[CodeGen] Use MIR syntax for MachineMemOperand printing

[thegameg]

Summary

Get rid of the "; mem:" suffix and use the one we use in MIR: ":: (load 2)".

[thegameg]

lib/CodeGen/MachineOperand.cpp

992

I just noticed this was printed before and won't be printed anymore after this patch.

IIUC, here we print the alignment of the base if it is affected by the offset, while in MIR and in this file at the line 999 we only care about the base alignment and only print the alignment of the whole reference.

Please correct me if I'm wrong.

[thegameg]

lib/CodeGen/MachineOperand.cpp

992

The only test in this case is CodeGen/AMDGPU/extload-align.ll.

The original is:

BUFFER_STORE_SHORT_OFFEN %13, %stack.0, %sgpr96_sgpr97_sgpr98_sgpr99, %sgpr3, 2, 0, 0, 0, implicit %exec; mem:ST2[FixedStack0(addrspace=5)(align=8)+2](align=2)

after this patch:

BUFFER_STORE_SHORT_OFFEN %13, %stack.0, %sgpr96_sgpr97_sgpr98_sgpr99, %sgpr3, 2, 0, 0, 0, implicit %exec :: (store 2 into %stack.0 + 2, align 8, addrspace 5)

The MIR parser expects that alignment to be the base alignment, while when reading it it looks like this is the alignment of the whole reference.

[thegameg]

Ping?

• Added support for addrspace in r323521.
• Rebase

[thegameg]

Ping? This is one of the latest changes for MIR / -debug unification.

[thegameg]

Ping.

[thegameg]

• Rebase
• Fix new tests:
  • test/CodeGen/AArch64/arm64-misched-memdep-bug.ll
  • test/CodeGen/ARM/misched-int-basic-thumb2.mir
  • test/CodeGen/ARM/single-issue-r52.mir
  • test/CodeGen/Hexagon/post-inc-aa-metadata.ll
  • test/CodeGen/PowerPC/byval-agg-info.ll

[bogner]

include/llvm/CodeGen/MachineMemOperand.h

299

Should this be SmallVectorImpl<StringRef> rather than forcing the size? Maybe even an ArrayRef?

lib/CodeGen/MachineInstr.cpp

1439–1448

This is just a dummy vector for the API right? We should be able to make it take less space. Though if we make the argument an ArrayRef we can just pass in None and save the variable entirely.

1446–1447

Would it be clearer to just use "else"?

lib/CodeGen/MachineOperand.cpp

469–480

I'd probably skip the extraneous curly braces here, since you aren't declaring any variables in these cases.

782

Is this an unrelated bug fix that snuck in or something? It isn't obvious to me why it's changing.

[thegameg]

Thanks for the review Justin!

include/llvm/CodeGen/MachineMemOperand.h

299

SmallVectorImpl<StringRef> is better, yes. I wish I could make it an ArrayRef but printSyncScope is caching the results from LLVMContext::getSyncScopeNames in SSNs, so we can't use ArrayRef here.

lib/CodeGen/MachineInstr.cpp

1439–1448

I made it <StringRef, 0>, my guess is that most of the time we will have SyncScope::System and won't print anything, but honestly I don't know much about it.

lib/CodeGen/MachineOperand.cpp

782

My bad, this shouldn't be here.

[bogner]

lg

[craig.topper]

llvm/trunk/lib/CodeGen/SelectionDAG/SelectionDAGDumper.cpp

428 ↗

(On Diff #138456)

G isn't guaranteed non-null here. There's a version of SDNode::dump() that passes nullptr. This is most often used when debugging in gdb/lldb. Though I found a use in the X86 backend that I fixed in r327744.

[thegameg]

llvm/trunk/lib/CodeGen/SelectionDAG/SelectionDAGDumper.cpp

428 ↗

(On Diff #138456)

Sorry about that. Put up a patch to fix this: https://reviews.llvm.org/D44611.

include/llvm/CodeGen/MachineMemOperand.h

public:
   /// Profile - Gather unique data for the object.
   ///
   void Profile(FoldingSetNodeID &ID) const;
 
   /// Support for operator<<.
   /// @{
   void print(raw_ostream &OS) const;
   void print(raw_ostream &OS, ModuleSlotTracker &MST) const;
+  void print(raw_ostream &OS, ModuleSlotTracker &MST,
+             SmallVectorImpl<StringRef> &SSNs, const LLVMContext &Context,
+             const MachineFrameInfo *MFI, const TargetInstrInfo *TII) const;
   /// @}
 
   friend bool operator==(const MachineMemOperand &LHS,
                          const MachineMemOperand &RHