ; RV64I-NEXT:    addi a1, sp, 6
 ; RV64I-NEXT:    li a3, 2
 ; RV64I-NEXT:    li a4, 2
 ; RV64I-NEXT:    call __atomic_compare_exchange_2@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: cmpxchg_i16_acquire_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a3, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    lui a4, 16
-; RV64IA-NEXT:    addiw a4, a4, -1
-; RV64IA-NEXT:    sllw a5, a4, a0
-; RV64IA-NEXT:    and a1, a1, a4
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:    and a2, a2, a4
-; RV64IA-NEXT:    sllw a0, a2, a0
-; RV64IA-NEXT:  .LBB12_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a2, (a3)
-; RV64IA-NEXT:    and a4, a2, a5
-; RV64IA-NEXT:    bne a4, a1, .LBB12_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB12_1 Depth=1
-; RV64IA-NEXT:    xor a4, a2, a0
-; RV64IA-NEXT:    and a4, a4, a5
-; RV64IA-NEXT:    xor a4, a2, a4
-; RV64IA-NEXT:    sc.w a4, a4, (a3)
-; RV64IA-NEXT:    bnez a4, .LBB12_1
-; RV64IA-NEXT:  .LBB12_3:
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: cmpxchg_i16_acquire_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a3, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    lui a4, 16
+; RV64IA-WMO-NEXT:    addiw a4, a4, -1
+; RV64IA-WMO-NEXT:    sllw a5, a4, a0
+; RV64IA-WMO-NEXT:    and a1, a1, a4
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:    and a2, a2, a4
+; RV64IA-WMO-NEXT:    sllw a0, a2, a0
+; RV64IA-WMO-NEXT:  .LBB12_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a2, (a3)
+; RV64IA-WMO-NEXT:    and a4, a2, a5
+; RV64IA-WMO-NEXT:    bne a4, a1, .LBB12_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB12_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a4, a2, a0
+; RV64IA-WMO-NEXT:    and a4, a4, a5
+; RV64IA-WMO-NEXT:    xor a4, a2, a4
+; RV64IA-WMO-NEXT:    sc.w a4, a4, (a3)
+; RV64IA-WMO-NEXT:    bnez a4, .LBB12_1
+; RV64IA-WMO-NEXT:  .LBB12_3:
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: cmpxchg_i16_acquire_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a3, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    lui a4, 16
+; RV64IA-TSO-NEXT:    addiw a4, a4, -1
+; RV64IA-TSO-NEXT:    sllw a5, a4, a0
+; RV64IA-TSO-NEXT:    and a1, a1, a4
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:    and a2, a2, a4
+; RV64IA-TSO-NEXT:    sllw a0, a2, a0
+; RV64IA-TSO-NEXT:  .LBB12_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a2, (a3)
+; RV64IA-TSO-NEXT:    and a4, a2, a5
+; RV64IA-TSO-NEXT:    bne a4, a1, .LBB12_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB12_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a4, a2, a0
+; RV64IA-TSO-NEXT:    and a4, a4, a5
+; RV64IA-TSO-NEXT:    xor a4, a2, a4
+; RV64IA-TSO-NEXT:    sc.w a4, a4, (a3)
+; RV64IA-TSO-NEXT:    bnez a4, .LBB12_1
+; RV64IA-TSO-NEXT:  .LBB12_3:
+; RV64IA-TSO-NEXT:    ret
   %res = cmpxchg ptr %ptr, i16 %cmp, i16 %val acquire acquire
   ret void
 }
 
 define void @cmpxchg_i16_release_monotonic(ptr %ptr, i16 %cmp, i16 %val) nounwind {
 ; RV32I-LABEL: cmpxchg_i16_release_monotonic:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    sh a1, 10(sp)
 ; RV32I-NEXT:    addi a1, sp, 10
 ; RV32I-NEXT:    li a3, 3
 ; RV32I-NEXT:    li a4, 0
 ; RV32I-NEXT:    call __atomic_compare_exchange_2@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: cmpxchg_i16_release_monotonic:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a3, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    lui a4, 16
-; RV32IA-NEXT:    addi a4, a4, -1
-; RV32IA-NEXT:    sll a5, a4, a0
-; RV32IA-NEXT:    and a1, a1, a4
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:    and a2, a2, a4
-; RV32IA-NEXT:    sll a0, a2, a0
-; RV32IA-NEXT:  .LBB13_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w a2, (a3)
-; RV32IA-NEXT:    and a4, a2, a5
-; RV32IA-NEXT:    bne a4, a1, .LBB13_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB13_1 Depth=1
-; RV32IA-NEXT:    xor a4, a2, a0
-; RV32IA-NEXT:    and a4, a4, a5
-; RV32IA-NEXT:    xor a4, a2, a4
-; RV32IA-NEXT:    sc.w.rl a4, a4, (a3)
-; RV32IA-NEXT:    bnez a4, .LBB13_1
-; RV32IA-NEXT:  .LBB13_3:
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: cmpxchg_i16_release_monotonic:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a3, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    lui a4, 16
+; RV32IA-WMO-NEXT:    addi a4, a4, -1
+; RV32IA-WMO-NEXT:    sll a5, a4, a0
+; RV32IA-WMO-NEXT:    and a1, a1, a4
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:    and a2, a2, a4
+; RV32IA-WMO-NEXT:    sll a0, a2, a0
+; RV32IA-WMO-NEXT:  .LBB13_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w a2, (a3)
+; RV32IA-WMO-NEXT:    and a4, a2, a5
+; RV32IA-WMO-NEXT:    bne a4, a1, .LBB13_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB13_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a4, a2, a0
+; RV32IA-WMO-NEXT:    and a4, a4, a5
+; RV32IA-WMO-NEXT:    xor a4, a2, a4
+; RV32IA-WMO-NEXT:    sc.w.rl a4, a4, (a3)
+; RV32IA-WMO-NEXT:    bnez a4, .LBB13_1
+; RV32IA-WMO-NEXT:  .LBB13_3:
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: cmpxchg_i16_release_monotonic:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a3, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    lui a4, 16
+; RV32IA-TSO-NEXT:    addi a4, a4, -1
+; RV32IA-TSO-NEXT:    sll a5, a4, a0
+; RV32IA-TSO-NEXT:    and a1, a1, a4
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:    and a2, a2, a4
+; RV32IA-TSO-NEXT:    sll a0, a2, a0
+; RV32IA-TSO-NEXT:  .LBB13_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a2, (a3)
+; RV32IA-TSO-NEXT:    and a4, a2, a5
+; RV32IA-TSO-NEXT:    bne a4, a1, .LBB13_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB13_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a4, a2, a0
+; RV32IA-TSO-NEXT:    and a4, a4, a5
+; RV32IA-TSO-NEXT:    xor a4, a2, a4
+; RV32IA-TSO-NEXT:    sc.w a4, a4, (a3)
+; RV32IA-TSO-NEXT:    bnez a4, .LBB13_1
+; RV32IA-TSO-NEXT:  .LBB13_3:
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: cmpxchg_i16_release_monotonic:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sh a1, 6(sp)
 ; RV64I-NEXT:    addi a1, sp, 6
 ; RV64I-NEXT:    li a3, 3
 ; RV64I-NEXT:    li a4, 0
 ; RV64I-NEXT:    call __atomic_compare_exchange_2@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: cmpxchg_i16_release_monotonic:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a3, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    lui a4, 16
-; RV64IA-NEXT:    addiw a4, a4, -1
-; RV64IA-NEXT:    sllw a5, a4, a0
-; RV64IA-NEXT:    and a1, a1, a4
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:    and a2, a2, a4
-; RV64IA-NEXT:    sllw a0, a2, a0
-; RV64IA-NEXT:  .LBB13_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w a2, (a3)
-; RV64IA-NEXT:    and a4, a2, a5
-; RV64IA-NEXT:    bne a4, a1, .LBB13_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB13_1 Depth=1
-; RV64IA-NEXT:    xor a4, a2, a0
-; RV64IA-NEXT:    and a4, a4, a5
-; RV64IA-NEXT:    xor a4, a2, a4
-; RV64IA-NEXT:    sc.w.rl a4, a4, (a3)
-; RV64IA-NEXT:    bnez a4, .LBB13_1
-; RV64IA-NEXT:  .LBB13_3:
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: cmpxchg_i16_release_monotonic:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a3, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    lui a4, 16
+; RV64IA-WMO-NEXT:    addiw a4, a4, -1
+; RV64IA-WMO-NEXT:    sllw a5, a4, a0
+; RV64IA-WMO-NEXT:    and a1, a1, a4
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:    and a2, a2, a4
+; RV64IA-WMO-NEXT:    sllw a0, a2, a0
+; RV64IA-WMO-NEXT:  .LBB13_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w a2, (a3)
+; RV64IA-WMO-NEXT:    and a4, a2, a5
+; RV64IA-WMO-NEXT:    bne a4, a1, .LBB13_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB13_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a4, a2, a0
+; RV64IA-WMO-NEXT:    and a4, a4, a5
+; RV64IA-WMO-NEXT:    xor a4, a2, a4
+; RV64IA-WMO-NEXT:    sc.w.rl a4, a4, (a3)
+; RV64IA-WMO-NEXT:    bnez a4, .LBB13_1
+; RV64IA-WMO-NEXT:  .LBB13_3:
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: cmpxchg_i16_release_monotonic:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a3, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    lui a4, 16
+; RV64IA-TSO-NEXT:    addiw a4, a4, -1
+; RV64IA-TSO-NEXT:    sllw a5, a4, a0
+; RV64IA-TSO-NEXT:    and a1, a1, a4
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:    and a2, a2, a4
+; RV64IA-TSO-NEXT:    sllw a0, a2, a0
+; RV64IA-TSO-NEXT:  .LBB13_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a2, (a3)
+; RV64IA-TSO-NEXT:    and a4, a2, a5
+; RV64IA-TSO-NEXT:    bne a4, a1, .LBB13_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB13_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a4, a2, a0
+; RV64IA-TSO-NEXT:    and a4, a4, a5
+; RV64IA-TSO-NEXT:    xor a4, a2, a4
+; RV64IA-TSO-NEXT:    sc.w a4, a4, (a3)
+; RV64IA-TSO-NEXT:    bnez a4, .LBB13_1
+; RV64IA-TSO-NEXT:  .LBB13_3:
+; RV64IA-TSO-NEXT:    ret
   %res = cmpxchg ptr %ptr, i16 %cmp, i16 %val release monotonic
   ret void
 }
 
 define void @cmpxchg_i16_release_acquire(ptr %ptr, i16 %cmp, i16 %val) nounwind {
 ; RV32I-LABEL: cmpxchg_i16_release_acquire:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    sh a1, 10(sp)
 ; RV32I-NEXT:    addi a1, sp, 10
 ; RV32I-NEXT:    li a3, 3
 ; RV32I-NEXT:    li a4, 2
 ; RV32I-NEXT:    call __atomic_compare_exchange_2@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: cmpxchg_i16_release_acquire:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a3, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    lui a4, 16
-; RV32IA-NEXT:    addi a4, a4, -1
-; RV32IA-NEXT:    sll a5, a4, a0
-; RV32IA-NEXT:    and a1, a1, a4
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:    and a2, a2, a4
-; RV32IA-NEXT:    sll a0, a2, a0
-; RV32IA-NEXT:  .LBB14_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a2, (a3)
-; RV32IA-NEXT:    and a4, a2, a5
-; RV32IA-NEXT:    bne a4, a1, .LBB14_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB14_1 Depth=1
-; RV32IA-NEXT:    xor a4, a2, a0
-; RV32IA-NEXT:    and a4, a4, a5
-; RV32IA-NEXT:    xor a4, a2, a4
-; RV32IA-NEXT:    sc.w.rl a4, a4, (a3)
-; RV32IA-NEXT:    bnez a4, .LBB14_1
-; RV32IA-NEXT:  .LBB14_3:
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: cmpxchg_i16_release_acquire:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a3, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    lui a4, 16
+; RV32IA-WMO-NEXT:    addi a4, a4, -1
+; RV32IA-WMO-NEXT:    sll a5, a4, a0
+; RV32IA-WMO-NEXT:    and a1, a1, a4
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:    and a2, a2, a4
+; RV32IA-WMO-NEXT:    sll a0, a2, a0
+; RV32IA-WMO-NEXT:  .LBB14_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a2, (a3)
+; RV32IA-WMO-NEXT:    and a4, a2, a5
+; RV32IA-WMO-NEXT:    bne a4, a1, .LBB14_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB14_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a4, a2, a0
+; RV32IA-WMO-NEXT:    and a4, a4, a5
+; RV32IA-WMO-NEXT:    xor a4, a2, a4
+; RV32IA-WMO-NEXT:    sc.w.rl a4, a4, (a3)
+; RV32IA-WMO-NEXT:    bnez a4, .LBB14_1
+; RV32IA-WMO-NEXT:  .LBB14_3:
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: cmpxchg_i16_release_acquire:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a3, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    lui a4, 16
+; RV32IA-TSO-NEXT:    addi a4, a4, -1
+; RV32IA-TSO-NEXT:    sll a5, a4, a0
+; RV32IA-TSO-NEXT:    and a1, a1, a4
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:    and a2, a2, a4
+; RV32IA-TSO-NEXT:    sll a0, a2, a0
+; RV32IA-TSO-NEXT:  .LBB14_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a2, (a3)
+; RV32IA-TSO-NEXT:    and a4, a2, a5
+; RV32IA-TSO-NEXT:    bne a4, a1, .LBB14_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB14_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a4, a2, a0
+; RV32IA-TSO-NEXT:    and a4, a4, a5
+; RV32IA-TSO-NEXT:    xor a4, a2, a4
+; RV32IA-TSO-NEXT:    sc.w a4, a4, (a3)
+; RV32IA-TSO-NEXT:    bnez a4, .LBB14_1
+; RV32IA-TSO-NEXT:  .LBB14_3:
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: cmpxchg_i16_release_acquire:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sh a1, 6(sp)
 ; RV64I-NEXT:    addi a1, sp, 6
 ; RV64I-NEXT:    li a3, 3
 ; RV64I-NEXT:    li a4, 2
 ; RV64I-NEXT:    call __atomic_compare_exchange_2@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: cmpxchg_i16_release_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a3, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    lui a4, 16
-; RV64IA-NEXT:    addiw a4, a4, -1
-; RV64IA-NEXT:    sllw a5, a4, a0
-; RV64IA-NEXT:    and a1, a1, a4
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:    and a2, a2, a4
-; RV64IA-NEXT:    sllw a0, a2, a0
-; RV64IA-NEXT:  .LBB14_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a2, (a3)
-; RV64IA-NEXT:    and a4, a2, a5
-; RV64IA-NEXT:    bne a4, a1, .LBB14_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB14_1 Depth=1
-; RV64IA-NEXT:    xor a4, a2, a0
-; RV64IA-NEXT:    and a4, a4, a5
-; RV64IA-NEXT:    xor a4, a2, a4
-; RV64IA-NEXT:    sc.w.rl a4, a4, (a3)
-; RV64IA-NEXT:    bnez a4, .LBB14_1
-; RV64IA-NEXT:  .LBB14_3:
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: cmpxchg_i16_release_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a3, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    lui a4, 16
+; RV64IA-WMO-NEXT:    addiw a4, a4, -1
+; RV64IA-WMO-NEXT:    sllw a5, a4, a0
+; RV64IA-WMO-NEXT:    and a1, a1, a4
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:    and a2, a2, a4
+; RV64IA-WMO-NEXT:    sllw a0, a2, a0
+; RV64IA-WMO-NEXT:  .LBB14_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a2, (a3)
+; RV64IA-WMO-NEXT:    and a4, a2, a5
+; RV64IA-WMO-NEXT:    bne a4, a1, .LBB14_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB14_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a4, a2, a0
+; RV64IA-WMO-NEXT:    and a4, a4, a5
+; RV64IA-WMO-NEXT:    xor a4, a2, a4
+; RV64IA-WMO-NEXT:    sc.w.rl a4, a4, (a3)
+; RV64IA-WMO-NEXT:    bnez a4, .LBB14_1
+; RV64IA-WMO-NEXT:  .LBB14_3:
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: cmpxchg_i16_release_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a3, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    lui a4, 16
+; RV64IA-TSO-NEXT:    addiw a4, a4, -1
+; RV64IA-TSO-NEXT:    sllw a5, a4, a0
+; RV64IA-TSO-NEXT:    and a1, a1, a4
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:    and a2, a2, a4
+; RV64IA-TSO-NEXT:    sllw a0, a2, a0
+; RV64IA-TSO-NEXT:  .LBB14_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a2, (a3)
+; RV64IA-TSO-NEXT:    and a4, a2, a5
+; RV64IA-TSO-NEXT:    bne a4, a1, .LBB14_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB14_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a4, a2, a0
+; RV64IA-TSO-NEXT:    and a4, a4, a5
+; RV64IA-TSO-NEXT:    xor a4, a2, a4
+; RV64IA-TSO-NEXT:    sc.w a4, a4, (a3)
+; RV64IA-TSO-NEXT:    bnez a4, .LBB14_1
+; RV64IA-TSO-NEXT:  .LBB14_3:
+; RV64IA-TSO-NEXT:    ret
   %res = cmpxchg ptr %ptr, i16 %cmp, i16 %val release acquire
   ret void
 }
 
 define void @cmpxchg_i16_acq_rel_monotonic(ptr %ptr, i16 %cmp, i16 %val) nounwind {
 ; RV32I-LABEL: cmpxchg_i16_acq_rel_monotonic:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    sh a1, 10(sp)
 ; RV32I-NEXT:    addi a1, sp, 10
 ; RV32I-NEXT:    li a3, 4
 ; RV32I-NEXT:    li a4, 0
 ; RV32I-NEXT:    call __atomic_compare_exchange_2@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: cmpxchg_i16_acq_rel_monotonic:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a3, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    lui a4, 16
-; RV32IA-NEXT:    addi a4, a4, -1
-; RV32IA-NEXT:    sll a5, a4, a0
-; RV32IA-NEXT:    and a1, a1, a4
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:    and a2, a2, a4
-; RV32IA-NEXT:    sll a0, a2, a0
-; RV32IA-NEXT:  .LBB15_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a2, (a3)
-; RV32IA-NEXT:    and a4, a2, a5
-; RV32IA-NEXT:    bne a4, a1, .LBB15_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB15_1 Depth=1
-; RV32IA-NEXT:    xor a4, a2, a0
-; RV32IA-NEXT:    and a4, a4, a5
-; RV32IA-NEXT:    xor a4, a2, a4
-; RV32IA-NEXT:    sc.w.rl a4, a4, (a3)
-; RV32IA-NEXT:    bnez a4, .LBB15_1
-; RV32IA-NEXT:  .LBB15_3:
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: cmpxchg_i16_acq_rel_monotonic:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a3, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    lui a4, 16
+; RV32IA-WMO-NEXT:    addi a4, a4, -1
+; RV32IA-WMO-NEXT:    sll a5, a4, a0
+; RV32IA-WMO-NEXT:    and a1, a1, a4
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:    and a2, a2, a4
+; RV32IA-WMO-NEXT:    sll a0, a2, a0
+; RV32IA-WMO-NEXT:  .LBB15_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a2, (a3)
+; RV32IA-WMO-NEXT:    and a4, a2, a5
+; RV32IA-WMO-NEXT:    bne a4, a1, .LBB15_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB15_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a4, a2, a0
+; RV32IA-WMO-NEXT:    and a4, a4, a5
+; RV32IA-WMO-NEXT:    xor a4, a2, a4
+; RV32IA-WMO-NEXT:    sc.w.rl a4, a4, (a3)
+; RV32IA-WMO-NEXT:    bnez a4, .LBB15_1
+; RV32IA-WMO-NEXT:  .LBB15_3:
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: cmpxchg_i16_acq_rel_monotonic:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a3, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    lui a4, 16
+; RV32IA-TSO-NEXT:    addi a4, a4, -1
+; RV32IA-TSO-NEXT:    sll a5, a4, a0
+; RV32IA-TSO-NEXT:    and a1, a1, a4
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:    and a2, a2, a4
+; RV32IA-TSO-NEXT:    sll a0, a2, a0
+; RV32IA-TSO-NEXT:  .LBB15_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a2, (a3)
+; RV32IA-TSO-NEXT:    and a4, a2, a5
+; RV32IA-TSO-NEXT:    bne a4, a1, .LBB15_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB15_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a4, a2, a0
+; RV32IA-TSO-NEXT:    and a4, a4, a5
+; RV32IA-TSO-NEXT:    xor a4, a2, a4
+; RV32IA-TSO-NEXT:    sc.w a4, a4, (a3)
+; RV32IA-TSO-NEXT:    bnez a4, .LBB15_1
+; RV32IA-TSO-NEXT:  .LBB15_3:
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: cmpxchg_i16_acq_rel_monotonic:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sh a1, 6(sp)
 ; RV64I-NEXT:    addi a1, sp, 6
 ; RV64I-NEXT:    li a3, 4
 ; RV64I-NEXT:    li a4, 0
 ; RV64I-NEXT:    call __atomic_compare_exchange_2@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: cmpxchg_i16_acq_rel_monotonic:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a3, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    lui a4, 16
-; RV64IA-NEXT:    addiw a4, a4, -1
-; RV64IA-NEXT:    sllw a5, a4, a0
-; RV64IA-NEXT:    and a1, a1, a4
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:    and a2, a2, a4
-; RV64IA-NEXT:    sllw a0, a2, a0
-; RV64IA-NEXT:  .LBB15_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a2, (a3)
-; RV64IA-NEXT:    and a4, a2, a5
-; RV64IA-NEXT:    bne a4, a1, .LBB15_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB15_1 Depth=1
-; RV64IA-NEXT:    xor a4, a2, a0
-; RV64IA-NEXT:    and a4, a4, a5
-; RV64IA-NEXT:    xor a4, a2, a4
-; RV64IA-NEXT:    sc.w.rl a4, a4, (a3)
-; RV64IA-NEXT:    bnez a4, .LBB15_1
-; RV64IA-NEXT:  .LBB15_3:
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: cmpxchg_i16_acq_rel_monotonic:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a3, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    lui a4, 16
+; RV64IA-WMO-NEXT:    addiw a4, a4, -1
+; RV64IA-WMO-NEXT:    sllw a5, a4, a0
+; RV64IA-WMO-NEXT:    and a1, a1, a4
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:    and a2, a2, a4
+; RV64IA-WMO-NEXT:    sllw a0, a2, a0
+; RV64IA-WMO-NEXT:  .LBB15_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a2, (a3)
+; RV64IA-WMO-NEXT:    and a4, a2, a5
+; RV64IA-WMO-NEXT:    bne a4, a1, .LBB15_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB15_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a4, a2, a0
+; RV64IA-WMO-NEXT:    and a4, a4, a5
+; RV64IA-WMO-NEXT:    xor a4, a2, a4
+; RV64IA-WMO-NEXT:    sc.w.rl a4, a4, (a3)
+; RV64IA-WMO-NEXT:    bnez a4, .LBB15_1
+; RV64IA-WMO-NEXT:  .LBB15_3:
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: cmpxchg_i16_acq_rel_monotonic:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a3, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    lui a4, 16
+; RV64IA-TSO-NEXT:    addiw a4, a4, -1
+; RV64IA-TSO-NEXT:    sllw a5, a4, a0
+; RV64IA-TSO-NEXT:    and a1, a1, a4
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:    and a2, a2, a4
+; RV64IA-TSO-NEXT:    sllw a0, a2, a0
+; RV64IA-TSO-NEXT:  .LBB15_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a2, (a3)
+; RV64IA-TSO-NEXT:    and a4, a2, a5
+; RV64IA-TSO-NEXT:    bne a4, a1, .LBB15_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB15_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a4, a2, a0
+; RV64IA-TSO-NEXT:    and a4, a4, a5
+; RV64IA-TSO-NEXT:    xor a4, a2, a4
+; RV64IA-TSO-NEXT:    sc.w a4, a4, (a3)
+; RV64IA-TSO-NEXT:    bnez a4, .LBB15_1
+; RV64IA-TSO-NEXT:  .LBB15_3:
+; RV64IA-TSO-NEXT:    ret
   %res = cmpxchg ptr %ptr, i16 %cmp, i16 %val acq_rel monotonic
   ret void
 }
 
 define void @cmpxchg_i16_acq_rel_acquire(ptr %ptr, i16 %cmp, i16 %val) nounwind {
 ; RV32I-LABEL: cmpxchg_i16_acq_rel_acquire:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    sh a1, 10(sp)
 ; RV32I-NEXT:    addi a1, sp, 10
 ; RV32I-NEXT:    li a3, 4
 ; RV32I-NEXT:    li a4, 2
 ; RV32I-NEXT:    call __atomic_compare_exchange_2@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: cmpxchg_i16_acq_rel_acquire:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a3, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    lui a4, 16
-; RV32IA-NEXT:    addi a4, a4, -1
-; RV32IA-NEXT:    sll a5, a4, a0
-; RV32IA-NEXT:    and a1, a1, a4
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:    and a2, a2, a4
-; RV32IA-NEXT:    sll a0, a2, a0
-; RV32IA-NEXT:  .LBB16_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a2, (a3)
-; RV32IA-NEXT:    and a4, a2, a5
-; RV32IA-NEXT:    bne a4, a1, .LBB16_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB16_1 Depth=1
-; RV32IA-NEXT:    xor a4, a2, a0
-; RV32IA-NEXT:    and a4, a4, a5
-; RV32IA-NEXT:    xor a4, a2, a4
-; RV32IA-NEXT:    sc.w.rl a4, a4, (a3)
-; RV32IA-NEXT:    bnez a4, .LBB16_1
-; RV32IA-NEXT:  .LBB16_3:
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: cmpxchg_i16_acq_rel_acquire:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a3, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    lui a4, 16
+; RV32IA-WMO-NEXT:    addi a4, a4, -1
+; RV32IA-WMO-NEXT:    sll a5, a4, a0
+; RV32IA-WMO-NEXT:    and a1, a1, a4
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:    and a2, a2, a4
+; RV32IA-WMO-NEXT:    sll a0, a2, a0
+; RV32IA-WMO-NEXT:  .LBB16_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a2, (a3)
+; RV32IA-WMO-NEXT:    and a4, a2, a5
+; RV32IA-WMO-NEXT:    bne a4, a1, .LBB16_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB16_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a4, a2, a0
+; RV32IA-WMO-NEXT:    and a4, a4, a5
+; RV32IA-WMO-NEXT:    xor a4, a2, a4
+; RV32IA-WMO-NEXT:    sc.w.rl a4, a4, (a3)
+; RV32IA-WMO-NEXT:    bnez a4, .LBB16_1
+; RV32IA-WMO-NEXT:  .LBB16_3:
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: cmpxchg_i16_acq_rel_acquire:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a3, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    lui a4, 16
+; RV32IA-TSO-NEXT:    addi a4, a4, -1
+; RV32IA-TSO-NEXT:    sll a5, a4, a0
+; RV32IA-TSO-NEXT:    and a1, a1, a4
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:    and a2, a2, a4
+; RV32IA-TSO-NEXT:    sll a0, a2, a0
+; RV32IA-TSO-NEXT:  .LBB16_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a2, (a3)
+; RV32IA-TSO-NEXT:    and a4, a2, a5
+; RV32IA-TSO-NEXT:    bne a4, a1, .LBB16_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB16_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a4, a2, a0
+; RV32IA-TSO-NEXT:    and a4, a4, a5
+; RV32IA-TSO-NEXT:    xor a4, a2, a4
+; RV32IA-TSO-NEXT:    sc.w a4, a4, (a3)
+; RV32IA-TSO-NEXT:    bnez a4, .LBB16_1
+; RV32IA-TSO-NEXT:  .LBB16_3:
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: cmpxchg_i16_acq_rel_acquire:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sh a1, 6(sp)
 ; RV64I-NEXT:    addi a1, sp, 6
 ; RV64I-NEXT:    li a3, 4
 ; RV64I-NEXT:    li a4, 2
 ; RV64I-NEXT:    call __atomic_compare_exchange_2@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: cmpxchg_i16_acq_rel_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a3, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    lui a4, 16
-; RV64IA-NEXT:    addiw a4, a4, -1
-; RV64IA-NEXT:    sllw a5, a4, a0
-; RV64IA-NEXT:    and a1, a1, a4
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:    and a2, a2, a4
-; RV64IA-NEXT:    sllw a0, a2, a0
-; RV64IA-NEXT:  .LBB16_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a2, (a3)
-; RV64IA-NEXT:    and a4, a2, a5
-; RV64IA-NEXT:    bne a4, a1, .LBB16_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB16_1 Depth=1
-; RV64IA-NEXT:    xor a4, a2, a0
-; RV64IA-NEXT:    and a4, a4, a5
-; RV64IA-NEXT:    xor a4, a2, a4
-; RV64IA-NEXT:    sc.w.rl a4, a4, (a3)
-; RV64IA-NEXT:    bnez a4, .LBB16_1
-; RV64IA-NEXT:  .LBB16_3:
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: cmpxchg_i16_acq_rel_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a3, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    lui a4, 16
+; RV64IA-WMO-NEXT:    addiw a4, a4, -1
+; RV64IA-WMO-NEXT:    sllw a5, a4, a0
+; RV64IA-WMO-NEXT:    and a1, a1, a4
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:    and a2, a2, a4
+; RV64IA-WMO-NEXT:    sllw a0, a2, a0
+; RV64IA-WMO-NEXT:  .LBB16_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a2, (a3)
+; RV64IA-WMO-NEXT:    and a4, a2, a5
+; RV64IA-WMO-NEXT:    bne a4, a1, .LBB16_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB16_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a4, a2, a0
+; RV64IA-WMO-NEXT:    and a4, a4, a5
+; RV64IA-WMO-NEXT:    xor a4, a2, a4
+; RV64IA-WMO-NEXT:    sc.w.rl a4, a4, (a3)
+; RV64IA-WMO-NEXT:    bnez a4, .LBB16_1
+; RV64IA-WMO-NEXT:  .LBB16_3:
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: cmpxchg_i16_acq_rel_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a3, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    lui a4, 16
+; RV64IA-TSO-NEXT:    addiw a4, a4, -1
+; RV64IA-TSO-NEXT:    sllw a5, a4, a0
+; RV64IA-TSO-NEXT:    and a1, a1, a4
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:    and a2, a2, a4
+; RV64IA-TSO-NEXT:    sllw a0, a2, a0
+; RV64IA-TSO-NEXT:  .LBB16_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a2, (a3)
+; RV64IA-TSO-NEXT:    and a4, a2, a5
+; RV64IA-TSO-NEXT:    bne a4, a1, .LBB16_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB16_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a4, a2, a0
+; RV64IA-TSO-NEXT:    and a4, a4, a5
+; RV64IA-TSO-NEXT:    xor a4, a2, a4
+; RV64IA-TSO-NEXT:    sc.w a4, a4, (a3)
+; RV64IA-TSO-NEXT:    bnez a4, .LBB16_1
+; RV64IA-TSO-NEXT:  .LBB16_3:
+; RV64IA-TSO-NEXT:    ret
   %res = cmpxchg ptr %ptr, i16 %cmp, i16 %val acq_rel acquire
   ret void
 }
 
 define void @cmpxchg_i16_seq_cst_monotonic(ptr %ptr, i16 %cmp, i16 %val) nounwind {
 ; RV32I-LABEL: cmpxchg_i16_seq_cst_monotonic:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16

 ; RV32I-NEXT:    addi a1, sp, 8
 ; RV32I-NEXT:    li a3, 2
 ; RV32I-NEXT:    li a4, 0
 ; RV32I-NEXT:    call __atomic_compare_exchange_4@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: cmpxchg_i32_acquire_monotonic:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:  .LBB21_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a3, (a0)
-; RV32IA-NEXT:    bne a3, a1, .LBB21_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB21_1 Depth=1
-; RV32IA-NEXT:    sc.w a4, a2, (a0)
-; RV32IA-NEXT:    bnez a4, .LBB21_1
-; RV32IA-NEXT:  .LBB21_3:
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: cmpxchg_i32_acquire_monotonic:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:  .LBB21_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a3, (a0)
+; RV32IA-WMO-NEXT:    bne a3, a1, .LBB21_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB21_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w a4, a2, (a0)
+; RV32IA-WMO-NEXT:    bnez a4, .LBB21_1
+; RV32IA-WMO-NEXT:  .LBB21_3:
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: cmpxchg_i32_acquire_monotonic:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:  .LBB21_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a0)
+; RV32IA-TSO-NEXT:    bne a3, a1, .LBB21_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB21_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a4, a2, (a0)
+; RV32IA-TSO-NEXT:    bnez a4, .LBB21_1
+; RV32IA-TSO-NEXT:  .LBB21_3:
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: cmpxchg_i32_acquire_monotonic:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sw a1, 4(sp)
 ; RV64I-NEXT:    addi a1, sp, 4
 ; RV64I-NEXT:    li a3, 2
 ; RV64I-NEXT:    li a4, 0
 ; RV64I-NEXT:    call __atomic_compare_exchange_4@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: cmpxchg_i32_acquire_monotonic:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    sext.w a1, a1
-; RV64IA-NEXT:  .LBB21_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a3, (a0)
-; RV64IA-NEXT:    bne a3, a1, .LBB21_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB21_1 Depth=1
-; RV64IA-NEXT:    sc.w a4, a2, (a0)
-; RV64IA-NEXT:    bnez a4, .LBB21_1
-; RV64IA-NEXT:  .LBB21_3:
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: cmpxchg_i32_acquire_monotonic:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    sext.w a1, a1
+; RV64IA-WMO-NEXT:  .LBB21_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a3, (a0)
+; RV64IA-WMO-NEXT:    bne a3, a1, .LBB21_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB21_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w a4, a2, (a0)
+; RV64IA-WMO-NEXT:    bnez a4, .LBB21_1
+; RV64IA-WMO-NEXT:  .LBB21_3:
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: cmpxchg_i32_acquire_monotonic:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    sext.w a1, a1
+; RV64IA-TSO-NEXT:  .LBB21_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a0)
+; RV64IA-TSO-NEXT:    bne a3, a1, .LBB21_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB21_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a4, a2, (a0)
+; RV64IA-TSO-NEXT:    bnez a4, .LBB21_1
+; RV64IA-TSO-NEXT:  .LBB21_3:
+; RV64IA-TSO-NEXT:    ret
   %res = cmpxchg ptr %ptr, i32 %cmp, i32 %val acquire monotonic
   ret void
 }
 
 define void @cmpxchg_i32_acquire_acquire(ptr %ptr, i32 %cmp, i32 %val) nounwind {
 ; RV32I-LABEL: cmpxchg_i32_acquire_acquire:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    sw a1, 8(sp)
 ; RV32I-NEXT:    addi a1, sp, 8
 ; RV32I-NEXT:    li a3, 2
 ; RV32I-NEXT:    li a4, 2
 ; RV32I-NEXT:    call __atomic_compare_exchange_4@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: cmpxchg_i32_acquire_acquire:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:  .LBB22_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a3, (a0)
-; RV32IA-NEXT:    bne a3, a1, .LBB22_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB22_1 Depth=1
-; RV32IA-NEXT:    sc.w a4, a2, (a0)
-; RV32IA-NEXT:    bnez a4, .LBB22_1
-; RV32IA-NEXT:  .LBB22_3:
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: cmpxchg_i32_acquire_acquire:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:  .LBB22_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a3, (a0)
+; RV32IA-WMO-NEXT:    bne a3, a1, .LBB22_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB22_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w a4, a2, (a0)
+; RV32IA-WMO-NEXT:    bnez a4, .LBB22_1
+; RV32IA-WMO-NEXT:  .LBB22_3:
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: cmpxchg_i32_acquire_acquire:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:  .LBB22_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a0)
+; RV32IA-TSO-NEXT:    bne a3, a1, .LBB22_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB22_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a4, a2, (a0)
+; RV32IA-TSO-NEXT:    bnez a4, .LBB22_1
+; RV32IA-TSO-NEXT:  .LBB22_3:
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: cmpxchg_i32_acquire_acquire:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sw a1, 4(sp)
 ; RV64I-NEXT:    addi a1, sp, 4
 ; RV64I-NEXT:    li a3, 2
 ; RV64I-NEXT:    li a4, 2
 ; RV64I-NEXT:    call __atomic_compare_exchange_4@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: cmpxchg_i32_acquire_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    sext.w a1, a1
-; RV64IA-NEXT:  .LBB22_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a3, (a0)
-; RV64IA-NEXT:    bne a3, a1, .LBB22_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB22_1 Depth=1
-; RV64IA-NEXT:    sc.w a4, a2, (a0)
-; RV64IA-NEXT:    bnez a4, .LBB22_1
-; RV64IA-NEXT:  .LBB22_3:
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: cmpxchg_i32_acquire_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    sext.w a1, a1
+; RV64IA-WMO-NEXT:  .LBB22_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a3, (a0)
+; RV64IA-WMO-NEXT:    bne a3, a1, .LBB22_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB22_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w a4, a2, (a0)
+; RV64IA-WMO-NEXT:    bnez a4, .LBB22_1
+; RV64IA-WMO-NEXT:  .LBB22_3:
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: cmpxchg_i32_acquire_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    sext.w a1, a1
+; RV64IA-TSO-NEXT:  .LBB22_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a0)
+; RV64IA-TSO-NEXT:    bne a3, a1, .LBB22_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB22_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a4, a2, (a0)
+; RV64IA-TSO-NEXT:    bnez a4, .LBB22_1
+; RV64IA-TSO-NEXT:  .LBB22_3:
+; RV64IA-TSO-NEXT:    ret
   %res = cmpxchg ptr %ptr, i32 %cmp, i32 %val acquire acquire
   ret void
 }
 
 define void @cmpxchg_i32_release_monotonic(ptr %ptr, i32 %cmp, i32 %val) nounwind {
 ; RV32I-LABEL: cmpxchg_i32_release_monotonic:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    sw a1, 8(sp)
 ; RV32I-NEXT:    addi a1, sp, 8
 ; RV32I-NEXT:    li a3, 3
 ; RV32I-NEXT:    li a4, 0
 ; RV32I-NEXT:    call __atomic_compare_exchange_4@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: cmpxchg_i32_release_monotonic:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:  .LBB23_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w a3, (a0)
-; RV32IA-NEXT:    bne a3, a1, .LBB23_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB23_1 Depth=1
-; RV32IA-NEXT:    sc.w.rl a4, a2, (a0)
-; RV32IA-NEXT:    bnez a4, .LBB23_1
-; RV32IA-NEXT:  .LBB23_3:
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: cmpxchg_i32_release_monotonic:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:  .LBB23_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w a3, (a0)
+; RV32IA-WMO-NEXT:    bne a3, a1, .LBB23_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB23_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w.rl a4, a2, (a0)
+; RV32IA-WMO-NEXT:    bnez a4, .LBB23_1
+; RV32IA-WMO-NEXT:  .LBB23_3:
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: cmpxchg_i32_release_monotonic:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:  .LBB23_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a0)
+; RV32IA-TSO-NEXT:    bne a3, a1, .LBB23_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB23_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a4, a2, (a0)
+; RV32IA-TSO-NEXT:    bnez a4, .LBB23_1
+; RV32IA-TSO-NEXT:  .LBB23_3:
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: cmpxchg_i32_release_monotonic:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sw a1, 4(sp)
 ; RV64I-NEXT:    addi a1, sp, 4
 ; RV64I-NEXT:    li a3, 3
 ; RV64I-NEXT:    li a4, 0
 ; RV64I-NEXT:    call __atomic_compare_exchange_4@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: cmpxchg_i32_release_monotonic:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    sext.w a1, a1
-; RV64IA-NEXT:  .LBB23_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w a3, (a0)
-; RV64IA-NEXT:    bne a3, a1, .LBB23_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB23_1 Depth=1
-; RV64IA-NEXT:    sc.w.rl a4, a2, (a0)
-; RV64IA-NEXT:    bnez a4, .LBB23_1
-; RV64IA-NEXT:  .LBB23_3:
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: cmpxchg_i32_release_monotonic:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    sext.w a1, a1
+; RV64IA-WMO-NEXT:  .LBB23_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w a3, (a0)
+; RV64IA-WMO-NEXT:    bne a3, a1, .LBB23_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB23_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w.rl a4, a2, (a0)
+; RV64IA-WMO-NEXT:    bnez a4, .LBB23_1
+; RV64IA-WMO-NEXT:  .LBB23_3:
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: cmpxchg_i32_release_monotonic:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    sext.w a1, a1
+; RV64IA-TSO-NEXT:  .LBB23_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a0)
+; RV64IA-TSO-NEXT:    bne a3, a1, .LBB23_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB23_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a4, a2, (a0)
+; RV64IA-TSO-NEXT:    bnez a4, .LBB23_1
+; RV64IA-TSO-NEXT:  .LBB23_3:
+; RV64IA-TSO-NEXT:    ret
   %res = cmpxchg ptr %ptr, i32 %cmp, i32 %val release monotonic
   ret void
 }
 
 define void @cmpxchg_i32_release_acquire(ptr %ptr, i32 %cmp, i32 %val) nounwind {
 ; RV32I-LABEL: cmpxchg_i32_release_acquire:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    sw a1, 8(sp)
 ; RV32I-NEXT:    addi a1, sp, 8
 ; RV32I-NEXT:    li a3, 3
 ; RV32I-NEXT:    li a4, 2
 ; RV32I-NEXT:    call __atomic_compare_exchange_4@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: cmpxchg_i32_release_acquire:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:  .LBB24_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a3, (a0)
-; RV32IA-NEXT:    bne a3, a1, .LBB24_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB24_1 Depth=1
-; RV32IA-NEXT:    sc.w.rl a4, a2, (a0)
-; RV32IA-NEXT:    bnez a4, .LBB24_1
-; RV32IA-NEXT:  .LBB24_3:
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: cmpxchg_i32_release_acquire:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:  .LBB24_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a3, (a0)
+; RV32IA-WMO-NEXT:    bne a3, a1, .LBB24_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB24_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w.rl a4, a2, (a0)
+; RV32IA-WMO-NEXT:    bnez a4, .LBB24_1
+; RV32IA-WMO-NEXT:  .LBB24_3:
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: cmpxchg_i32_release_acquire:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:  .LBB24_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a0)
+; RV32IA-TSO-NEXT:    bne a3, a1, .LBB24_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB24_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a4, a2, (a0)
+; RV32IA-TSO-NEXT:    bnez a4, .LBB24_1
+; RV32IA-TSO-NEXT:  .LBB24_3:
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: cmpxchg_i32_release_acquire:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sw a1, 4(sp)
 ; RV64I-NEXT:    addi a1, sp, 4
 ; RV64I-NEXT:    li a3, 3
 ; RV64I-NEXT:    li a4, 2
 ; RV64I-NEXT:    call __atomic_compare_exchange_4@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: cmpxchg_i32_release_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    sext.w a1, a1
-; RV64IA-NEXT:  .LBB24_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a3, (a0)
-; RV64IA-NEXT:    bne a3, a1, .LBB24_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB24_1 Depth=1
-; RV64IA-NEXT:    sc.w.rl a4, a2, (a0)
-; RV64IA-NEXT:    bnez a4, .LBB24_1
-; RV64IA-NEXT:  .LBB24_3:
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: cmpxchg_i32_release_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    sext.w a1, a1
+; RV64IA-WMO-NEXT:  .LBB24_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a3, (a0)
+; RV64IA-WMO-NEXT:    bne a3, a1, .LBB24_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB24_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w.rl a4, a2, (a0)
+; RV64IA-WMO-NEXT:    bnez a4, .LBB24_1
+; RV64IA-WMO-NEXT:  .LBB24_3:
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: cmpxchg_i32_release_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    sext.w a1, a1
+; RV64IA-TSO-NEXT:  .LBB24_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a0)
+; RV64IA-TSO-NEXT:    bne a3, a1, .LBB24_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB24_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a4, a2, (a0)
+; RV64IA-TSO-NEXT:    bnez a4, .LBB24_1
+; RV64IA-TSO-NEXT:  .LBB24_3:
+; RV64IA-TSO-NEXT:    ret
   %res = cmpxchg ptr %ptr, i32 %cmp, i32 %val release acquire
   ret void
 }
 
 define void @cmpxchg_i32_acq_rel_monotonic(ptr %ptr, i32 %cmp, i32 %val) nounwind {
 ; RV32I-LABEL: cmpxchg_i32_acq_rel_monotonic:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    sw a1, 8(sp)
 ; RV32I-NEXT:    addi a1, sp, 8
 ; RV32I-NEXT:    li a3, 4
 ; RV32I-NEXT:    li a4, 0
 ; RV32I-NEXT:    call __atomic_compare_exchange_4@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: cmpxchg_i32_acq_rel_monotonic:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:  .LBB25_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a3, (a0)
-; RV32IA-NEXT:    bne a3, a1, .LBB25_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB25_1 Depth=1
-; RV32IA-NEXT:    sc.w.rl a4, a2, (a0)
-; RV32IA-NEXT:    bnez a4, .LBB25_1
-; RV32IA-NEXT:  .LBB25_3:
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: cmpxchg_i32_acq_rel_monotonic:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:  .LBB25_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a3, (a0)
+; RV32IA-WMO-NEXT:    bne a3, a1, .LBB25_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB25_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w.rl a4, a2, (a0)
+; RV32IA-WMO-NEXT:    bnez a4, .LBB25_1
+; RV32IA-WMO-NEXT:  .LBB25_3:
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: cmpxchg_i32_acq_rel_monotonic:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:  .LBB25_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a0)
+; RV32IA-TSO-NEXT:    bne a3, a1, .LBB25_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB25_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a4, a2, (a0)
+; RV32IA-TSO-NEXT:    bnez a4, .LBB25_1
+; RV32IA-TSO-NEXT:  .LBB25_3:
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: cmpxchg_i32_acq_rel_monotonic:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sw a1, 4(sp)
 ; RV64I-NEXT:    addi a1, sp, 4
 ; RV64I-NEXT:    li a3, 4
 ; RV64I-NEXT:    li a4, 0
 ; RV64I-NEXT:    call __atomic_compare_exchange_4@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: cmpxchg_i32_acq_rel_monotonic:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    sext.w a1, a1
-; RV64IA-NEXT:  .LBB25_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a3, (a0)
-; RV64IA-NEXT:    bne a3, a1, .LBB25_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB25_1 Depth=1
-; RV64IA-NEXT:    sc.w.rl a4, a2, (a0)
-; RV64IA-NEXT:    bnez a4, .LBB25_1
-; RV64IA-NEXT:  .LBB25_3:
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: cmpxchg_i32_acq_rel_monotonic:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    sext.w a1, a1
+; RV64IA-WMO-NEXT:  .LBB25_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a3, (a0)
+; RV64IA-WMO-NEXT:    bne a3, a1, .LBB25_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB25_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w.rl a4, a2, (a0)
+; RV64IA-WMO-NEXT:    bnez a4, .LBB25_1
+; RV64IA-WMO-NEXT:  .LBB25_3:
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: cmpxchg_i32_acq_rel_monotonic:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    sext.w a1, a1
+; RV64IA-TSO-NEXT:  .LBB25_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a0)
+; RV64IA-TSO-NEXT:    bne a3, a1, .LBB25_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB25_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a4, a2, (a0)
+; RV64IA-TSO-NEXT:    bnez a4, .LBB25_1
+; RV64IA-TSO-NEXT:  .LBB25_3:
+; RV64IA-TSO-NEXT:    ret
   %res = cmpxchg ptr %ptr, i32 %cmp, i32 %val acq_rel monotonic
   ret void
 }
 
 define void @cmpxchg_i32_acq_rel_acquire(ptr %ptr, i32 %cmp, i32 %val) nounwind {
 ; RV32I-LABEL: cmpxchg_i32_acq_rel_acquire:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    sw a1, 8(sp)
 ; RV32I-NEXT:    addi a1, sp, 8
 ; RV32I-NEXT:    li a3, 4
 ; RV32I-NEXT:    li a4, 2
 ; RV32I-NEXT:    call __atomic_compare_exchange_4@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: cmpxchg_i32_acq_rel_acquire:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:  .LBB26_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a3, (a0)
-; RV32IA-NEXT:    bne a3, a1, .LBB26_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB26_1 Depth=1
-; RV32IA-NEXT:    sc.w.rl a4, a2, (a0)
-; RV32IA-NEXT:    bnez a4, .LBB26_1
-; RV32IA-NEXT:  .LBB26_3:
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: cmpxchg_i32_acq_rel_acquire:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:  .LBB26_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a3, (a0)
+; RV32IA-WMO-NEXT:    bne a3, a1, .LBB26_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB26_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w.rl a4, a2, (a0)
+; RV32IA-WMO-NEXT:    bnez a4, .LBB26_1
+; RV32IA-WMO-NEXT:  .LBB26_3:
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: cmpxchg_i32_acq_rel_acquire:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:  .LBB26_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a0)
+; RV32IA-TSO-NEXT:    bne a3, a1, .LBB26_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB26_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a4, a2, (a0)
+; RV32IA-TSO-NEXT:    bnez a4, .LBB26_1
+; RV32IA-TSO-NEXT:  .LBB26_3:
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: cmpxchg_i32_acq_rel_acquire:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sw a1, 4(sp)
 ; RV64I-NEXT:    addi a1, sp, 4
 ; RV64I-NEXT:    li a3, 4
 ; RV64I-NEXT:    li a4, 2
 ; RV64I-NEXT:    call __atomic_compare_exchange_4@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: cmpxchg_i32_acq_rel_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    sext.w a1, a1
-; RV64IA-NEXT:  .LBB26_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a3, (a0)
-; RV64IA-NEXT:    bne a3, a1, .LBB26_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB26_1 Depth=1
-; RV64IA-NEXT:    sc.w.rl a4, a2, (a0)
-; RV64IA-NEXT:    bnez a4, .LBB26_1
-; RV64IA-NEXT:  .LBB26_3:
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: cmpxchg_i32_acq_rel_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    sext.w a1, a1
+; RV64IA-WMO-NEXT:  .LBB26_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a3, (a0)
+; RV64IA-WMO-NEXT:    bne a3, a1, .LBB26_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB26_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w.rl a4, a2, (a0)
+; RV64IA-WMO-NEXT:    bnez a4, .LBB26_1
+; RV64IA-WMO-NEXT:  .LBB26_3:
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: cmpxchg_i32_acq_rel_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    sext.w a1, a1
+; RV64IA-TSO-NEXT:  .LBB26_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a0)
+; RV64IA-TSO-NEXT:    bne a3, a1, .LBB26_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB26_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a4, a2, (a0)
+; RV64IA-TSO-NEXT:    bnez a4, .LBB26_1
+; RV64IA-TSO-NEXT:  .LBB26_3:
+; RV64IA-TSO-NEXT:    ret
   %res = cmpxchg ptr %ptr, i32 %cmp, i32 %val acq_rel acquire
   ret void
 }
 
 define void @cmpxchg_i32_seq_cst_monotonic(ptr %ptr, i32 %cmp, i32 %val) nounwind {
 ; RV32I-LABEL: cmpxchg_i32_seq_cst_monotonic:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16

 ; RV64I-NEXT:    mv a1, sp
 ; RV64I-NEXT:    li a3, 2
 ; RV64I-NEXT:    li a4, 0
 ; RV64I-NEXT:    call __atomic_compare_exchange_8@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: cmpxchg_i64_acquire_monotonic:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:  .LBB31_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.d.aq a3, (a0)
-; RV64IA-NEXT:    bne a3, a1, .LBB31_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB31_1 Depth=1
-; RV64IA-NEXT:    sc.d a4, a2, (a0)
-; RV64IA-NEXT:    bnez a4, .LBB31_1
-; RV64IA-NEXT:  .LBB31_3:
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: cmpxchg_i64_acquire_monotonic:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:  .LBB31_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.d.aq a3, (a0)
+; RV64IA-WMO-NEXT:    bne a3, a1, .LBB31_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB31_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.d a4, a2, (a0)
+; RV64IA-WMO-NEXT:    bnez a4, .LBB31_1
+; RV64IA-WMO-NEXT:  .LBB31_3:
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: cmpxchg_i64_acquire_monotonic:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:  .LBB31_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.d a3, (a0)
+; RV64IA-TSO-NEXT:    bne a3, a1, .LBB31_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB31_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.d a4, a2, (a0)
+; RV64IA-TSO-NEXT:    bnez a4, .LBB31_1
+; RV64IA-TSO-NEXT:  .LBB31_3:
+; RV64IA-TSO-NEXT:    ret
   %res = cmpxchg ptr %ptr, i64 %cmp, i64 %val acquire monotonic
   ret void
 }
 
 define void @cmpxchg_i64_acquire_acquire(ptr %ptr, i64 %cmp, i64 %val) nounwind {
 ; RV32I-LABEL: cmpxchg_i64_acquire_acquire:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16

 ; RV64I-NEXT:    mv a1, sp
 ; RV64I-NEXT:    li a3, 2
 ; RV64I-NEXT:    li a4, 2
 ; RV64I-NEXT:    call __atomic_compare_exchange_8@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: cmpxchg_i64_acquire_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:  .LBB32_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.d.aq a3, (a0)
-; RV64IA-NEXT:    bne a3, a1, .LBB32_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB32_1 Depth=1
-; RV64IA-NEXT:    sc.d a4, a2, (a0)
-; RV64IA-NEXT:    bnez a4, .LBB32_1
-; RV64IA-NEXT:  .LBB32_3:
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: cmpxchg_i64_acquire_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:  .LBB32_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.d.aq a3, (a0)
+; RV64IA-WMO-NEXT:    bne a3, a1, .LBB32_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB32_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.d a4, a2, (a0)
+; RV64IA-WMO-NEXT:    bnez a4, .LBB32_1
+; RV64IA-WMO-NEXT:  .LBB32_3:
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: cmpxchg_i64_acquire_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:  .LBB32_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.d a3, (a0)
+; RV64IA-TSO-NEXT:    bne a3, a1, .LBB32_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB32_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.d a4, a2, (a0)
+; RV64IA-TSO-NEXT:    bnez a4, .LBB32_1
+; RV64IA-TSO-NEXT:  .LBB32_3:
+; RV64IA-TSO-NEXT:    ret
   %res = cmpxchg ptr %ptr, i64 %cmp, i64 %val acquire acquire
   ret void
 }
 
 define void @cmpxchg_i64_release_monotonic(ptr %ptr, i64 %cmp, i64 %val) nounwind {
 ; RV32I-LABEL: cmpxchg_i64_release_monotonic:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16

 ; RV64I-NEXT:    mv a1, sp
 ; RV64I-NEXT:    li a3, 3
 ; RV64I-NEXT:    li a4, 0
 ; RV64I-NEXT:    call __atomic_compare_exchange_8@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: cmpxchg_i64_release_monotonic:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:  .LBB33_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.d a3, (a0)
-; RV64IA-NEXT:    bne a3, a1, .LBB33_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB33_1 Depth=1
-; RV64IA-NEXT:    sc.d.rl a4, a2, (a0)
-; RV64IA-NEXT:    bnez a4, .LBB33_1
-; RV64IA-NEXT:  .LBB33_3:
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: cmpxchg_i64_release_monotonic:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:  .LBB33_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.d a3, (a0)
+; RV64IA-WMO-NEXT:    bne a3, a1, .LBB33_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB33_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.d.rl a4, a2, (a0)
+; RV64IA-WMO-NEXT:    bnez a4, .LBB33_1
+; RV64IA-WMO-NEXT:  .LBB33_3:
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: cmpxchg_i64_release_monotonic:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:  .LBB33_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.d a3, (a0)
+; RV64IA-TSO-NEXT:    bne a3, a1, .LBB33_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB33_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.d a4, a2, (a0)
+; RV64IA-TSO-NEXT:    bnez a4, .LBB33_1
+; RV64IA-TSO-NEXT:  .LBB33_3:
+; RV64IA-TSO-NEXT:    ret
   %res = cmpxchg ptr %ptr, i64 %cmp, i64 %val release monotonic
   ret void
 }
 
 define void @cmpxchg_i64_release_acquire(ptr %ptr, i64 %cmp, i64 %val) nounwind {
 ; RV32I-LABEL: cmpxchg_i64_release_acquire:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16

 ; RV64I-NEXT:    mv a1, sp
 ; RV64I-NEXT:    li a3, 3
 ; RV64I-NEXT:    li a4, 2
 ; RV64I-NEXT:    call __atomic_compare_exchange_8@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: cmpxchg_i64_release_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:  .LBB34_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.d.aq a3, (a0)
-; RV64IA-NEXT:    bne a3, a1, .LBB34_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB34_1 Depth=1
-; RV64IA-NEXT:    sc.d.rl a4, a2, (a0)
-; RV64IA-NEXT:    bnez a4, .LBB34_1
-; RV64IA-NEXT:  .LBB34_3:
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: cmpxchg_i64_release_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:  .LBB34_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.d.aq a3, (a0)
+; RV64IA-WMO-NEXT:    bne a3, a1, .LBB34_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB34_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.d.rl a4, a2, (a0)
+; RV64IA-WMO-NEXT:    bnez a4, .LBB34_1
+; RV64IA-WMO-NEXT:  .LBB34_3:
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: cmpxchg_i64_release_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:  .LBB34_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.d a3, (a0)
+; RV64IA-TSO-NEXT:    bne a3, a1, .LBB34_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB34_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.d a4, a2, (a0)
+; RV64IA-TSO-NEXT:    bnez a4, .LBB34_1
+; RV64IA-TSO-NEXT:  .LBB34_3:
+; RV64IA-TSO-NEXT:    ret
   %res = cmpxchg ptr %ptr, i64 %cmp, i64 %val release acquire
   ret void
 }
 
 define void @cmpxchg_i64_acq_rel_monotonic(ptr %ptr, i64 %cmp, i64 %val) nounwind {
 ; RV32I-LABEL: cmpxchg_i64_acq_rel_monotonic:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16

 ; RV64I-NEXT:    mv a1, sp
 ; RV64I-NEXT:    li a3, 4
 ; RV64I-NEXT:    li a4, 0
 ; RV64I-NEXT:    call __atomic_compare_exchange_8@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: cmpxchg_i64_acq_rel_monotonic:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:  .LBB35_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.d.aq a3, (a0)
-; RV64IA-NEXT:    bne a3, a1, .LBB35_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB35_1 Depth=1
-; RV64IA-NEXT:    sc.d.rl a4, a2, (a0)
-; RV64IA-NEXT:    bnez a4, .LBB35_1
-; RV64IA-NEXT:  .LBB35_3:
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: cmpxchg_i64_acq_rel_monotonic:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:  .LBB35_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.d.aq a3, (a0)
+; RV64IA-WMO-NEXT:    bne a3, a1, .LBB35_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB35_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.d.rl a4, a2, (a0)
+; RV64IA-WMO-NEXT:    bnez a4, .LBB35_1
+; RV64IA-WMO-NEXT:  .LBB35_3:
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: cmpxchg_i64_acq_rel_monotonic:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:  .LBB35_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.d a3, (a0)
+; RV64IA-TSO-NEXT:    bne a3, a1, .LBB35_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB35_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.d a4, a2, (a0)
+; RV64IA-TSO-NEXT:    bnez a4, .LBB35_1
+; RV64IA-TSO-NEXT:  .LBB35_3:
+; RV64IA-TSO-NEXT:    ret
   %res = cmpxchg ptr %ptr, i64 %cmp, i64 %val acq_rel monotonic
   ret void
 }
 
 define void @cmpxchg_i64_acq_rel_acquire(ptr %ptr, i64 %cmp, i64 %val) nounwind {
 ; RV32I-LABEL: cmpxchg_i64_acq_rel_acquire:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16

 ; RV64I-NEXT:    mv a1, sp
 ; RV64I-NEXT:    li a3, 4
 ; RV64I-NEXT:    li a4, 2
 ; RV64I-NEXT:    call __atomic_compare_exchange_8@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: cmpxchg_i64_acq_rel_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:  .LBB36_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.d.aq a3, (a0)
-; RV64IA-NEXT:    bne a3, a1, .LBB36_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB36_1 Depth=1
-; RV64IA-NEXT:    sc.d.rl a4, a2, (a0)
-; RV64IA-NEXT:    bnez a4, .LBB36_1
-; RV64IA-NEXT:  .LBB36_3:
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: cmpxchg_i64_acq_rel_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:  .LBB36_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.d.aq a3, (a0)
+; RV64IA-WMO-NEXT:    bne a3, a1, .LBB36_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB36_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.d.rl a4, a2, (a0)
+; RV64IA-WMO-NEXT:    bnez a4, .LBB36_1
+; RV64IA-WMO-NEXT:  .LBB36_3:
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: cmpxchg_i64_acq_rel_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:  .LBB36_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.d a3, (a0)
+; RV64IA-TSO-NEXT:    bne a3, a1, .LBB36_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB36_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.d a4, a2, (a0)
+; RV64IA-TSO-NEXT:    bnez a4, .LBB36_1
+; RV64IA-TSO-NEXT:  .LBB36_3:
+; RV64IA-TSO-NEXT:    ret
   %res = cmpxchg ptr %ptr, i64 %cmp, i64 %val acq_rel acquire
   ret void
 }
 
 define void @cmpxchg_i64_seq_cst_monotonic(ptr %ptr, i64 %cmp, i64 %val) nounwind {
 ; RV32I-LABEL: cmpxchg_i64_seq_cst_monotonic:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16

llvm/test/CodeGen/RISCV/atomic-load-store.ll

 ; RV32IA-WMO:       # %bb.0:
 ; RV32IA-WMO-NEXT:    fence rw, w
 ; RV32IA-WMO-NEXT:    sb a1, 0(a0)
 ; RV32IA-WMO-NEXT:    ret
 ;
 ; RV32IA-TSO-LABEL: atomic_store_i8_seq_cst:
 ; RV32IA-TSO:       # %bb.0:
 ; RV32IA-TSO-NEXT:    sb a1, 0(a0)
+; RV32IA-TSO-NEXT:    fence rw, rw
 ; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomic_store_i8_seq_cst:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 5
 ; RV64I-NEXT:    call __atomic_store_1@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
 ; RV64IA-WMO-LABEL: atomic_store_i8_seq_cst:
 ; RV64IA-WMO:       # %bb.0:
 ; RV64IA-WMO-NEXT:    fence rw, w
 ; RV64IA-WMO-NEXT:    sb a1, 0(a0)
 ; RV64IA-WMO-NEXT:    ret
 ;
 ; RV64IA-TSO-LABEL: atomic_store_i8_seq_cst:
 ; RV64IA-TSO:       # %bb.0:
 ; RV64IA-TSO-NEXT:    sb a1, 0(a0)
+; RV64IA-TSO-NEXT:    fence rw, rw
 ; RV64IA-TSO-NEXT:    ret
 ;
 ; RV32IA-WMO-TRAILING-FENCE-LABEL: atomic_store_i8_seq_cst:
 ; RV32IA-WMO-TRAILING-FENCE:       # %bb.0:
 ; RV32IA-WMO-TRAILING-FENCE-NEXT:    fence rw, w
 ; RV32IA-WMO-TRAILING-FENCE-NEXT:    sb a1, 0(a0)
 ; RV32IA-WMO-TRAILING-FENCE-NEXT:    fence rw, rw
 ; RV32IA-WMO-TRAILING-FENCE-NEXT:    ret
 ;
 ; RV32IA-TSO-TRAILING-FENCE-LABEL: atomic_store_i8_seq_cst:
 ; RV32IA-TSO-TRAILING-FENCE:       # %bb.0:
 ; RV32IA-TSO-TRAILING-FENCE-NEXT:    sb a1, 0(a0)
+; RV32IA-TSO-TRAILING-FENCE-NEXT:    fence rw, rw
 ; RV32IA-TSO-TRAILING-FENCE-NEXT:    ret
 ;
 ; RV64IA-WMO-TRAILING-FENCE-LABEL: atomic_store_i8_seq_cst:
 ; RV64IA-WMO-TRAILING-FENCE:       # %bb.0:
 ; RV64IA-WMO-TRAILING-FENCE-NEXT:    fence rw, w
 ; RV64IA-WMO-TRAILING-FENCE-NEXT:    sb a1, 0(a0)
 ; RV64IA-WMO-TRAILING-FENCE-NEXT:    fence rw, rw
 ; RV64IA-WMO-TRAILING-FENCE-NEXT:    ret
 ;
 ; RV64IA-TSO-TRAILING-FENCE-LABEL: atomic_store_i8_seq_cst:
 ; RV64IA-TSO-TRAILING-FENCE:       # %bb.0:
 ; RV64IA-TSO-TRAILING-FENCE-NEXT:    sb a1, 0(a0)
+; RV64IA-TSO-TRAILING-FENCE-NEXT:    fence rw, rw
 ; RV64IA-TSO-TRAILING-FENCE-NEXT:    ret
   store atomic i8 %b, ptr %a seq_cst, align 1
   ret void
 }
 
 define void @atomic_store_i16_unordered(ptr %a, i16 %b) nounwind {
 ; RV32I-LABEL: atomic_store_i16_unordered:
 ; RV32I:       # %bb.0:

 ; RV32IA-WMO:       # %bb.0:
 ; RV32IA-WMO-NEXT:    fence rw, w
 ; RV32IA-WMO-NEXT:    sh a1, 0(a0)
 ; RV32IA-WMO-NEXT:    ret
 ;
 ; RV32IA-TSO-LABEL: atomic_store_i16_seq_cst:
 ; RV32IA-TSO:       # %bb.0:
 ; RV32IA-TSO-NEXT:    sh a1, 0(a0)
+; RV32IA-TSO-NEXT:    fence rw, rw
 ; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomic_store_i16_seq_cst:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 5
 ; RV64I-NEXT:    call __atomic_store_2@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
 ; RV64IA-WMO-LABEL: atomic_store_i16_seq_cst:
 ; RV64IA-WMO:       # %bb.0:
 ; RV64IA-WMO-NEXT:    fence rw, w
 ; RV64IA-WMO-NEXT:    sh a1, 0(a0)
 ; RV64IA-WMO-NEXT:    ret
 ;
 ; RV64IA-TSO-LABEL: atomic_store_i16_seq_cst:
 ; RV64IA-TSO:       # %bb.0:
 ; RV64IA-TSO-NEXT:    sh a1, 0(a0)
+; RV64IA-TSO-NEXT:    fence rw, rw
 ; RV64IA-TSO-NEXT:    ret
 ;
 ; RV32IA-WMO-TRAILING-FENCE-LABEL: atomic_store_i16_seq_cst:
 ; RV32IA-WMO-TRAILING-FENCE:       # %bb.0:
 ; RV32IA-WMO-TRAILING-FENCE-NEXT:    fence rw, w
 ; RV32IA-WMO-TRAILING-FENCE-NEXT:    sh a1, 0(a0)
 ; RV32IA-WMO-TRAILING-FENCE-NEXT:    fence rw, rw
 ; RV32IA-WMO-TRAILING-FENCE-NEXT:    ret
 ;
 ; RV32IA-TSO-TRAILING-FENCE-LABEL: atomic_store_i16_seq_cst:
 ; RV32IA-TSO-TRAILING-FENCE:       # %bb.0:
 ; RV32IA-TSO-TRAILING-FENCE-NEXT:    sh a1, 0(a0)
+; RV32IA-TSO-TRAILING-FENCE-NEXT:    fence rw, rw
 ; RV32IA-TSO-TRAILING-FENCE-NEXT:    ret
 ;
 ; RV64IA-WMO-TRAILING-FENCE-LABEL: atomic_store_i16_seq_cst:
 ; RV64IA-WMO-TRAILING-FENCE:       # %bb.0:
 ; RV64IA-WMO-TRAILING-FENCE-NEXT:    fence rw, w
 ; RV64IA-WMO-TRAILING-FENCE-NEXT:    sh a1, 0(a0)
 ; RV64IA-WMO-TRAILING-FENCE-NEXT:    fence rw, rw
 ; RV64IA-WMO-TRAILING-FENCE-NEXT:    ret
 ;
 ; RV64IA-TSO-TRAILING-FENCE-LABEL: atomic_store_i16_seq_cst:
 ; RV64IA-TSO-TRAILING-FENCE:       # %bb.0:
 ; RV64IA-TSO-TRAILING-FENCE-NEXT:    sh a1, 0(a0)
+; RV64IA-TSO-TRAILING-FENCE-NEXT:    fence rw, rw
 ; RV64IA-TSO-TRAILING-FENCE-NEXT:    ret
   store atomic i16 %b, ptr %a seq_cst, align 2
   ret void
 }
 
 define void @atomic_store_i32_unordered(ptr %a, i32 %b) nounwind {
 ; RV32I-LABEL: atomic_store_i32_unordered:
 ; RV32I:       # %bb.0:

 ; RV32IA-WMO:       # %bb.0:
 ; RV32IA-WMO-NEXT:    fence rw, w
 ; RV32IA-WMO-NEXT:    sw a1, 0(a0)
 ; RV32IA-WMO-NEXT:    ret
 ;
 ; RV32IA-TSO-LABEL: atomic_store_i32_seq_cst:
 ; RV32IA-TSO:       # %bb.0:
 ; RV32IA-TSO-NEXT:    sw a1, 0(a0)
+; RV32IA-TSO-NEXT:    fence rw, rw
 ; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomic_store_i32_seq_cst:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 5
 ; RV64I-NEXT:    call __atomic_store_4@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
 ; RV64IA-WMO-LABEL: atomic_store_i32_seq_cst:
 ; RV64IA-WMO:       # %bb.0:
 ; RV64IA-WMO-NEXT:    fence rw, w
 ; RV64IA-WMO-NEXT:    sw a1, 0(a0)
 ; RV64IA-WMO-NEXT:    ret
 ;
 ; RV64IA-TSO-LABEL: atomic_store_i32_seq_cst:
 ; RV64IA-TSO:       # %bb.0:
 ; RV64IA-TSO-NEXT:    sw a1, 0(a0)
+; RV64IA-TSO-NEXT:    fence rw, rw
 ; RV64IA-TSO-NEXT:    ret
 ;
 ; RV32IA-WMO-TRAILING-FENCE-LABEL: atomic_store_i32_seq_cst:
 ; RV32IA-WMO-TRAILING-FENCE:       # %bb.0:
 ; RV32IA-WMO-TRAILING-FENCE-NEXT:    fence rw, w
 ; RV32IA-WMO-TRAILING-FENCE-NEXT:    sw a1, 0(a0)
 ; RV32IA-WMO-TRAILING-FENCE-NEXT:    fence rw, rw
 ; RV32IA-WMO-TRAILING-FENCE-NEXT:    ret
 ;
 ; RV32IA-TSO-TRAILING-FENCE-LABEL: atomic_store_i32_seq_cst:
 ; RV32IA-TSO-TRAILING-FENCE:       # %bb.0:
 ; RV32IA-TSO-TRAILING-FENCE-NEXT:    sw a1, 0(a0)
+; RV32IA-TSO-TRAILING-FENCE-NEXT:    fence rw, rw
 ; RV32IA-TSO-TRAILING-FENCE-NEXT:    ret
 ;
 ; RV64IA-WMO-TRAILING-FENCE-LABEL: atomic_store_i32_seq_cst:
 ; RV64IA-WMO-TRAILING-FENCE:       # %bb.0:
 ; RV64IA-WMO-TRAILING-FENCE-NEXT:    fence rw, w
 ; RV64IA-WMO-TRAILING-FENCE-NEXT:    sw a1, 0(a0)
 ; RV64IA-WMO-TRAILING-FENCE-NEXT:    fence rw, rw
 ; RV64IA-WMO-TRAILING-FENCE-NEXT:    ret
 ;
 ; RV64IA-TSO-TRAILING-FENCE-LABEL: atomic_store_i32_seq_cst:
 ; RV64IA-TSO-TRAILING-FENCE:       # %bb.0:
 ; RV64IA-TSO-TRAILING-FENCE-NEXT:    sw a1, 0(a0)
+; RV64IA-TSO-TRAILING-FENCE-NEXT:    fence rw, rw
 ; RV64IA-TSO-TRAILING-FENCE-NEXT:    ret
   store atomic i32 %b, ptr %a seq_cst, align 4
   ret void
 }
 
 define void @atomic_store_i64_unordered(ptr %a, i64 %b) nounwind {
 ; RV32I-LABEL: atomic_store_i64_unordered:
 ; RV32I:       # %bb.0:

 ; RV64IA-WMO:       # %bb.0:
 ; RV64IA-WMO-NEXT:    fence rw, w
 ; RV64IA-WMO-NEXT:    sd a1, 0(a0)
 ; RV64IA-WMO-NEXT:    ret
 ;
 ; RV64IA-TSO-LABEL: atomic_store_i64_seq_cst:
 ; RV64IA-TSO:       # %bb.0:
 ; RV64IA-TSO-NEXT:    sd a1, 0(a0)
+; RV64IA-TSO-NEXT:    fence rw, rw
 ; RV64IA-TSO-NEXT:    ret
 ;
 ; RV64IA-WMO-TRAILING-FENCE-LABEL: atomic_store_i64_seq_cst:
 ; RV64IA-WMO-TRAILING-FENCE:       # %bb.0:
 ; RV64IA-WMO-TRAILING-FENCE-NEXT:    fence rw, w
 ; RV64IA-WMO-TRAILING-FENCE-NEXT:    sd a1, 0(a0)
 ; RV64IA-WMO-TRAILING-FENCE-NEXT:    fence rw, rw
 ; RV64IA-WMO-TRAILING-FENCE-NEXT:    ret
 ;
 ; RV64IA-TSO-TRAILING-FENCE-LABEL: atomic_store_i64_seq_cst:
 ; RV64IA-TSO-TRAILING-FENCE:       # %bb.0:
 ; RV64IA-TSO-TRAILING-FENCE-NEXT:    sd a1, 0(a0)
+; RV64IA-TSO-TRAILING-FENCE-NEXT:    fence rw, rw
 ; RV64IA-TSO-TRAILING-FENCE-NEXT:    ret
   store atomic i64 %b, ptr %a seq_cst, align 8
   ret void
 }

llvm/test/CodeGen/RISCV/atomic-rmw.ll

 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc -mtriple=riscv32 -verify-machineinstrs < %s \
 ; RUN:   | FileCheck -check-prefix=RV32I %s
 ; RUN: llc -mtriple=riscv32 -mattr=+a -verify-machineinstrs < %s \
-; RUN:   | FileCheck -check-prefix=RV32IA %s
+; RUN:   | FileCheck -check-prefixes=RV32IA,RV32IA-WMO %s
 ; RUN: llc -mtriple=riscv32 -mattr=+a,+experimental-ztso -verify-machineinstrs < %s \
-; RUN:   | FileCheck -check-prefix=RV32IA %s
+; RUN:   | FileCheck -check-prefixes=RV32IA,RV32IA-TSO %s
 ; RUN: llc -mtriple=riscv64 -verify-machineinstrs < %s \
 ; RUN:   | FileCheck -check-prefix=RV64I %s
 ; RUN: llc -mtriple=riscv64 -mattr=+a -verify-machineinstrs < %s \
-; RUN:   | FileCheck -check-prefix=RV64IA %s
+; RUN:   | FileCheck -check-prefixes=RV64IA,RV64IA-WMO %s
 ; RUN: llc -mtriple=riscv64 -mattr=+a,+experimental-ztso -verify-machineinstrs < %s \
-; RUN:   | FileCheck -check-prefix=RV64IA %s
+; RUN:   | FileCheck -check-prefixes=RV64IA,RV64IA-TSO %s
 
 define i8 @atomicrmw_xchg_i8_monotonic(ptr %a, i8 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_xchg_i8_monotonic:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    li a2, 0
 ; RV32I-NEXT:    call __atomic_exchange_1@plt

 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    li a2, 2
 ; RV32I-NEXT:    call __atomic_exchange_1@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_xchg_i8_acquire:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    li a3, 255
-; RV32IA-NEXT:    sll a3, a3, a0
-; RV32IA-NEXT:    andi a1, a1, 255
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB1_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a4, (a2)
-; RV32IA-NEXT:    mv a5, a1
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:    and a5, a5, a3
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:    sc.w a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB1_1
-; RV32IA-NEXT:  # %bb.2:
-; RV32IA-NEXT:    srl a0, a4, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_xchg_i8_acquire:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    li a3, 255
+; RV32IA-WMO-NEXT:    sll a3, a3, a0
+; RV32IA-WMO-NEXT:    andi a1, a1, 255
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB1_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a4, (a2)
+; RV32IA-WMO-NEXT:    mv a5, a1
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:    and a5, a5, a3
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB1_1
+; RV32IA-WMO-NEXT:  # %bb.2:
+; RV32IA-WMO-NEXT:    srl a0, a4, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_xchg_i8_acquire:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    li a3, 255
+; RV32IA-TSO-NEXT:    sll a3, a3, a0
+; RV32IA-TSO-NEXT:    andi a1, a1, 255
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB1_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a4, (a2)
+; RV32IA-TSO-NEXT:    mv a5, a1
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:    and a5, a5, a3
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB1_1
+; RV32IA-TSO-NEXT:  # %bb.2:
+; RV32IA-TSO-NEXT:    srl a0, a4, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_xchg_i8_acquire:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 2
 ; RV64I-NEXT:    call __atomic_exchange_1@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_xchg_i8_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    li a3, 255
-; RV64IA-NEXT:    sllw a3, a3, a0
-; RV64IA-NEXT:    andi a1, a1, 255
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB1_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a4, (a2)
-; RV64IA-NEXT:    mv a5, a1
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:    and a5, a5, a3
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:    sc.w a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB1_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    srlw a0, a4, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_xchg_i8_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    li a3, 255
+; RV64IA-WMO-NEXT:    sllw a3, a3, a0
+; RV64IA-WMO-NEXT:    andi a1, a1, 255
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB1_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a4, (a2)
+; RV64IA-WMO-NEXT:    mv a5, a1
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:    and a5, a5, a3
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB1_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    srlw a0, a4, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_xchg_i8_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    li a3, 255
+; RV64IA-TSO-NEXT:    sllw a3, a3, a0
+; RV64IA-TSO-NEXT:    andi a1, a1, 255
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB1_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a4, (a2)
+; RV64IA-TSO-NEXT:    mv a5, a1
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:    and a5, a5, a3
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB1_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    srlw a0, a4, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw xchg ptr %a, i8 %b acquire
   ret i8 %1
 }
 
 define i8 @atomicrmw_xchg_i8_release(ptr %a, i8 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_xchg_i8_release:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    li a2, 3
 ; RV32I-NEXT:    call __atomic_exchange_1@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_xchg_i8_release:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    li a3, 255
-; RV32IA-NEXT:    sll a3, a3, a0
-; RV32IA-NEXT:    andi a1, a1, 255
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB2_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w a4, (a2)
-; RV32IA-NEXT:    mv a5, a1
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:    and a5, a5, a3
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB2_1
-; RV32IA-NEXT:  # %bb.2:
-; RV32IA-NEXT:    srl a0, a4, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_xchg_i8_release:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    li a3, 255
+; RV32IA-WMO-NEXT:    sll a3, a3, a0
+; RV32IA-WMO-NEXT:    andi a1, a1, 255
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB2_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w a4, (a2)
+; RV32IA-WMO-NEXT:    mv a5, a1
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:    and a5, a5, a3
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB2_1
+; RV32IA-WMO-NEXT:  # %bb.2:
+; RV32IA-WMO-NEXT:    srl a0, a4, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_xchg_i8_release:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    li a3, 255
+; RV32IA-TSO-NEXT:    sll a3, a3, a0
+; RV32IA-TSO-NEXT:    andi a1, a1, 255
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB2_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a4, (a2)
+; RV32IA-TSO-NEXT:    mv a5, a1
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:    and a5, a5, a3
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB2_1
+; RV32IA-TSO-NEXT:  # %bb.2:
+; RV32IA-TSO-NEXT:    srl a0, a4, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_xchg_i8_release:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 3
 ; RV64I-NEXT:    call __atomic_exchange_1@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_xchg_i8_release:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    li a3, 255
-; RV64IA-NEXT:    sllw a3, a3, a0
-; RV64IA-NEXT:    andi a1, a1, 255
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB2_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w a4, (a2)
-; RV64IA-NEXT:    mv a5, a1
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:    and a5, a5, a3
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB2_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    srlw a0, a4, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_xchg_i8_release:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    li a3, 255
+; RV64IA-WMO-NEXT:    sllw a3, a3, a0
+; RV64IA-WMO-NEXT:    andi a1, a1, 255
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB2_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w a4, (a2)
+; RV64IA-WMO-NEXT:    mv a5, a1
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:    and a5, a5, a3
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB2_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    srlw a0, a4, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_xchg_i8_release:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    li a3, 255
+; RV64IA-TSO-NEXT:    sllw a3, a3, a0
+; RV64IA-TSO-NEXT:    andi a1, a1, 255
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB2_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a4, (a2)
+; RV64IA-TSO-NEXT:    mv a5, a1
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:    and a5, a5, a3
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB2_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    srlw a0, a4, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw xchg ptr %a, i8 %b release
   ret i8 %1
 }
 
 define i8 @atomicrmw_xchg_i8_acq_rel(ptr %a, i8 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_xchg_i8_acq_rel:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    li a2, 4
 ; RV32I-NEXT:    call __atomic_exchange_1@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_xchg_i8_acq_rel:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    li a3, 255
-; RV32IA-NEXT:    sll a3, a3, a0
-; RV32IA-NEXT:    andi a1, a1, 255
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB3_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a4, (a2)
-; RV32IA-NEXT:    mv a5, a1
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:    and a5, a5, a3
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB3_1
-; RV32IA-NEXT:  # %bb.2:
-; RV32IA-NEXT:    srl a0, a4, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_xchg_i8_acq_rel:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    li a3, 255
+; RV32IA-WMO-NEXT:    sll a3, a3, a0
+; RV32IA-WMO-NEXT:    andi a1, a1, 255
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB3_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a4, (a2)
+; RV32IA-WMO-NEXT:    mv a5, a1
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:    and a5, a5, a3
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB3_1
+; RV32IA-WMO-NEXT:  # %bb.2:
+; RV32IA-WMO-NEXT:    srl a0, a4, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_xchg_i8_acq_rel:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    li a3, 255
+; RV32IA-TSO-NEXT:    sll a3, a3, a0
+; RV32IA-TSO-NEXT:    andi a1, a1, 255
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB3_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a4, (a2)
+; RV32IA-TSO-NEXT:    mv a5, a1
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:    and a5, a5, a3
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB3_1
+; RV32IA-TSO-NEXT:  # %bb.2:
+; RV32IA-TSO-NEXT:    srl a0, a4, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_xchg_i8_acq_rel:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 4
 ; RV64I-NEXT:    call __atomic_exchange_1@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_xchg_i8_acq_rel:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    li a3, 255
-; RV64IA-NEXT:    sllw a3, a3, a0
-; RV64IA-NEXT:    andi a1, a1, 255
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB3_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a4, (a2)
-; RV64IA-NEXT:    mv a5, a1
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:    and a5, a5, a3
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB3_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    srlw a0, a4, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_xchg_i8_acq_rel:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    li a3, 255
+; RV64IA-WMO-NEXT:    sllw a3, a3, a0
+; RV64IA-WMO-NEXT:    andi a1, a1, 255
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB3_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a4, (a2)
+; RV64IA-WMO-NEXT:    mv a5, a1
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:    and a5, a5, a3
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB3_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    srlw a0, a4, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_xchg_i8_acq_rel:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    li a3, 255
+; RV64IA-TSO-NEXT:    sllw a3, a3, a0
+; RV64IA-TSO-NEXT:    andi a1, a1, 255
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB3_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a4, (a2)
+; RV64IA-TSO-NEXT:    mv a5, a1
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:    and a5, a5, a3
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB3_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    srlw a0, a4, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw xchg ptr %a, i8 %b acq_rel
   ret i8 %1
 }
 
 define i8 @atomicrmw_xchg_i8_seq_cst(ptr %a, i8 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_xchg_i8_seq_cst:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16

 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    li a2, 2
 ; RV32I-NEXT:    call __atomic_fetch_add_1@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_add_i8_acquire:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    li a3, 255
-; RV32IA-NEXT:    sll a3, a3, a0
-; RV32IA-NEXT:    andi a1, a1, 255
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB6_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a4, (a2)
-; RV32IA-NEXT:    add a5, a4, a1
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:    and a5, a5, a3
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:    sc.w a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB6_1
-; RV32IA-NEXT:  # %bb.2:
-; RV32IA-NEXT:    srl a0, a4, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_add_i8_acquire:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    li a3, 255
+; RV32IA-WMO-NEXT:    sll a3, a3, a0
+; RV32IA-WMO-NEXT:    andi a1, a1, 255
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB6_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a4, (a2)
+; RV32IA-WMO-NEXT:    add a5, a4, a1
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:    and a5, a5, a3
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB6_1
+; RV32IA-WMO-NEXT:  # %bb.2:
+; RV32IA-WMO-NEXT:    srl a0, a4, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_add_i8_acquire:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    li a3, 255
+; RV32IA-TSO-NEXT:    sll a3, a3, a0
+; RV32IA-TSO-NEXT:    andi a1, a1, 255
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB6_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a4, (a2)
+; RV32IA-TSO-NEXT:    add a5, a4, a1
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:    and a5, a5, a3
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB6_1
+; RV32IA-TSO-NEXT:  # %bb.2:
+; RV32IA-TSO-NEXT:    srl a0, a4, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_add_i8_acquire:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 2
 ; RV64I-NEXT:    call __atomic_fetch_add_1@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_add_i8_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    li a3, 255
-; RV64IA-NEXT:    sllw a3, a3, a0
-; RV64IA-NEXT:    andi a1, a1, 255
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB6_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a4, (a2)
-; RV64IA-NEXT:    add a5, a4, a1
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:    and a5, a5, a3
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:    sc.w a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB6_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    srlw a0, a4, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_add_i8_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    li a3, 255
+; RV64IA-WMO-NEXT:    sllw a3, a3, a0
+; RV64IA-WMO-NEXT:    andi a1, a1, 255
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB6_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a4, (a2)
+; RV64IA-WMO-NEXT:    add a5, a4, a1
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:    and a5, a5, a3
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB6_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    srlw a0, a4, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_add_i8_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    li a3, 255
+; RV64IA-TSO-NEXT:    sllw a3, a3, a0
+; RV64IA-TSO-NEXT:    andi a1, a1, 255
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB6_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a4, (a2)
+; RV64IA-TSO-NEXT:    add a5, a4, a1
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:    and a5, a5, a3
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB6_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    srlw a0, a4, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw add ptr %a, i8 %b acquire
   ret i8 %1
 }
 
 define i8 @atomicrmw_add_i8_release(ptr %a, i8 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_add_i8_release:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    li a2, 3
 ; RV32I-NEXT:    call __atomic_fetch_add_1@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_add_i8_release:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    li a3, 255
-; RV32IA-NEXT:    sll a3, a3, a0
-; RV32IA-NEXT:    andi a1, a1, 255
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB7_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w a4, (a2)
-; RV32IA-NEXT:    add a5, a4, a1
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:    and a5, a5, a3
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB7_1
-; RV32IA-NEXT:  # %bb.2:
-; RV32IA-NEXT:    srl a0, a4, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_add_i8_release:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    li a3, 255
+; RV32IA-WMO-NEXT:    sll a3, a3, a0
+; RV32IA-WMO-NEXT:    andi a1, a1, 255
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB7_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w a4, (a2)
+; RV32IA-WMO-NEXT:    add a5, a4, a1
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:    and a5, a5, a3
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB7_1
+; RV32IA-WMO-NEXT:  # %bb.2:
+; RV32IA-WMO-NEXT:    srl a0, a4, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_add_i8_release:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    li a3, 255
+; RV32IA-TSO-NEXT:    sll a3, a3, a0
+; RV32IA-TSO-NEXT:    andi a1, a1, 255
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB7_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a4, (a2)
+; RV32IA-TSO-NEXT:    add a5, a4, a1
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:    and a5, a5, a3
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB7_1
+; RV32IA-TSO-NEXT:  # %bb.2:
+; RV32IA-TSO-NEXT:    srl a0, a4, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_add_i8_release:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 3
 ; RV64I-NEXT:    call __atomic_fetch_add_1@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_add_i8_release:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    li a3, 255
-; RV64IA-NEXT:    sllw a3, a3, a0
-; RV64IA-NEXT:    andi a1, a1, 255
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB7_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w a4, (a2)
-; RV64IA-NEXT:    add a5, a4, a1
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:    and a5, a5, a3
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB7_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    srlw a0, a4, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_add_i8_release:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    li a3, 255
+; RV64IA-WMO-NEXT:    sllw a3, a3, a0
+; RV64IA-WMO-NEXT:    andi a1, a1, 255
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB7_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w a4, (a2)
+; RV64IA-WMO-NEXT:    add a5, a4, a1
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:    and a5, a5, a3
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB7_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    srlw a0, a4, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_add_i8_release:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    li a3, 255
+; RV64IA-TSO-NEXT:    sllw a3, a3, a0
+; RV64IA-TSO-NEXT:    andi a1, a1, 255
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB7_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a4, (a2)
+; RV64IA-TSO-NEXT:    add a5, a4, a1
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:    and a5, a5, a3
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB7_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    srlw a0, a4, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw add ptr %a, i8 %b release
   ret i8 %1
 }
 
 define i8 @atomicrmw_add_i8_acq_rel(ptr %a, i8 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_add_i8_acq_rel:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    li a2, 4
 ; RV32I-NEXT:    call __atomic_fetch_add_1@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_add_i8_acq_rel:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    li a3, 255
-; RV32IA-NEXT:    sll a3, a3, a0
-; RV32IA-NEXT:    andi a1, a1, 255
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB8_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a4, (a2)
-; RV32IA-NEXT:    add a5, a4, a1
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:    and a5, a5, a3
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB8_1
-; RV32IA-NEXT:  # %bb.2:
-; RV32IA-NEXT:    srl a0, a4, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_add_i8_acq_rel:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    li a3, 255
+; RV32IA-WMO-NEXT:    sll a3, a3, a0
+; RV32IA-WMO-NEXT:    andi a1, a1, 255
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB8_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a4, (a2)
+; RV32IA-WMO-NEXT:    add a5, a4, a1
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:    and a5, a5, a3
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB8_1
+; RV32IA-WMO-NEXT:  # %bb.2:
+; RV32IA-WMO-NEXT:    srl a0, a4, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_add_i8_acq_rel:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    li a3, 255
+; RV32IA-TSO-NEXT:    sll a3, a3, a0
+; RV32IA-TSO-NEXT:    andi a1, a1, 255
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB8_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a4, (a2)
+; RV32IA-TSO-NEXT:    add a5, a4, a1
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:    and a5, a5, a3
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB8_1
+; RV32IA-TSO-NEXT:  # %bb.2:
+; RV32IA-TSO-NEXT:    srl a0, a4, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_add_i8_acq_rel:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 4
 ; RV64I-NEXT:    call __atomic_fetch_add_1@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_add_i8_acq_rel:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    li a3, 255
-; RV64IA-NEXT:    sllw a3, a3, a0
-; RV64IA-NEXT:    andi a1, a1, 255
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB8_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a4, (a2)
-; RV64IA-NEXT:    add a5, a4, a1
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:    and a5, a5, a3
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB8_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    srlw a0, a4, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_add_i8_acq_rel:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    li a3, 255
+; RV64IA-WMO-NEXT:    sllw a3, a3, a0
+; RV64IA-WMO-NEXT:    andi a1, a1, 255
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB8_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a4, (a2)
+; RV64IA-WMO-NEXT:    add a5, a4, a1
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:    and a5, a5, a3
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB8_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    srlw a0, a4, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_add_i8_acq_rel:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    li a3, 255
+; RV64IA-TSO-NEXT:    sllw a3, a3, a0
+; RV64IA-TSO-NEXT:    andi a1, a1, 255
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB8_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a4, (a2)
+; RV64IA-TSO-NEXT:    add a5, a4, a1
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:    and a5, a5, a3
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB8_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    srlw a0, a4, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw add ptr %a, i8 %b acq_rel
   ret i8 %1
 }
 
 define i8 @atomicrmw_add_i8_seq_cst(ptr %a, i8 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_add_i8_seq_cst:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16

 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    li a2, 2
 ; RV32I-NEXT:    call __atomic_fetch_sub_1@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_sub_i8_acquire:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    li a3, 255
-; RV32IA-NEXT:    sll a3, a3, a0
-; RV32IA-NEXT:    andi a1, a1, 255
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB11_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a4, (a2)
-; RV32IA-NEXT:    sub a5, a4, a1
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:    and a5, a5, a3
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:    sc.w a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB11_1
-; RV32IA-NEXT:  # %bb.2:
-; RV32IA-NEXT:    srl a0, a4, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_sub_i8_acquire:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    li a3, 255
+; RV32IA-WMO-NEXT:    sll a3, a3, a0
+; RV32IA-WMO-NEXT:    andi a1, a1, 255
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB11_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a4, (a2)
+; RV32IA-WMO-NEXT:    sub a5, a4, a1
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:    and a5, a5, a3
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB11_1
+; RV32IA-WMO-NEXT:  # %bb.2:
+; RV32IA-WMO-NEXT:    srl a0, a4, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_sub_i8_acquire:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    li a3, 255
+; RV32IA-TSO-NEXT:    sll a3, a3, a0
+; RV32IA-TSO-NEXT:    andi a1, a1, 255
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB11_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a4, (a2)
+; RV32IA-TSO-NEXT:    sub a5, a4, a1
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:    and a5, a5, a3
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB11_1
+; RV32IA-TSO-NEXT:  # %bb.2:
+; RV32IA-TSO-NEXT:    srl a0, a4, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_sub_i8_acquire:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 2
 ; RV64I-NEXT:    call __atomic_fetch_sub_1@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_sub_i8_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    li a3, 255
-; RV64IA-NEXT:    sllw a3, a3, a0
-; RV64IA-NEXT:    andi a1, a1, 255
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB11_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a4, (a2)
-; RV64IA-NEXT:    sub a5, a4, a1
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:    and a5, a5, a3
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:    sc.w a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB11_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    srlw a0, a4, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_sub_i8_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    li a3, 255
+; RV64IA-WMO-NEXT:    sllw a3, a3, a0
+; RV64IA-WMO-NEXT:    andi a1, a1, 255
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB11_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a4, (a2)
+; RV64IA-WMO-NEXT:    sub a5, a4, a1
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:    and a5, a5, a3
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB11_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    srlw a0, a4, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_sub_i8_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    li a3, 255
+; RV64IA-TSO-NEXT:    sllw a3, a3, a0
+; RV64IA-TSO-NEXT:    andi a1, a1, 255
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB11_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a4, (a2)
+; RV64IA-TSO-NEXT:    sub a5, a4, a1
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:    and a5, a5, a3
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB11_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    srlw a0, a4, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw sub ptr %a, i8 %b acquire
   ret i8 %1
 }
 
 define i8 @atomicrmw_sub_i8_release(ptr %a, i8 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_sub_i8_release:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    li a2, 3
 ; RV32I-NEXT:    call __atomic_fetch_sub_1@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_sub_i8_release:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    li a3, 255
-; RV32IA-NEXT:    sll a3, a3, a0
-; RV32IA-NEXT:    andi a1, a1, 255
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB12_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w a4, (a2)
-; RV32IA-NEXT:    sub a5, a4, a1
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:    and a5, a5, a3
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB12_1
-; RV32IA-NEXT:  # %bb.2:
-; RV32IA-NEXT:    srl a0, a4, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_sub_i8_release:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    li a3, 255
+; RV32IA-WMO-NEXT:    sll a3, a3, a0
+; RV32IA-WMO-NEXT:    andi a1, a1, 255
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB12_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w a4, (a2)
+; RV32IA-WMO-NEXT:    sub a5, a4, a1
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:    and a5, a5, a3
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB12_1
+; RV32IA-WMO-NEXT:  # %bb.2:
+; RV32IA-WMO-NEXT:    srl a0, a4, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_sub_i8_release:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    li a3, 255
+; RV32IA-TSO-NEXT:    sll a3, a3, a0
+; RV32IA-TSO-NEXT:    andi a1, a1, 255
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB12_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a4, (a2)
+; RV32IA-TSO-NEXT:    sub a5, a4, a1
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:    and a5, a5, a3
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB12_1
+; RV32IA-TSO-NEXT:  # %bb.2:
+; RV32IA-TSO-NEXT:    srl a0, a4, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_sub_i8_release:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 3
 ; RV64I-NEXT:    call __atomic_fetch_sub_1@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_sub_i8_release:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    li a3, 255
-; RV64IA-NEXT:    sllw a3, a3, a0
-; RV64IA-NEXT:    andi a1, a1, 255
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB12_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w a4, (a2)
-; RV64IA-NEXT:    sub a5, a4, a1
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:    and a5, a5, a3
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB12_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    srlw a0, a4, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_sub_i8_release:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    li a3, 255
+; RV64IA-WMO-NEXT:    sllw a3, a3, a0
+; RV64IA-WMO-NEXT:    andi a1, a1, 255
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB12_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w a4, (a2)
+; RV64IA-WMO-NEXT:    sub a5, a4, a1
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:    and a5, a5, a3
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB12_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    srlw a0, a4, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_sub_i8_release:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    li a3, 255
+; RV64IA-TSO-NEXT:    sllw a3, a3, a0
+; RV64IA-TSO-NEXT:    andi a1, a1, 255
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB12_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a4, (a2)
+; RV64IA-TSO-NEXT:    sub a5, a4, a1
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:    and a5, a5, a3
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB12_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    srlw a0, a4, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw sub ptr %a, i8 %b release
   ret i8 %1
 }
 
 define i8 @atomicrmw_sub_i8_acq_rel(ptr %a, i8 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_sub_i8_acq_rel:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    li a2, 4
 ; RV32I-NEXT:    call __atomic_fetch_sub_1@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_sub_i8_acq_rel:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    li a3, 255
-; RV32IA-NEXT:    sll a3, a3, a0
-; RV32IA-NEXT:    andi a1, a1, 255
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB13_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a4, (a2)
-; RV32IA-NEXT:    sub a5, a4, a1
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:    and a5, a5, a3
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB13_1
-; RV32IA-NEXT:  # %bb.2:
-; RV32IA-NEXT:    srl a0, a4, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_sub_i8_acq_rel:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    li a3, 255
+; RV32IA-WMO-NEXT:    sll a3, a3, a0
+; RV32IA-WMO-NEXT:    andi a1, a1, 255
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB13_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a4, (a2)
+; RV32IA-WMO-NEXT:    sub a5, a4, a1
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:    and a5, a5, a3
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB13_1
+; RV32IA-WMO-NEXT:  # %bb.2:
+; RV32IA-WMO-NEXT:    srl a0, a4, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_sub_i8_acq_rel:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    li a3, 255
+; RV32IA-TSO-NEXT:    sll a3, a3, a0
+; RV32IA-TSO-NEXT:    andi a1, a1, 255
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB13_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a4, (a2)
+; RV32IA-TSO-NEXT:    sub a5, a4, a1
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:    and a5, a5, a3
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB13_1
+; RV32IA-TSO-NEXT:  # %bb.2:
+; RV32IA-TSO-NEXT:    srl a0, a4, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_sub_i8_acq_rel:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 4
 ; RV64I-NEXT:    call __atomic_fetch_sub_1@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_sub_i8_acq_rel:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    li a3, 255
-; RV64IA-NEXT:    sllw a3, a3, a0
-; RV64IA-NEXT:    andi a1, a1, 255
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB13_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a4, (a2)
-; RV64IA-NEXT:    sub a5, a4, a1
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:    and a5, a5, a3
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB13_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    srlw a0, a4, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_sub_i8_acq_rel:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    li a3, 255
+; RV64IA-WMO-NEXT:    sllw a3, a3, a0
+; RV64IA-WMO-NEXT:    andi a1, a1, 255
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB13_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a4, (a2)
+; RV64IA-WMO-NEXT:    sub a5, a4, a1
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:    and a5, a5, a3
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB13_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    srlw a0, a4, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_sub_i8_acq_rel:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    li a3, 255
+; RV64IA-TSO-NEXT:    sllw a3, a3, a0
+; RV64IA-TSO-NEXT:    andi a1, a1, 255
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB13_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a4, (a2)
+; RV64IA-TSO-NEXT:    sub a5, a4, a1
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:    and a5, a5, a3
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB13_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    srlw a0, a4, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw sub ptr %a, i8 %b acq_rel
   ret i8 %1
 }
 
 define i8 @atomicrmw_sub_i8_seq_cst(ptr %a, i8 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_sub_i8_seq_cst:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16

 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    li a2, 2
 ; RV32I-NEXT:    call __atomic_fetch_nand_1@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_nand_i8_acquire:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    li a3, 255
-; RV32IA-NEXT:    sll a3, a3, a0
-; RV32IA-NEXT:    andi a1, a1, 255
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB21_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a4, (a2)
-; RV32IA-NEXT:    and a5, a4, a1
-; RV32IA-NEXT:    not a5, a5
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:    and a5, a5, a3
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:    sc.w a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB21_1
-; RV32IA-NEXT:  # %bb.2:
-; RV32IA-NEXT:    srl a0, a4, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_nand_i8_acquire:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    li a3, 255
+; RV32IA-WMO-NEXT:    sll a3, a3, a0
+; RV32IA-WMO-NEXT:    andi a1, a1, 255
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB21_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a4, (a2)
+; RV32IA-WMO-NEXT:    and a5, a4, a1
+; RV32IA-WMO-NEXT:    not a5, a5
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:    and a5, a5, a3
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB21_1
+; RV32IA-WMO-NEXT:  # %bb.2:
+; RV32IA-WMO-NEXT:    srl a0, a4, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_nand_i8_acquire:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    li a3, 255
+; RV32IA-TSO-NEXT:    sll a3, a3, a0
+; RV32IA-TSO-NEXT:    andi a1, a1, 255
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB21_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a4, (a2)
+; RV32IA-TSO-NEXT:    and a5, a4, a1
+; RV32IA-TSO-NEXT:    not a5, a5
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:    and a5, a5, a3
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB21_1
+; RV32IA-TSO-NEXT:  # %bb.2:
+; RV32IA-TSO-NEXT:    srl a0, a4, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_nand_i8_acquire:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 2
 ; RV64I-NEXT:    call __atomic_fetch_nand_1@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_nand_i8_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    li a3, 255
-; RV64IA-NEXT:    sllw a3, a3, a0
-; RV64IA-NEXT:    andi a1, a1, 255
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB21_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a4, (a2)
-; RV64IA-NEXT:    and a5, a4, a1
-; RV64IA-NEXT:    not a5, a5
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:    and a5, a5, a3
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:    sc.w a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB21_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    srlw a0, a4, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_nand_i8_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    li a3, 255
+; RV64IA-WMO-NEXT:    sllw a3, a3, a0
+; RV64IA-WMO-NEXT:    andi a1, a1, 255
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB21_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a4, (a2)
+; RV64IA-WMO-NEXT:    and a5, a4, a1
+; RV64IA-WMO-NEXT:    not a5, a5
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:    and a5, a5, a3
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB21_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    srlw a0, a4, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_nand_i8_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    li a3, 255
+; RV64IA-TSO-NEXT:    sllw a3, a3, a0
+; RV64IA-TSO-NEXT:    andi a1, a1, 255
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB21_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a4, (a2)
+; RV64IA-TSO-NEXT:    and a5, a4, a1
+; RV64IA-TSO-NEXT:    not a5, a5
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:    and a5, a5, a3
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB21_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    srlw a0, a4, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw nand ptr %a, i8 %b acquire
   ret i8 %1
 }
 
 define i8 @atomicrmw_nand_i8_release(ptr %a, i8 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_nand_i8_release:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    li a2, 3
 ; RV32I-NEXT:    call __atomic_fetch_nand_1@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_nand_i8_release:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    li a3, 255
-; RV32IA-NEXT:    sll a3, a3, a0
-; RV32IA-NEXT:    andi a1, a1, 255
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB22_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w a4, (a2)
-; RV32IA-NEXT:    and a5, a4, a1
-; RV32IA-NEXT:    not a5, a5
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:    and a5, a5, a3
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB22_1
-; RV32IA-NEXT:  # %bb.2:
-; RV32IA-NEXT:    srl a0, a4, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_nand_i8_release:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    li a3, 255
+; RV32IA-WMO-NEXT:    sll a3, a3, a0
+; RV32IA-WMO-NEXT:    andi a1, a1, 255
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB22_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w a4, (a2)
+; RV32IA-WMO-NEXT:    and a5, a4, a1
+; RV32IA-WMO-NEXT:    not a5, a5
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:    and a5, a5, a3
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB22_1
+; RV32IA-WMO-NEXT:  # %bb.2:
+; RV32IA-WMO-NEXT:    srl a0, a4, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_nand_i8_release:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    li a3, 255
+; RV32IA-TSO-NEXT:    sll a3, a3, a0
+; RV32IA-TSO-NEXT:    andi a1, a1, 255
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB22_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a4, (a2)
+; RV32IA-TSO-NEXT:    and a5, a4, a1
+; RV32IA-TSO-NEXT:    not a5, a5
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:    and a5, a5, a3
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB22_1
+; RV32IA-TSO-NEXT:  # %bb.2:
+; RV32IA-TSO-NEXT:    srl a0, a4, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_nand_i8_release:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 3
 ; RV64I-NEXT:    call __atomic_fetch_nand_1@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_nand_i8_release:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    li a3, 255
-; RV64IA-NEXT:    sllw a3, a3, a0
-; RV64IA-NEXT:    andi a1, a1, 255
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB22_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w a4, (a2)
-; RV64IA-NEXT:    and a5, a4, a1
-; RV64IA-NEXT:    not a5, a5
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:    and a5, a5, a3
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB22_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    srlw a0, a4, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_nand_i8_release:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    li a3, 255
+; RV64IA-WMO-NEXT:    sllw a3, a3, a0
+; RV64IA-WMO-NEXT:    andi a1, a1, 255
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB22_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w a4, (a2)
+; RV64IA-WMO-NEXT:    and a5, a4, a1
+; RV64IA-WMO-NEXT:    not a5, a5
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:    and a5, a5, a3
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB22_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    srlw a0, a4, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_nand_i8_release:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    li a3, 255
+; RV64IA-TSO-NEXT:    sllw a3, a3, a0
+; RV64IA-TSO-NEXT:    andi a1, a1, 255
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB22_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a4, (a2)
+; RV64IA-TSO-NEXT:    and a5, a4, a1
+; RV64IA-TSO-NEXT:    not a5, a5
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:    and a5, a5, a3
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB22_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    srlw a0, a4, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw nand ptr %a, i8 %b release
   ret i8 %1
 }
 
 define i8 @atomicrmw_nand_i8_acq_rel(ptr %a, i8 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_nand_i8_acq_rel:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    li a2, 4
 ; RV32I-NEXT:    call __atomic_fetch_nand_1@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_nand_i8_acq_rel:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    li a3, 255
-; RV32IA-NEXT:    sll a3, a3, a0
-; RV32IA-NEXT:    andi a1, a1, 255
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB23_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a4, (a2)
-; RV32IA-NEXT:    and a5, a4, a1
-; RV32IA-NEXT:    not a5, a5
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:    and a5, a5, a3
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB23_1
-; RV32IA-NEXT:  # %bb.2:
-; RV32IA-NEXT:    srl a0, a4, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_nand_i8_acq_rel:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    li a3, 255
+; RV32IA-WMO-NEXT:    sll a3, a3, a0
+; RV32IA-WMO-NEXT:    andi a1, a1, 255
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB23_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a4, (a2)
+; RV32IA-WMO-NEXT:    and a5, a4, a1
+; RV32IA-WMO-NEXT:    not a5, a5
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:    and a5, a5, a3
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB23_1
+; RV32IA-WMO-NEXT:  # %bb.2:
+; RV32IA-WMO-NEXT:    srl a0, a4, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_nand_i8_acq_rel:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    li a3, 255
+; RV32IA-TSO-NEXT:    sll a3, a3, a0
+; RV32IA-TSO-NEXT:    andi a1, a1, 255
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB23_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a4, (a2)
+; RV32IA-TSO-NEXT:    and a5, a4, a1
+; RV32IA-TSO-NEXT:    not a5, a5
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:    and a5, a5, a3
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB23_1
+; RV32IA-TSO-NEXT:  # %bb.2:
+; RV32IA-TSO-NEXT:    srl a0, a4, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_nand_i8_acq_rel:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 4
 ; RV64I-NEXT:    call __atomic_fetch_nand_1@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_nand_i8_acq_rel:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    li a3, 255
-; RV64IA-NEXT:    sllw a3, a3, a0
-; RV64IA-NEXT:    andi a1, a1, 255
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB23_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a4, (a2)
-; RV64IA-NEXT:    and a5, a4, a1
-; RV64IA-NEXT:    not a5, a5
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:    and a5, a5, a3
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB23_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    srlw a0, a4, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_nand_i8_acq_rel:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    li a3, 255
+; RV64IA-WMO-NEXT:    sllw a3, a3, a0
+; RV64IA-WMO-NEXT:    andi a1, a1, 255
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB23_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a4, (a2)
+; RV64IA-WMO-NEXT:    and a5, a4, a1
+; RV64IA-WMO-NEXT:    not a5, a5
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:    and a5, a5, a3
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB23_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    srlw a0, a4, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_nand_i8_acq_rel:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    li a3, 255
+; RV64IA-TSO-NEXT:    sllw a3, a3, a0
+; RV64IA-TSO-NEXT:    andi a1, a1, 255
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB23_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a4, (a2)
+; RV64IA-TSO-NEXT:    and a5, a4, a1
+; RV64IA-TSO-NEXT:    not a5, a5
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:    and a5, a5, a3
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB23_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    srlw a0, a4, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw nand ptr %a, i8 %b acq_rel
   ret i8 %1
 }
 
 define i8 @atomicrmw_nand_i8_seq_cst(ptr %a, i8 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_nand_i8_seq_cst:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16

 ; RV32I-NEXT:    mv a0, a3
 ; RV32I-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s2, 16(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 32
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_max_i8_acquire:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    andi a3, a0, 24
-; RV32IA-NEXT:    li a4, 255
-; RV32IA-NEXT:    sll a4, a4, a0
-; RV32IA-NEXT:    slli a1, a1, 24
-; RV32IA-NEXT:    srai a1, a1, 24
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:    xori a3, a3, 24
-; RV32IA-NEXT:  .LBB36_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a5, (a2)
-; RV32IA-NEXT:    and a7, a5, a4
-; RV32IA-NEXT:    mv a6, a5
-; RV32IA-NEXT:    sll a7, a7, a3
-; RV32IA-NEXT:    sra a7, a7, a3
-; RV32IA-NEXT:    bge a7, a1, .LBB36_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB36_1 Depth=1
-; RV32IA-NEXT:    xor a6, a5, a1
-; RV32IA-NEXT:    and a6, a6, a4
-; RV32IA-NEXT:    xor a6, a5, a6
-; RV32IA-NEXT:  .LBB36_3: # in Loop: Header=BB36_1 Depth=1
-; RV32IA-NEXT:    sc.w a6, a6, (a2)
-; RV32IA-NEXT:    bnez a6, .LBB36_1
-; RV32IA-NEXT:  # %bb.4:
-; RV32IA-NEXT:    srl a0, a5, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_max_i8_acquire:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    andi a3, a0, 24
+; RV32IA-WMO-NEXT:    li a4, 255
+; RV32IA-WMO-NEXT:    sll a4, a4, a0
+; RV32IA-WMO-NEXT:    slli a1, a1, 24
+; RV32IA-WMO-NEXT:    srai a1, a1, 24
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:    xori a3, a3, 24
+; RV32IA-WMO-NEXT:  .LBB36_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a5, (a2)
+; RV32IA-WMO-NEXT:    and a7, a5, a4
+; RV32IA-WMO-NEXT:    mv a6, a5
+; RV32IA-WMO-NEXT:    sll a7, a7, a3
+; RV32IA-WMO-NEXT:    sra a7, a7, a3
+; RV32IA-WMO-NEXT:    bge a7, a1, .LBB36_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB36_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a6, a5, a1
+; RV32IA-WMO-NEXT:    and a6, a6, a4
+; RV32IA-WMO-NEXT:    xor a6, a5, a6
+; RV32IA-WMO-NEXT:  .LBB36_3: # in Loop: Header=BB36_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w a6, a6, (a2)
+; RV32IA-WMO-NEXT:    bnez a6, .LBB36_1
+; RV32IA-WMO-NEXT:  # %bb.4:
+; RV32IA-WMO-NEXT:    srl a0, a5, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_max_i8_acquire:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    andi a3, a0, 24
+; RV32IA-TSO-NEXT:    li a4, 255
+; RV32IA-TSO-NEXT:    sll a4, a4, a0
+; RV32IA-TSO-NEXT:    slli a1, a1, 24
+; RV32IA-TSO-NEXT:    srai a1, a1, 24
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:    xori a3, a3, 24
+; RV32IA-TSO-NEXT:  .LBB36_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a5, (a2)
+; RV32IA-TSO-NEXT:    and a7, a5, a4
+; RV32IA-TSO-NEXT:    mv a6, a5
+; RV32IA-TSO-NEXT:    sll a7, a7, a3
+; RV32IA-TSO-NEXT:    sra a7, a7, a3
+; RV32IA-TSO-NEXT:    bge a7, a1, .LBB36_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB36_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a6, a5, a1
+; RV32IA-TSO-NEXT:    and a6, a6, a4
+; RV32IA-TSO-NEXT:    xor a6, a5, a6
+; RV32IA-TSO-NEXT:  .LBB36_3: # in Loop: Header=BB36_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a6, a6, (a2)
+; RV32IA-TSO-NEXT:    bnez a6, .LBB36_1
+; RV32IA-TSO-NEXT:  # %bb.4:
+; RV32IA-TSO-NEXT:    srl a0, a5, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_max_i8_acquire:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -48
 ; RV64I-NEXT:    sd ra, 40(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s0, 32(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s1, 24(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s2, 16(sp) # 8-byte Folded Spill

 ; RV64I-NEXT:    mv a0, a3
 ; RV64I-NEXT:    ld ra, 40(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s0, 32(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s1, 24(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s2, 16(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 48
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_max_i8_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    andi a3, a0, 24
-; RV64IA-NEXT:    li a4, 255
-; RV64IA-NEXT:    sllw a4, a4, a0
-; RV64IA-NEXT:    slli a1, a1, 56
-; RV64IA-NEXT:    srai a1, a1, 56
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:    xori a3, a3, 56
-; RV64IA-NEXT:  .LBB36_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a5, (a2)
-; RV64IA-NEXT:    and a7, a5, a4
-; RV64IA-NEXT:    mv a6, a5
-; RV64IA-NEXT:    sll a7, a7, a3
-; RV64IA-NEXT:    sra a7, a7, a3
-; RV64IA-NEXT:    bge a7, a1, .LBB36_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB36_1 Depth=1
-; RV64IA-NEXT:    xor a6, a5, a1
-; RV64IA-NEXT:    and a6, a6, a4
-; RV64IA-NEXT:    xor a6, a5, a6
-; RV64IA-NEXT:  .LBB36_3: # in Loop: Header=BB36_1 Depth=1
-; RV64IA-NEXT:    sc.w a6, a6, (a2)
-; RV64IA-NEXT:    bnez a6, .LBB36_1
-; RV64IA-NEXT:  # %bb.4:
-; RV64IA-NEXT:    srlw a0, a5, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_max_i8_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    andi a3, a0, 24
+; RV64IA-WMO-NEXT:    li a4, 255
+; RV64IA-WMO-NEXT:    sllw a4, a4, a0
+; RV64IA-WMO-NEXT:    slli a1, a1, 56
+; RV64IA-WMO-NEXT:    srai a1, a1, 56
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:    xori a3, a3, 56
+; RV64IA-WMO-NEXT:  .LBB36_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a5, (a2)
+; RV64IA-WMO-NEXT:    and a7, a5, a4
+; RV64IA-WMO-NEXT:    mv a6, a5
+; RV64IA-WMO-NEXT:    sll a7, a7, a3
+; RV64IA-WMO-NEXT:    sra a7, a7, a3
+; RV64IA-WMO-NEXT:    bge a7, a1, .LBB36_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB36_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a6, a5, a1
+; RV64IA-WMO-NEXT:    and a6, a6, a4
+; RV64IA-WMO-NEXT:    xor a6, a5, a6
+; RV64IA-WMO-NEXT:  .LBB36_3: # in Loop: Header=BB36_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w a6, a6, (a2)
+; RV64IA-WMO-NEXT:    bnez a6, .LBB36_1
+; RV64IA-WMO-NEXT:  # %bb.4:
+; RV64IA-WMO-NEXT:    srlw a0, a5, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_max_i8_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    andi a3, a0, 24
+; RV64IA-TSO-NEXT:    li a4, 255
+; RV64IA-TSO-NEXT:    sllw a4, a4, a0
+; RV64IA-TSO-NEXT:    slli a1, a1, 56
+; RV64IA-TSO-NEXT:    srai a1, a1, 56
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:    xori a3, a3, 56
+; RV64IA-TSO-NEXT:  .LBB36_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a5, (a2)
+; RV64IA-TSO-NEXT:    and a7, a5, a4
+; RV64IA-TSO-NEXT:    mv a6, a5
+; RV64IA-TSO-NEXT:    sll a7, a7, a3
+; RV64IA-TSO-NEXT:    sra a7, a7, a3
+; RV64IA-TSO-NEXT:    bge a7, a1, .LBB36_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB36_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a6, a5, a1
+; RV64IA-TSO-NEXT:    and a6, a6, a4
+; RV64IA-TSO-NEXT:    xor a6, a5, a6
+; RV64IA-TSO-NEXT:  .LBB36_3: # in Loop: Header=BB36_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a6, a6, (a2)
+; RV64IA-TSO-NEXT:    bnez a6, .LBB36_1
+; RV64IA-TSO-NEXT:  # %bb.4:
+; RV64IA-TSO-NEXT:    srlw a0, a5, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw max ptr %a, i8 %b acquire
   ret i8 %1
 }
 
 define i8 @atomicrmw_max_i8_release(ptr %a, i8 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_max_i8_release:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -32

 ; RV32I-NEXT:    mv a0, a3
 ; RV32I-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s2, 16(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 32
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_max_i8_release:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    andi a3, a0, 24
-; RV32IA-NEXT:    li a4, 255
-; RV32IA-NEXT:    sll a4, a4, a0
-; RV32IA-NEXT:    slli a1, a1, 24
-; RV32IA-NEXT:    srai a1, a1, 24
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:    xori a3, a3, 24
-; RV32IA-NEXT:  .LBB37_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w a5, (a2)
-; RV32IA-NEXT:    and a7, a5, a4
-; RV32IA-NEXT:    mv a6, a5
-; RV32IA-NEXT:    sll a7, a7, a3
-; RV32IA-NEXT:    sra a7, a7, a3
-; RV32IA-NEXT:    bge a7, a1, .LBB37_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB37_1 Depth=1
-; RV32IA-NEXT:    xor a6, a5, a1
-; RV32IA-NEXT:    and a6, a6, a4
-; RV32IA-NEXT:    xor a6, a5, a6
-; RV32IA-NEXT:  .LBB37_3: # in Loop: Header=BB37_1 Depth=1
-; RV32IA-NEXT:    sc.w.rl a6, a6, (a2)
-; RV32IA-NEXT:    bnez a6, .LBB37_1
-; RV32IA-NEXT:  # %bb.4:
-; RV32IA-NEXT:    srl a0, a5, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_max_i8_release:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    andi a3, a0, 24
+; RV32IA-WMO-NEXT:    li a4, 255
+; RV32IA-WMO-NEXT:    sll a4, a4, a0
+; RV32IA-WMO-NEXT:    slli a1, a1, 24
+; RV32IA-WMO-NEXT:    srai a1, a1, 24
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:    xori a3, a3, 24
+; RV32IA-WMO-NEXT:  .LBB37_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w a5, (a2)
+; RV32IA-WMO-NEXT:    and a7, a5, a4
+; RV32IA-WMO-NEXT:    mv a6, a5
+; RV32IA-WMO-NEXT:    sll a7, a7, a3
+; RV32IA-WMO-NEXT:    sra a7, a7, a3
+; RV32IA-WMO-NEXT:    bge a7, a1, .LBB37_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB37_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a6, a5, a1
+; RV32IA-WMO-NEXT:    and a6, a6, a4
+; RV32IA-WMO-NEXT:    xor a6, a5, a6
+; RV32IA-WMO-NEXT:  .LBB37_3: # in Loop: Header=BB37_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w.rl a6, a6, (a2)
+; RV32IA-WMO-NEXT:    bnez a6, .LBB37_1
+; RV32IA-WMO-NEXT:  # %bb.4:
+; RV32IA-WMO-NEXT:    srl a0, a5, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_max_i8_release:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    andi a3, a0, 24
+; RV32IA-TSO-NEXT:    li a4, 255
+; RV32IA-TSO-NEXT:    sll a4, a4, a0
+; RV32IA-TSO-NEXT:    slli a1, a1, 24
+; RV32IA-TSO-NEXT:    srai a1, a1, 24
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:    xori a3, a3, 24
+; RV32IA-TSO-NEXT:  .LBB37_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a5, (a2)
+; RV32IA-TSO-NEXT:    and a7, a5, a4
+; RV32IA-TSO-NEXT:    mv a6, a5
+; RV32IA-TSO-NEXT:    sll a7, a7, a3
+; RV32IA-TSO-NEXT:    sra a7, a7, a3
+; RV32IA-TSO-NEXT:    bge a7, a1, .LBB37_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB37_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a6, a5, a1
+; RV32IA-TSO-NEXT:    and a6, a6, a4
+; RV32IA-TSO-NEXT:    xor a6, a5, a6
+; RV32IA-TSO-NEXT:  .LBB37_3: # in Loop: Header=BB37_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a6, a6, (a2)
+; RV32IA-TSO-NEXT:    bnez a6, .LBB37_1
+; RV32IA-TSO-NEXT:  # %bb.4:
+; RV32IA-TSO-NEXT:    srl a0, a5, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_max_i8_release:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -48
 ; RV64I-NEXT:    sd ra, 40(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s0, 32(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s1, 24(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s2, 16(sp) # 8-byte Folded Spill

 ; RV64I-NEXT:    mv a0, a3
 ; RV64I-NEXT:    ld ra, 40(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s0, 32(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s1, 24(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s2, 16(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 48
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_max_i8_release:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    andi a3, a0, 24
-; RV64IA-NEXT:    li a4, 255
-; RV64IA-NEXT:    sllw a4, a4, a0
-; RV64IA-NEXT:    slli a1, a1, 56
-; RV64IA-NEXT:    srai a1, a1, 56
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:    xori a3, a3, 56
-; RV64IA-NEXT:  .LBB37_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w a5, (a2)
-; RV64IA-NEXT:    and a7, a5, a4
-; RV64IA-NEXT:    mv a6, a5
-; RV64IA-NEXT:    sll a7, a7, a3
-; RV64IA-NEXT:    sra a7, a7, a3
-; RV64IA-NEXT:    bge a7, a1, .LBB37_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB37_1 Depth=1
-; RV64IA-NEXT:    xor a6, a5, a1
-; RV64IA-NEXT:    and a6, a6, a4
-; RV64IA-NEXT:    xor a6, a5, a6
-; RV64IA-NEXT:  .LBB37_3: # in Loop: Header=BB37_1 Depth=1
-; RV64IA-NEXT:    sc.w.rl a6, a6, (a2)
-; RV64IA-NEXT:    bnez a6, .LBB37_1
-; RV64IA-NEXT:  # %bb.4:
-; RV64IA-NEXT:    srlw a0, a5, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_max_i8_release:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    andi a3, a0, 24
+; RV64IA-WMO-NEXT:    li a4, 255
+; RV64IA-WMO-NEXT:    sllw a4, a4, a0
+; RV64IA-WMO-NEXT:    slli a1, a1, 56
+; RV64IA-WMO-NEXT:    srai a1, a1, 56
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:    xori a3, a3, 56
+; RV64IA-WMO-NEXT:  .LBB37_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w a5, (a2)
+; RV64IA-WMO-NEXT:    and a7, a5, a4
+; RV64IA-WMO-NEXT:    mv a6, a5
+; RV64IA-WMO-NEXT:    sll a7, a7, a3
+; RV64IA-WMO-NEXT:    sra a7, a7, a3
+; RV64IA-WMO-NEXT:    bge a7, a1, .LBB37_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB37_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a6, a5, a1
+; RV64IA-WMO-NEXT:    and a6, a6, a4
+; RV64IA-WMO-NEXT:    xor a6, a5, a6
+; RV64IA-WMO-NEXT:  .LBB37_3: # in Loop: Header=BB37_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w.rl a6, a6, (a2)
+; RV64IA-WMO-NEXT:    bnez a6, .LBB37_1
+; RV64IA-WMO-NEXT:  # %bb.4:
+; RV64IA-WMO-NEXT:    srlw a0, a5, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_max_i8_release:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    andi a3, a0, 24
+; RV64IA-TSO-NEXT:    li a4, 255
+; RV64IA-TSO-NEXT:    sllw a4, a4, a0
+; RV64IA-TSO-NEXT:    slli a1, a1, 56
+; RV64IA-TSO-NEXT:    srai a1, a1, 56
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:    xori a3, a3, 56
+; RV64IA-TSO-NEXT:  .LBB37_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a5, (a2)
+; RV64IA-TSO-NEXT:    and a7, a5, a4
+; RV64IA-TSO-NEXT:    mv a6, a5
+; RV64IA-TSO-NEXT:    sll a7, a7, a3
+; RV64IA-TSO-NEXT:    sra a7, a7, a3
+; RV64IA-TSO-NEXT:    bge a7, a1, .LBB37_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB37_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a6, a5, a1
+; RV64IA-TSO-NEXT:    and a6, a6, a4
+; RV64IA-TSO-NEXT:    xor a6, a5, a6
+; RV64IA-TSO-NEXT:  .LBB37_3: # in Loop: Header=BB37_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a6, a6, (a2)
+; RV64IA-TSO-NEXT:    bnez a6, .LBB37_1
+; RV64IA-TSO-NEXT:  # %bb.4:
+; RV64IA-TSO-NEXT:    srlw a0, a5, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw max ptr %a, i8 %b release
   ret i8 %1
 }
 
 define i8 @atomicrmw_max_i8_acq_rel(ptr %a, i8 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_max_i8_acq_rel:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -32

 ; RV32I-NEXT:    mv a0, a3
 ; RV32I-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s2, 16(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 32
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_max_i8_acq_rel:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    andi a3, a0, 24
-; RV32IA-NEXT:    li a4, 255
-; RV32IA-NEXT:    sll a4, a4, a0
-; RV32IA-NEXT:    slli a1, a1, 24
-; RV32IA-NEXT:    srai a1, a1, 24
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:    xori a3, a3, 24
-; RV32IA-NEXT:  .LBB38_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a5, (a2)
-; RV32IA-NEXT:    and a7, a5, a4
-; RV32IA-NEXT:    mv a6, a5
-; RV32IA-NEXT:    sll a7, a7, a3
-; RV32IA-NEXT:    sra a7, a7, a3
-; RV32IA-NEXT:    bge a7, a1, .LBB38_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB38_1 Depth=1
-; RV32IA-NEXT:    xor a6, a5, a1
-; RV32IA-NEXT:    and a6, a6, a4
-; RV32IA-NEXT:    xor a6, a5, a6
-; RV32IA-NEXT:  .LBB38_3: # in Loop: Header=BB38_1 Depth=1
-; RV32IA-NEXT:    sc.w.rl a6, a6, (a2)
-; RV32IA-NEXT:    bnez a6, .LBB38_1
-; RV32IA-NEXT:  # %bb.4:
-; RV32IA-NEXT:    srl a0, a5, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_max_i8_acq_rel:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    andi a3, a0, 24
+; RV32IA-WMO-NEXT:    li a4, 255
+; RV32IA-WMO-NEXT:    sll a4, a4, a0
+; RV32IA-WMO-NEXT:    slli a1, a1, 24
+; RV32IA-WMO-NEXT:    srai a1, a1, 24
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:    xori a3, a3, 24
+; RV32IA-WMO-NEXT:  .LBB38_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a5, (a2)
+; RV32IA-WMO-NEXT:    and a7, a5, a4
+; RV32IA-WMO-NEXT:    mv a6, a5
+; RV32IA-WMO-NEXT:    sll a7, a7, a3
+; RV32IA-WMO-NEXT:    sra a7, a7, a3
+; RV32IA-WMO-NEXT:    bge a7, a1, .LBB38_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB38_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a6, a5, a1
+; RV32IA-WMO-NEXT:    and a6, a6, a4
+; RV32IA-WMO-NEXT:    xor a6, a5, a6
+; RV32IA-WMO-NEXT:  .LBB38_3: # in Loop: Header=BB38_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w.rl a6, a6, (a2)
+; RV32IA-WMO-NEXT:    bnez a6, .LBB38_1
+; RV32IA-WMO-NEXT:  # %bb.4:
+; RV32IA-WMO-NEXT:    srl a0, a5, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_max_i8_acq_rel:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    andi a3, a0, 24
+; RV32IA-TSO-NEXT:    li a4, 255
+; RV32IA-TSO-NEXT:    sll a4, a4, a0
+; RV32IA-TSO-NEXT:    slli a1, a1, 24
+; RV32IA-TSO-NEXT:    srai a1, a1, 24
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:    xori a3, a3, 24
+; RV32IA-TSO-NEXT:  .LBB38_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a5, (a2)
+; RV32IA-TSO-NEXT:    and a7, a5, a4
+; RV32IA-TSO-NEXT:    mv a6, a5
+; RV32IA-TSO-NEXT:    sll a7, a7, a3
+; RV32IA-TSO-NEXT:    sra a7, a7, a3
+; RV32IA-TSO-NEXT:    bge a7, a1, .LBB38_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB38_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a6, a5, a1
+; RV32IA-TSO-NEXT:    and a6, a6, a4
+; RV32IA-TSO-NEXT:    xor a6, a5, a6
+; RV32IA-TSO-NEXT:  .LBB38_3: # in Loop: Header=BB38_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a6, a6, (a2)
+; RV32IA-TSO-NEXT:    bnez a6, .LBB38_1
+; RV32IA-TSO-NEXT:  # %bb.4:
+; RV32IA-TSO-NEXT:    srl a0, a5, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_max_i8_acq_rel:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -48
 ; RV64I-NEXT:    sd ra, 40(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s0, 32(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s1, 24(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s2, 16(sp) # 8-byte Folded Spill

 ; RV64I-NEXT:    mv a0, a3
 ; RV64I-NEXT:    ld ra, 40(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s0, 32(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s1, 24(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s2, 16(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 48
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_max_i8_acq_rel:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    andi a3, a0, 24
-; RV64IA-NEXT:    li a4, 255
-; RV64IA-NEXT:    sllw a4, a4, a0
-; RV64IA-NEXT:    slli a1, a1, 56
-; RV64IA-NEXT:    srai a1, a1, 56
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:    xori a3, a3, 56
-; RV64IA-NEXT:  .LBB38_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a5, (a2)
-; RV64IA-NEXT:    and a7, a5, a4
-; RV64IA-NEXT:    mv a6, a5
-; RV64IA-NEXT:    sll a7, a7, a3
-; RV64IA-NEXT:    sra a7, a7, a3
-; RV64IA-NEXT:    bge a7, a1, .LBB38_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB38_1 Depth=1
-; RV64IA-NEXT:    xor a6, a5, a1
-; RV64IA-NEXT:    and a6, a6, a4
-; RV64IA-NEXT:    xor a6, a5, a6
-; RV64IA-NEXT:  .LBB38_3: # in Loop: Header=BB38_1 Depth=1
-; RV64IA-NEXT:    sc.w.rl a6, a6, (a2)
-; RV64IA-NEXT:    bnez a6, .LBB38_1
-; RV64IA-NEXT:  # %bb.4:
-; RV64IA-NEXT:    srlw a0, a5, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_max_i8_acq_rel:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    andi a3, a0, 24
+; RV64IA-WMO-NEXT:    li a4, 255
+; RV64IA-WMO-NEXT:    sllw a4, a4, a0
+; RV64IA-WMO-NEXT:    slli a1, a1, 56
+; RV64IA-WMO-NEXT:    srai a1, a1, 56
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:    xori a3, a3, 56
+; RV64IA-WMO-NEXT:  .LBB38_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a5, (a2)
+; RV64IA-WMO-NEXT:    and a7, a5, a4
+; RV64IA-WMO-NEXT:    mv a6, a5
+; RV64IA-WMO-NEXT:    sll a7, a7, a3
+; RV64IA-WMO-NEXT:    sra a7, a7, a3
+; RV64IA-WMO-NEXT:    bge a7, a1, .LBB38_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB38_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a6, a5, a1
+; RV64IA-WMO-NEXT:    and a6, a6, a4
+; RV64IA-WMO-NEXT:    xor a6, a5, a6
+; RV64IA-WMO-NEXT:  .LBB38_3: # in Loop: Header=BB38_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w.rl a6, a6, (a2)
+; RV64IA-WMO-NEXT:    bnez a6, .LBB38_1
+; RV64IA-WMO-NEXT:  # %bb.4:
+; RV64IA-WMO-NEXT:    srlw a0, a5, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_max_i8_acq_rel:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    andi a3, a0, 24
+; RV64IA-TSO-NEXT:    li a4, 255
+; RV64IA-TSO-NEXT:    sllw a4, a4, a0
+; RV64IA-TSO-NEXT:    slli a1, a1, 56
+; RV64IA-TSO-NEXT:    srai a1, a1, 56
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:    xori a3, a3, 56
+; RV64IA-TSO-NEXT:  .LBB38_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a5, (a2)
+; RV64IA-TSO-NEXT:    and a7, a5, a4
+; RV64IA-TSO-NEXT:    mv a6, a5
+; RV64IA-TSO-NEXT:    sll a7, a7, a3
+; RV64IA-TSO-NEXT:    sra a7, a7, a3
+; RV64IA-TSO-NEXT:    bge a7, a1, .LBB38_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB38_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a6, a5, a1
+; RV64IA-TSO-NEXT:    and a6, a6, a4
+; RV64IA-TSO-NEXT:    xor a6, a5, a6
+; RV64IA-TSO-NEXT:  .LBB38_3: # in Loop: Header=BB38_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a6, a6, (a2)
+; RV64IA-TSO-NEXT:    bnez a6, .LBB38_1
+; RV64IA-TSO-NEXT:  # %bb.4:
+; RV64IA-TSO-NEXT:    srlw a0, a5, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw max ptr %a, i8 %b acq_rel
   ret i8 %1
 }
 
 define i8 @atomicrmw_max_i8_seq_cst(ptr %a, i8 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_max_i8_seq_cst:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -32

 ; RV32I-NEXT:    mv a0, a3
 ; RV32I-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s2, 16(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 32
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_min_i8_acquire:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    andi a3, a0, 24
-; RV32IA-NEXT:    li a4, 255
-; RV32IA-NEXT:    sll a4, a4, a0
-; RV32IA-NEXT:    slli a1, a1, 24
-; RV32IA-NEXT:    srai a1, a1, 24
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:    xori a3, a3, 24
-; RV32IA-NEXT:  .LBB41_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a5, (a2)
-; RV32IA-NEXT:    and a7, a5, a4
-; RV32IA-NEXT:    mv a6, a5
-; RV32IA-NEXT:    sll a7, a7, a3
-; RV32IA-NEXT:    sra a7, a7, a3
-; RV32IA-NEXT:    bge a1, a7, .LBB41_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB41_1 Depth=1
-; RV32IA-NEXT:    xor a6, a5, a1
-; RV32IA-NEXT:    and a6, a6, a4
-; RV32IA-NEXT:    xor a6, a5, a6
-; RV32IA-NEXT:  .LBB41_3: # in Loop: Header=BB41_1 Depth=1
-; RV32IA-NEXT:    sc.w a6, a6, (a2)
-; RV32IA-NEXT:    bnez a6, .LBB41_1
-; RV32IA-NEXT:  # %bb.4:
-; RV32IA-NEXT:    srl a0, a5, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_min_i8_acquire:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    andi a3, a0, 24
+; RV32IA-WMO-NEXT:    li a4, 255
+; RV32IA-WMO-NEXT:    sll a4, a4, a0
+; RV32IA-WMO-NEXT:    slli a1, a1, 24
+; RV32IA-WMO-NEXT:    srai a1, a1, 24
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:    xori a3, a3, 24
+; RV32IA-WMO-NEXT:  .LBB41_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a5, (a2)
+; RV32IA-WMO-NEXT:    and a7, a5, a4
+; RV32IA-WMO-NEXT:    mv a6, a5
+; RV32IA-WMO-NEXT:    sll a7, a7, a3
+; RV32IA-WMO-NEXT:    sra a7, a7, a3
+; RV32IA-WMO-NEXT:    bge a1, a7, .LBB41_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB41_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a6, a5, a1
+; RV32IA-WMO-NEXT:    and a6, a6, a4
+; RV32IA-WMO-NEXT:    xor a6, a5, a6
+; RV32IA-WMO-NEXT:  .LBB41_3: # in Loop: Header=BB41_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w a6, a6, (a2)
+; RV32IA-WMO-NEXT:    bnez a6, .LBB41_1
+; RV32IA-WMO-NEXT:  # %bb.4:
+; RV32IA-WMO-NEXT:    srl a0, a5, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_min_i8_acquire:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    andi a3, a0, 24
+; RV32IA-TSO-NEXT:    li a4, 255
+; RV32IA-TSO-NEXT:    sll a4, a4, a0
+; RV32IA-TSO-NEXT:    slli a1, a1, 24
+; RV32IA-TSO-NEXT:    srai a1, a1, 24
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:    xori a3, a3, 24
+; RV32IA-TSO-NEXT:  .LBB41_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a5, (a2)
+; RV32IA-TSO-NEXT:    and a7, a5, a4
+; RV32IA-TSO-NEXT:    mv a6, a5
+; RV32IA-TSO-NEXT:    sll a7, a7, a3
+; RV32IA-TSO-NEXT:    sra a7, a7, a3
+; RV32IA-TSO-NEXT:    bge a1, a7, .LBB41_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB41_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a6, a5, a1
+; RV32IA-TSO-NEXT:    and a6, a6, a4
+; RV32IA-TSO-NEXT:    xor a6, a5, a6
+; RV32IA-TSO-NEXT:  .LBB41_3: # in Loop: Header=BB41_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a6, a6, (a2)
+; RV32IA-TSO-NEXT:    bnez a6, .LBB41_1
+; RV32IA-TSO-NEXT:  # %bb.4:
+; RV32IA-TSO-NEXT:    srl a0, a5, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_min_i8_acquire:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -48
 ; RV64I-NEXT:    sd ra, 40(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s0, 32(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s1, 24(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s2, 16(sp) # 8-byte Folded Spill

 ; RV64I-NEXT:    mv a0, a3
 ; RV64I-NEXT:    ld ra, 40(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s0, 32(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s1, 24(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s2, 16(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 48
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_min_i8_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    andi a3, a0, 24
-; RV64IA-NEXT:    li a4, 255
-; RV64IA-NEXT:    sllw a4, a4, a0
-; RV64IA-NEXT:    slli a1, a1, 56
-; RV64IA-NEXT:    srai a1, a1, 56
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:    xori a3, a3, 56
-; RV64IA-NEXT:  .LBB41_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a5, (a2)
-; RV64IA-NEXT:    and a7, a5, a4
-; RV64IA-NEXT:    mv a6, a5
-; RV64IA-NEXT:    sll a7, a7, a3
-; RV64IA-NEXT:    sra a7, a7, a3
-; RV64IA-NEXT:    bge a1, a7, .LBB41_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB41_1 Depth=1
-; RV64IA-NEXT:    xor a6, a5, a1
-; RV64IA-NEXT:    and a6, a6, a4
-; RV64IA-NEXT:    xor a6, a5, a6
-; RV64IA-NEXT:  .LBB41_3: # in Loop: Header=BB41_1 Depth=1
-; RV64IA-NEXT:    sc.w a6, a6, (a2)
-; RV64IA-NEXT:    bnez a6, .LBB41_1
-; RV64IA-NEXT:  # %bb.4:
-; RV64IA-NEXT:    srlw a0, a5, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_min_i8_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    andi a3, a0, 24
+; RV64IA-WMO-NEXT:    li a4, 255
+; RV64IA-WMO-NEXT:    sllw a4, a4, a0
+; RV64IA-WMO-NEXT:    slli a1, a1, 56
+; RV64IA-WMO-NEXT:    srai a1, a1, 56
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:    xori a3, a3, 56
+; RV64IA-WMO-NEXT:  .LBB41_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a5, (a2)
+; RV64IA-WMO-NEXT:    and a7, a5, a4
+; RV64IA-WMO-NEXT:    mv a6, a5
+; RV64IA-WMO-NEXT:    sll a7, a7, a3
+; RV64IA-WMO-NEXT:    sra a7, a7, a3
+; RV64IA-WMO-NEXT:    bge a1, a7, .LBB41_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB41_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a6, a5, a1
+; RV64IA-WMO-NEXT:    and a6, a6, a4
+; RV64IA-WMO-NEXT:    xor a6, a5, a6
+; RV64IA-WMO-NEXT:  .LBB41_3: # in Loop: Header=BB41_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w a6, a6, (a2)
+; RV64IA-WMO-NEXT:    bnez a6, .LBB41_1
+; RV64IA-WMO-NEXT:  # %bb.4:
+; RV64IA-WMO-NEXT:    srlw a0, a5, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_min_i8_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    andi a3, a0, 24
+; RV64IA-TSO-NEXT:    li a4, 255
+; RV64IA-TSO-NEXT:    sllw a4, a4, a0
+; RV64IA-TSO-NEXT:    slli a1, a1, 56
+; RV64IA-TSO-NEXT:    srai a1, a1, 56
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:    xori a3, a3, 56
+; RV64IA-TSO-NEXT:  .LBB41_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a5, (a2)
+; RV64IA-TSO-NEXT:    and a7, a5, a4
+; RV64IA-TSO-NEXT:    mv a6, a5
+; RV64IA-TSO-NEXT:    sll a7, a7, a3
+; RV64IA-TSO-NEXT:    sra a7, a7, a3
+; RV64IA-TSO-NEXT:    bge a1, a7, .LBB41_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB41_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a6, a5, a1
+; RV64IA-TSO-NEXT:    and a6, a6, a4
+; RV64IA-TSO-NEXT:    xor a6, a5, a6
+; RV64IA-TSO-NEXT:  .LBB41_3: # in Loop: Header=BB41_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a6, a6, (a2)
+; RV64IA-TSO-NEXT:    bnez a6, .LBB41_1
+; RV64IA-TSO-NEXT:  # %bb.4:
+; RV64IA-TSO-NEXT:    srlw a0, a5, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw min ptr %a, i8 %b acquire
   ret i8 %1
 }
 
 define i8 @atomicrmw_min_i8_release(ptr %a, i8 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_min_i8_release:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -32

 ; RV32I-NEXT:    mv a0, a3
 ; RV32I-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s2, 16(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 32
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_min_i8_release:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    andi a3, a0, 24
-; RV32IA-NEXT:    li a4, 255
-; RV32IA-NEXT:    sll a4, a4, a0
-; RV32IA-NEXT:    slli a1, a1, 24
-; RV32IA-NEXT:    srai a1, a1, 24
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:    xori a3, a3, 24
-; RV32IA-NEXT:  .LBB42_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w a5, (a2)
-; RV32IA-NEXT:    and a7, a5, a4
-; RV32IA-NEXT:    mv a6, a5
-; RV32IA-NEXT:    sll a7, a7, a3
-; RV32IA-NEXT:    sra a7, a7, a3
-; RV32IA-NEXT:    bge a1, a7, .LBB42_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB42_1 Depth=1
-; RV32IA-NEXT:    xor a6, a5, a1
-; RV32IA-NEXT:    and a6, a6, a4
-; RV32IA-NEXT:    xor a6, a5, a6
-; RV32IA-NEXT:  .LBB42_3: # in Loop: Header=BB42_1 Depth=1
-; RV32IA-NEXT:    sc.w.rl a6, a6, (a2)
-; RV32IA-NEXT:    bnez a6, .LBB42_1
-; RV32IA-NEXT:  # %bb.4:
-; RV32IA-NEXT:    srl a0, a5, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_min_i8_release:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    andi a3, a0, 24
+; RV32IA-WMO-NEXT:    li a4, 255
+; RV32IA-WMO-NEXT:    sll a4, a4, a0
+; RV32IA-WMO-NEXT:    slli a1, a1, 24
+; RV32IA-WMO-NEXT:    srai a1, a1, 24
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:    xori a3, a3, 24
+; RV32IA-WMO-NEXT:  .LBB42_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w a5, (a2)
+; RV32IA-WMO-NEXT:    and a7, a5, a4
+; RV32IA-WMO-NEXT:    mv a6, a5
+; RV32IA-WMO-NEXT:    sll a7, a7, a3
+; RV32IA-WMO-NEXT:    sra a7, a7, a3
+; RV32IA-WMO-NEXT:    bge a1, a7, .LBB42_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB42_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a6, a5, a1
+; RV32IA-WMO-NEXT:    and a6, a6, a4
+; RV32IA-WMO-NEXT:    xor a6, a5, a6
+; RV32IA-WMO-NEXT:  .LBB42_3: # in Loop: Header=BB42_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w.rl a6, a6, (a2)
+; RV32IA-WMO-NEXT:    bnez a6, .LBB42_1
+; RV32IA-WMO-NEXT:  # %bb.4:
+; RV32IA-WMO-NEXT:    srl a0, a5, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_min_i8_release:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    andi a3, a0, 24
+; RV32IA-TSO-NEXT:    li a4, 255
+; RV32IA-TSO-NEXT:    sll a4, a4, a0
+; RV32IA-TSO-NEXT:    slli a1, a1, 24
+; RV32IA-TSO-NEXT:    srai a1, a1, 24
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:    xori a3, a3, 24
+; RV32IA-TSO-NEXT:  .LBB42_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a5, (a2)
+; RV32IA-TSO-NEXT:    and a7, a5, a4
+; RV32IA-TSO-NEXT:    mv a6, a5
+; RV32IA-TSO-NEXT:    sll a7, a7, a3
+; RV32IA-TSO-NEXT:    sra a7, a7, a3
+; RV32IA-TSO-NEXT:    bge a1, a7, .LBB42_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB42_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a6, a5, a1
+; RV32IA-TSO-NEXT:    and a6, a6, a4
+; RV32IA-TSO-NEXT:    xor a6, a5, a6
+; RV32IA-TSO-NEXT:  .LBB42_3: # in Loop: Header=BB42_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a6, a6, (a2)
+; RV32IA-TSO-NEXT:    bnez a6, .LBB42_1
+; RV32IA-TSO-NEXT:  # %bb.4:
+; RV32IA-TSO-NEXT:    srl a0, a5, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_min_i8_release:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -48
 ; RV64I-NEXT:    sd ra, 40(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s0, 32(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s1, 24(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s2, 16(sp) # 8-byte Folded Spill

 ; RV64I-NEXT:    mv a0, a3
 ; RV64I-NEXT:    ld ra, 40(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s0, 32(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s1, 24(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s2, 16(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 48
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_min_i8_release:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    andi a3, a0, 24
-; RV64IA-NEXT:    li a4, 255
-; RV64IA-NEXT:    sllw a4, a4, a0
-; RV64IA-NEXT:    slli a1, a1, 56
-; RV64IA-NEXT:    srai a1, a1, 56
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:    xori a3, a3, 56
-; RV64IA-NEXT:  .LBB42_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w a5, (a2)
-; RV64IA-NEXT:    and a7, a5, a4
-; RV64IA-NEXT:    mv a6, a5
-; RV64IA-NEXT:    sll a7, a7, a3
-; RV64IA-NEXT:    sra a7, a7, a3
-; RV64IA-NEXT:    bge a1, a7, .LBB42_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB42_1 Depth=1
-; RV64IA-NEXT:    xor a6, a5, a1
-; RV64IA-NEXT:    and a6, a6, a4
-; RV64IA-NEXT:    xor a6, a5, a6
-; RV64IA-NEXT:  .LBB42_3: # in Loop: Header=BB42_1 Depth=1
-; RV64IA-NEXT:    sc.w.rl a6, a6, (a2)
-; RV64IA-NEXT:    bnez a6, .LBB42_1
-; RV64IA-NEXT:  # %bb.4:
-; RV64IA-NEXT:    srlw a0, a5, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_min_i8_release:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    andi a3, a0, 24
+; RV64IA-WMO-NEXT:    li a4, 255
+; RV64IA-WMO-NEXT:    sllw a4, a4, a0
+; RV64IA-WMO-NEXT:    slli a1, a1, 56
+; RV64IA-WMO-NEXT:    srai a1, a1, 56
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:    xori a3, a3, 56
+; RV64IA-WMO-NEXT:  .LBB42_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w a5, (a2)
+; RV64IA-WMO-NEXT:    and a7, a5, a4
+; RV64IA-WMO-NEXT:    mv a6, a5
+; RV64IA-WMO-NEXT:    sll a7, a7, a3
+; RV64IA-WMO-NEXT:    sra a7, a7, a3
+; RV64IA-WMO-NEXT:    bge a1, a7, .LBB42_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB42_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a6, a5, a1
+; RV64IA-WMO-NEXT:    and a6, a6, a4
+; RV64IA-WMO-NEXT:    xor a6, a5, a6
+; RV64IA-WMO-NEXT:  .LBB42_3: # in Loop: Header=BB42_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w.rl a6, a6, (a2)
+; RV64IA-WMO-NEXT:    bnez a6, .LBB42_1
+; RV64IA-WMO-NEXT:  # %bb.4:
+; RV64IA-WMO-NEXT:    srlw a0, a5, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_min_i8_release:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    andi a3, a0, 24
+; RV64IA-TSO-NEXT:    li a4, 255
+; RV64IA-TSO-NEXT:    sllw a4, a4, a0
+; RV64IA-TSO-NEXT:    slli a1, a1, 56
+; RV64IA-TSO-NEXT:    srai a1, a1, 56
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:    xori a3, a3, 56
+; RV64IA-TSO-NEXT:  .LBB42_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a5, (a2)
+; RV64IA-TSO-NEXT:    and a7, a5, a4
+; RV64IA-TSO-NEXT:    mv a6, a5
+; RV64IA-TSO-NEXT:    sll a7, a7, a3
+; RV64IA-TSO-NEXT:    sra a7, a7, a3
+; RV64IA-TSO-NEXT:    bge a1, a7, .LBB42_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB42_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a6, a5, a1
+; RV64IA-TSO-NEXT:    and a6, a6, a4
+; RV64IA-TSO-NEXT:    xor a6, a5, a6
+; RV64IA-TSO-NEXT:  .LBB42_3: # in Loop: Header=BB42_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a6, a6, (a2)
+; RV64IA-TSO-NEXT:    bnez a6, .LBB42_1
+; RV64IA-TSO-NEXT:  # %bb.4:
+; RV64IA-TSO-NEXT:    srlw a0, a5, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw min ptr %a, i8 %b release
   ret i8 %1
 }
 
 define i8 @atomicrmw_min_i8_acq_rel(ptr %a, i8 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_min_i8_acq_rel:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -32

 ; RV32I-NEXT:    mv a0, a3
 ; RV32I-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s2, 16(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 32
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_min_i8_acq_rel:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    andi a3, a0, 24
-; RV32IA-NEXT:    li a4, 255
-; RV32IA-NEXT:    sll a4, a4, a0
-; RV32IA-NEXT:    slli a1, a1, 24
-; RV32IA-NEXT:    srai a1, a1, 24
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:    xori a3, a3, 24
-; RV32IA-NEXT:  .LBB43_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a5, (a2)
-; RV32IA-NEXT:    and a7, a5, a4
-; RV32IA-NEXT:    mv a6, a5
-; RV32IA-NEXT:    sll a7, a7, a3
-; RV32IA-NEXT:    sra a7, a7, a3
-; RV32IA-NEXT:    bge a1, a7, .LBB43_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB43_1 Depth=1
-; RV32IA-NEXT:    xor a6, a5, a1
-; RV32IA-NEXT:    and a6, a6, a4
-; RV32IA-NEXT:    xor a6, a5, a6
-; RV32IA-NEXT:  .LBB43_3: # in Loop: Header=BB43_1 Depth=1
-; RV32IA-NEXT:    sc.w.rl a6, a6, (a2)
-; RV32IA-NEXT:    bnez a6, .LBB43_1
-; RV32IA-NEXT:  # %bb.4:
-; RV32IA-NEXT:    srl a0, a5, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_min_i8_acq_rel:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    andi a3, a0, 24
+; RV32IA-WMO-NEXT:    li a4, 255
+; RV32IA-WMO-NEXT:    sll a4, a4, a0
+; RV32IA-WMO-NEXT:    slli a1, a1, 24
+; RV32IA-WMO-NEXT:    srai a1, a1, 24
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:    xori a3, a3, 24
+; RV32IA-WMO-NEXT:  .LBB43_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a5, (a2)
+; RV32IA-WMO-NEXT:    and a7, a5, a4
+; RV32IA-WMO-NEXT:    mv a6, a5
+; RV32IA-WMO-NEXT:    sll a7, a7, a3
+; RV32IA-WMO-NEXT:    sra a7, a7, a3
+; RV32IA-WMO-NEXT:    bge a1, a7, .LBB43_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB43_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a6, a5, a1
+; RV32IA-WMO-NEXT:    and a6, a6, a4
+; RV32IA-WMO-NEXT:    xor a6, a5, a6
+; RV32IA-WMO-NEXT:  .LBB43_3: # in Loop: Header=BB43_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w.rl a6, a6, (a2)
+; RV32IA-WMO-NEXT:    bnez a6, .LBB43_1
+; RV32IA-WMO-NEXT:  # %bb.4:
+; RV32IA-WMO-NEXT:    srl a0, a5, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_min_i8_acq_rel:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    andi a3, a0, 24
+; RV32IA-TSO-NEXT:    li a4, 255
+; RV32IA-TSO-NEXT:    sll a4, a4, a0
+; RV32IA-TSO-NEXT:    slli a1, a1, 24
+; RV32IA-TSO-NEXT:    srai a1, a1, 24
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:    xori a3, a3, 24
+; RV32IA-TSO-NEXT:  .LBB43_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a5, (a2)
+; RV32IA-TSO-NEXT:    and a7, a5, a4
+; RV32IA-TSO-NEXT:    mv a6, a5
+; RV32IA-TSO-NEXT:    sll a7, a7, a3
+; RV32IA-TSO-NEXT:    sra a7, a7, a3
+; RV32IA-TSO-NEXT:    bge a1, a7, .LBB43_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB43_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a6, a5, a1
+; RV32IA-TSO-NEXT:    and a6, a6, a4
+; RV32IA-TSO-NEXT:    xor a6, a5, a6
+; RV32IA-TSO-NEXT:  .LBB43_3: # in Loop: Header=BB43_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a6, a6, (a2)
+; RV32IA-TSO-NEXT:    bnez a6, .LBB43_1
+; RV32IA-TSO-NEXT:  # %bb.4:
+; RV32IA-TSO-NEXT:    srl a0, a5, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_min_i8_acq_rel:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -48
 ; RV64I-NEXT:    sd ra, 40(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s0, 32(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s1, 24(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s2, 16(sp) # 8-byte Folded Spill

 ; RV64I-NEXT:    mv a0, a3
 ; RV64I-NEXT:    ld ra, 40(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s0, 32(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s1, 24(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s2, 16(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 48
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_min_i8_acq_rel:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    andi a3, a0, 24
-; RV64IA-NEXT:    li a4, 255
-; RV64IA-NEXT:    sllw a4, a4, a0
-; RV64IA-NEXT:    slli a1, a1, 56
-; RV64IA-NEXT:    srai a1, a1, 56
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:    xori a3, a3, 56
-; RV64IA-NEXT:  .LBB43_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a5, (a2)
-; RV64IA-NEXT:    and a7, a5, a4
-; RV64IA-NEXT:    mv a6, a5
-; RV64IA-NEXT:    sll a7, a7, a3
-; RV64IA-NEXT:    sra a7, a7, a3
-; RV64IA-NEXT:    bge a1, a7, .LBB43_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB43_1 Depth=1
-; RV64IA-NEXT:    xor a6, a5, a1
-; RV64IA-NEXT:    and a6, a6, a4
-; RV64IA-NEXT:    xor a6, a5, a6
-; RV64IA-NEXT:  .LBB43_3: # in Loop: Header=BB43_1 Depth=1
-; RV64IA-NEXT:    sc.w.rl a6, a6, (a2)
-; RV64IA-NEXT:    bnez a6, .LBB43_1
-; RV64IA-NEXT:  # %bb.4:
-; RV64IA-NEXT:    srlw a0, a5, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_min_i8_acq_rel:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    andi a3, a0, 24
+; RV64IA-WMO-NEXT:    li a4, 255
+; RV64IA-WMO-NEXT:    sllw a4, a4, a0
+; RV64IA-WMO-NEXT:    slli a1, a1, 56
+; RV64IA-WMO-NEXT:    srai a1, a1, 56
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:    xori a3, a3, 56
+; RV64IA-WMO-NEXT:  .LBB43_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a5, (a2)
+; RV64IA-WMO-NEXT:    and a7, a5, a4
+; RV64IA-WMO-NEXT:    mv a6, a5
+; RV64IA-WMO-NEXT:    sll a7, a7, a3
+; RV64IA-WMO-NEXT:    sra a7, a7, a3
+; RV64IA-WMO-NEXT:    bge a1, a7, .LBB43_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB43_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a6, a5, a1
+; RV64IA-WMO-NEXT:    and a6, a6, a4
+; RV64IA-WMO-NEXT:    xor a6, a5, a6
+; RV64IA-WMO-NEXT:  .LBB43_3: # in Loop: Header=BB43_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w.rl a6, a6, (a2)
+; RV64IA-WMO-NEXT:    bnez a6, .LBB43_1
+; RV64IA-WMO-NEXT:  # %bb.4:
+; RV64IA-WMO-NEXT:    srlw a0, a5, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_min_i8_acq_rel:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    andi a3, a0, 24
+; RV64IA-TSO-NEXT:    li a4, 255
+; RV64IA-TSO-NEXT:    sllw a4, a4, a0
+; RV64IA-TSO-NEXT:    slli a1, a1, 56
+; RV64IA-TSO-NEXT:    srai a1, a1, 56
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:    xori a3, a3, 56
+; RV64IA-TSO-NEXT:  .LBB43_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a5, (a2)
+; RV64IA-TSO-NEXT:    and a7, a5, a4
+; RV64IA-TSO-NEXT:    mv a6, a5
+; RV64IA-TSO-NEXT:    sll a7, a7, a3
+; RV64IA-TSO-NEXT:    sra a7, a7, a3
+; RV64IA-TSO-NEXT:    bge a1, a7, .LBB43_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB43_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a6, a5, a1
+; RV64IA-TSO-NEXT:    and a6, a6, a4
+; RV64IA-TSO-NEXT:    xor a6, a5, a6
+; RV64IA-TSO-NEXT:  .LBB43_3: # in Loop: Header=BB43_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a6, a6, (a2)
+; RV64IA-TSO-NEXT:    bnez a6, .LBB43_1
+; RV64IA-TSO-NEXT:  # %bb.4:
+; RV64IA-TSO-NEXT:    srlw a0, a5, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw min ptr %a, i8 %b acq_rel
   ret i8 %1
 }
 
 define i8 @atomicrmw_min_i8_seq_cst(ptr %a, i8 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_min_i8_seq_cst:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -32

 ; RV32I-NEXT:    mv a0, a3
 ; RV32I-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s2, 16(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 32
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_umax_i8_acquire:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    li a3, 255
-; RV32IA-NEXT:    sll a3, a3, a0
-; RV32IA-NEXT:    andi a1, a1, 255
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB46_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a4, (a2)
-; RV32IA-NEXT:    and a6, a4, a3
-; RV32IA-NEXT:    mv a5, a4
-; RV32IA-NEXT:    bgeu a6, a1, .LBB46_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB46_1 Depth=1
-; RV32IA-NEXT:    xor a5, a4, a1
-; RV32IA-NEXT:    and a5, a5, a3
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:  .LBB46_3: # in Loop: Header=BB46_1 Depth=1
-; RV32IA-NEXT:    sc.w a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB46_1
-; RV32IA-NEXT:  # %bb.4:
-; RV32IA-NEXT:    srl a0, a4, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_umax_i8_acquire:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    li a3, 255
+; RV32IA-WMO-NEXT:    sll a3, a3, a0
+; RV32IA-WMO-NEXT:    andi a1, a1, 255
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB46_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a4, (a2)
+; RV32IA-WMO-NEXT:    and a6, a4, a3
+; RV32IA-WMO-NEXT:    mv a5, a4
+; RV32IA-WMO-NEXT:    bgeu a6, a1, .LBB46_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB46_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a5, a4, a1
+; RV32IA-WMO-NEXT:    and a5, a5, a3
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:  .LBB46_3: # in Loop: Header=BB46_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB46_1
+; RV32IA-WMO-NEXT:  # %bb.4:
+; RV32IA-WMO-NEXT:    srl a0, a4, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_umax_i8_acquire:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    li a3, 255
+; RV32IA-TSO-NEXT:    sll a3, a3, a0
+; RV32IA-TSO-NEXT:    andi a1, a1, 255
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB46_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a4, (a2)
+; RV32IA-TSO-NEXT:    and a6, a4, a3
+; RV32IA-TSO-NEXT:    mv a5, a4
+; RV32IA-TSO-NEXT:    bgeu a6, a1, .LBB46_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB46_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a5, a4, a1
+; RV32IA-TSO-NEXT:    and a5, a5, a3
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:  .LBB46_3: # in Loop: Header=BB46_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB46_1
+; RV32IA-TSO-NEXT:  # %bb.4:
+; RV32IA-TSO-NEXT:    srl a0, a4, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_umax_i8_acquire:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -48
 ; RV64I-NEXT:    sd ra, 40(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s0, 32(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s1, 24(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s2, 16(sp) # 8-byte Folded Spill

 ; RV64I-NEXT:    mv a0, a3
 ; RV64I-NEXT:    ld ra, 40(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s0, 32(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s1, 24(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s2, 16(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 48
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_umax_i8_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    li a3, 255
-; RV64IA-NEXT:    sllw a3, a3, a0
-; RV64IA-NEXT:    andi a1, a1, 255
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB46_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a4, (a2)
-; RV64IA-NEXT:    and a6, a4, a3
-; RV64IA-NEXT:    mv a5, a4
-; RV64IA-NEXT:    bgeu a6, a1, .LBB46_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB46_1 Depth=1
-; RV64IA-NEXT:    xor a5, a4, a1
-; RV64IA-NEXT:    and a5, a5, a3
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:  .LBB46_3: # in Loop: Header=BB46_1 Depth=1
-; RV64IA-NEXT:    sc.w a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB46_1
-; RV64IA-NEXT:  # %bb.4:
-; RV64IA-NEXT:    srlw a0, a4, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_umax_i8_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    li a3, 255
+; RV64IA-WMO-NEXT:    sllw a3, a3, a0
+; RV64IA-WMO-NEXT:    andi a1, a1, 255
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB46_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a4, (a2)
+; RV64IA-WMO-NEXT:    and a6, a4, a3
+; RV64IA-WMO-NEXT:    mv a5, a4
+; RV64IA-WMO-NEXT:    bgeu a6, a1, .LBB46_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB46_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a5, a4, a1
+; RV64IA-WMO-NEXT:    and a5, a5, a3
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:  .LBB46_3: # in Loop: Header=BB46_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB46_1
+; RV64IA-WMO-NEXT:  # %bb.4:
+; RV64IA-WMO-NEXT:    srlw a0, a4, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_umax_i8_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    li a3, 255
+; RV64IA-TSO-NEXT:    sllw a3, a3, a0
+; RV64IA-TSO-NEXT:    andi a1, a1, 255
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB46_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a4, (a2)
+; RV64IA-TSO-NEXT:    and a6, a4, a3
+; RV64IA-TSO-NEXT:    mv a5, a4
+; RV64IA-TSO-NEXT:    bgeu a6, a1, .LBB46_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB46_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a5, a4, a1
+; RV64IA-TSO-NEXT:    and a5, a5, a3
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:  .LBB46_3: # in Loop: Header=BB46_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB46_1
+; RV64IA-TSO-NEXT:  # %bb.4:
+; RV64IA-TSO-NEXT:    srlw a0, a4, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw umax ptr %a, i8 %b acquire
   ret i8 %1
 }
 
 define i8 @atomicrmw_umax_i8_release(ptr %a, i8 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_umax_i8_release:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -32

 ; RV32I-NEXT:    mv a0, a3
 ; RV32I-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s2, 16(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 32
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_umax_i8_release:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    li a3, 255
-; RV32IA-NEXT:    sll a3, a3, a0
-; RV32IA-NEXT:    andi a1, a1, 255
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB47_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w a4, (a2)
-; RV32IA-NEXT:    and a6, a4, a3
-; RV32IA-NEXT:    mv a5, a4
-; RV32IA-NEXT:    bgeu a6, a1, .LBB47_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB47_1 Depth=1
-; RV32IA-NEXT:    xor a5, a4, a1
-; RV32IA-NEXT:    and a5, a5, a3
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:  .LBB47_3: # in Loop: Header=BB47_1 Depth=1
-; RV32IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB47_1
-; RV32IA-NEXT:  # %bb.4:
-; RV32IA-NEXT:    srl a0, a4, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_umax_i8_release:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    li a3, 255
+; RV32IA-WMO-NEXT:    sll a3, a3, a0
+; RV32IA-WMO-NEXT:    andi a1, a1, 255
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB47_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w a4, (a2)
+; RV32IA-WMO-NEXT:    and a6, a4, a3
+; RV32IA-WMO-NEXT:    mv a5, a4
+; RV32IA-WMO-NEXT:    bgeu a6, a1, .LBB47_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB47_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a5, a4, a1
+; RV32IA-WMO-NEXT:    and a5, a5, a3
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:  .LBB47_3: # in Loop: Header=BB47_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB47_1
+; RV32IA-WMO-NEXT:  # %bb.4:
+; RV32IA-WMO-NEXT:    srl a0, a4, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_umax_i8_release:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    li a3, 255
+; RV32IA-TSO-NEXT:    sll a3, a3, a0
+; RV32IA-TSO-NEXT:    andi a1, a1, 255
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB47_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a4, (a2)
+; RV32IA-TSO-NEXT:    and a6, a4, a3
+; RV32IA-TSO-NEXT:    mv a5, a4
+; RV32IA-TSO-NEXT:    bgeu a6, a1, .LBB47_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB47_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a5, a4, a1
+; RV32IA-TSO-NEXT:    and a5, a5, a3
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:  .LBB47_3: # in Loop: Header=BB47_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB47_1
+; RV32IA-TSO-NEXT:  # %bb.4:
+; RV32IA-TSO-NEXT:    srl a0, a4, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_umax_i8_release:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -48
 ; RV64I-NEXT:    sd ra, 40(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s0, 32(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s1, 24(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s2, 16(sp) # 8-byte Folded Spill

 ; RV64I-NEXT:    mv a0, a3
 ; RV64I-NEXT:    ld ra, 40(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s0, 32(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s1, 24(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s2, 16(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 48
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_umax_i8_release:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    li a3, 255
-; RV64IA-NEXT:    sllw a3, a3, a0
-; RV64IA-NEXT:    andi a1, a1, 255
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB47_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w a4, (a2)
-; RV64IA-NEXT:    and a6, a4, a3
-; RV64IA-NEXT:    mv a5, a4
-; RV64IA-NEXT:    bgeu a6, a1, .LBB47_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB47_1 Depth=1
-; RV64IA-NEXT:    xor a5, a4, a1
-; RV64IA-NEXT:    and a5, a5, a3
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:  .LBB47_3: # in Loop: Header=BB47_1 Depth=1
-; RV64IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB47_1
-; RV64IA-NEXT:  # %bb.4:
-; RV64IA-NEXT:    srlw a0, a4, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_umax_i8_release:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    li a3, 255
+; RV64IA-WMO-NEXT:    sllw a3, a3, a0
+; RV64IA-WMO-NEXT:    andi a1, a1, 255
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB47_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w a4, (a2)
+; RV64IA-WMO-NEXT:    and a6, a4, a3
+; RV64IA-WMO-NEXT:    mv a5, a4
+; RV64IA-WMO-NEXT:    bgeu a6, a1, .LBB47_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB47_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a5, a4, a1
+; RV64IA-WMO-NEXT:    and a5, a5, a3
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:  .LBB47_3: # in Loop: Header=BB47_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB47_1
+; RV64IA-WMO-NEXT:  # %bb.4:
+; RV64IA-WMO-NEXT:    srlw a0, a4, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_umax_i8_release:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    li a3, 255
+; RV64IA-TSO-NEXT:    sllw a3, a3, a0
+; RV64IA-TSO-NEXT:    andi a1, a1, 255
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB47_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a4, (a2)
+; RV64IA-TSO-NEXT:    and a6, a4, a3
+; RV64IA-TSO-NEXT:    mv a5, a4
+; RV64IA-TSO-NEXT:    bgeu a6, a1, .LBB47_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB47_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a5, a4, a1
+; RV64IA-TSO-NEXT:    and a5, a5, a3
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:  .LBB47_3: # in Loop: Header=BB47_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB47_1
+; RV64IA-TSO-NEXT:  # %bb.4:
+; RV64IA-TSO-NEXT:    srlw a0, a4, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw umax ptr %a, i8 %b release
   ret i8 %1
 }
 
 define i8 @atomicrmw_umax_i8_acq_rel(ptr %a, i8 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_umax_i8_acq_rel:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -32

 ; RV32I-NEXT:    mv a0, a3
 ; RV32I-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s2, 16(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 32
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_umax_i8_acq_rel:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    li a3, 255
-; RV32IA-NEXT:    sll a3, a3, a0
-; RV32IA-NEXT:    andi a1, a1, 255
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB48_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a4, (a2)
-; RV32IA-NEXT:    and a6, a4, a3
-; RV32IA-NEXT:    mv a5, a4
-; RV32IA-NEXT:    bgeu a6, a1, .LBB48_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB48_1 Depth=1
-; RV32IA-NEXT:    xor a5, a4, a1
-; RV32IA-NEXT:    and a5, a5, a3
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:  .LBB48_3: # in Loop: Header=BB48_1 Depth=1
-; RV32IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB48_1
-; RV32IA-NEXT:  # %bb.4:
-; RV32IA-NEXT:    srl a0, a4, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_umax_i8_acq_rel:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    li a3, 255
+; RV32IA-WMO-NEXT:    sll a3, a3, a0
+; RV32IA-WMO-NEXT:    andi a1, a1, 255
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB48_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a4, (a2)
+; RV32IA-WMO-NEXT:    and a6, a4, a3
+; RV32IA-WMO-NEXT:    mv a5, a4
+; RV32IA-WMO-NEXT:    bgeu a6, a1, .LBB48_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB48_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a5, a4, a1
+; RV32IA-WMO-NEXT:    and a5, a5, a3
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:  .LBB48_3: # in Loop: Header=BB48_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB48_1
+; RV32IA-WMO-NEXT:  # %bb.4:
+; RV32IA-WMO-NEXT:    srl a0, a4, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_umax_i8_acq_rel:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    li a3, 255
+; RV32IA-TSO-NEXT:    sll a3, a3, a0
+; RV32IA-TSO-NEXT:    andi a1, a1, 255
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB48_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a4, (a2)
+; RV32IA-TSO-NEXT:    and a6, a4, a3
+; RV32IA-TSO-NEXT:    mv a5, a4
+; RV32IA-TSO-NEXT:    bgeu a6, a1, .LBB48_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB48_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a5, a4, a1
+; RV32IA-TSO-NEXT:    and a5, a5, a3
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:  .LBB48_3: # in Loop: Header=BB48_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB48_1
+; RV32IA-TSO-NEXT:  # %bb.4:
+; RV32IA-TSO-NEXT:    srl a0, a4, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_umax_i8_acq_rel:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -48
 ; RV64I-NEXT:    sd ra, 40(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s0, 32(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s1, 24(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s2, 16(sp) # 8-byte Folded Spill

 ; RV64I-NEXT:    mv a0, a3
 ; RV64I-NEXT:    ld ra, 40(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s0, 32(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s1, 24(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s2, 16(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 48
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_umax_i8_acq_rel:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    li a3, 255
-; RV64IA-NEXT:    sllw a3, a3, a0
-; RV64IA-NEXT:    andi a1, a1, 255
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB48_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a4, (a2)
-; RV64IA-NEXT:    and a6, a4, a3
-; RV64IA-NEXT:    mv a5, a4
-; RV64IA-NEXT:    bgeu a6, a1, .LBB48_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB48_1 Depth=1
-; RV64IA-NEXT:    xor a5, a4, a1
-; RV64IA-NEXT:    and a5, a5, a3
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:  .LBB48_3: # in Loop: Header=BB48_1 Depth=1
-; RV64IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB48_1
-; RV64IA-NEXT:  # %bb.4:
-; RV64IA-NEXT:    srlw a0, a4, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_umax_i8_acq_rel:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    li a3, 255
+; RV64IA-WMO-NEXT:    sllw a3, a3, a0
+; RV64IA-WMO-NEXT:    andi a1, a1, 255
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB48_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a4, (a2)
+; RV64IA-WMO-NEXT:    and a6, a4, a3
+; RV64IA-WMO-NEXT:    mv a5, a4
+; RV64IA-WMO-NEXT:    bgeu a6, a1, .LBB48_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB48_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a5, a4, a1
+; RV64IA-WMO-NEXT:    and a5, a5, a3
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:  .LBB48_3: # in Loop: Header=BB48_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB48_1
+; RV64IA-WMO-NEXT:  # %bb.4:
+; RV64IA-WMO-NEXT:    srlw a0, a4, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_umax_i8_acq_rel:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    li a3, 255
+; RV64IA-TSO-NEXT:    sllw a3, a3, a0
+; RV64IA-TSO-NEXT:    andi a1, a1, 255
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB48_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a4, (a2)
+; RV64IA-TSO-NEXT:    and a6, a4, a3
+; RV64IA-TSO-NEXT:    mv a5, a4
+; RV64IA-TSO-NEXT:    bgeu a6, a1, .LBB48_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB48_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a5, a4, a1
+; RV64IA-TSO-NEXT:    and a5, a5, a3
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:  .LBB48_3: # in Loop: Header=BB48_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB48_1
+; RV64IA-TSO-NEXT:  # %bb.4:
+; RV64IA-TSO-NEXT:    srlw a0, a4, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw umax ptr %a, i8 %b acq_rel
   ret i8 %1
 }
 
 define i8 @atomicrmw_umax_i8_seq_cst(ptr %a, i8 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_umax_i8_seq_cst:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -32

 ; RV32I-NEXT:    mv a0, a3
 ; RV32I-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s2, 16(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 32
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_umin_i8_acquire:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    li a3, 255
-; RV32IA-NEXT:    sll a3, a3, a0
-; RV32IA-NEXT:    andi a1, a1, 255
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB51_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a4, (a2)
-; RV32IA-NEXT:    and a6, a4, a3
-; RV32IA-NEXT:    mv a5, a4
-; RV32IA-NEXT:    bgeu a1, a6, .LBB51_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB51_1 Depth=1
-; RV32IA-NEXT:    xor a5, a4, a1
-; RV32IA-NEXT:    and a5, a5, a3
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:  .LBB51_3: # in Loop: Header=BB51_1 Depth=1
-; RV32IA-NEXT:    sc.w a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB51_1
-; RV32IA-NEXT:  # %bb.4:
-; RV32IA-NEXT:    srl a0, a4, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_umin_i8_acquire:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    li a3, 255
+; RV32IA-WMO-NEXT:    sll a3, a3, a0
+; RV32IA-WMO-NEXT:    andi a1, a1, 255
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB51_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a4, (a2)
+; RV32IA-WMO-NEXT:    and a6, a4, a3
+; RV32IA-WMO-NEXT:    mv a5, a4
+; RV32IA-WMO-NEXT:    bgeu a1, a6, .LBB51_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB51_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a5, a4, a1
+; RV32IA-WMO-NEXT:    and a5, a5, a3
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:  .LBB51_3: # in Loop: Header=BB51_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB51_1
+; RV32IA-WMO-NEXT:  # %bb.4:
+; RV32IA-WMO-NEXT:    srl a0, a4, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_umin_i8_acquire:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    li a3, 255
+; RV32IA-TSO-NEXT:    sll a3, a3, a0
+; RV32IA-TSO-NEXT:    andi a1, a1, 255
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB51_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a4, (a2)
+; RV32IA-TSO-NEXT:    and a6, a4, a3
+; RV32IA-TSO-NEXT:    mv a5, a4
+; RV32IA-TSO-NEXT:    bgeu a1, a6, .LBB51_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB51_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a5, a4, a1
+; RV32IA-TSO-NEXT:    and a5, a5, a3
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:  .LBB51_3: # in Loop: Header=BB51_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB51_1
+; RV32IA-TSO-NEXT:  # %bb.4:
+; RV32IA-TSO-NEXT:    srl a0, a4, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_umin_i8_acquire:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -48
 ; RV64I-NEXT:    sd ra, 40(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s0, 32(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s1, 24(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s2, 16(sp) # 8-byte Folded Spill

 ; RV64I-NEXT:    mv a0, a3
 ; RV64I-NEXT:    ld ra, 40(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s0, 32(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s1, 24(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s2, 16(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 48
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_umin_i8_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    li a3, 255
-; RV64IA-NEXT:    sllw a3, a3, a0
-; RV64IA-NEXT:    andi a1, a1, 255
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB51_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a4, (a2)
-; RV64IA-NEXT:    and a6, a4, a3
-; RV64IA-NEXT:    mv a5, a4
-; RV64IA-NEXT:    bgeu a1, a6, .LBB51_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB51_1 Depth=1
-; RV64IA-NEXT:    xor a5, a4, a1
-; RV64IA-NEXT:    and a5, a5, a3
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:  .LBB51_3: # in Loop: Header=BB51_1 Depth=1
-; RV64IA-NEXT:    sc.w a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB51_1
-; RV64IA-NEXT:  # %bb.4:
-; RV64IA-NEXT:    srlw a0, a4, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_umin_i8_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    li a3, 255
+; RV64IA-WMO-NEXT:    sllw a3, a3, a0
+; RV64IA-WMO-NEXT:    andi a1, a1, 255
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB51_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a4, (a2)
+; RV64IA-WMO-NEXT:    and a6, a4, a3
+; RV64IA-WMO-NEXT:    mv a5, a4
+; RV64IA-WMO-NEXT:    bgeu a1, a6, .LBB51_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB51_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a5, a4, a1
+; RV64IA-WMO-NEXT:    and a5, a5, a3
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:  .LBB51_3: # in Loop: Header=BB51_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB51_1
+; RV64IA-WMO-NEXT:  # %bb.4:
+; RV64IA-WMO-NEXT:    srlw a0, a4, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_umin_i8_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    li a3, 255
+; RV64IA-TSO-NEXT:    sllw a3, a3, a0
+; RV64IA-TSO-NEXT:    andi a1, a1, 255
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB51_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a4, (a2)
+; RV64IA-TSO-NEXT:    and a6, a4, a3
+; RV64IA-TSO-NEXT:    mv a5, a4
+; RV64IA-TSO-NEXT:    bgeu a1, a6, .LBB51_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB51_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a5, a4, a1
+; RV64IA-TSO-NEXT:    and a5, a5, a3
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:  .LBB51_3: # in Loop: Header=BB51_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB51_1
+; RV64IA-TSO-NEXT:  # %bb.4:
+; RV64IA-TSO-NEXT:    srlw a0, a4, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw umin ptr %a, i8 %b acquire
   ret i8 %1
 }
 
 define i8 @atomicrmw_umin_i8_release(ptr %a, i8 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_umin_i8_release:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -32

 ; RV32I-NEXT:    mv a0, a3
 ; RV32I-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s2, 16(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 32
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_umin_i8_release:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    li a3, 255
-; RV32IA-NEXT:    sll a3, a3, a0
-; RV32IA-NEXT:    andi a1, a1, 255
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB52_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w a4, (a2)
-; RV32IA-NEXT:    and a6, a4, a3
-; RV32IA-NEXT:    mv a5, a4
-; RV32IA-NEXT:    bgeu a1, a6, .LBB52_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB52_1 Depth=1
-; RV32IA-NEXT:    xor a5, a4, a1
-; RV32IA-NEXT:    and a5, a5, a3
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:  .LBB52_3: # in Loop: Header=BB52_1 Depth=1
-; RV32IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB52_1
-; RV32IA-NEXT:  # %bb.4:
-; RV32IA-NEXT:    srl a0, a4, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_umin_i8_release:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    li a3, 255
+; RV32IA-WMO-NEXT:    sll a3, a3, a0
+; RV32IA-WMO-NEXT:    andi a1, a1, 255
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB52_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w a4, (a2)
+; RV32IA-WMO-NEXT:    and a6, a4, a3
+; RV32IA-WMO-NEXT:    mv a5, a4
+; RV32IA-WMO-NEXT:    bgeu a1, a6, .LBB52_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB52_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a5, a4, a1
+; RV32IA-WMO-NEXT:    and a5, a5, a3
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:  .LBB52_3: # in Loop: Header=BB52_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB52_1
+; RV32IA-WMO-NEXT:  # %bb.4:
+; RV32IA-WMO-NEXT:    srl a0, a4, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_umin_i8_release:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    li a3, 255
+; RV32IA-TSO-NEXT:    sll a3, a3, a0
+; RV32IA-TSO-NEXT:    andi a1, a1, 255
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB52_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a4, (a2)
+; RV32IA-TSO-NEXT:    and a6, a4, a3
+; RV32IA-TSO-NEXT:    mv a5, a4
+; RV32IA-TSO-NEXT:    bgeu a1, a6, .LBB52_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB52_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a5, a4, a1
+; RV32IA-TSO-NEXT:    and a5, a5, a3
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:  .LBB52_3: # in Loop: Header=BB52_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB52_1
+; RV32IA-TSO-NEXT:  # %bb.4:
+; RV32IA-TSO-NEXT:    srl a0, a4, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_umin_i8_release:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -48
 ; RV64I-NEXT:    sd ra, 40(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s0, 32(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s1, 24(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s2, 16(sp) # 8-byte Folded Spill

 ; RV64I-NEXT:    mv a0, a3
 ; RV64I-NEXT:    ld ra, 40(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s0, 32(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s1, 24(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s2, 16(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 48
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_umin_i8_release:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    li a3, 255
-; RV64IA-NEXT:    sllw a3, a3, a0
-; RV64IA-NEXT:    andi a1, a1, 255
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB52_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w a4, (a2)
-; RV64IA-NEXT:    and a6, a4, a3
-; RV64IA-NEXT:    mv a5, a4
-; RV64IA-NEXT:    bgeu a1, a6, .LBB52_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB52_1 Depth=1
-; RV64IA-NEXT:    xor a5, a4, a1
-; RV64IA-NEXT:    and a5, a5, a3
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:  .LBB52_3: # in Loop: Header=BB52_1 Depth=1
-; RV64IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB52_1
-; RV64IA-NEXT:  # %bb.4:
-; RV64IA-NEXT:    srlw a0, a4, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_umin_i8_release:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    li a3, 255
+; RV64IA-WMO-NEXT:    sllw a3, a3, a0
+; RV64IA-WMO-NEXT:    andi a1, a1, 255
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB52_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w a4, (a2)
+; RV64IA-WMO-NEXT:    and a6, a4, a3
+; RV64IA-WMO-NEXT:    mv a5, a4
+; RV64IA-WMO-NEXT:    bgeu a1, a6, .LBB52_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB52_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a5, a4, a1
+; RV64IA-WMO-NEXT:    and a5, a5, a3
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:  .LBB52_3: # in Loop: Header=BB52_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB52_1
+; RV64IA-WMO-NEXT:  # %bb.4:
+; RV64IA-WMO-NEXT:    srlw a0, a4, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_umin_i8_release:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    li a3, 255
+; RV64IA-TSO-NEXT:    sllw a3, a3, a0
+; RV64IA-TSO-NEXT:    andi a1, a1, 255
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB52_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a4, (a2)
+; RV64IA-TSO-NEXT:    and a6, a4, a3
+; RV64IA-TSO-NEXT:    mv a5, a4
+; RV64IA-TSO-NEXT:    bgeu a1, a6, .LBB52_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB52_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a5, a4, a1
+; RV64IA-TSO-NEXT:    and a5, a5, a3
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:  .LBB52_3: # in Loop: Header=BB52_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB52_1
+; RV64IA-TSO-NEXT:  # %bb.4:
+; RV64IA-TSO-NEXT:    srlw a0, a4, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw umin ptr %a, i8 %b release
   ret i8 %1
 }
 
 define i8 @atomicrmw_umin_i8_acq_rel(ptr %a, i8 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_umin_i8_acq_rel:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -32

 ; RV32I-NEXT:    mv a0, a3
 ; RV32I-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s2, 16(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 32
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_umin_i8_acq_rel:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    li a3, 255
-; RV32IA-NEXT:    sll a3, a3, a0
-; RV32IA-NEXT:    andi a1, a1, 255
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB53_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a4, (a2)
-; RV32IA-NEXT:    and a6, a4, a3
-; RV32IA-NEXT:    mv a5, a4
-; RV32IA-NEXT:    bgeu a1, a6, .LBB53_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB53_1 Depth=1
-; RV32IA-NEXT:    xor a5, a4, a1
-; RV32IA-NEXT:    and a5, a5, a3
-; RV32IA-NEXT:    xor a5, a4, a5
-; RV32IA-NEXT:  .LBB53_3: # in Loop: Header=BB53_1 Depth=1
-; RV32IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB53_1
-; RV32IA-NEXT:  # %bb.4:
-; RV32IA-NEXT:    srl a0, a4, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_umin_i8_acq_rel:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    li a3, 255
+; RV32IA-WMO-NEXT:    sll a3, a3, a0
+; RV32IA-WMO-NEXT:    andi a1, a1, 255
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB53_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a4, (a2)
+; RV32IA-WMO-NEXT:    and a6, a4, a3
+; RV32IA-WMO-NEXT:    mv a5, a4
+; RV32IA-WMO-NEXT:    bgeu a1, a6, .LBB53_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB53_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a5, a4, a1
+; RV32IA-WMO-NEXT:    and a5, a5, a3
+; RV32IA-WMO-NEXT:    xor a5, a4, a5
+; RV32IA-WMO-NEXT:  .LBB53_3: # in Loop: Header=BB53_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB53_1
+; RV32IA-WMO-NEXT:  # %bb.4:
+; RV32IA-WMO-NEXT:    srl a0, a4, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_umin_i8_acq_rel:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    li a3, 255
+; RV32IA-TSO-NEXT:    sll a3, a3, a0
+; RV32IA-TSO-NEXT:    andi a1, a1, 255
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB53_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a4, (a2)
+; RV32IA-TSO-NEXT:    and a6, a4, a3
+; RV32IA-TSO-NEXT:    mv a5, a4
+; RV32IA-TSO-NEXT:    bgeu a1, a6, .LBB53_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB53_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a5, a4, a1
+; RV32IA-TSO-NEXT:    and a5, a5, a3
+; RV32IA-TSO-NEXT:    xor a5, a4, a5
+; RV32IA-TSO-NEXT:  .LBB53_3: # in Loop: Header=BB53_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB53_1
+; RV32IA-TSO-NEXT:  # %bb.4:
+; RV32IA-TSO-NEXT:    srl a0, a4, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_umin_i8_acq_rel:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -48
 ; RV64I-NEXT:    sd ra, 40(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s0, 32(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s1, 24(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s2, 16(sp) # 8-byte Folded Spill

 ; RV64I-NEXT:    mv a0, a3
 ; RV64I-NEXT:    ld ra, 40(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s0, 32(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s1, 24(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s2, 16(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 48
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_umin_i8_acq_rel:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    li a3, 255
-; RV64IA-NEXT:    sllw a3, a3, a0
-; RV64IA-NEXT:    andi a1, a1, 255
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB53_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a4, (a2)
-; RV64IA-NEXT:    and a6, a4, a3
-; RV64IA-NEXT:    mv a5, a4
-; RV64IA-NEXT:    bgeu a1, a6, .LBB53_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB53_1 Depth=1
-; RV64IA-NEXT:    xor a5, a4, a1
-; RV64IA-NEXT:    and a5, a5, a3
-; RV64IA-NEXT:    xor a5, a4, a5
-; RV64IA-NEXT:  .LBB53_3: # in Loop: Header=BB53_1 Depth=1
-; RV64IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB53_1
-; RV64IA-NEXT:  # %bb.4:
-; RV64IA-NEXT:    srlw a0, a4, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_umin_i8_acq_rel:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    li a3, 255
+; RV64IA-WMO-NEXT:    sllw a3, a3, a0
+; RV64IA-WMO-NEXT:    andi a1, a1, 255
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB53_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a4, (a2)
+; RV64IA-WMO-NEXT:    and a6, a4, a3
+; RV64IA-WMO-NEXT:    mv a5, a4
+; RV64IA-WMO-NEXT:    bgeu a1, a6, .LBB53_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB53_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a5, a4, a1
+; RV64IA-WMO-NEXT:    and a5, a5, a3
+; RV64IA-WMO-NEXT:    xor a5, a4, a5
+; RV64IA-WMO-NEXT:  .LBB53_3: # in Loop: Header=BB53_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB53_1
+; RV64IA-WMO-NEXT:  # %bb.4:
+; RV64IA-WMO-NEXT:    srlw a0, a4, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_umin_i8_acq_rel:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    li a3, 255
+; RV64IA-TSO-NEXT:    sllw a3, a3, a0
+; RV64IA-TSO-NEXT:    andi a1, a1, 255
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB53_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a4, (a2)
+; RV64IA-TSO-NEXT:    and a6, a4, a3
+; RV64IA-TSO-NEXT:    mv a5, a4
+; RV64IA-TSO-NEXT:    bgeu a1, a6, .LBB53_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB53_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a5, a4, a1
+; RV64IA-TSO-NEXT:    and a5, a5, a3
+; RV64IA-TSO-NEXT:    xor a5, a4, a5
+; RV64IA-TSO-NEXT:  .LBB53_3: # in Loop: Header=BB53_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB53_1
+; RV64IA-TSO-NEXT:  # %bb.4:
+; RV64IA-TSO-NEXT:    srlw a0, a4, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw umin ptr %a, i8 %b acq_rel
   ret i8 %1
 }
 
 define i8 @atomicrmw_umin_i8_seq_cst(ptr %a, i8 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_umin_i8_seq_cst:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -32

 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    li a2, 2
 ; RV32I-NEXT:    call __atomic_exchange_2@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_xchg_i16_acquire:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    lui a3, 16
-; RV32IA-NEXT:    addi a3, a3, -1
-; RV32IA-NEXT:    sll a4, a3, a0
-; RV32IA-NEXT:    and a1, a1, a3
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB56_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a3, (a2)
-; RV32IA-NEXT:    mv a5, a1
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:    and a5, a5, a4
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:    sc.w a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB56_1
-; RV32IA-NEXT:  # %bb.2:
-; RV32IA-NEXT:    srl a0, a3, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_xchg_i16_acquire:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    lui a3, 16
+; RV32IA-WMO-NEXT:    addi a3, a3, -1
+; RV32IA-WMO-NEXT:    sll a4, a3, a0
+; RV32IA-WMO-NEXT:    and a1, a1, a3
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB56_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV32IA-WMO-NEXT:    mv a5, a1
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:    and a5, a5, a4
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB56_1
+; RV32IA-WMO-NEXT:  # %bb.2:
+; RV32IA-WMO-NEXT:    srl a0, a3, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_xchg_i16_acquire:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    lui a3, 16
+; RV32IA-TSO-NEXT:    addi a3, a3, -1
+; RV32IA-TSO-NEXT:    sll a4, a3, a0
+; RV32IA-TSO-NEXT:    and a1, a1, a3
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB56_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a2)
+; RV32IA-TSO-NEXT:    mv a5, a1
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:    and a5, a5, a4
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB56_1
+; RV32IA-TSO-NEXT:  # %bb.2:
+; RV32IA-TSO-NEXT:    srl a0, a3, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_xchg_i16_acquire:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 2
 ; RV64I-NEXT:    call __atomic_exchange_2@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_xchg_i16_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    lui a3, 16
-; RV64IA-NEXT:    addiw a3, a3, -1
-; RV64IA-NEXT:    sllw a4, a3, a0
-; RV64IA-NEXT:    and a1, a1, a3
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB56_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a3, (a2)
-; RV64IA-NEXT:    mv a5, a1
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:    and a5, a5, a4
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:    sc.w a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB56_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    srlw a0, a3, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_xchg_i16_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    lui a3, 16
+; RV64IA-WMO-NEXT:    addiw a3, a3, -1
+; RV64IA-WMO-NEXT:    sllw a4, a3, a0
+; RV64IA-WMO-NEXT:    and a1, a1, a3
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB56_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV64IA-WMO-NEXT:    mv a5, a1
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:    and a5, a5, a4
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB56_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    srlw a0, a3, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_xchg_i16_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    lui a3, 16
+; RV64IA-TSO-NEXT:    addiw a3, a3, -1
+; RV64IA-TSO-NEXT:    sllw a4, a3, a0
+; RV64IA-TSO-NEXT:    and a1, a1, a3
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB56_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a2)
+; RV64IA-TSO-NEXT:    mv a5, a1
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:    and a5, a5, a4
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB56_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    srlw a0, a3, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw xchg ptr %a, i16 %b acquire
   ret i16 %1
 }
 
 define i16 @atomicrmw_xchg_i16_release(ptr %a, i16 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_xchg_i16_release:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    li a2, 3
 ; RV32I-NEXT:    call __atomic_exchange_2@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_xchg_i16_release:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    lui a3, 16
-; RV32IA-NEXT:    addi a3, a3, -1
-; RV32IA-NEXT:    sll a4, a3, a0
-; RV32IA-NEXT:    and a1, a1, a3
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB57_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w a3, (a2)
-; RV32IA-NEXT:    mv a5, a1
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:    and a5, a5, a4
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB57_1
-; RV32IA-NEXT:  # %bb.2:
-; RV32IA-NEXT:    srl a0, a3, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_xchg_i16_release:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    lui a3, 16
+; RV32IA-WMO-NEXT:    addi a3, a3, -1
+; RV32IA-WMO-NEXT:    sll a4, a3, a0
+; RV32IA-WMO-NEXT:    and a1, a1, a3
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB57_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w a3, (a2)
+; RV32IA-WMO-NEXT:    mv a5, a1
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:    and a5, a5, a4
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB57_1
+; RV32IA-WMO-NEXT:  # %bb.2:
+; RV32IA-WMO-NEXT:    srl a0, a3, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_xchg_i16_release:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    lui a3, 16
+; RV32IA-TSO-NEXT:    addi a3, a3, -1
+; RV32IA-TSO-NEXT:    sll a4, a3, a0
+; RV32IA-TSO-NEXT:    and a1, a1, a3
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB57_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a2)
+; RV32IA-TSO-NEXT:    mv a5, a1
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:    and a5, a5, a4
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB57_1
+; RV32IA-TSO-NEXT:  # %bb.2:
+; RV32IA-TSO-NEXT:    srl a0, a3, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_xchg_i16_release:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 3
 ; RV64I-NEXT:    call __atomic_exchange_2@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_xchg_i16_release:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    lui a3, 16
-; RV64IA-NEXT:    addiw a3, a3, -1
-; RV64IA-NEXT:    sllw a4, a3, a0
-; RV64IA-NEXT:    and a1, a1, a3
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB57_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w a3, (a2)
-; RV64IA-NEXT:    mv a5, a1
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:    and a5, a5, a4
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB57_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    srlw a0, a3, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_xchg_i16_release:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    lui a3, 16
+; RV64IA-WMO-NEXT:    addiw a3, a3, -1
+; RV64IA-WMO-NEXT:    sllw a4, a3, a0
+; RV64IA-WMO-NEXT:    and a1, a1, a3
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB57_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w a3, (a2)
+; RV64IA-WMO-NEXT:    mv a5, a1
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:    and a5, a5, a4
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB57_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    srlw a0, a3, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_xchg_i16_release:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    lui a3, 16
+; RV64IA-TSO-NEXT:    addiw a3, a3, -1
+; RV64IA-TSO-NEXT:    sllw a4, a3, a0
+; RV64IA-TSO-NEXT:    and a1, a1, a3
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB57_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a2)
+; RV64IA-TSO-NEXT:    mv a5, a1
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:    and a5, a5, a4
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB57_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    srlw a0, a3, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw xchg ptr %a, i16 %b release
   ret i16 %1
 }
 
 define i16 @atomicrmw_xchg_i16_acq_rel(ptr %a, i16 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_xchg_i16_acq_rel:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    li a2, 4
 ; RV32I-NEXT:    call __atomic_exchange_2@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_xchg_i16_acq_rel:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    lui a3, 16
-; RV32IA-NEXT:    addi a3, a3, -1
-; RV32IA-NEXT:    sll a4, a3, a0
-; RV32IA-NEXT:    and a1, a1, a3
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB58_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a3, (a2)
-; RV32IA-NEXT:    mv a5, a1
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:    and a5, a5, a4
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB58_1
-; RV32IA-NEXT:  # %bb.2:
-; RV32IA-NEXT:    srl a0, a3, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_xchg_i16_acq_rel:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    lui a3, 16
+; RV32IA-WMO-NEXT:    addi a3, a3, -1
+; RV32IA-WMO-NEXT:    sll a4, a3, a0
+; RV32IA-WMO-NEXT:    and a1, a1, a3
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB58_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV32IA-WMO-NEXT:    mv a5, a1
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:    and a5, a5, a4
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB58_1
+; RV32IA-WMO-NEXT:  # %bb.2:
+; RV32IA-WMO-NEXT:    srl a0, a3, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_xchg_i16_acq_rel:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    lui a3, 16
+; RV32IA-TSO-NEXT:    addi a3, a3, -1
+; RV32IA-TSO-NEXT:    sll a4, a3, a0
+; RV32IA-TSO-NEXT:    and a1, a1, a3
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB58_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a2)
+; RV32IA-TSO-NEXT:    mv a5, a1
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:    and a5, a5, a4
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB58_1
+; RV32IA-TSO-NEXT:  # %bb.2:
+; RV32IA-TSO-NEXT:    srl a0, a3, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_xchg_i16_acq_rel:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 4
 ; RV64I-NEXT:    call __atomic_exchange_2@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_xchg_i16_acq_rel:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    lui a3, 16
-; RV64IA-NEXT:    addiw a3, a3, -1
-; RV64IA-NEXT:    sllw a4, a3, a0
-; RV64IA-NEXT:    and a1, a1, a3
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB58_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a3, (a2)
-; RV64IA-NEXT:    mv a5, a1
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:    and a5, a5, a4
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB58_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    srlw a0, a3, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_xchg_i16_acq_rel:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    lui a3, 16
+; RV64IA-WMO-NEXT:    addiw a3, a3, -1
+; RV64IA-WMO-NEXT:    sllw a4, a3, a0
+; RV64IA-WMO-NEXT:    and a1, a1, a3
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB58_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV64IA-WMO-NEXT:    mv a5, a1
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:    and a5, a5, a4
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB58_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    srlw a0, a3, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_xchg_i16_acq_rel:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    lui a3, 16
+; RV64IA-TSO-NEXT:    addiw a3, a3, -1
+; RV64IA-TSO-NEXT:    sllw a4, a3, a0
+; RV64IA-TSO-NEXT:    and a1, a1, a3
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB58_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a2)
+; RV64IA-TSO-NEXT:    mv a5, a1
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:    and a5, a5, a4
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB58_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    srlw a0, a3, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw xchg ptr %a, i16 %b acq_rel
   ret i16 %1
 }
 
 define i16 @atomicrmw_xchg_i16_seq_cst(ptr %a, i16 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_xchg_i16_seq_cst:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16

 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    li a2, 2
 ; RV32I-NEXT:    call __atomic_fetch_add_2@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_add_i16_acquire:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    lui a3, 16
-; RV32IA-NEXT:    addi a3, a3, -1
-; RV32IA-NEXT:    sll a4, a3, a0
-; RV32IA-NEXT:    and a1, a1, a3
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB61_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a3, (a2)
-; RV32IA-NEXT:    add a5, a3, a1
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:    and a5, a5, a4
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:    sc.w a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB61_1
-; RV32IA-NEXT:  # %bb.2:
-; RV32IA-NEXT:    srl a0, a3, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_add_i16_acquire:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    lui a3, 16
+; RV32IA-WMO-NEXT:    addi a3, a3, -1
+; RV32IA-WMO-NEXT:    sll a4, a3, a0
+; RV32IA-WMO-NEXT:    and a1, a1, a3
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB61_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV32IA-WMO-NEXT:    add a5, a3, a1
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:    and a5, a5, a4
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB61_1
+; RV32IA-WMO-NEXT:  # %bb.2:
+; RV32IA-WMO-NEXT:    srl a0, a3, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_add_i16_acquire:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    lui a3, 16
+; RV32IA-TSO-NEXT:    addi a3, a3, -1
+; RV32IA-TSO-NEXT:    sll a4, a3, a0
+; RV32IA-TSO-NEXT:    and a1, a1, a3
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB61_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a2)
+; RV32IA-TSO-NEXT:    add a5, a3, a1
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:    and a5, a5, a4
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB61_1
+; RV32IA-TSO-NEXT:  # %bb.2:
+; RV32IA-TSO-NEXT:    srl a0, a3, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_add_i16_acquire:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 2
 ; RV64I-NEXT:    call __atomic_fetch_add_2@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_add_i16_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    lui a3, 16
-; RV64IA-NEXT:    addiw a3, a3, -1
-; RV64IA-NEXT:    sllw a4, a3, a0
-; RV64IA-NEXT:    and a1, a1, a3
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB61_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a3, (a2)
-; RV64IA-NEXT:    add a5, a3, a1
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:    and a5, a5, a4
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:    sc.w a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB61_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    srlw a0, a3, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_add_i16_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    lui a3, 16
+; RV64IA-WMO-NEXT:    addiw a3, a3, -1
+; RV64IA-WMO-NEXT:    sllw a4, a3, a0
+; RV64IA-WMO-NEXT:    and a1, a1, a3
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB61_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV64IA-WMO-NEXT:    add a5, a3, a1
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:    and a5, a5, a4
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB61_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    srlw a0, a3, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_add_i16_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    lui a3, 16
+; RV64IA-TSO-NEXT:    addiw a3, a3, -1
+; RV64IA-TSO-NEXT:    sllw a4, a3, a0
+; RV64IA-TSO-NEXT:    and a1, a1, a3
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB61_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a2)
+; RV64IA-TSO-NEXT:    add a5, a3, a1
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:    and a5, a5, a4
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB61_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    srlw a0, a3, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw add ptr %a, i16 %b acquire
   ret i16 %1
 }
 
 define i16 @atomicrmw_add_i16_release(ptr %a, i16 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_add_i16_release:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    li a2, 3
 ; RV32I-NEXT:    call __atomic_fetch_add_2@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_add_i16_release:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    lui a3, 16
-; RV32IA-NEXT:    addi a3, a3, -1
-; RV32IA-NEXT:    sll a4, a3, a0
-; RV32IA-NEXT:    and a1, a1, a3
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB62_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w a3, (a2)
-; RV32IA-NEXT:    add a5, a3, a1
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:    and a5, a5, a4
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB62_1
-; RV32IA-NEXT:  # %bb.2:
-; RV32IA-NEXT:    srl a0, a3, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_add_i16_release:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    lui a3, 16
+; RV32IA-WMO-NEXT:    addi a3, a3, -1
+; RV32IA-WMO-NEXT:    sll a4, a3, a0
+; RV32IA-WMO-NEXT:    and a1, a1, a3
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB62_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w a3, (a2)
+; RV32IA-WMO-NEXT:    add a5, a3, a1
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:    and a5, a5, a4
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB62_1
+; RV32IA-WMO-NEXT:  # %bb.2:
+; RV32IA-WMO-NEXT:    srl a0, a3, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_add_i16_release:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    lui a3, 16
+; RV32IA-TSO-NEXT:    addi a3, a3, -1
+; RV32IA-TSO-NEXT:    sll a4, a3, a0
+; RV32IA-TSO-NEXT:    and a1, a1, a3
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB62_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a2)
+; RV32IA-TSO-NEXT:    add a5, a3, a1
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:    and a5, a5, a4
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB62_1
+; RV32IA-TSO-NEXT:  # %bb.2:
+; RV32IA-TSO-NEXT:    srl a0, a3, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_add_i16_release:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 3
 ; RV64I-NEXT:    call __atomic_fetch_add_2@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_add_i16_release:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    lui a3, 16
-; RV64IA-NEXT:    addiw a3, a3, -1
-; RV64IA-NEXT:    sllw a4, a3, a0
-; RV64IA-NEXT:    and a1, a1, a3
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB62_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w a3, (a2)
-; RV64IA-NEXT:    add a5, a3, a1
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:    and a5, a5, a4
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB62_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    srlw a0, a3, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_add_i16_release:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    lui a3, 16
+; RV64IA-WMO-NEXT:    addiw a3, a3, -1
+; RV64IA-WMO-NEXT:    sllw a4, a3, a0
+; RV64IA-WMO-NEXT:    and a1, a1, a3
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB62_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w a3, (a2)
+; RV64IA-WMO-NEXT:    add a5, a3, a1
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:    and a5, a5, a4
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB62_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    srlw a0, a3, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_add_i16_release:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    lui a3, 16
+; RV64IA-TSO-NEXT:    addiw a3, a3, -1
+; RV64IA-TSO-NEXT:    sllw a4, a3, a0
+; RV64IA-TSO-NEXT:    and a1, a1, a3
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB62_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a2)
+; RV64IA-TSO-NEXT:    add a5, a3, a1
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:    and a5, a5, a4
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB62_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    srlw a0, a3, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw add ptr %a, i16 %b release
   ret i16 %1
 }
 
 define i16 @atomicrmw_add_i16_acq_rel(ptr %a, i16 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_add_i16_acq_rel:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    li a2, 4
 ; RV32I-NEXT:    call __atomic_fetch_add_2@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_add_i16_acq_rel:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    lui a3, 16
-; RV32IA-NEXT:    addi a3, a3, -1
-; RV32IA-NEXT:    sll a4, a3, a0
-; RV32IA-NEXT:    and a1, a1, a3
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB63_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a3, (a2)
-; RV32IA-NEXT:    add a5, a3, a1
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:    and a5, a5, a4
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB63_1
-; RV32IA-NEXT:  # %bb.2:
-; RV32IA-NEXT:    srl a0, a3, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_add_i16_acq_rel:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    lui a3, 16
+; RV32IA-WMO-NEXT:    addi a3, a3, -1
+; RV32IA-WMO-NEXT:    sll a4, a3, a0
+; RV32IA-WMO-NEXT:    and a1, a1, a3
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB63_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV32IA-WMO-NEXT:    add a5, a3, a1
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:    and a5, a5, a4
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB63_1
+; RV32IA-WMO-NEXT:  # %bb.2:
+; RV32IA-WMO-NEXT:    srl a0, a3, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_add_i16_acq_rel:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    lui a3, 16
+; RV32IA-TSO-NEXT:    addi a3, a3, -1
+; RV32IA-TSO-NEXT:    sll a4, a3, a0
+; RV32IA-TSO-NEXT:    and a1, a1, a3
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB63_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a2)
+; RV32IA-TSO-NEXT:    add a5, a3, a1
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:    and a5, a5, a4
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB63_1
+; RV32IA-TSO-NEXT:  # %bb.2:
+; RV32IA-TSO-NEXT:    srl a0, a3, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_add_i16_acq_rel:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 4
 ; RV64I-NEXT:    call __atomic_fetch_add_2@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_add_i16_acq_rel:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    lui a3, 16
-; RV64IA-NEXT:    addiw a3, a3, -1
-; RV64IA-NEXT:    sllw a4, a3, a0
-; RV64IA-NEXT:    and a1, a1, a3
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB63_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a3, (a2)
-; RV64IA-NEXT:    add a5, a3, a1
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:    and a5, a5, a4
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB63_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    srlw a0, a3, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_add_i16_acq_rel:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    lui a3, 16
+; RV64IA-WMO-NEXT:    addiw a3, a3, -1
+; RV64IA-WMO-NEXT:    sllw a4, a3, a0
+; RV64IA-WMO-NEXT:    and a1, a1, a3
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB63_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV64IA-WMO-NEXT:    add a5, a3, a1
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:    and a5, a5, a4
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB63_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    srlw a0, a3, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_add_i16_acq_rel:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    lui a3, 16
+; RV64IA-TSO-NEXT:    addiw a3, a3, -1
+; RV64IA-TSO-NEXT:    sllw a4, a3, a0
+; RV64IA-TSO-NEXT:    and a1, a1, a3
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB63_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a2)
+; RV64IA-TSO-NEXT:    add a5, a3, a1
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:    and a5, a5, a4
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB63_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    srlw a0, a3, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw add ptr %a, i16 %b acq_rel
   ret i16 %1
 }
 
 define i16 @atomicrmw_add_i16_seq_cst(ptr %a, i16 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_add_i16_seq_cst:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16

 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    li a2, 2
 ; RV32I-NEXT:    call __atomic_fetch_sub_2@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_sub_i16_acquire:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    lui a3, 16
-; RV32IA-NEXT:    addi a3, a3, -1
-; RV32IA-NEXT:    sll a4, a3, a0
-; RV32IA-NEXT:    and a1, a1, a3
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB66_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a3, (a2)
-; RV32IA-NEXT:    sub a5, a3, a1
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:    and a5, a5, a4
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:    sc.w a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB66_1
-; RV32IA-NEXT:  # %bb.2:
-; RV32IA-NEXT:    srl a0, a3, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_sub_i16_acquire:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    lui a3, 16
+; RV32IA-WMO-NEXT:    addi a3, a3, -1
+; RV32IA-WMO-NEXT:    sll a4, a3, a0
+; RV32IA-WMO-NEXT:    and a1, a1, a3
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB66_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV32IA-WMO-NEXT:    sub a5, a3, a1
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:    and a5, a5, a4
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB66_1
+; RV32IA-WMO-NEXT:  # %bb.2:
+; RV32IA-WMO-NEXT:    srl a0, a3, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_sub_i16_acquire:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    lui a3, 16
+; RV32IA-TSO-NEXT:    addi a3, a3, -1
+; RV32IA-TSO-NEXT:    sll a4, a3, a0
+; RV32IA-TSO-NEXT:    and a1, a1, a3
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB66_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a2)
+; RV32IA-TSO-NEXT:    sub a5, a3, a1
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:    and a5, a5, a4
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB66_1
+; RV32IA-TSO-NEXT:  # %bb.2:
+; RV32IA-TSO-NEXT:    srl a0, a3, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_sub_i16_acquire:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 2
 ; RV64I-NEXT:    call __atomic_fetch_sub_2@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_sub_i16_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    lui a3, 16
-; RV64IA-NEXT:    addiw a3, a3, -1
-; RV64IA-NEXT:    sllw a4, a3, a0
-; RV64IA-NEXT:    and a1, a1, a3
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB66_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a3, (a2)
-; RV64IA-NEXT:    sub a5, a3, a1
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:    and a5, a5, a4
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:    sc.w a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB66_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    srlw a0, a3, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_sub_i16_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    lui a3, 16
+; RV64IA-WMO-NEXT:    addiw a3, a3, -1
+; RV64IA-WMO-NEXT:    sllw a4, a3, a0
+; RV64IA-WMO-NEXT:    and a1, a1, a3
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB66_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV64IA-WMO-NEXT:    sub a5, a3, a1
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:    and a5, a5, a4
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB66_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    srlw a0, a3, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_sub_i16_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    lui a3, 16
+; RV64IA-TSO-NEXT:    addiw a3, a3, -1
+; RV64IA-TSO-NEXT:    sllw a4, a3, a0
+; RV64IA-TSO-NEXT:    and a1, a1, a3
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB66_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a2)
+; RV64IA-TSO-NEXT:    sub a5, a3, a1
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:    and a5, a5, a4
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB66_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    srlw a0, a3, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw sub ptr %a, i16 %b acquire
   ret i16 %1
 }
 
 define i16 @atomicrmw_sub_i16_release(ptr %a, i16 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_sub_i16_release:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    li a2, 3
 ; RV32I-NEXT:    call __atomic_fetch_sub_2@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_sub_i16_release:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    lui a3, 16
-; RV32IA-NEXT:    addi a3, a3, -1
-; RV32IA-NEXT:    sll a4, a3, a0
-; RV32IA-NEXT:    and a1, a1, a3
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB67_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w a3, (a2)
-; RV32IA-NEXT:    sub a5, a3, a1
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:    and a5, a5, a4
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB67_1
-; RV32IA-NEXT:  # %bb.2:
-; RV32IA-NEXT:    srl a0, a3, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_sub_i16_release:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    lui a3, 16
+; RV32IA-WMO-NEXT:    addi a3, a3, -1
+; RV32IA-WMO-NEXT:    sll a4, a3, a0
+; RV32IA-WMO-NEXT:    and a1, a1, a3
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB67_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w a3, (a2)
+; RV32IA-WMO-NEXT:    sub a5, a3, a1
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:    and a5, a5, a4
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB67_1
+; RV32IA-WMO-NEXT:  # %bb.2:
+; RV32IA-WMO-NEXT:    srl a0, a3, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_sub_i16_release:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    lui a3, 16
+; RV32IA-TSO-NEXT:    addi a3, a3, -1
+; RV32IA-TSO-NEXT:    sll a4, a3, a0
+; RV32IA-TSO-NEXT:    and a1, a1, a3
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB67_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a2)
+; RV32IA-TSO-NEXT:    sub a5, a3, a1
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:    and a5, a5, a4
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB67_1
+; RV32IA-TSO-NEXT:  # %bb.2:
+; RV32IA-TSO-NEXT:    srl a0, a3, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_sub_i16_release:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 3
 ; RV64I-NEXT:    call __atomic_fetch_sub_2@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_sub_i16_release:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    lui a3, 16
-; RV64IA-NEXT:    addiw a3, a3, -1
-; RV64IA-NEXT:    sllw a4, a3, a0
-; RV64IA-NEXT:    and a1, a1, a3
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB67_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w a3, (a2)
-; RV64IA-NEXT:    sub a5, a3, a1
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:    and a5, a5, a4
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB67_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    srlw a0, a3, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_sub_i16_release:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    lui a3, 16
+; RV64IA-WMO-NEXT:    addiw a3, a3, -1
+; RV64IA-WMO-NEXT:    sllw a4, a3, a0
+; RV64IA-WMO-NEXT:    and a1, a1, a3
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB67_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w a3, (a2)
+; RV64IA-WMO-NEXT:    sub a5, a3, a1
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:    and a5, a5, a4
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB67_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    srlw a0, a3, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_sub_i16_release:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    lui a3, 16
+; RV64IA-TSO-NEXT:    addiw a3, a3, -1
+; RV64IA-TSO-NEXT:    sllw a4, a3, a0
+; RV64IA-TSO-NEXT:    and a1, a1, a3
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB67_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a2)
+; RV64IA-TSO-NEXT:    sub a5, a3, a1
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:    and a5, a5, a4
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB67_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    srlw a0, a3, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw sub ptr %a, i16 %b release
   ret i16 %1
 }
 
 define i16 @atomicrmw_sub_i16_acq_rel(ptr %a, i16 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_sub_i16_acq_rel:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    li a2, 4
 ; RV32I-NEXT:    call __atomic_fetch_sub_2@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_sub_i16_acq_rel:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    lui a3, 16
-; RV32IA-NEXT:    addi a3, a3, -1
-; RV32IA-NEXT:    sll a4, a3, a0
-; RV32IA-NEXT:    and a1, a1, a3
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB68_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a3, (a2)
-; RV32IA-NEXT:    sub a5, a3, a1
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:    and a5, a5, a4
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB68_1
-; RV32IA-NEXT:  # %bb.2:
-; RV32IA-NEXT:    srl a0, a3, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_sub_i16_acq_rel:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    lui a3, 16
+; RV32IA-WMO-NEXT:    addi a3, a3, -1
+; RV32IA-WMO-NEXT:    sll a4, a3, a0
+; RV32IA-WMO-NEXT:    and a1, a1, a3
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB68_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV32IA-WMO-NEXT:    sub a5, a3, a1
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:    and a5, a5, a4
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB68_1
+; RV32IA-WMO-NEXT:  # %bb.2:
+; RV32IA-WMO-NEXT:    srl a0, a3, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_sub_i16_acq_rel:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    lui a3, 16
+; RV32IA-TSO-NEXT:    addi a3, a3, -1
+; RV32IA-TSO-NEXT:    sll a4, a3, a0
+; RV32IA-TSO-NEXT:    and a1, a1, a3
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB68_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a2)
+; RV32IA-TSO-NEXT:    sub a5, a3, a1
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:    and a5, a5, a4
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB68_1
+; RV32IA-TSO-NEXT:  # %bb.2:
+; RV32IA-TSO-NEXT:    srl a0, a3, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_sub_i16_acq_rel:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 4
 ; RV64I-NEXT:    call __atomic_fetch_sub_2@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_sub_i16_acq_rel:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    lui a3, 16
-; RV64IA-NEXT:    addiw a3, a3, -1
-; RV64IA-NEXT:    sllw a4, a3, a0
-; RV64IA-NEXT:    and a1, a1, a3
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB68_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a3, (a2)
-; RV64IA-NEXT:    sub a5, a3, a1
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:    and a5, a5, a4
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB68_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    srlw a0, a3, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_sub_i16_acq_rel:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    lui a3, 16
+; RV64IA-WMO-NEXT:    addiw a3, a3, -1
+; RV64IA-WMO-NEXT:    sllw a4, a3, a0
+; RV64IA-WMO-NEXT:    and a1, a1, a3
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB68_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV64IA-WMO-NEXT:    sub a5, a3, a1
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:    and a5, a5, a4
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB68_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    srlw a0, a3, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_sub_i16_acq_rel:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    lui a3, 16
+; RV64IA-TSO-NEXT:    addiw a3, a3, -1
+; RV64IA-TSO-NEXT:    sllw a4, a3, a0
+; RV64IA-TSO-NEXT:    and a1, a1, a3
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB68_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a2)
+; RV64IA-TSO-NEXT:    sub a5, a3, a1
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:    and a5, a5, a4
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB68_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    srlw a0, a3, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw sub ptr %a, i16 %b acq_rel
   ret i16 %1
 }
 
 define i16 @atomicrmw_sub_i16_seq_cst(ptr %a, i16 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_sub_i16_seq_cst:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16

 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    li a2, 2
 ; RV32I-NEXT:    call __atomic_fetch_nand_2@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_nand_i16_acquire:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    lui a3, 16
-; RV32IA-NEXT:    addi a3, a3, -1
-; RV32IA-NEXT:    sll a4, a3, a0
-; RV32IA-NEXT:    and a1, a1, a3
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB76_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a3, (a2)
-; RV32IA-NEXT:    and a5, a3, a1
-; RV32IA-NEXT:    not a5, a5
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:    and a5, a5, a4
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:    sc.w a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB76_1
-; RV32IA-NEXT:  # %bb.2:
-; RV32IA-NEXT:    srl a0, a3, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_nand_i16_acquire:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    lui a3, 16
+; RV32IA-WMO-NEXT:    addi a3, a3, -1
+; RV32IA-WMO-NEXT:    sll a4, a3, a0
+; RV32IA-WMO-NEXT:    and a1, a1, a3
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB76_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV32IA-WMO-NEXT:    and a5, a3, a1
+; RV32IA-WMO-NEXT:    not a5, a5
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:    and a5, a5, a4
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB76_1
+; RV32IA-WMO-NEXT:  # %bb.2:
+; RV32IA-WMO-NEXT:    srl a0, a3, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_nand_i16_acquire:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    lui a3, 16
+; RV32IA-TSO-NEXT:    addi a3, a3, -1
+; RV32IA-TSO-NEXT:    sll a4, a3, a0
+; RV32IA-TSO-NEXT:    and a1, a1, a3
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB76_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a2)
+; RV32IA-TSO-NEXT:    and a5, a3, a1
+; RV32IA-TSO-NEXT:    not a5, a5
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:    and a5, a5, a4
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB76_1
+; RV32IA-TSO-NEXT:  # %bb.2:
+; RV32IA-TSO-NEXT:    srl a0, a3, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_nand_i16_acquire:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 2
 ; RV64I-NEXT:    call __atomic_fetch_nand_2@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_nand_i16_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    lui a3, 16
-; RV64IA-NEXT:    addiw a3, a3, -1
-; RV64IA-NEXT:    sllw a4, a3, a0
-; RV64IA-NEXT:    and a1, a1, a3
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB76_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a3, (a2)
-; RV64IA-NEXT:    and a5, a3, a1
-; RV64IA-NEXT:    not a5, a5
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:    and a5, a5, a4
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:    sc.w a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB76_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    srlw a0, a3, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_nand_i16_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    lui a3, 16
+; RV64IA-WMO-NEXT:    addiw a3, a3, -1
+; RV64IA-WMO-NEXT:    sllw a4, a3, a0
+; RV64IA-WMO-NEXT:    and a1, a1, a3
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB76_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV64IA-WMO-NEXT:    and a5, a3, a1
+; RV64IA-WMO-NEXT:    not a5, a5
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:    and a5, a5, a4
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB76_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    srlw a0, a3, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_nand_i16_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    lui a3, 16
+; RV64IA-TSO-NEXT:    addiw a3, a3, -1
+; RV64IA-TSO-NEXT:    sllw a4, a3, a0
+; RV64IA-TSO-NEXT:    and a1, a1, a3
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB76_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a2)
+; RV64IA-TSO-NEXT:    and a5, a3, a1
+; RV64IA-TSO-NEXT:    not a5, a5
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:    and a5, a5, a4
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB76_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    srlw a0, a3, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw nand ptr %a, i16 %b acquire
   ret i16 %1
 }
 
 define i16 @atomicrmw_nand_i16_release(ptr %a, i16 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_nand_i16_release:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    li a2, 3
 ; RV32I-NEXT:    call __atomic_fetch_nand_2@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_nand_i16_release:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    lui a3, 16
-; RV32IA-NEXT:    addi a3, a3, -1
-; RV32IA-NEXT:    sll a4, a3, a0
-; RV32IA-NEXT:    and a1, a1, a3
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB77_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w a3, (a2)
-; RV32IA-NEXT:    and a5, a3, a1
-; RV32IA-NEXT:    not a5, a5
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:    and a5, a5, a4
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB77_1
-; RV32IA-NEXT:  # %bb.2:
-; RV32IA-NEXT:    srl a0, a3, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_nand_i16_release:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    lui a3, 16
+; RV32IA-WMO-NEXT:    addi a3, a3, -1
+; RV32IA-WMO-NEXT:    sll a4, a3, a0
+; RV32IA-WMO-NEXT:    and a1, a1, a3
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB77_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w a3, (a2)
+; RV32IA-WMO-NEXT:    and a5, a3, a1
+; RV32IA-WMO-NEXT:    not a5, a5
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:    and a5, a5, a4
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB77_1
+; RV32IA-WMO-NEXT:  # %bb.2:
+; RV32IA-WMO-NEXT:    srl a0, a3, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_nand_i16_release:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    lui a3, 16
+; RV32IA-TSO-NEXT:    addi a3, a3, -1
+; RV32IA-TSO-NEXT:    sll a4, a3, a0
+; RV32IA-TSO-NEXT:    and a1, a1, a3
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB77_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a2)
+; RV32IA-TSO-NEXT:    and a5, a3, a1
+; RV32IA-TSO-NEXT:    not a5, a5
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:    and a5, a5, a4
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB77_1
+; RV32IA-TSO-NEXT:  # %bb.2:
+; RV32IA-TSO-NEXT:    srl a0, a3, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_nand_i16_release:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 3
 ; RV64I-NEXT:    call __atomic_fetch_nand_2@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_nand_i16_release:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    lui a3, 16
-; RV64IA-NEXT:    addiw a3, a3, -1
-; RV64IA-NEXT:    sllw a4, a3, a0
-; RV64IA-NEXT:    and a1, a1, a3
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB77_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w a3, (a2)
-; RV64IA-NEXT:    and a5, a3, a1
-; RV64IA-NEXT:    not a5, a5
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:    and a5, a5, a4
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB77_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    srlw a0, a3, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_nand_i16_release:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    lui a3, 16
+; RV64IA-WMO-NEXT:    addiw a3, a3, -1
+; RV64IA-WMO-NEXT:    sllw a4, a3, a0
+; RV64IA-WMO-NEXT:    and a1, a1, a3
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB77_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w a3, (a2)
+; RV64IA-WMO-NEXT:    and a5, a3, a1
+; RV64IA-WMO-NEXT:    not a5, a5
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:    and a5, a5, a4
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB77_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    srlw a0, a3, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_nand_i16_release:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    lui a3, 16
+; RV64IA-TSO-NEXT:    addiw a3, a3, -1
+; RV64IA-TSO-NEXT:    sllw a4, a3, a0
+; RV64IA-TSO-NEXT:    and a1, a1, a3
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB77_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a2)
+; RV64IA-TSO-NEXT:    and a5, a3, a1
+; RV64IA-TSO-NEXT:    not a5, a5
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:    and a5, a5, a4
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB77_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    srlw a0, a3, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw nand ptr %a, i16 %b release
   ret i16 %1
 }
 
 define i16 @atomicrmw_nand_i16_acq_rel(ptr %a, i16 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_nand_i16_acq_rel:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    li a2, 4
 ; RV32I-NEXT:    call __atomic_fetch_nand_2@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_nand_i16_acq_rel:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    lui a3, 16
-; RV32IA-NEXT:    addi a3, a3, -1
-; RV32IA-NEXT:    sll a4, a3, a0
-; RV32IA-NEXT:    and a1, a1, a3
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB78_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a3, (a2)
-; RV32IA-NEXT:    and a5, a3, a1
-; RV32IA-NEXT:    not a5, a5
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:    and a5, a5, a4
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB78_1
-; RV32IA-NEXT:  # %bb.2:
-; RV32IA-NEXT:    srl a0, a3, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_nand_i16_acq_rel:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    lui a3, 16
+; RV32IA-WMO-NEXT:    addi a3, a3, -1
+; RV32IA-WMO-NEXT:    sll a4, a3, a0
+; RV32IA-WMO-NEXT:    and a1, a1, a3
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB78_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV32IA-WMO-NEXT:    and a5, a3, a1
+; RV32IA-WMO-NEXT:    not a5, a5
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:    and a5, a5, a4
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB78_1
+; RV32IA-WMO-NEXT:  # %bb.2:
+; RV32IA-WMO-NEXT:    srl a0, a3, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_nand_i16_acq_rel:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    lui a3, 16
+; RV32IA-TSO-NEXT:    addi a3, a3, -1
+; RV32IA-TSO-NEXT:    sll a4, a3, a0
+; RV32IA-TSO-NEXT:    and a1, a1, a3
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB78_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a2)
+; RV32IA-TSO-NEXT:    and a5, a3, a1
+; RV32IA-TSO-NEXT:    not a5, a5
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:    and a5, a5, a4
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB78_1
+; RV32IA-TSO-NEXT:  # %bb.2:
+; RV32IA-TSO-NEXT:    srl a0, a3, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_nand_i16_acq_rel:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 4
 ; RV64I-NEXT:    call __atomic_fetch_nand_2@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_nand_i16_acq_rel:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    lui a3, 16
-; RV64IA-NEXT:    addiw a3, a3, -1
-; RV64IA-NEXT:    sllw a4, a3, a0
-; RV64IA-NEXT:    and a1, a1, a3
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB78_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a3, (a2)
-; RV64IA-NEXT:    and a5, a3, a1
-; RV64IA-NEXT:    not a5, a5
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:    and a5, a5, a4
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB78_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    srlw a0, a3, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_nand_i16_acq_rel:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    lui a3, 16
+; RV64IA-WMO-NEXT:    addiw a3, a3, -1
+; RV64IA-WMO-NEXT:    sllw a4, a3, a0
+; RV64IA-WMO-NEXT:    and a1, a1, a3
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB78_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV64IA-WMO-NEXT:    and a5, a3, a1
+; RV64IA-WMO-NEXT:    not a5, a5
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:    and a5, a5, a4
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB78_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    srlw a0, a3, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_nand_i16_acq_rel:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    lui a3, 16
+; RV64IA-TSO-NEXT:    addiw a3, a3, -1
+; RV64IA-TSO-NEXT:    sllw a4, a3, a0
+; RV64IA-TSO-NEXT:    and a1, a1, a3
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB78_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a2)
+; RV64IA-TSO-NEXT:    and a5, a3, a1
+; RV64IA-TSO-NEXT:    not a5, a5
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:    and a5, a5, a4
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB78_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    srlw a0, a3, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw nand ptr %a, i16 %b acq_rel
   ret i16 %1
 }
 
 define i16 @atomicrmw_nand_i16_seq_cst(ptr %a, i16 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_nand_i16_seq_cst:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16

 ; RV32I-NEXT:    mv a0, a3
 ; RV32I-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s2, 16(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 32
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_max_i16_acquire:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    andi a3, a0, 24
-; RV32IA-NEXT:    lui a4, 16
-; RV32IA-NEXT:    addi a4, a4, -1
-; RV32IA-NEXT:    sll a4, a4, a0
-; RV32IA-NEXT:    slli a1, a1, 16
-; RV32IA-NEXT:    srai a1, a1, 16
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:    li a5, 16
-; RV32IA-NEXT:    sub a5, a5, a3
-; RV32IA-NEXT:  .LBB91_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a3, (a2)
-; RV32IA-NEXT:    and a7, a3, a4
-; RV32IA-NEXT:    mv a6, a3
-; RV32IA-NEXT:    sll a7, a7, a5
-; RV32IA-NEXT:    sra a7, a7, a5
-; RV32IA-NEXT:    bge a7, a1, .LBB91_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB91_1 Depth=1
-; RV32IA-NEXT:    xor a6, a3, a1
-; RV32IA-NEXT:    and a6, a6, a4
-; RV32IA-NEXT:    xor a6, a3, a6
-; RV32IA-NEXT:  .LBB91_3: # in Loop: Header=BB91_1 Depth=1
-; RV32IA-NEXT:    sc.w a6, a6, (a2)
-; RV32IA-NEXT:    bnez a6, .LBB91_1
-; RV32IA-NEXT:  # %bb.4:
-; RV32IA-NEXT:    srl a0, a3, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_max_i16_acquire:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    andi a3, a0, 24
+; RV32IA-WMO-NEXT:    lui a4, 16
+; RV32IA-WMO-NEXT:    addi a4, a4, -1
+; RV32IA-WMO-NEXT:    sll a4, a4, a0
+; RV32IA-WMO-NEXT:    slli a1, a1, 16
+; RV32IA-WMO-NEXT:    srai a1, a1, 16
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:    li a5, 16
+; RV32IA-WMO-NEXT:    sub a5, a5, a3
+; RV32IA-WMO-NEXT:  .LBB91_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV32IA-WMO-NEXT:    and a7, a3, a4
+; RV32IA-WMO-NEXT:    mv a6, a3
+; RV32IA-WMO-NEXT:    sll a7, a7, a5
+; RV32IA-WMO-NEXT:    sra a7, a7, a5
+; RV32IA-WMO-NEXT:    bge a7, a1, .LBB91_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB91_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a6, a3, a1
+; RV32IA-WMO-NEXT:    and a6, a6, a4
+; RV32IA-WMO-NEXT:    xor a6, a3, a6
+; RV32IA-WMO-NEXT:  .LBB91_3: # in Loop: Header=BB91_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w a6, a6, (a2)
+; RV32IA-WMO-NEXT:    bnez a6, .LBB91_1
+; RV32IA-WMO-NEXT:  # %bb.4:
+; RV32IA-WMO-NEXT:    srl a0, a3, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_max_i16_acquire:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    andi a3, a0, 24
+; RV32IA-TSO-NEXT:    lui a4, 16
+; RV32IA-TSO-NEXT:    addi a4, a4, -1
+; RV32IA-TSO-NEXT:    sll a4, a4, a0
+; RV32IA-TSO-NEXT:    slli a1, a1, 16
+; RV32IA-TSO-NEXT:    srai a1, a1, 16
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:    li a5, 16
+; RV32IA-TSO-NEXT:    sub a5, a5, a3
+; RV32IA-TSO-NEXT:  .LBB91_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a2)
+; RV32IA-TSO-NEXT:    and a7, a3, a4
+; RV32IA-TSO-NEXT:    mv a6, a3
+; RV32IA-TSO-NEXT:    sll a7, a7, a5
+; RV32IA-TSO-NEXT:    sra a7, a7, a5
+; RV32IA-TSO-NEXT:    bge a7, a1, .LBB91_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB91_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a6, a3, a1
+; RV32IA-TSO-NEXT:    and a6, a6, a4
+; RV32IA-TSO-NEXT:    xor a6, a3, a6
+; RV32IA-TSO-NEXT:  .LBB91_3: # in Loop: Header=BB91_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a6, a6, (a2)
+; RV32IA-TSO-NEXT:    bnez a6, .LBB91_1
+; RV32IA-TSO-NEXT:  # %bb.4:
+; RV32IA-TSO-NEXT:    srl a0, a3, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_max_i16_acquire:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -48
 ; RV64I-NEXT:    sd ra, 40(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s0, 32(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s1, 24(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s2, 16(sp) # 8-byte Folded Spill

 ; RV64I-NEXT:    mv a0, a3
 ; RV64I-NEXT:    ld ra, 40(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s0, 32(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s1, 24(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s2, 16(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 48
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_max_i16_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    andi a3, a0, 24
-; RV64IA-NEXT:    lui a4, 16
-; RV64IA-NEXT:    addiw a4, a4, -1
-; RV64IA-NEXT:    sllw a4, a4, a0
-; RV64IA-NEXT:    slli a1, a1, 48
-; RV64IA-NEXT:    srai a1, a1, 48
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:    li a5, 48
-; RV64IA-NEXT:    sub a5, a5, a3
-; RV64IA-NEXT:  .LBB91_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a3, (a2)
-; RV64IA-NEXT:    and a7, a3, a4
-; RV64IA-NEXT:    mv a6, a3
-; RV64IA-NEXT:    sll a7, a7, a5
-; RV64IA-NEXT:    sra a7, a7, a5
-; RV64IA-NEXT:    bge a7, a1, .LBB91_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB91_1 Depth=1
-; RV64IA-NEXT:    xor a6, a3, a1
-; RV64IA-NEXT:    and a6, a6, a4
-; RV64IA-NEXT:    xor a6, a3, a6
-; RV64IA-NEXT:  .LBB91_3: # in Loop: Header=BB91_1 Depth=1
-; RV64IA-NEXT:    sc.w a6, a6, (a2)
-; RV64IA-NEXT:    bnez a6, .LBB91_1
-; RV64IA-NEXT:  # %bb.4:
-; RV64IA-NEXT:    srlw a0, a3, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_max_i16_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    andi a3, a0, 24
+; RV64IA-WMO-NEXT:    lui a4, 16
+; RV64IA-WMO-NEXT:    addiw a4, a4, -1
+; RV64IA-WMO-NEXT:    sllw a4, a4, a0
+; RV64IA-WMO-NEXT:    slli a1, a1, 48
+; RV64IA-WMO-NEXT:    srai a1, a1, 48
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:    li a5, 48
+; RV64IA-WMO-NEXT:    sub a5, a5, a3
+; RV64IA-WMO-NEXT:  .LBB91_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV64IA-WMO-NEXT:    and a7, a3, a4
+; RV64IA-WMO-NEXT:    mv a6, a3
+; RV64IA-WMO-NEXT:    sll a7, a7, a5
+; RV64IA-WMO-NEXT:    sra a7, a7, a5
+; RV64IA-WMO-NEXT:    bge a7, a1, .LBB91_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB91_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a6, a3, a1
+; RV64IA-WMO-NEXT:    and a6, a6, a4
+; RV64IA-WMO-NEXT:    xor a6, a3, a6
+; RV64IA-WMO-NEXT:  .LBB91_3: # in Loop: Header=BB91_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w a6, a6, (a2)
+; RV64IA-WMO-NEXT:    bnez a6, .LBB91_1
+; RV64IA-WMO-NEXT:  # %bb.4:
+; RV64IA-WMO-NEXT:    srlw a0, a3, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_max_i16_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    andi a3, a0, 24
+; RV64IA-TSO-NEXT:    lui a4, 16
+; RV64IA-TSO-NEXT:    addiw a4, a4, -1
+; RV64IA-TSO-NEXT:    sllw a4, a4, a0
+; RV64IA-TSO-NEXT:    slli a1, a1, 48
+; RV64IA-TSO-NEXT:    srai a1, a1, 48
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:    li a5, 48
+; RV64IA-TSO-NEXT:    sub a5, a5, a3
+; RV64IA-TSO-NEXT:  .LBB91_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a2)
+; RV64IA-TSO-NEXT:    and a7, a3, a4
+; RV64IA-TSO-NEXT:    mv a6, a3
+; RV64IA-TSO-NEXT:    sll a7, a7, a5
+; RV64IA-TSO-NEXT:    sra a7, a7, a5
+; RV64IA-TSO-NEXT:    bge a7, a1, .LBB91_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB91_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a6, a3, a1
+; RV64IA-TSO-NEXT:    and a6, a6, a4
+; RV64IA-TSO-NEXT:    xor a6, a3, a6
+; RV64IA-TSO-NEXT:  .LBB91_3: # in Loop: Header=BB91_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a6, a6, (a2)
+; RV64IA-TSO-NEXT:    bnez a6, .LBB91_1
+; RV64IA-TSO-NEXT:  # %bb.4:
+; RV64IA-TSO-NEXT:    srlw a0, a3, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw max ptr %a, i16 %b acquire
   ret i16 %1
 }
 
 define i16 @atomicrmw_max_i16_release(ptr %a, i16 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_max_i16_release:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -32

 ; RV32I-NEXT:    mv a0, a3
 ; RV32I-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s2, 16(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 32
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_max_i16_release:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    andi a3, a0, 24
-; RV32IA-NEXT:    lui a4, 16
-; RV32IA-NEXT:    addi a4, a4, -1
-; RV32IA-NEXT:    sll a4, a4, a0
-; RV32IA-NEXT:    slli a1, a1, 16
-; RV32IA-NEXT:    srai a1, a1, 16
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:    li a5, 16
-; RV32IA-NEXT:    sub a5, a5, a3
-; RV32IA-NEXT:  .LBB92_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w a3, (a2)
-; RV32IA-NEXT:    and a7, a3, a4
-; RV32IA-NEXT:    mv a6, a3
-; RV32IA-NEXT:    sll a7, a7, a5
-; RV32IA-NEXT:    sra a7, a7, a5
-; RV32IA-NEXT:    bge a7, a1, .LBB92_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB92_1 Depth=1
-; RV32IA-NEXT:    xor a6, a3, a1
-; RV32IA-NEXT:    and a6, a6, a4
-; RV32IA-NEXT:    xor a6, a3, a6
-; RV32IA-NEXT:  .LBB92_3: # in Loop: Header=BB92_1 Depth=1
-; RV32IA-NEXT:    sc.w.rl a6, a6, (a2)
-; RV32IA-NEXT:    bnez a6, .LBB92_1
-; RV32IA-NEXT:  # %bb.4:
-; RV32IA-NEXT:    srl a0, a3, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_max_i16_release:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    andi a3, a0, 24
+; RV32IA-WMO-NEXT:    lui a4, 16
+; RV32IA-WMO-NEXT:    addi a4, a4, -1
+; RV32IA-WMO-NEXT:    sll a4, a4, a0
+; RV32IA-WMO-NEXT:    slli a1, a1, 16
+; RV32IA-WMO-NEXT:    srai a1, a1, 16
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:    li a5, 16
+; RV32IA-WMO-NEXT:    sub a5, a5, a3
+; RV32IA-WMO-NEXT:  .LBB92_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w a3, (a2)
+; RV32IA-WMO-NEXT:    and a7, a3, a4
+; RV32IA-WMO-NEXT:    mv a6, a3
+; RV32IA-WMO-NEXT:    sll a7, a7, a5
+; RV32IA-WMO-NEXT:    sra a7, a7, a5
+; RV32IA-WMO-NEXT:    bge a7, a1, .LBB92_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB92_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a6, a3, a1
+; RV32IA-WMO-NEXT:    and a6, a6, a4
+; RV32IA-WMO-NEXT:    xor a6, a3, a6
+; RV32IA-WMO-NEXT:  .LBB92_3: # in Loop: Header=BB92_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w.rl a6, a6, (a2)
+; RV32IA-WMO-NEXT:    bnez a6, .LBB92_1
+; RV32IA-WMO-NEXT:  # %bb.4:
+; RV32IA-WMO-NEXT:    srl a0, a3, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_max_i16_release:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    andi a3, a0, 24
+; RV32IA-TSO-NEXT:    lui a4, 16
+; RV32IA-TSO-NEXT:    addi a4, a4, -1
+; RV32IA-TSO-NEXT:    sll a4, a4, a0
+; RV32IA-TSO-NEXT:    slli a1, a1, 16
+; RV32IA-TSO-NEXT:    srai a1, a1, 16
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:    li a5, 16
+; RV32IA-TSO-NEXT:    sub a5, a5, a3
+; RV32IA-TSO-NEXT:  .LBB92_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a2)
+; RV32IA-TSO-NEXT:    and a7, a3, a4
+; RV32IA-TSO-NEXT:    mv a6, a3
+; RV32IA-TSO-NEXT:    sll a7, a7, a5
+; RV32IA-TSO-NEXT:    sra a7, a7, a5
+; RV32IA-TSO-NEXT:    bge a7, a1, .LBB92_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB92_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a6, a3, a1
+; RV32IA-TSO-NEXT:    and a6, a6, a4
+; RV32IA-TSO-NEXT:    xor a6, a3, a6
+; RV32IA-TSO-NEXT:  .LBB92_3: # in Loop: Header=BB92_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a6, a6, (a2)
+; RV32IA-TSO-NEXT:    bnez a6, .LBB92_1
+; RV32IA-TSO-NEXT:  # %bb.4:
+; RV32IA-TSO-NEXT:    srl a0, a3, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_max_i16_release:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -48
 ; RV64I-NEXT:    sd ra, 40(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s0, 32(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s1, 24(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s2, 16(sp) # 8-byte Folded Spill

 ; RV64I-NEXT:    mv a0, a3
 ; RV64I-NEXT:    ld ra, 40(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s0, 32(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s1, 24(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s2, 16(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 48
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_max_i16_release:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    andi a3, a0, 24
-; RV64IA-NEXT:    lui a4, 16
-; RV64IA-NEXT:    addiw a4, a4, -1
-; RV64IA-NEXT:    sllw a4, a4, a0
-; RV64IA-NEXT:    slli a1, a1, 48
-; RV64IA-NEXT:    srai a1, a1, 48
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:    li a5, 48
-; RV64IA-NEXT:    sub a5, a5, a3
-; RV64IA-NEXT:  .LBB92_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w a3, (a2)
-; RV64IA-NEXT:    and a7, a3, a4
-; RV64IA-NEXT:    mv a6, a3
-; RV64IA-NEXT:    sll a7, a7, a5
-; RV64IA-NEXT:    sra a7, a7, a5
-; RV64IA-NEXT:    bge a7, a1, .LBB92_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB92_1 Depth=1
-; RV64IA-NEXT:    xor a6, a3, a1
-; RV64IA-NEXT:    and a6, a6, a4
-; RV64IA-NEXT:    xor a6, a3, a6
-; RV64IA-NEXT:  .LBB92_3: # in Loop: Header=BB92_1 Depth=1
-; RV64IA-NEXT:    sc.w.rl a6, a6, (a2)
-; RV64IA-NEXT:    bnez a6, .LBB92_1
-; RV64IA-NEXT:  # %bb.4:
-; RV64IA-NEXT:    srlw a0, a3, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_max_i16_release:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    andi a3, a0, 24
+; RV64IA-WMO-NEXT:    lui a4, 16
+; RV64IA-WMO-NEXT:    addiw a4, a4, -1
+; RV64IA-WMO-NEXT:    sllw a4, a4, a0
+; RV64IA-WMO-NEXT:    slli a1, a1, 48
+; RV64IA-WMO-NEXT:    srai a1, a1, 48
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:    li a5, 48
+; RV64IA-WMO-NEXT:    sub a5, a5, a3
+; RV64IA-WMO-NEXT:  .LBB92_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w a3, (a2)
+; RV64IA-WMO-NEXT:    and a7, a3, a4
+; RV64IA-WMO-NEXT:    mv a6, a3
+; RV64IA-WMO-NEXT:    sll a7, a7, a5
+; RV64IA-WMO-NEXT:    sra a7, a7, a5
+; RV64IA-WMO-NEXT:    bge a7, a1, .LBB92_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB92_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a6, a3, a1
+; RV64IA-WMO-NEXT:    and a6, a6, a4
+; RV64IA-WMO-NEXT:    xor a6, a3, a6
+; RV64IA-WMO-NEXT:  .LBB92_3: # in Loop: Header=BB92_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w.rl a6, a6, (a2)
+; RV64IA-WMO-NEXT:    bnez a6, .LBB92_1
+; RV64IA-WMO-NEXT:  # %bb.4:
+; RV64IA-WMO-NEXT:    srlw a0, a3, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_max_i16_release:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    andi a3, a0, 24
+; RV64IA-TSO-NEXT:    lui a4, 16
+; RV64IA-TSO-NEXT:    addiw a4, a4, -1
+; RV64IA-TSO-NEXT:    sllw a4, a4, a0
+; RV64IA-TSO-NEXT:    slli a1, a1, 48
+; RV64IA-TSO-NEXT:    srai a1, a1, 48
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:    li a5, 48
+; RV64IA-TSO-NEXT:    sub a5, a5, a3
+; RV64IA-TSO-NEXT:  .LBB92_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a2)
+; RV64IA-TSO-NEXT:    and a7, a3, a4
+; RV64IA-TSO-NEXT:    mv a6, a3
+; RV64IA-TSO-NEXT:    sll a7, a7, a5
+; RV64IA-TSO-NEXT:    sra a7, a7, a5
+; RV64IA-TSO-NEXT:    bge a7, a1, .LBB92_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB92_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a6, a3, a1
+; RV64IA-TSO-NEXT:    and a6, a6, a4
+; RV64IA-TSO-NEXT:    xor a6, a3, a6
+; RV64IA-TSO-NEXT:  .LBB92_3: # in Loop: Header=BB92_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a6, a6, (a2)
+; RV64IA-TSO-NEXT:    bnez a6, .LBB92_1
+; RV64IA-TSO-NEXT:  # %bb.4:
+; RV64IA-TSO-NEXT:    srlw a0, a3, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw max ptr %a, i16 %b release
   ret i16 %1
 }
 
 define i16 @atomicrmw_max_i16_acq_rel(ptr %a, i16 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_max_i16_acq_rel:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -32

 ; RV32I-NEXT:    mv a0, a3
 ; RV32I-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s2, 16(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 32
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_max_i16_acq_rel:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    andi a3, a0, 24
-; RV32IA-NEXT:    lui a4, 16
-; RV32IA-NEXT:    addi a4, a4, -1
-; RV32IA-NEXT:    sll a4, a4, a0
-; RV32IA-NEXT:    slli a1, a1, 16
-; RV32IA-NEXT:    srai a1, a1, 16
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:    li a5, 16
-; RV32IA-NEXT:    sub a5, a5, a3
-; RV32IA-NEXT:  .LBB93_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a3, (a2)
-; RV32IA-NEXT:    and a7, a3, a4
-; RV32IA-NEXT:    mv a6, a3
-; RV32IA-NEXT:    sll a7, a7, a5
-; RV32IA-NEXT:    sra a7, a7, a5
-; RV32IA-NEXT:    bge a7, a1, .LBB93_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB93_1 Depth=1
-; RV32IA-NEXT:    xor a6, a3, a1
-; RV32IA-NEXT:    and a6, a6, a4
-; RV32IA-NEXT:    xor a6, a3, a6
-; RV32IA-NEXT:  .LBB93_3: # in Loop: Header=BB93_1 Depth=1
-; RV32IA-NEXT:    sc.w.rl a6, a6, (a2)
-; RV32IA-NEXT:    bnez a6, .LBB93_1
-; RV32IA-NEXT:  # %bb.4:
-; RV32IA-NEXT:    srl a0, a3, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_max_i16_acq_rel:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    andi a3, a0, 24
+; RV32IA-WMO-NEXT:    lui a4, 16
+; RV32IA-WMO-NEXT:    addi a4, a4, -1
+; RV32IA-WMO-NEXT:    sll a4, a4, a0
+; RV32IA-WMO-NEXT:    slli a1, a1, 16
+; RV32IA-WMO-NEXT:    srai a1, a1, 16
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:    li a5, 16
+; RV32IA-WMO-NEXT:    sub a5, a5, a3
+; RV32IA-WMO-NEXT:  .LBB93_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV32IA-WMO-NEXT:    and a7, a3, a4
+; RV32IA-WMO-NEXT:    mv a6, a3
+; RV32IA-WMO-NEXT:    sll a7, a7, a5
+; RV32IA-WMO-NEXT:    sra a7, a7, a5
+; RV32IA-WMO-NEXT:    bge a7, a1, .LBB93_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB93_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a6, a3, a1
+; RV32IA-WMO-NEXT:    and a6, a6, a4
+; RV32IA-WMO-NEXT:    xor a6, a3, a6
+; RV32IA-WMO-NEXT:  .LBB93_3: # in Loop: Header=BB93_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w.rl a6, a6, (a2)
+; RV32IA-WMO-NEXT:    bnez a6, .LBB93_1
+; RV32IA-WMO-NEXT:  # %bb.4:
+; RV32IA-WMO-NEXT:    srl a0, a3, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_max_i16_acq_rel:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    andi a3, a0, 24
+; RV32IA-TSO-NEXT:    lui a4, 16
+; RV32IA-TSO-NEXT:    addi a4, a4, -1
+; RV32IA-TSO-NEXT:    sll a4, a4, a0
+; RV32IA-TSO-NEXT:    slli a1, a1, 16
+; RV32IA-TSO-NEXT:    srai a1, a1, 16
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:    li a5, 16
+; RV32IA-TSO-NEXT:    sub a5, a5, a3
+; RV32IA-TSO-NEXT:  .LBB93_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a2)
+; RV32IA-TSO-NEXT:    and a7, a3, a4
+; RV32IA-TSO-NEXT:    mv a6, a3
+; RV32IA-TSO-NEXT:    sll a7, a7, a5
+; RV32IA-TSO-NEXT:    sra a7, a7, a5
+; RV32IA-TSO-NEXT:    bge a7, a1, .LBB93_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB93_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a6, a3, a1
+; RV32IA-TSO-NEXT:    and a6, a6, a4
+; RV32IA-TSO-NEXT:    xor a6, a3, a6
+; RV32IA-TSO-NEXT:  .LBB93_3: # in Loop: Header=BB93_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a6, a6, (a2)
+; RV32IA-TSO-NEXT:    bnez a6, .LBB93_1
+; RV32IA-TSO-NEXT:  # %bb.4:
+; RV32IA-TSO-NEXT:    srl a0, a3, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_max_i16_acq_rel:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -48
 ; RV64I-NEXT:    sd ra, 40(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s0, 32(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s1, 24(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s2, 16(sp) # 8-byte Folded Spill

 ; RV64I-NEXT:    mv a0, a3
 ; RV64I-NEXT:    ld ra, 40(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s0, 32(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s1, 24(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s2, 16(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 48
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_max_i16_acq_rel:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    andi a3, a0, 24
-; RV64IA-NEXT:    lui a4, 16
-; RV64IA-NEXT:    addiw a4, a4, -1
-; RV64IA-NEXT:    sllw a4, a4, a0
-; RV64IA-NEXT:    slli a1, a1, 48
-; RV64IA-NEXT:    srai a1, a1, 48
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:    li a5, 48
-; RV64IA-NEXT:    sub a5, a5, a3
-; RV64IA-NEXT:  .LBB93_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a3, (a2)
-; RV64IA-NEXT:    and a7, a3, a4
-; RV64IA-NEXT:    mv a6, a3
-; RV64IA-NEXT:    sll a7, a7, a5
-; RV64IA-NEXT:    sra a7, a7, a5
-; RV64IA-NEXT:    bge a7, a1, .LBB93_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB93_1 Depth=1
-; RV64IA-NEXT:    xor a6, a3, a1
-; RV64IA-NEXT:    and a6, a6, a4
-; RV64IA-NEXT:    xor a6, a3, a6
-; RV64IA-NEXT:  .LBB93_3: # in Loop: Header=BB93_1 Depth=1
-; RV64IA-NEXT:    sc.w.rl a6, a6, (a2)
-; RV64IA-NEXT:    bnez a6, .LBB93_1
-; RV64IA-NEXT:  # %bb.4:
-; RV64IA-NEXT:    srlw a0, a3, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_max_i16_acq_rel:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    andi a3, a0, 24
+; RV64IA-WMO-NEXT:    lui a4, 16
+; RV64IA-WMO-NEXT:    addiw a4, a4, -1
+; RV64IA-WMO-NEXT:    sllw a4, a4, a0
+; RV64IA-WMO-NEXT:    slli a1, a1, 48
+; RV64IA-WMO-NEXT:    srai a1, a1, 48
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:    li a5, 48
+; RV64IA-WMO-NEXT:    sub a5, a5, a3
+; RV64IA-WMO-NEXT:  .LBB93_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV64IA-WMO-NEXT:    and a7, a3, a4
+; RV64IA-WMO-NEXT:    mv a6, a3
+; RV64IA-WMO-NEXT:    sll a7, a7, a5
+; RV64IA-WMO-NEXT:    sra a7, a7, a5
+; RV64IA-WMO-NEXT:    bge a7, a1, .LBB93_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB93_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a6, a3, a1
+; RV64IA-WMO-NEXT:    and a6, a6, a4
+; RV64IA-WMO-NEXT:    xor a6, a3, a6
+; RV64IA-WMO-NEXT:  .LBB93_3: # in Loop: Header=BB93_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w.rl a6, a6, (a2)
+; RV64IA-WMO-NEXT:    bnez a6, .LBB93_1
+; RV64IA-WMO-NEXT:  # %bb.4:
+; RV64IA-WMO-NEXT:    srlw a0, a3, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_max_i16_acq_rel:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    andi a3, a0, 24
+; RV64IA-TSO-NEXT:    lui a4, 16
+; RV64IA-TSO-NEXT:    addiw a4, a4, -1
+; RV64IA-TSO-NEXT:    sllw a4, a4, a0
+; RV64IA-TSO-NEXT:    slli a1, a1, 48
+; RV64IA-TSO-NEXT:    srai a1, a1, 48
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:    li a5, 48
+; RV64IA-TSO-NEXT:    sub a5, a5, a3
+; RV64IA-TSO-NEXT:  .LBB93_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a2)
+; RV64IA-TSO-NEXT:    and a7, a3, a4
+; RV64IA-TSO-NEXT:    mv a6, a3
+; RV64IA-TSO-NEXT:    sll a7, a7, a5
+; RV64IA-TSO-NEXT:    sra a7, a7, a5
+; RV64IA-TSO-NEXT:    bge a7, a1, .LBB93_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB93_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a6, a3, a1
+; RV64IA-TSO-NEXT:    and a6, a6, a4
+; RV64IA-TSO-NEXT:    xor a6, a3, a6
+; RV64IA-TSO-NEXT:  .LBB93_3: # in Loop: Header=BB93_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a6, a6, (a2)
+; RV64IA-TSO-NEXT:    bnez a6, .LBB93_1
+; RV64IA-TSO-NEXT:  # %bb.4:
+; RV64IA-TSO-NEXT:    srlw a0, a3, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw max ptr %a, i16 %b acq_rel
   ret i16 %1
 }
 
 define i16 @atomicrmw_max_i16_seq_cst(ptr %a, i16 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_max_i16_seq_cst:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -32

 ; RV32I-NEXT:    mv a0, a3
 ; RV32I-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s2, 16(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 32
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_min_i16_acquire:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    andi a3, a0, 24
-; RV32IA-NEXT:    lui a4, 16
-; RV32IA-NEXT:    addi a4, a4, -1
-; RV32IA-NEXT:    sll a4, a4, a0
-; RV32IA-NEXT:    slli a1, a1, 16
-; RV32IA-NEXT:    srai a1, a1, 16
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:    li a5, 16
-; RV32IA-NEXT:    sub a5, a5, a3
-; RV32IA-NEXT:  .LBB96_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a3, (a2)
-; RV32IA-NEXT:    and a7, a3, a4
-; RV32IA-NEXT:    mv a6, a3
-; RV32IA-NEXT:    sll a7, a7, a5
-; RV32IA-NEXT:    sra a7, a7, a5
-; RV32IA-NEXT:    bge a1, a7, .LBB96_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB96_1 Depth=1
-; RV32IA-NEXT:    xor a6, a3, a1
-; RV32IA-NEXT:    and a6, a6, a4
-; RV32IA-NEXT:    xor a6, a3, a6
-; RV32IA-NEXT:  .LBB96_3: # in Loop: Header=BB96_1 Depth=1
-; RV32IA-NEXT:    sc.w a6, a6, (a2)
-; RV32IA-NEXT:    bnez a6, .LBB96_1
-; RV32IA-NEXT:  # %bb.4:
-; RV32IA-NEXT:    srl a0, a3, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_min_i16_acquire:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    andi a3, a0, 24
+; RV32IA-WMO-NEXT:    lui a4, 16
+; RV32IA-WMO-NEXT:    addi a4, a4, -1
+; RV32IA-WMO-NEXT:    sll a4, a4, a0
+; RV32IA-WMO-NEXT:    slli a1, a1, 16
+; RV32IA-WMO-NEXT:    srai a1, a1, 16
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:    li a5, 16
+; RV32IA-WMO-NEXT:    sub a5, a5, a3
+; RV32IA-WMO-NEXT:  .LBB96_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV32IA-WMO-NEXT:    and a7, a3, a4
+; RV32IA-WMO-NEXT:    mv a6, a3
+; RV32IA-WMO-NEXT:    sll a7, a7, a5
+; RV32IA-WMO-NEXT:    sra a7, a7, a5
+; RV32IA-WMO-NEXT:    bge a1, a7, .LBB96_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB96_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a6, a3, a1
+; RV32IA-WMO-NEXT:    and a6, a6, a4
+; RV32IA-WMO-NEXT:    xor a6, a3, a6
+; RV32IA-WMO-NEXT:  .LBB96_3: # in Loop: Header=BB96_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w a6, a6, (a2)
+; RV32IA-WMO-NEXT:    bnez a6, .LBB96_1
+; RV32IA-WMO-NEXT:  # %bb.4:
+; RV32IA-WMO-NEXT:    srl a0, a3, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_min_i16_acquire:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    andi a3, a0, 24
+; RV32IA-TSO-NEXT:    lui a4, 16
+; RV32IA-TSO-NEXT:    addi a4, a4, -1
+; RV32IA-TSO-NEXT:    sll a4, a4, a0
+; RV32IA-TSO-NEXT:    slli a1, a1, 16
+; RV32IA-TSO-NEXT:    srai a1, a1, 16
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:    li a5, 16
+; RV32IA-TSO-NEXT:    sub a5, a5, a3
+; RV32IA-TSO-NEXT:  .LBB96_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a2)
+; RV32IA-TSO-NEXT:    and a7, a3, a4
+; RV32IA-TSO-NEXT:    mv a6, a3
+; RV32IA-TSO-NEXT:    sll a7, a7, a5
+; RV32IA-TSO-NEXT:    sra a7, a7, a5
+; RV32IA-TSO-NEXT:    bge a1, a7, .LBB96_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB96_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a6, a3, a1
+; RV32IA-TSO-NEXT:    and a6, a6, a4
+; RV32IA-TSO-NEXT:    xor a6, a3, a6
+; RV32IA-TSO-NEXT:  .LBB96_3: # in Loop: Header=BB96_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a6, a6, (a2)
+; RV32IA-TSO-NEXT:    bnez a6, .LBB96_1
+; RV32IA-TSO-NEXT:  # %bb.4:
+; RV32IA-TSO-NEXT:    srl a0, a3, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_min_i16_acquire:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -48
 ; RV64I-NEXT:    sd ra, 40(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s0, 32(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s1, 24(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s2, 16(sp) # 8-byte Folded Spill

 ; RV64I-NEXT:    mv a0, a3
 ; RV64I-NEXT:    ld ra, 40(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s0, 32(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s1, 24(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s2, 16(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 48
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_min_i16_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    andi a3, a0, 24
-; RV64IA-NEXT:    lui a4, 16
-; RV64IA-NEXT:    addiw a4, a4, -1
-; RV64IA-NEXT:    sllw a4, a4, a0
-; RV64IA-NEXT:    slli a1, a1, 48
-; RV64IA-NEXT:    srai a1, a1, 48
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:    li a5, 48
-; RV64IA-NEXT:    sub a5, a5, a3
-; RV64IA-NEXT:  .LBB96_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a3, (a2)
-; RV64IA-NEXT:    and a7, a3, a4
-; RV64IA-NEXT:    mv a6, a3
-; RV64IA-NEXT:    sll a7, a7, a5
-; RV64IA-NEXT:    sra a7, a7, a5
-; RV64IA-NEXT:    bge a1, a7, .LBB96_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB96_1 Depth=1
-; RV64IA-NEXT:    xor a6, a3, a1
-; RV64IA-NEXT:    and a6, a6, a4
-; RV64IA-NEXT:    xor a6, a3, a6
-; RV64IA-NEXT:  .LBB96_3: # in Loop: Header=BB96_1 Depth=1
-; RV64IA-NEXT:    sc.w a6, a6, (a2)
-; RV64IA-NEXT:    bnez a6, .LBB96_1
-; RV64IA-NEXT:  # %bb.4:
-; RV64IA-NEXT:    srlw a0, a3, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_min_i16_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    andi a3, a0, 24
+; RV64IA-WMO-NEXT:    lui a4, 16
+; RV64IA-WMO-NEXT:    addiw a4, a4, -1
+; RV64IA-WMO-NEXT:    sllw a4, a4, a0
+; RV64IA-WMO-NEXT:    slli a1, a1, 48
+; RV64IA-WMO-NEXT:    srai a1, a1, 48
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:    li a5, 48
+; RV64IA-WMO-NEXT:    sub a5, a5, a3
+; RV64IA-WMO-NEXT:  .LBB96_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV64IA-WMO-NEXT:    and a7, a3, a4
+; RV64IA-WMO-NEXT:    mv a6, a3
+; RV64IA-WMO-NEXT:    sll a7, a7, a5
+; RV64IA-WMO-NEXT:    sra a7, a7, a5
+; RV64IA-WMO-NEXT:    bge a1, a7, .LBB96_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB96_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a6, a3, a1
+; RV64IA-WMO-NEXT:    and a6, a6, a4
+; RV64IA-WMO-NEXT:    xor a6, a3, a6
+; RV64IA-WMO-NEXT:  .LBB96_3: # in Loop: Header=BB96_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w a6, a6, (a2)
+; RV64IA-WMO-NEXT:    bnez a6, .LBB96_1
+; RV64IA-WMO-NEXT:  # %bb.4:
+; RV64IA-WMO-NEXT:    srlw a0, a3, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_min_i16_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    andi a3, a0, 24
+; RV64IA-TSO-NEXT:    lui a4, 16
+; RV64IA-TSO-NEXT:    addiw a4, a4, -1
+; RV64IA-TSO-NEXT:    sllw a4, a4, a0
+; RV64IA-TSO-NEXT:    slli a1, a1, 48
+; RV64IA-TSO-NEXT:    srai a1, a1, 48
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:    li a5, 48
+; RV64IA-TSO-NEXT:    sub a5, a5, a3
+; RV64IA-TSO-NEXT:  .LBB96_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a2)
+; RV64IA-TSO-NEXT:    and a7, a3, a4
+; RV64IA-TSO-NEXT:    mv a6, a3
+; RV64IA-TSO-NEXT:    sll a7, a7, a5
+; RV64IA-TSO-NEXT:    sra a7, a7, a5
+; RV64IA-TSO-NEXT:    bge a1, a7, .LBB96_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB96_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a6, a3, a1
+; RV64IA-TSO-NEXT:    and a6, a6, a4
+; RV64IA-TSO-NEXT:    xor a6, a3, a6
+; RV64IA-TSO-NEXT:  .LBB96_3: # in Loop: Header=BB96_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a6, a6, (a2)
+; RV64IA-TSO-NEXT:    bnez a6, .LBB96_1
+; RV64IA-TSO-NEXT:  # %bb.4:
+; RV64IA-TSO-NEXT:    srlw a0, a3, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw min ptr %a, i16 %b acquire
   ret i16 %1
 }
 
 define i16 @atomicrmw_min_i16_release(ptr %a, i16 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_min_i16_release:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -32

 ; RV32I-NEXT:    mv a0, a3
 ; RV32I-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s2, 16(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 32
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_min_i16_release:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    andi a3, a0, 24
-; RV32IA-NEXT:    lui a4, 16
-; RV32IA-NEXT:    addi a4, a4, -1
-; RV32IA-NEXT:    sll a4, a4, a0
-; RV32IA-NEXT:    slli a1, a1, 16
-; RV32IA-NEXT:    srai a1, a1, 16
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:    li a5, 16
-; RV32IA-NEXT:    sub a5, a5, a3
-; RV32IA-NEXT:  .LBB97_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w a3, (a2)
-; RV32IA-NEXT:    and a7, a3, a4
-; RV32IA-NEXT:    mv a6, a3
-; RV32IA-NEXT:    sll a7, a7, a5
-; RV32IA-NEXT:    sra a7, a7, a5
-; RV32IA-NEXT:    bge a1, a7, .LBB97_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB97_1 Depth=1
-; RV32IA-NEXT:    xor a6, a3, a1
-; RV32IA-NEXT:    and a6, a6, a4
-; RV32IA-NEXT:    xor a6, a3, a6
-; RV32IA-NEXT:  .LBB97_3: # in Loop: Header=BB97_1 Depth=1
-; RV32IA-NEXT:    sc.w.rl a6, a6, (a2)
-; RV32IA-NEXT:    bnez a6, .LBB97_1
-; RV32IA-NEXT:  # %bb.4:
-; RV32IA-NEXT:    srl a0, a3, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_min_i16_release:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    andi a3, a0, 24
+; RV32IA-WMO-NEXT:    lui a4, 16
+; RV32IA-WMO-NEXT:    addi a4, a4, -1
+; RV32IA-WMO-NEXT:    sll a4, a4, a0
+; RV32IA-WMO-NEXT:    slli a1, a1, 16
+; RV32IA-WMO-NEXT:    srai a1, a1, 16
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:    li a5, 16
+; RV32IA-WMO-NEXT:    sub a5, a5, a3
+; RV32IA-WMO-NEXT:  .LBB97_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w a3, (a2)
+; RV32IA-WMO-NEXT:    and a7, a3, a4
+; RV32IA-WMO-NEXT:    mv a6, a3
+; RV32IA-WMO-NEXT:    sll a7, a7, a5
+; RV32IA-WMO-NEXT:    sra a7, a7, a5
+; RV32IA-WMO-NEXT:    bge a1, a7, .LBB97_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB97_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a6, a3, a1
+; RV32IA-WMO-NEXT:    and a6, a6, a4
+; RV32IA-WMO-NEXT:    xor a6, a3, a6
+; RV32IA-WMO-NEXT:  .LBB97_3: # in Loop: Header=BB97_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w.rl a6, a6, (a2)
+; RV32IA-WMO-NEXT:    bnez a6, .LBB97_1
+; RV32IA-WMO-NEXT:  # %bb.4:
+; RV32IA-WMO-NEXT:    srl a0, a3, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_min_i16_release:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    andi a3, a0, 24
+; RV32IA-TSO-NEXT:    lui a4, 16
+; RV32IA-TSO-NEXT:    addi a4, a4, -1
+; RV32IA-TSO-NEXT:    sll a4, a4, a0
+; RV32IA-TSO-NEXT:    slli a1, a1, 16
+; RV32IA-TSO-NEXT:    srai a1, a1, 16
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:    li a5, 16
+; RV32IA-TSO-NEXT:    sub a5, a5, a3
+; RV32IA-TSO-NEXT:  .LBB97_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a2)
+; RV32IA-TSO-NEXT:    and a7, a3, a4
+; RV32IA-TSO-NEXT:    mv a6, a3
+; RV32IA-TSO-NEXT:    sll a7, a7, a5
+; RV32IA-TSO-NEXT:    sra a7, a7, a5
+; RV32IA-TSO-NEXT:    bge a1, a7, .LBB97_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB97_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a6, a3, a1
+; RV32IA-TSO-NEXT:    and a6, a6, a4
+; RV32IA-TSO-NEXT:    xor a6, a3, a6
+; RV32IA-TSO-NEXT:  .LBB97_3: # in Loop: Header=BB97_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a6, a6, (a2)
+; RV32IA-TSO-NEXT:    bnez a6, .LBB97_1
+; RV32IA-TSO-NEXT:  # %bb.4:
+; RV32IA-TSO-NEXT:    srl a0, a3, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_min_i16_release:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -48
 ; RV64I-NEXT:    sd ra, 40(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s0, 32(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s1, 24(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s2, 16(sp) # 8-byte Folded Spill

 ; RV64I-NEXT:    mv a0, a3
 ; RV64I-NEXT:    ld ra, 40(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s0, 32(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s1, 24(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s2, 16(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 48
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_min_i16_release:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    andi a3, a0, 24
-; RV64IA-NEXT:    lui a4, 16
-; RV64IA-NEXT:    addiw a4, a4, -1
-; RV64IA-NEXT:    sllw a4, a4, a0
-; RV64IA-NEXT:    slli a1, a1, 48
-; RV64IA-NEXT:    srai a1, a1, 48
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:    li a5, 48
-; RV64IA-NEXT:    sub a5, a5, a3
-; RV64IA-NEXT:  .LBB97_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w a3, (a2)
-; RV64IA-NEXT:    and a7, a3, a4
-; RV64IA-NEXT:    mv a6, a3
-; RV64IA-NEXT:    sll a7, a7, a5
-; RV64IA-NEXT:    sra a7, a7, a5
-; RV64IA-NEXT:    bge a1, a7, .LBB97_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB97_1 Depth=1
-; RV64IA-NEXT:    xor a6, a3, a1
-; RV64IA-NEXT:    and a6, a6, a4
-; RV64IA-NEXT:    xor a6, a3, a6
-; RV64IA-NEXT:  .LBB97_3: # in Loop: Header=BB97_1 Depth=1
-; RV64IA-NEXT:    sc.w.rl a6, a6, (a2)
-; RV64IA-NEXT:    bnez a6, .LBB97_1
-; RV64IA-NEXT:  # %bb.4:
-; RV64IA-NEXT:    srlw a0, a3, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_min_i16_release:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    andi a3, a0, 24
+; RV64IA-WMO-NEXT:    lui a4, 16
+; RV64IA-WMO-NEXT:    addiw a4, a4, -1
+; RV64IA-WMO-NEXT:    sllw a4, a4, a0
+; RV64IA-WMO-NEXT:    slli a1, a1, 48
+; RV64IA-WMO-NEXT:    srai a1, a1, 48
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:    li a5, 48
+; RV64IA-WMO-NEXT:    sub a5, a5, a3
+; RV64IA-WMO-NEXT:  .LBB97_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w a3, (a2)
+; RV64IA-WMO-NEXT:    and a7, a3, a4
+; RV64IA-WMO-NEXT:    mv a6, a3
+; RV64IA-WMO-NEXT:    sll a7, a7, a5
+; RV64IA-WMO-NEXT:    sra a7, a7, a5
+; RV64IA-WMO-NEXT:    bge a1, a7, .LBB97_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB97_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a6, a3, a1
+; RV64IA-WMO-NEXT:    and a6, a6, a4
+; RV64IA-WMO-NEXT:    xor a6, a3, a6
+; RV64IA-WMO-NEXT:  .LBB97_3: # in Loop: Header=BB97_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w.rl a6, a6, (a2)
+; RV64IA-WMO-NEXT:    bnez a6, .LBB97_1
+; RV64IA-WMO-NEXT:  # %bb.4:
+; RV64IA-WMO-NEXT:    srlw a0, a3, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_min_i16_release:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    andi a3, a0, 24
+; RV64IA-TSO-NEXT:    lui a4, 16
+; RV64IA-TSO-NEXT:    addiw a4, a4, -1
+; RV64IA-TSO-NEXT:    sllw a4, a4, a0
+; RV64IA-TSO-NEXT:    slli a1, a1, 48
+; RV64IA-TSO-NEXT:    srai a1, a1, 48
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:    li a5, 48
+; RV64IA-TSO-NEXT:    sub a5, a5, a3
+; RV64IA-TSO-NEXT:  .LBB97_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a2)
+; RV64IA-TSO-NEXT:    and a7, a3, a4
+; RV64IA-TSO-NEXT:    mv a6, a3
+; RV64IA-TSO-NEXT:    sll a7, a7, a5
+; RV64IA-TSO-NEXT:    sra a7, a7, a5
+; RV64IA-TSO-NEXT:    bge a1, a7, .LBB97_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB97_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a6, a3, a1
+; RV64IA-TSO-NEXT:    and a6, a6, a4
+; RV64IA-TSO-NEXT:    xor a6, a3, a6
+; RV64IA-TSO-NEXT:  .LBB97_3: # in Loop: Header=BB97_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a6, a6, (a2)
+; RV64IA-TSO-NEXT:    bnez a6, .LBB97_1
+; RV64IA-TSO-NEXT:  # %bb.4:
+; RV64IA-TSO-NEXT:    srlw a0, a3, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw min ptr %a, i16 %b release
   ret i16 %1
 }
 
 define i16 @atomicrmw_min_i16_acq_rel(ptr %a, i16 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_min_i16_acq_rel:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -32

 ; RV32I-NEXT:    mv a0, a3
 ; RV32I-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s2, 16(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 32
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_min_i16_acq_rel:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    andi a3, a0, 24
-; RV32IA-NEXT:    lui a4, 16
-; RV32IA-NEXT:    addi a4, a4, -1
-; RV32IA-NEXT:    sll a4, a4, a0
-; RV32IA-NEXT:    slli a1, a1, 16
-; RV32IA-NEXT:    srai a1, a1, 16
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:    li a5, 16
-; RV32IA-NEXT:    sub a5, a5, a3
-; RV32IA-NEXT:  .LBB98_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a3, (a2)
-; RV32IA-NEXT:    and a7, a3, a4
-; RV32IA-NEXT:    mv a6, a3
-; RV32IA-NEXT:    sll a7, a7, a5
-; RV32IA-NEXT:    sra a7, a7, a5
-; RV32IA-NEXT:    bge a1, a7, .LBB98_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB98_1 Depth=1
-; RV32IA-NEXT:    xor a6, a3, a1
-; RV32IA-NEXT:    and a6, a6, a4
-; RV32IA-NEXT:    xor a6, a3, a6
-; RV32IA-NEXT:  .LBB98_3: # in Loop: Header=BB98_1 Depth=1
-; RV32IA-NEXT:    sc.w.rl a6, a6, (a2)
-; RV32IA-NEXT:    bnez a6, .LBB98_1
-; RV32IA-NEXT:  # %bb.4:
-; RV32IA-NEXT:    srl a0, a3, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_min_i16_acq_rel:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    andi a3, a0, 24
+; RV32IA-WMO-NEXT:    lui a4, 16
+; RV32IA-WMO-NEXT:    addi a4, a4, -1
+; RV32IA-WMO-NEXT:    sll a4, a4, a0
+; RV32IA-WMO-NEXT:    slli a1, a1, 16
+; RV32IA-WMO-NEXT:    srai a1, a1, 16
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:    li a5, 16
+; RV32IA-WMO-NEXT:    sub a5, a5, a3
+; RV32IA-WMO-NEXT:  .LBB98_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV32IA-WMO-NEXT:    and a7, a3, a4
+; RV32IA-WMO-NEXT:    mv a6, a3
+; RV32IA-WMO-NEXT:    sll a7, a7, a5
+; RV32IA-WMO-NEXT:    sra a7, a7, a5
+; RV32IA-WMO-NEXT:    bge a1, a7, .LBB98_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB98_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a6, a3, a1
+; RV32IA-WMO-NEXT:    and a6, a6, a4
+; RV32IA-WMO-NEXT:    xor a6, a3, a6
+; RV32IA-WMO-NEXT:  .LBB98_3: # in Loop: Header=BB98_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w.rl a6, a6, (a2)
+; RV32IA-WMO-NEXT:    bnez a6, .LBB98_1
+; RV32IA-WMO-NEXT:  # %bb.4:
+; RV32IA-WMO-NEXT:    srl a0, a3, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_min_i16_acq_rel:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    andi a3, a0, 24
+; RV32IA-TSO-NEXT:    lui a4, 16
+; RV32IA-TSO-NEXT:    addi a4, a4, -1
+; RV32IA-TSO-NEXT:    sll a4, a4, a0
+; RV32IA-TSO-NEXT:    slli a1, a1, 16
+; RV32IA-TSO-NEXT:    srai a1, a1, 16
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:    li a5, 16
+; RV32IA-TSO-NEXT:    sub a5, a5, a3
+; RV32IA-TSO-NEXT:  .LBB98_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a2)
+; RV32IA-TSO-NEXT:    and a7, a3, a4
+; RV32IA-TSO-NEXT:    mv a6, a3
+; RV32IA-TSO-NEXT:    sll a7, a7, a5
+; RV32IA-TSO-NEXT:    sra a7, a7, a5
+; RV32IA-TSO-NEXT:    bge a1, a7, .LBB98_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB98_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a6, a3, a1
+; RV32IA-TSO-NEXT:    and a6, a6, a4
+; RV32IA-TSO-NEXT:    xor a6, a3, a6
+; RV32IA-TSO-NEXT:  .LBB98_3: # in Loop: Header=BB98_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a6, a6, (a2)
+; RV32IA-TSO-NEXT:    bnez a6, .LBB98_1
+; RV32IA-TSO-NEXT:  # %bb.4:
+; RV32IA-TSO-NEXT:    srl a0, a3, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_min_i16_acq_rel:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -48
 ; RV64I-NEXT:    sd ra, 40(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s0, 32(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s1, 24(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s2, 16(sp) # 8-byte Folded Spill

 ; RV64I-NEXT:    mv a0, a3
 ; RV64I-NEXT:    ld ra, 40(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s0, 32(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s1, 24(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s2, 16(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 48
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_min_i16_acq_rel:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    andi a3, a0, 24
-; RV64IA-NEXT:    lui a4, 16
-; RV64IA-NEXT:    addiw a4, a4, -1
-; RV64IA-NEXT:    sllw a4, a4, a0
-; RV64IA-NEXT:    slli a1, a1, 48
-; RV64IA-NEXT:    srai a1, a1, 48
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:    li a5, 48
-; RV64IA-NEXT:    sub a5, a5, a3
-; RV64IA-NEXT:  .LBB98_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a3, (a2)
-; RV64IA-NEXT:    and a7, a3, a4
-; RV64IA-NEXT:    mv a6, a3
-; RV64IA-NEXT:    sll a7, a7, a5
-; RV64IA-NEXT:    sra a7, a7, a5
-; RV64IA-NEXT:    bge a1, a7, .LBB98_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB98_1 Depth=1
-; RV64IA-NEXT:    xor a6, a3, a1
-; RV64IA-NEXT:    and a6, a6, a4
-; RV64IA-NEXT:    xor a6, a3, a6
-; RV64IA-NEXT:  .LBB98_3: # in Loop: Header=BB98_1 Depth=1
-; RV64IA-NEXT:    sc.w.rl a6, a6, (a2)
-; RV64IA-NEXT:    bnez a6, .LBB98_1
-; RV64IA-NEXT:  # %bb.4:
-; RV64IA-NEXT:    srlw a0, a3, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_min_i16_acq_rel:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    andi a3, a0, 24
+; RV64IA-WMO-NEXT:    lui a4, 16
+; RV64IA-WMO-NEXT:    addiw a4, a4, -1
+; RV64IA-WMO-NEXT:    sllw a4, a4, a0
+; RV64IA-WMO-NEXT:    slli a1, a1, 48
+; RV64IA-WMO-NEXT:    srai a1, a1, 48
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:    li a5, 48
+; RV64IA-WMO-NEXT:    sub a5, a5, a3
+; RV64IA-WMO-NEXT:  .LBB98_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV64IA-WMO-NEXT:    and a7, a3, a4
+; RV64IA-WMO-NEXT:    mv a6, a3
+; RV64IA-WMO-NEXT:    sll a7, a7, a5
+; RV64IA-WMO-NEXT:    sra a7, a7, a5
+; RV64IA-WMO-NEXT:    bge a1, a7, .LBB98_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB98_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a6, a3, a1
+; RV64IA-WMO-NEXT:    and a6, a6, a4
+; RV64IA-WMO-NEXT:    xor a6, a3, a6
+; RV64IA-WMO-NEXT:  .LBB98_3: # in Loop: Header=BB98_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w.rl a6, a6, (a2)
+; RV64IA-WMO-NEXT:    bnez a6, .LBB98_1
+; RV64IA-WMO-NEXT:  # %bb.4:
+; RV64IA-WMO-NEXT:    srlw a0, a3, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_min_i16_acq_rel:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    andi a3, a0, 24
+; RV64IA-TSO-NEXT:    lui a4, 16
+; RV64IA-TSO-NEXT:    addiw a4, a4, -1
+; RV64IA-TSO-NEXT:    sllw a4, a4, a0
+; RV64IA-TSO-NEXT:    slli a1, a1, 48
+; RV64IA-TSO-NEXT:    srai a1, a1, 48
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:    li a5, 48
+; RV64IA-TSO-NEXT:    sub a5, a5, a3
+; RV64IA-TSO-NEXT:  .LBB98_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a2)
+; RV64IA-TSO-NEXT:    and a7, a3, a4
+; RV64IA-TSO-NEXT:    mv a6, a3
+; RV64IA-TSO-NEXT:    sll a7, a7, a5
+; RV64IA-TSO-NEXT:    sra a7, a7, a5
+; RV64IA-TSO-NEXT:    bge a1, a7, .LBB98_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB98_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a6, a3, a1
+; RV64IA-TSO-NEXT:    and a6, a6, a4
+; RV64IA-TSO-NEXT:    xor a6, a3, a6
+; RV64IA-TSO-NEXT:  .LBB98_3: # in Loop: Header=BB98_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a6, a6, (a2)
+; RV64IA-TSO-NEXT:    bnez a6, .LBB98_1
+; RV64IA-TSO-NEXT:  # %bb.4:
+; RV64IA-TSO-NEXT:    srlw a0, a3, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw min ptr %a, i16 %b acq_rel
   ret i16 %1
 }
 
 define i16 @atomicrmw_min_i16_seq_cst(ptr %a, i16 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_min_i16_seq_cst:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -32

 ; RV32I-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s2, 16(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s3, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 32
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_umax_i16_acquire:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    lui a3, 16
-; RV32IA-NEXT:    addi a3, a3, -1
-; RV32IA-NEXT:    sll a4, a3, a0
-; RV32IA-NEXT:    and a1, a1, a3
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB101_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a3, (a2)
-; RV32IA-NEXT:    and a6, a3, a4
-; RV32IA-NEXT:    mv a5, a3
-; RV32IA-NEXT:    bgeu a6, a1, .LBB101_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB101_1 Depth=1
-; RV32IA-NEXT:    xor a5, a3, a1
-; RV32IA-NEXT:    and a5, a5, a4
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:  .LBB101_3: # in Loop: Header=BB101_1 Depth=1
-; RV32IA-NEXT:    sc.w a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB101_1
-; RV32IA-NEXT:  # %bb.4:
-; RV32IA-NEXT:    srl a0, a3, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_umax_i16_acquire:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    lui a3, 16
+; RV32IA-WMO-NEXT:    addi a3, a3, -1
+; RV32IA-WMO-NEXT:    sll a4, a3, a0
+; RV32IA-WMO-NEXT:    and a1, a1, a3
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB101_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV32IA-WMO-NEXT:    and a6, a3, a4
+; RV32IA-WMO-NEXT:    mv a5, a3
+; RV32IA-WMO-NEXT:    bgeu a6, a1, .LBB101_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB101_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a5, a3, a1
+; RV32IA-WMO-NEXT:    and a5, a5, a4
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:  .LBB101_3: # in Loop: Header=BB101_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB101_1
+; RV32IA-WMO-NEXT:  # %bb.4:
+; RV32IA-WMO-NEXT:    srl a0, a3, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_umax_i16_acquire:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    lui a3, 16
+; RV32IA-TSO-NEXT:    addi a3, a3, -1
+; RV32IA-TSO-NEXT:    sll a4, a3, a0
+; RV32IA-TSO-NEXT:    and a1, a1, a3
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB101_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a2)
+; RV32IA-TSO-NEXT:    and a6, a3, a4
+; RV32IA-TSO-NEXT:    mv a5, a3
+; RV32IA-TSO-NEXT:    bgeu a6, a1, .LBB101_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB101_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a5, a3, a1
+; RV32IA-TSO-NEXT:    and a5, a5, a4
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:  .LBB101_3: # in Loop: Header=BB101_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB101_1
+; RV32IA-TSO-NEXT:  # %bb.4:
+; RV32IA-TSO-NEXT:    srl a0, a3, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_umax_i16_acquire:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -48
 ; RV64I-NEXT:    sd ra, 40(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s0, 32(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s1, 24(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s2, 16(sp) # 8-byte Folded Spill

 ; RV64I-NEXT:    ld ra, 40(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s0, 32(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s1, 24(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s2, 16(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s3, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 48
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_umax_i16_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    lui a3, 16
-; RV64IA-NEXT:    addiw a3, a3, -1
-; RV64IA-NEXT:    sllw a4, a3, a0
-; RV64IA-NEXT:    and a1, a1, a3
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB101_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a3, (a2)
-; RV64IA-NEXT:    and a6, a3, a4
-; RV64IA-NEXT:    mv a5, a3
-; RV64IA-NEXT:    bgeu a6, a1, .LBB101_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB101_1 Depth=1
-; RV64IA-NEXT:    xor a5, a3, a1
-; RV64IA-NEXT:    and a5, a5, a4
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:  .LBB101_3: # in Loop: Header=BB101_1 Depth=1
-; RV64IA-NEXT:    sc.w a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB101_1
-; RV64IA-NEXT:  # %bb.4:
-; RV64IA-NEXT:    srlw a0, a3, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_umax_i16_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    lui a3, 16
+; RV64IA-WMO-NEXT:    addiw a3, a3, -1
+; RV64IA-WMO-NEXT:    sllw a4, a3, a0
+; RV64IA-WMO-NEXT:    and a1, a1, a3
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB101_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV64IA-WMO-NEXT:    and a6, a3, a4
+; RV64IA-WMO-NEXT:    mv a5, a3
+; RV64IA-WMO-NEXT:    bgeu a6, a1, .LBB101_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB101_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a5, a3, a1
+; RV64IA-WMO-NEXT:    and a5, a5, a4
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:  .LBB101_3: # in Loop: Header=BB101_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB101_1
+; RV64IA-WMO-NEXT:  # %bb.4:
+; RV64IA-WMO-NEXT:    srlw a0, a3, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_umax_i16_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    lui a3, 16
+; RV64IA-TSO-NEXT:    addiw a3, a3, -1
+; RV64IA-TSO-NEXT:    sllw a4, a3, a0
+; RV64IA-TSO-NEXT:    and a1, a1, a3
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB101_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a2)
+; RV64IA-TSO-NEXT:    and a6, a3, a4
+; RV64IA-TSO-NEXT:    mv a5, a3
+; RV64IA-TSO-NEXT:    bgeu a6, a1, .LBB101_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB101_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a5, a3, a1
+; RV64IA-TSO-NEXT:    and a5, a5, a4
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:  .LBB101_3: # in Loop: Header=BB101_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB101_1
+; RV64IA-TSO-NEXT:  # %bb.4:
+; RV64IA-TSO-NEXT:    srlw a0, a3, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw umax ptr %a, i16 %b acquire
   ret i16 %1
 }
 
 define i16 @atomicrmw_umax_i16_release(ptr %a, i16 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_umax_i16_release:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -32

 ; RV32I-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s2, 16(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s3, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 32
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_umax_i16_release:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    lui a3, 16
-; RV32IA-NEXT:    addi a3, a3, -1
-; RV32IA-NEXT:    sll a4, a3, a0
-; RV32IA-NEXT:    and a1, a1, a3
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB102_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w a3, (a2)
-; RV32IA-NEXT:    and a6, a3, a4
-; RV32IA-NEXT:    mv a5, a3
-; RV32IA-NEXT:    bgeu a6, a1, .LBB102_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB102_1 Depth=1
-; RV32IA-NEXT:    xor a5, a3, a1
-; RV32IA-NEXT:    and a5, a5, a4
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:  .LBB102_3: # in Loop: Header=BB102_1 Depth=1
-; RV32IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB102_1
-; RV32IA-NEXT:  # %bb.4:
-; RV32IA-NEXT:    srl a0, a3, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_umax_i16_release:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    lui a3, 16
+; RV32IA-WMO-NEXT:    addi a3, a3, -1
+; RV32IA-WMO-NEXT:    sll a4, a3, a0
+; RV32IA-WMO-NEXT:    and a1, a1, a3
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB102_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w a3, (a2)
+; RV32IA-WMO-NEXT:    and a6, a3, a4
+; RV32IA-WMO-NEXT:    mv a5, a3
+; RV32IA-WMO-NEXT:    bgeu a6, a1, .LBB102_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB102_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a5, a3, a1
+; RV32IA-WMO-NEXT:    and a5, a5, a4
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:  .LBB102_3: # in Loop: Header=BB102_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB102_1
+; RV32IA-WMO-NEXT:  # %bb.4:
+; RV32IA-WMO-NEXT:    srl a0, a3, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_umax_i16_release:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    lui a3, 16
+; RV32IA-TSO-NEXT:    addi a3, a3, -1
+; RV32IA-TSO-NEXT:    sll a4, a3, a0
+; RV32IA-TSO-NEXT:    and a1, a1, a3
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB102_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a2)
+; RV32IA-TSO-NEXT:    and a6, a3, a4
+; RV32IA-TSO-NEXT:    mv a5, a3
+; RV32IA-TSO-NEXT:    bgeu a6, a1, .LBB102_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB102_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a5, a3, a1
+; RV32IA-TSO-NEXT:    and a5, a5, a4
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:  .LBB102_3: # in Loop: Header=BB102_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB102_1
+; RV32IA-TSO-NEXT:  # %bb.4:
+; RV32IA-TSO-NEXT:    srl a0, a3, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_umax_i16_release:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -48
 ; RV64I-NEXT:    sd ra, 40(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s0, 32(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s1, 24(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s2, 16(sp) # 8-byte Folded Spill

 ; RV64I-NEXT:    ld ra, 40(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s0, 32(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s1, 24(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s2, 16(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s3, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 48
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_umax_i16_release:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    lui a3, 16
-; RV64IA-NEXT:    addiw a3, a3, -1
-; RV64IA-NEXT:    sllw a4, a3, a0
-; RV64IA-NEXT:    and a1, a1, a3
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB102_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w a3, (a2)
-; RV64IA-NEXT:    and a6, a3, a4
-; RV64IA-NEXT:    mv a5, a3
-; RV64IA-NEXT:    bgeu a6, a1, .LBB102_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB102_1 Depth=1
-; RV64IA-NEXT:    xor a5, a3, a1
-; RV64IA-NEXT:    and a5, a5, a4
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:  .LBB102_3: # in Loop: Header=BB102_1 Depth=1
-; RV64IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB102_1
-; RV64IA-NEXT:  # %bb.4:
-; RV64IA-NEXT:    srlw a0, a3, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_umax_i16_release:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    lui a3, 16
+; RV64IA-WMO-NEXT:    addiw a3, a3, -1
+; RV64IA-WMO-NEXT:    sllw a4, a3, a0
+; RV64IA-WMO-NEXT:    and a1, a1, a3
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB102_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w a3, (a2)
+; RV64IA-WMO-NEXT:    and a6, a3, a4
+; RV64IA-WMO-NEXT:    mv a5, a3
+; RV64IA-WMO-NEXT:    bgeu a6, a1, .LBB102_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB102_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a5, a3, a1
+; RV64IA-WMO-NEXT:    and a5, a5, a4
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:  .LBB102_3: # in Loop: Header=BB102_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB102_1
+; RV64IA-WMO-NEXT:  # %bb.4:
+; RV64IA-WMO-NEXT:    srlw a0, a3, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_umax_i16_release:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    lui a3, 16
+; RV64IA-TSO-NEXT:    addiw a3, a3, -1
+; RV64IA-TSO-NEXT:    sllw a4, a3, a0
+; RV64IA-TSO-NEXT:    and a1, a1, a3
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB102_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a2)
+; RV64IA-TSO-NEXT:    and a6, a3, a4
+; RV64IA-TSO-NEXT:    mv a5, a3
+; RV64IA-TSO-NEXT:    bgeu a6, a1, .LBB102_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB102_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a5, a3, a1
+; RV64IA-TSO-NEXT:    and a5, a5, a4
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:  .LBB102_3: # in Loop: Header=BB102_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB102_1
+; RV64IA-TSO-NEXT:  # %bb.4:
+; RV64IA-TSO-NEXT:    srlw a0, a3, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw umax ptr %a, i16 %b release
   ret i16 %1
 }
 
 define i16 @atomicrmw_umax_i16_acq_rel(ptr %a, i16 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_umax_i16_acq_rel:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -32

 ; RV32I-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s2, 16(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s3, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 32
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_umax_i16_acq_rel:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    lui a3, 16
-; RV32IA-NEXT:    addi a3, a3, -1
-; RV32IA-NEXT:    sll a4, a3, a0
-; RV32IA-NEXT:    and a1, a1, a3
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB103_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a3, (a2)
-; RV32IA-NEXT:    and a6, a3, a4
-; RV32IA-NEXT:    mv a5, a3
-; RV32IA-NEXT:    bgeu a6, a1, .LBB103_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB103_1 Depth=1
-; RV32IA-NEXT:    xor a5, a3, a1
-; RV32IA-NEXT:    and a5, a5, a4
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:  .LBB103_3: # in Loop: Header=BB103_1 Depth=1
-; RV32IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB103_1
-; RV32IA-NEXT:  # %bb.4:
-; RV32IA-NEXT:    srl a0, a3, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_umax_i16_acq_rel:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    lui a3, 16
+; RV32IA-WMO-NEXT:    addi a3, a3, -1
+; RV32IA-WMO-NEXT:    sll a4, a3, a0
+; RV32IA-WMO-NEXT:    and a1, a1, a3
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB103_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV32IA-WMO-NEXT:    and a6, a3, a4
+; RV32IA-WMO-NEXT:    mv a5, a3
+; RV32IA-WMO-NEXT:    bgeu a6, a1, .LBB103_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB103_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a5, a3, a1
+; RV32IA-WMO-NEXT:    and a5, a5, a4
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:  .LBB103_3: # in Loop: Header=BB103_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB103_1
+; RV32IA-WMO-NEXT:  # %bb.4:
+; RV32IA-WMO-NEXT:    srl a0, a3, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_umax_i16_acq_rel:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    lui a3, 16
+; RV32IA-TSO-NEXT:    addi a3, a3, -1
+; RV32IA-TSO-NEXT:    sll a4, a3, a0
+; RV32IA-TSO-NEXT:    and a1, a1, a3
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB103_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a2)
+; RV32IA-TSO-NEXT:    and a6, a3, a4
+; RV32IA-TSO-NEXT:    mv a5, a3
+; RV32IA-TSO-NEXT:    bgeu a6, a1, .LBB103_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB103_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a5, a3, a1
+; RV32IA-TSO-NEXT:    and a5, a5, a4
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:  .LBB103_3: # in Loop: Header=BB103_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB103_1
+; RV32IA-TSO-NEXT:  # %bb.4:
+; RV32IA-TSO-NEXT:    srl a0, a3, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_umax_i16_acq_rel:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -48
 ; RV64I-NEXT:    sd ra, 40(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s0, 32(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s1, 24(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s2, 16(sp) # 8-byte Folded Spill

 ; RV64I-NEXT:    ld ra, 40(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s0, 32(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s1, 24(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s2, 16(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s3, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 48
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_umax_i16_acq_rel:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    lui a3, 16
-; RV64IA-NEXT:    addiw a3, a3, -1
-; RV64IA-NEXT:    sllw a4, a3, a0
-; RV64IA-NEXT:    and a1, a1, a3
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB103_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a3, (a2)
-; RV64IA-NEXT:    and a6, a3, a4
-; RV64IA-NEXT:    mv a5, a3
-; RV64IA-NEXT:    bgeu a6, a1, .LBB103_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB103_1 Depth=1
-; RV64IA-NEXT:    xor a5, a3, a1
-; RV64IA-NEXT:    and a5, a5, a4
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:  .LBB103_3: # in Loop: Header=BB103_1 Depth=1
-; RV64IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB103_1
-; RV64IA-NEXT:  # %bb.4:
-; RV64IA-NEXT:    srlw a0, a3, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_umax_i16_acq_rel:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    lui a3, 16
+; RV64IA-WMO-NEXT:    addiw a3, a3, -1
+; RV64IA-WMO-NEXT:    sllw a4, a3, a0
+; RV64IA-WMO-NEXT:    and a1, a1, a3
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB103_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV64IA-WMO-NEXT:    and a6, a3, a4
+; RV64IA-WMO-NEXT:    mv a5, a3
+; RV64IA-WMO-NEXT:    bgeu a6, a1, .LBB103_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB103_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a5, a3, a1
+; RV64IA-WMO-NEXT:    and a5, a5, a4
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:  .LBB103_3: # in Loop: Header=BB103_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB103_1
+; RV64IA-WMO-NEXT:  # %bb.4:
+; RV64IA-WMO-NEXT:    srlw a0, a3, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_umax_i16_acq_rel:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    lui a3, 16
+; RV64IA-TSO-NEXT:    addiw a3, a3, -1
+; RV64IA-TSO-NEXT:    sllw a4, a3, a0
+; RV64IA-TSO-NEXT:    and a1, a1, a3
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB103_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a2)
+; RV64IA-TSO-NEXT:    and a6, a3, a4
+; RV64IA-TSO-NEXT:    mv a5, a3
+; RV64IA-TSO-NEXT:    bgeu a6, a1, .LBB103_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB103_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a5, a3, a1
+; RV64IA-TSO-NEXT:    and a5, a5, a4
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:  .LBB103_3: # in Loop: Header=BB103_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB103_1
+; RV64IA-TSO-NEXT:  # %bb.4:
+; RV64IA-TSO-NEXT:    srlw a0, a3, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw umax ptr %a, i16 %b acq_rel
   ret i16 %1
 }
 
 define i16 @atomicrmw_umax_i16_seq_cst(ptr %a, i16 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_umax_i16_seq_cst:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -32

 ; RV32I-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s2, 16(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s3, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 32
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_umin_i16_acquire:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    lui a3, 16
-; RV32IA-NEXT:    addi a3, a3, -1
-; RV32IA-NEXT:    sll a4, a3, a0
-; RV32IA-NEXT:    and a1, a1, a3
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB106_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a3, (a2)
-; RV32IA-NEXT:    and a6, a3, a4
-; RV32IA-NEXT:    mv a5, a3
-; RV32IA-NEXT:    bgeu a1, a6, .LBB106_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB106_1 Depth=1
-; RV32IA-NEXT:    xor a5, a3, a1
-; RV32IA-NEXT:    and a5, a5, a4
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:  .LBB106_3: # in Loop: Header=BB106_1 Depth=1
-; RV32IA-NEXT:    sc.w a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB106_1
-; RV32IA-NEXT:  # %bb.4:
-; RV32IA-NEXT:    srl a0, a3, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_umin_i16_acquire:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    lui a3, 16
+; RV32IA-WMO-NEXT:    addi a3, a3, -1
+; RV32IA-WMO-NEXT:    sll a4, a3, a0
+; RV32IA-WMO-NEXT:    and a1, a1, a3
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB106_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV32IA-WMO-NEXT:    and a6, a3, a4
+; RV32IA-WMO-NEXT:    mv a5, a3
+; RV32IA-WMO-NEXT:    bgeu a1, a6, .LBB106_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB106_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a5, a3, a1
+; RV32IA-WMO-NEXT:    and a5, a5, a4
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:  .LBB106_3: # in Loop: Header=BB106_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB106_1
+; RV32IA-WMO-NEXT:  # %bb.4:
+; RV32IA-WMO-NEXT:    srl a0, a3, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_umin_i16_acquire:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    lui a3, 16
+; RV32IA-TSO-NEXT:    addi a3, a3, -1
+; RV32IA-TSO-NEXT:    sll a4, a3, a0
+; RV32IA-TSO-NEXT:    and a1, a1, a3
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB106_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a2)
+; RV32IA-TSO-NEXT:    and a6, a3, a4
+; RV32IA-TSO-NEXT:    mv a5, a3
+; RV32IA-TSO-NEXT:    bgeu a1, a6, .LBB106_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB106_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a5, a3, a1
+; RV32IA-TSO-NEXT:    and a5, a5, a4
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:  .LBB106_3: # in Loop: Header=BB106_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB106_1
+; RV32IA-TSO-NEXT:  # %bb.4:
+; RV32IA-TSO-NEXT:    srl a0, a3, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_umin_i16_acquire:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -48
 ; RV64I-NEXT:    sd ra, 40(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s0, 32(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s1, 24(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s2, 16(sp) # 8-byte Folded Spill

 ; RV64I-NEXT:    ld ra, 40(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s0, 32(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s1, 24(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s2, 16(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s3, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 48
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_umin_i16_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    lui a3, 16
-; RV64IA-NEXT:    addiw a3, a3, -1
-; RV64IA-NEXT:    sllw a4, a3, a0
-; RV64IA-NEXT:    and a1, a1, a3
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB106_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a3, (a2)
-; RV64IA-NEXT:    and a6, a3, a4
-; RV64IA-NEXT:    mv a5, a3
-; RV64IA-NEXT:    bgeu a1, a6, .LBB106_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB106_1 Depth=1
-; RV64IA-NEXT:    xor a5, a3, a1
-; RV64IA-NEXT:    and a5, a5, a4
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:  .LBB106_3: # in Loop: Header=BB106_1 Depth=1
-; RV64IA-NEXT:    sc.w a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB106_1
-; RV64IA-NEXT:  # %bb.4:
-; RV64IA-NEXT:    srlw a0, a3, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_umin_i16_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    lui a3, 16
+; RV64IA-WMO-NEXT:    addiw a3, a3, -1
+; RV64IA-WMO-NEXT:    sllw a4, a3, a0
+; RV64IA-WMO-NEXT:    and a1, a1, a3
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB106_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV64IA-WMO-NEXT:    and a6, a3, a4
+; RV64IA-WMO-NEXT:    mv a5, a3
+; RV64IA-WMO-NEXT:    bgeu a1, a6, .LBB106_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB106_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a5, a3, a1
+; RV64IA-WMO-NEXT:    and a5, a5, a4
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:  .LBB106_3: # in Loop: Header=BB106_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB106_1
+; RV64IA-WMO-NEXT:  # %bb.4:
+; RV64IA-WMO-NEXT:    srlw a0, a3, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_umin_i16_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    lui a3, 16
+; RV64IA-TSO-NEXT:    addiw a3, a3, -1
+; RV64IA-TSO-NEXT:    sllw a4, a3, a0
+; RV64IA-TSO-NEXT:    and a1, a1, a3
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB106_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a2)
+; RV64IA-TSO-NEXT:    and a6, a3, a4
+; RV64IA-TSO-NEXT:    mv a5, a3
+; RV64IA-TSO-NEXT:    bgeu a1, a6, .LBB106_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB106_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a5, a3, a1
+; RV64IA-TSO-NEXT:    and a5, a5, a4
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:  .LBB106_3: # in Loop: Header=BB106_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB106_1
+; RV64IA-TSO-NEXT:  # %bb.4:
+; RV64IA-TSO-NEXT:    srlw a0, a3, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw umin ptr %a, i16 %b acquire
   ret i16 %1
 }
 
 define i16 @atomicrmw_umin_i16_release(ptr %a, i16 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_umin_i16_release:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -32

 ; RV32I-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s2, 16(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s3, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 32
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_umin_i16_release:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    lui a3, 16
-; RV32IA-NEXT:    addi a3, a3, -1
-; RV32IA-NEXT:    sll a4, a3, a0
-; RV32IA-NEXT:    and a1, a1, a3
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB107_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w a3, (a2)
-; RV32IA-NEXT:    and a6, a3, a4
-; RV32IA-NEXT:    mv a5, a3
-; RV32IA-NEXT:    bgeu a1, a6, .LBB107_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB107_1 Depth=1
-; RV32IA-NEXT:    xor a5, a3, a1
-; RV32IA-NEXT:    and a5, a5, a4
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:  .LBB107_3: # in Loop: Header=BB107_1 Depth=1
-; RV32IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB107_1
-; RV32IA-NEXT:  # %bb.4:
-; RV32IA-NEXT:    srl a0, a3, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_umin_i16_release:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    lui a3, 16
+; RV32IA-WMO-NEXT:    addi a3, a3, -1
+; RV32IA-WMO-NEXT:    sll a4, a3, a0
+; RV32IA-WMO-NEXT:    and a1, a1, a3
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB107_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w a3, (a2)
+; RV32IA-WMO-NEXT:    and a6, a3, a4
+; RV32IA-WMO-NEXT:    mv a5, a3
+; RV32IA-WMO-NEXT:    bgeu a1, a6, .LBB107_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB107_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a5, a3, a1
+; RV32IA-WMO-NEXT:    and a5, a5, a4
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:  .LBB107_3: # in Loop: Header=BB107_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB107_1
+; RV32IA-WMO-NEXT:  # %bb.4:
+; RV32IA-WMO-NEXT:    srl a0, a3, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_umin_i16_release:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    lui a3, 16
+; RV32IA-TSO-NEXT:    addi a3, a3, -1
+; RV32IA-TSO-NEXT:    sll a4, a3, a0
+; RV32IA-TSO-NEXT:    and a1, a1, a3
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB107_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a2)
+; RV32IA-TSO-NEXT:    and a6, a3, a4
+; RV32IA-TSO-NEXT:    mv a5, a3
+; RV32IA-TSO-NEXT:    bgeu a1, a6, .LBB107_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB107_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a5, a3, a1
+; RV32IA-TSO-NEXT:    and a5, a5, a4
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:  .LBB107_3: # in Loop: Header=BB107_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB107_1
+; RV32IA-TSO-NEXT:  # %bb.4:
+; RV32IA-TSO-NEXT:    srl a0, a3, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_umin_i16_release:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -48
 ; RV64I-NEXT:    sd ra, 40(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s0, 32(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s1, 24(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s2, 16(sp) # 8-byte Folded Spill

 ; RV64I-NEXT:    ld ra, 40(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s0, 32(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s1, 24(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s2, 16(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s3, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 48
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_umin_i16_release:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    lui a3, 16
-; RV64IA-NEXT:    addiw a3, a3, -1
-; RV64IA-NEXT:    sllw a4, a3, a0
-; RV64IA-NEXT:    and a1, a1, a3
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB107_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w a3, (a2)
-; RV64IA-NEXT:    and a6, a3, a4
-; RV64IA-NEXT:    mv a5, a3
-; RV64IA-NEXT:    bgeu a1, a6, .LBB107_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB107_1 Depth=1
-; RV64IA-NEXT:    xor a5, a3, a1
-; RV64IA-NEXT:    and a5, a5, a4
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:  .LBB107_3: # in Loop: Header=BB107_1 Depth=1
-; RV64IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB107_1
-; RV64IA-NEXT:  # %bb.4:
-; RV64IA-NEXT:    srlw a0, a3, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_umin_i16_release:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    lui a3, 16
+; RV64IA-WMO-NEXT:    addiw a3, a3, -1
+; RV64IA-WMO-NEXT:    sllw a4, a3, a0
+; RV64IA-WMO-NEXT:    and a1, a1, a3
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB107_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w a3, (a2)
+; RV64IA-WMO-NEXT:    and a6, a3, a4
+; RV64IA-WMO-NEXT:    mv a5, a3
+; RV64IA-WMO-NEXT:    bgeu a1, a6, .LBB107_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB107_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a5, a3, a1
+; RV64IA-WMO-NEXT:    and a5, a5, a4
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:  .LBB107_3: # in Loop: Header=BB107_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB107_1
+; RV64IA-WMO-NEXT:  # %bb.4:
+; RV64IA-WMO-NEXT:    srlw a0, a3, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_umin_i16_release:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    lui a3, 16
+; RV64IA-TSO-NEXT:    addiw a3, a3, -1
+; RV64IA-TSO-NEXT:    sllw a4, a3, a0
+; RV64IA-TSO-NEXT:    and a1, a1, a3
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB107_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a2)
+; RV64IA-TSO-NEXT:    and a6, a3, a4
+; RV64IA-TSO-NEXT:    mv a5, a3
+; RV64IA-TSO-NEXT:    bgeu a1, a6, .LBB107_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB107_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a5, a3, a1
+; RV64IA-TSO-NEXT:    and a5, a5, a4
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:  .LBB107_3: # in Loop: Header=BB107_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB107_1
+; RV64IA-TSO-NEXT:  # %bb.4:
+; RV64IA-TSO-NEXT:    srlw a0, a3, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw umin ptr %a, i16 %b release
   ret i16 %1
 }
 
 define i16 @atomicrmw_umin_i16_acq_rel(ptr %a, i16 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_umin_i16_acq_rel:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -32

 ; RV32I-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s2, 16(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    lw s3, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 32
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_umin_i16_acq_rel:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:    andi a2, a0, -4
-; RV32IA-NEXT:    slli a0, a0, 3
-; RV32IA-NEXT:    lui a3, 16
-; RV32IA-NEXT:    addi a3, a3, -1
-; RV32IA-NEXT:    sll a4, a3, a0
-; RV32IA-NEXT:    and a1, a1, a3
-; RV32IA-NEXT:    sll a1, a1, a0
-; RV32IA-NEXT:  .LBB108_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a3, (a2)
-; RV32IA-NEXT:    and a6, a3, a4
-; RV32IA-NEXT:    mv a5, a3
-; RV32IA-NEXT:    bgeu a1, a6, .LBB108_3
-; RV32IA-NEXT:  # %bb.2: # in Loop: Header=BB108_1 Depth=1
-; RV32IA-NEXT:    xor a5, a3, a1
-; RV32IA-NEXT:    and a5, a5, a4
-; RV32IA-NEXT:    xor a5, a3, a5
-; RV32IA-NEXT:  .LBB108_3: # in Loop: Header=BB108_1 Depth=1
-; RV32IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV32IA-NEXT:    bnez a5, .LBB108_1
-; RV32IA-NEXT:  # %bb.4:
-; RV32IA-NEXT:    srl a0, a3, a0
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_umin_i16_acq_rel:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:    andi a2, a0, -4
+; RV32IA-WMO-NEXT:    slli a0, a0, 3
+; RV32IA-WMO-NEXT:    lui a3, 16
+; RV32IA-WMO-NEXT:    addi a3, a3, -1
+; RV32IA-WMO-NEXT:    sll a4, a3, a0
+; RV32IA-WMO-NEXT:    and a1, a1, a3
+; RV32IA-WMO-NEXT:    sll a1, a1, a0
+; RV32IA-WMO-NEXT:  .LBB108_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV32IA-WMO-NEXT:    and a6, a3, a4
+; RV32IA-WMO-NEXT:    mv a5, a3
+; RV32IA-WMO-NEXT:    bgeu a1, a6, .LBB108_3
+; RV32IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB108_1 Depth=1
+; RV32IA-WMO-NEXT:    xor a5, a3, a1
+; RV32IA-WMO-NEXT:    and a5, a5, a4
+; RV32IA-WMO-NEXT:    xor a5, a3, a5
+; RV32IA-WMO-NEXT:  .LBB108_3: # in Loop: Header=BB108_1 Depth=1
+; RV32IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV32IA-WMO-NEXT:    bnez a5, .LBB108_1
+; RV32IA-WMO-NEXT:  # %bb.4:
+; RV32IA-WMO-NEXT:    srl a0, a3, a0
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_umin_i16_acq_rel:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:    andi a2, a0, -4
+; RV32IA-TSO-NEXT:    slli a0, a0, 3
+; RV32IA-TSO-NEXT:    lui a3, 16
+; RV32IA-TSO-NEXT:    addi a3, a3, -1
+; RV32IA-TSO-NEXT:    sll a4, a3, a0
+; RV32IA-TSO-NEXT:    and a1, a1, a3
+; RV32IA-TSO-NEXT:    sll a1, a1, a0
+; RV32IA-TSO-NEXT:  .LBB108_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a3, (a2)
+; RV32IA-TSO-NEXT:    and a6, a3, a4
+; RV32IA-TSO-NEXT:    mv a5, a3
+; RV32IA-TSO-NEXT:    bgeu a1, a6, .LBB108_3
+; RV32IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB108_1 Depth=1
+; RV32IA-TSO-NEXT:    xor a5, a3, a1
+; RV32IA-TSO-NEXT:    and a5, a5, a4
+; RV32IA-TSO-NEXT:    xor a5, a3, a5
+; RV32IA-TSO-NEXT:  .LBB108_3: # in Loop: Header=BB108_1 Depth=1
+; RV32IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV32IA-TSO-NEXT:    bnez a5, .LBB108_1
+; RV32IA-TSO-NEXT:  # %bb.4:
+; RV32IA-TSO-NEXT:    srl a0, a3, a0
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_umin_i16_acq_rel:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -48
 ; RV64I-NEXT:    sd ra, 40(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s0, 32(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s1, 24(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    sd s2, 16(sp) # 8-byte Folded Spill

 ; RV64I-NEXT:    ld ra, 40(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s0, 32(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s1, 24(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s2, 16(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    ld s3, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 48
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_umin_i16_acq_rel:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:    andi a2, a0, -4
-; RV64IA-NEXT:    slli a0, a0, 3
-; RV64IA-NEXT:    lui a3, 16
-; RV64IA-NEXT:    addiw a3, a3, -1
-; RV64IA-NEXT:    sllw a4, a3, a0
-; RV64IA-NEXT:    and a1, a1, a3
-; RV64IA-NEXT:    sllw a1, a1, a0
-; RV64IA-NEXT:  .LBB108_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a3, (a2)
-; RV64IA-NEXT:    and a6, a3, a4
-; RV64IA-NEXT:    mv a5, a3
-; RV64IA-NEXT:    bgeu a1, a6, .LBB108_3
-; RV64IA-NEXT:  # %bb.2: # in Loop: Header=BB108_1 Depth=1
-; RV64IA-NEXT:    xor a5, a3, a1
-; RV64IA-NEXT:    and a5, a5, a4
-; RV64IA-NEXT:    xor a5, a3, a5
-; RV64IA-NEXT:  .LBB108_3: # in Loop: Header=BB108_1 Depth=1
-; RV64IA-NEXT:    sc.w.rl a5, a5, (a2)
-; RV64IA-NEXT:    bnez a5, .LBB108_1
-; RV64IA-NEXT:  # %bb.4:
-; RV64IA-NEXT:    srlw a0, a3, a0
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_umin_i16_acq_rel:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:    andi a2, a0, -4
+; RV64IA-WMO-NEXT:    slli a0, a0, 3
+; RV64IA-WMO-NEXT:    lui a3, 16
+; RV64IA-WMO-NEXT:    addiw a3, a3, -1
+; RV64IA-WMO-NEXT:    sllw a4, a3, a0
+; RV64IA-WMO-NEXT:    and a1, a1, a3
+; RV64IA-WMO-NEXT:    sllw a1, a1, a0
+; RV64IA-WMO-NEXT:  .LBB108_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a3, (a2)
+; RV64IA-WMO-NEXT:    and a6, a3, a4
+; RV64IA-WMO-NEXT:    mv a5, a3
+; RV64IA-WMO-NEXT:    bgeu a1, a6, .LBB108_3
+; RV64IA-WMO-NEXT:  # %bb.2: # in Loop: Header=BB108_1 Depth=1
+; RV64IA-WMO-NEXT:    xor a5, a3, a1
+; RV64IA-WMO-NEXT:    and a5, a5, a4
+; RV64IA-WMO-NEXT:    xor a5, a3, a5
+; RV64IA-WMO-NEXT:  .LBB108_3: # in Loop: Header=BB108_1 Depth=1
+; RV64IA-WMO-NEXT:    sc.w.rl a5, a5, (a2)
+; RV64IA-WMO-NEXT:    bnez a5, .LBB108_1
+; RV64IA-WMO-NEXT:  # %bb.4:
+; RV64IA-WMO-NEXT:    srlw a0, a3, a0
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_umin_i16_acq_rel:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:    andi a2, a0, -4
+; RV64IA-TSO-NEXT:    slli a0, a0, 3
+; RV64IA-TSO-NEXT:    lui a3, 16
+; RV64IA-TSO-NEXT:    addiw a3, a3, -1
+; RV64IA-TSO-NEXT:    sllw a4, a3, a0
+; RV64IA-TSO-NEXT:    and a1, a1, a3
+; RV64IA-TSO-NEXT:    sllw a1, a1, a0
+; RV64IA-TSO-NEXT:  .LBB108_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a3, (a2)
+; RV64IA-TSO-NEXT:    and a6, a3, a4
+; RV64IA-TSO-NEXT:    mv a5, a3
+; RV64IA-TSO-NEXT:    bgeu a1, a6, .LBB108_3
+; RV64IA-TSO-NEXT:  # %bb.2: # in Loop: Header=BB108_1 Depth=1
+; RV64IA-TSO-NEXT:    xor a5, a3, a1
+; RV64IA-TSO-NEXT:    and a5, a5, a4
+; RV64IA-TSO-NEXT:    xor a5, a3, a5
+; RV64IA-TSO-NEXT:  .LBB108_3: # in Loop: Header=BB108_1 Depth=1
+; RV64IA-TSO-NEXT:    sc.w a5, a5, (a2)
+; RV64IA-TSO-NEXT:    bnez a5, .LBB108_1
+; RV64IA-TSO-NEXT:  # %bb.4:
+; RV64IA-TSO-NEXT:    srlw a0, a3, a0
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw umin ptr %a, i16 %b acq_rel
   ret i16 %1
 }
 
 define i16 @atomicrmw_umin_i16_seq_cst(ptr %a, i16 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_umin_i16_seq_cst:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -32

 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    li a2, 2
 ; RV32I-NEXT:    call __atomic_fetch_nand_4@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_nand_i32_acquire:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:  .LBB131_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a2, (a0)
-; RV32IA-NEXT:    and a3, a2, a1
-; RV32IA-NEXT:    not a3, a3
-; RV32IA-NEXT:    sc.w a3, a3, (a0)
-; RV32IA-NEXT:    bnez a3, .LBB131_1
-; RV32IA-NEXT:  # %bb.2:
-; RV32IA-NEXT:    mv a0, a2
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_nand_i32_acquire:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:  .LBB131_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a2, (a0)
+; RV32IA-WMO-NEXT:    and a3, a2, a1
+; RV32IA-WMO-NEXT:    not a3, a3
+; RV32IA-WMO-NEXT:    sc.w a3, a3, (a0)
+; RV32IA-WMO-NEXT:    bnez a3, .LBB131_1
+; RV32IA-WMO-NEXT:  # %bb.2:
+; RV32IA-WMO-NEXT:    mv a0, a2
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_nand_i32_acquire:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:  .LBB131_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a2, (a0)
+; RV32IA-TSO-NEXT:    and a3, a2, a1
+; RV32IA-TSO-NEXT:    not a3, a3
+; RV32IA-TSO-NEXT:    sc.w a3, a3, (a0)
+; RV32IA-TSO-NEXT:    bnez a3, .LBB131_1
+; RV32IA-TSO-NEXT:  # %bb.2:
+; RV32IA-TSO-NEXT:    mv a0, a2
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_nand_i32_acquire:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 2
 ; RV64I-NEXT:    call __atomic_fetch_nand_4@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_nand_i32_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:  .LBB131_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a2, (a0)
-; RV64IA-NEXT:    and a3, a2, a1
-; RV64IA-NEXT:    not a3, a3
-; RV64IA-NEXT:    sc.w a3, a3, (a0)
-; RV64IA-NEXT:    bnez a3, .LBB131_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    mv a0, a2
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_nand_i32_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:  .LBB131_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a2, (a0)
+; RV64IA-WMO-NEXT:    and a3, a2, a1
+; RV64IA-WMO-NEXT:    not a3, a3
+; RV64IA-WMO-NEXT:    sc.w a3, a3, (a0)
+; RV64IA-WMO-NEXT:    bnez a3, .LBB131_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    mv a0, a2
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_nand_i32_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:  .LBB131_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a2, (a0)
+; RV64IA-TSO-NEXT:    and a3, a2, a1
+; RV64IA-TSO-NEXT:    not a3, a3
+; RV64IA-TSO-NEXT:    sc.w a3, a3, (a0)
+; RV64IA-TSO-NEXT:    bnez a3, .LBB131_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    mv a0, a2
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw nand ptr %a, i32 %b acquire
   ret i32 %1
 }
 
 define i32 @atomicrmw_nand_i32_release(ptr %a, i32 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_nand_i32_release:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    li a2, 3
 ; RV32I-NEXT:    call __atomic_fetch_nand_4@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_nand_i32_release:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:  .LBB132_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w a2, (a0)
-; RV32IA-NEXT:    and a3, a2, a1
-; RV32IA-NEXT:    not a3, a3
-; RV32IA-NEXT:    sc.w.rl a3, a3, (a0)
-; RV32IA-NEXT:    bnez a3, .LBB132_1
-; RV32IA-NEXT:  # %bb.2:
-; RV32IA-NEXT:    mv a0, a2
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_nand_i32_release:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:  .LBB132_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w a2, (a0)
+; RV32IA-WMO-NEXT:    and a3, a2, a1
+; RV32IA-WMO-NEXT:    not a3, a3
+; RV32IA-WMO-NEXT:    sc.w.rl a3, a3, (a0)
+; RV32IA-WMO-NEXT:    bnez a3, .LBB132_1
+; RV32IA-WMO-NEXT:  # %bb.2:
+; RV32IA-WMO-NEXT:    mv a0, a2
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_nand_i32_release:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:  .LBB132_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a2, (a0)
+; RV32IA-TSO-NEXT:    and a3, a2, a1
+; RV32IA-TSO-NEXT:    not a3, a3
+; RV32IA-TSO-NEXT:    sc.w a3, a3, (a0)
+; RV32IA-TSO-NEXT:    bnez a3, .LBB132_1
+; RV32IA-TSO-NEXT:  # %bb.2:
+; RV32IA-TSO-NEXT:    mv a0, a2
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_nand_i32_release:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 3
 ; RV64I-NEXT:    call __atomic_fetch_nand_4@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_nand_i32_release:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:  .LBB132_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w a2, (a0)
-; RV64IA-NEXT:    and a3, a2, a1
-; RV64IA-NEXT:    not a3, a3
-; RV64IA-NEXT:    sc.w.rl a3, a3, (a0)
-; RV64IA-NEXT:    bnez a3, .LBB132_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    mv a0, a2
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_nand_i32_release:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:  .LBB132_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w a2, (a0)
+; RV64IA-WMO-NEXT:    and a3, a2, a1
+; RV64IA-WMO-NEXT:    not a3, a3
+; RV64IA-WMO-NEXT:    sc.w.rl a3, a3, (a0)
+; RV64IA-WMO-NEXT:    bnez a3, .LBB132_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    mv a0, a2
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_nand_i32_release:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:  .LBB132_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a2, (a0)
+; RV64IA-TSO-NEXT:    and a3, a2, a1
+; RV64IA-TSO-NEXT:    not a3, a3
+; RV64IA-TSO-NEXT:    sc.w a3, a3, (a0)
+; RV64IA-TSO-NEXT:    bnez a3, .LBB132_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    mv a0, a2
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw nand ptr %a, i32 %b release
   ret i32 %1
 }
 
 define i32 @atomicrmw_nand_i32_acq_rel(ptr %a, i32 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_nand_i32_acq_rel:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16
 ; RV32I-NEXT:    sw ra, 12(sp) # 4-byte Folded Spill
 ; RV32I-NEXT:    li a2, 4
 ; RV32I-NEXT:    call __atomic_fetch_nand_4@plt
 ; RV32I-NEXT:    lw ra, 12(sp) # 4-byte Folded Reload
 ; RV32I-NEXT:    addi sp, sp, 16
 ; RV32I-NEXT:    ret
 ;
-; RV32IA-LABEL: atomicrmw_nand_i32_acq_rel:
-; RV32IA:       # %bb.0:
-; RV32IA-NEXT:  .LBB133_1: # =>This Inner Loop Header: Depth=1
-; RV32IA-NEXT:    lr.w.aq a2, (a0)
-; RV32IA-NEXT:    and a3, a2, a1
-; RV32IA-NEXT:    not a3, a3
-; RV32IA-NEXT:    sc.w.rl a3, a3, (a0)
-; RV32IA-NEXT:    bnez a3, .LBB133_1
-; RV32IA-NEXT:  # %bb.2:
-; RV32IA-NEXT:    mv a0, a2
-; RV32IA-NEXT:    ret
+; RV32IA-WMO-LABEL: atomicrmw_nand_i32_acq_rel:
+; RV32IA-WMO:       # %bb.0:
+; RV32IA-WMO-NEXT:  .LBB133_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-WMO-NEXT:    lr.w.aq a2, (a0)
+; RV32IA-WMO-NEXT:    and a3, a2, a1
+; RV32IA-WMO-NEXT:    not a3, a3
+; RV32IA-WMO-NEXT:    sc.w.rl a3, a3, (a0)
+; RV32IA-WMO-NEXT:    bnez a3, .LBB133_1
+; RV32IA-WMO-NEXT:  # %bb.2:
+; RV32IA-WMO-NEXT:    mv a0, a2
+; RV32IA-WMO-NEXT:    ret
+;
+; RV32IA-TSO-LABEL: atomicrmw_nand_i32_acq_rel:
+; RV32IA-TSO:       # %bb.0:
+; RV32IA-TSO-NEXT:  .LBB133_1: # =>This Inner Loop Header: Depth=1
+; RV32IA-TSO-NEXT:    lr.w a2, (a0)
+; RV32IA-TSO-NEXT:    and a3, a2, a1
+; RV32IA-TSO-NEXT:    not a3, a3
+; RV32IA-TSO-NEXT:    sc.w a3, a3, (a0)
+; RV32IA-TSO-NEXT:    bnez a3, .LBB133_1
+; RV32IA-TSO-NEXT:  # %bb.2:
+; RV32IA-TSO-NEXT:    mv a0, a2
+; RV32IA-TSO-NEXT:    ret
 ;
 ; RV64I-LABEL: atomicrmw_nand_i32_acq_rel:
 ; RV64I:       # %bb.0:
 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 4
 ; RV64I-NEXT:    call __atomic_fetch_nand_4@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_nand_i32_acq_rel:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:  .LBB133_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.w.aq a2, (a0)
-; RV64IA-NEXT:    and a3, a2, a1
-; RV64IA-NEXT:    not a3, a3
-; RV64IA-NEXT:    sc.w.rl a3, a3, (a0)
-; RV64IA-NEXT:    bnez a3, .LBB133_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    mv a0, a2
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_nand_i32_acq_rel:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:  .LBB133_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.w.aq a2, (a0)
+; RV64IA-WMO-NEXT:    and a3, a2, a1
+; RV64IA-WMO-NEXT:    not a3, a3
+; RV64IA-WMO-NEXT:    sc.w.rl a3, a3, (a0)
+; RV64IA-WMO-NEXT:    bnez a3, .LBB133_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    mv a0, a2
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_nand_i32_acq_rel:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:  .LBB133_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.w a2, (a0)
+; RV64IA-TSO-NEXT:    and a3, a2, a1
+; RV64IA-TSO-NEXT:    not a3, a3
+; RV64IA-TSO-NEXT:    sc.w a3, a3, (a0)
+; RV64IA-TSO-NEXT:    bnez a3, .LBB133_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    mv a0, a2
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw nand ptr %a, i32 %b acq_rel
   ret i32 %1
 }
 
 define i32 @atomicrmw_nand_i32_seq_cst(ptr %a, i32 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_nand_i32_seq_cst:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16

 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 2
 ; RV64I-NEXT:    call __atomic_fetch_nand_8@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_nand_i64_acquire:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:  .LBB186_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.d.aq a2, (a0)
-; RV64IA-NEXT:    and a3, a2, a1
-; RV64IA-NEXT:    not a3, a3
-; RV64IA-NEXT:    sc.d a3, a3, (a0)
-; RV64IA-NEXT:    bnez a3, .LBB186_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    mv a0, a2
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_nand_i64_acquire:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:  .LBB186_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.d.aq a2, (a0)
+; RV64IA-WMO-NEXT:    and a3, a2, a1
+; RV64IA-WMO-NEXT:    not a3, a3
+; RV64IA-WMO-NEXT:    sc.d a3, a3, (a0)
+; RV64IA-WMO-NEXT:    bnez a3, .LBB186_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    mv a0, a2
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_nand_i64_acquire:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:  .LBB186_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.d a2, (a0)
+; RV64IA-TSO-NEXT:    and a3, a2, a1
+; RV64IA-TSO-NEXT:    not a3, a3
+; RV64IA-TSO-NEXT:    sc.d a3, a3, (a0)
+; RV64IA-TSO-NEXT:    bnez a3, .LBB186_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    mv a0, a2
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw nand ptr %a, i64 %b acquire
   ret i64 %1
 }
 
 define i64 @atomicrmw_nand_i64_release(ptr %a, i64 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_nand_i64_release:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16

 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 3
 ; RV64I-NEXT:    call __atomic_fetch_nand_8@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_nand_i64_release:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:  .LBB187_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.d a2, (a0)
-; RV64IA-NEXT:    and a3, a2, a1
-; RV64IA-NEXT:    not a3, a3
-; RV64IA-NEXT:    sc.d.rl a3, a3, (a0)
-; RV64IA-NEXT:    bnez a3, .LBB187_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    mv a0, a2
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_nand_i64_release:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:  .LBB187_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.d a2, (a0)
+; RV64IA-WMO-NEXT:    and a3, a2, a1
+; RV64IA-WMO-NEXT:    not a3, a3
+; RV64IA-WMO-NEXT:    sc.d.rl a3, a3, (a0)
+; RV64IA-WMO-NEXT:    bnez a3, .LBB187_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    mv a0, a2
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_nand_i64_release:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:  .LBB187_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.d a2, (a0)
+; RV64IA-TSO-NEXT:    and a3, a2, a1
+; RV64IA-TSO-NEXT:    not a3, a3
+; RV64IA-TSO-NEXT:    sc.d a3, a3, (a0)
+; RV64IA-TSO-NEXT:    bnez a3, .LBB187_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    mv a0, a2
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw nand ptr %a, i64 %b release
   ret i64 %1
 }
 
 define i64 @atomicrmw_nand_i64_acq_rel(ptr %a, i64 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_nand_i64_acq_rel:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16

 ; RV64I-NEXT:    addi sp, sp, -16
 ; RV64I-NEXT:    sd ra, 8(sp) # 8-byte Folded Spill
 ; RV64I-NEXT:    li a2, 4
 ; RV64I-NEXT:    call __atomic_fetch_nand_8@plt
 ; RV64I-NEXT:    ld ra, 8(sp) # 8-byte Folded Reload
 ; RV64I-NEXT:    addi sp, sp, 16
 ; RV64I-NEXT:    ret
 ;
-; RV64IA-LABEL: atomicrmw_nand_i64_acq_rel:
-; RV64IA:       # %bb.0:
-; RV64IA-NEXT:  .LBB188_1: # =>This Inner Loop Header: Depth=1
-; RV64IA-NEXT:    lr.d.aq a2, (a0)
-; RV64IA-NEXT:    and a3, a2, a1
-; RV64IA-NEXT:    not a3, a3
-; RV64IA-NEXT:    sc.d.rl a3, a3, (a0)
-; RV64IA-NEXT:    bnez a3, .LBB188_1
-; RV64IA-NEXT:  # %bb.2:
-; RV64IA-NEXT:    mv a0, a2
-; RV64IA-NEXT:    ret
+; RV64IA-WMO-LABEL: atomicrmw_nand_i64_acq_rel:
+; RV64IA-WMO:       # %bb.0:
+; RV64IA-WMO-NEXT:  .LBB188_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-WMO-NEXT:    lr.d.aq a2, (a0)
+; RV64IA-WMO-NEXT:    and a3, a2, a1
+; RV64IA-WMO-NEXT:    not a3, a3
+; RV64IA-WMO-NEXT:    sc.d.rl a3, a3, (a0)
+; RV64IA-WMO-NEXT:    bnez a3, .LBB188_1
+; RV64IA-WMO-NEXT:  # %bb.2:
+; RV64IA-WMO-NEXT:    mv a0, a2
+; RV64IA-WMO-NEXT:    ret
+;
+; RV64IA-TSO-LABEL: atomicrmw_nand_i64_acq_rel:
+; RV64IA-TSO:       # %bb.0:
+; RV64IA-TSO-NEXT:  .LBB188_1: # =>This Inner Loop Header: Depth=1
+; RV64IA-TSO-NEXT:    lr.d a2, (a0)
+; RV64IA-TSO-NEXT:    and a3, a2, a1
+; RV64IA-TSO-NEXT:    not a3, a3
+; RV64IA-TSO-NEXT:    sc.d a3, a3, (a0)
+; RV64IA-TSO-NEXT:    bnez a3, .LBB188_1
+; RV64IA-TSO-NEXT:  # %bb.2:
+; RV64IA-TSO-NEXT:    mv a0, a2
+; RV64IA-TSO-NEXT:    ret
   %1 = atomicrmw nand ptr %a, i64 %b acq_rel
   ret i64 %1
 }
 
 define i64 @atomicrmw_nand_i64_seq_cst(ptr %a, i64 %b) nounwind {
 ; RV32I-LABEL: atomicrmw_nand_i64_seq_cst:
 ; RV32I:       # %bb.0:
 ; RV32I-NEXT:    addi sp, sp, -16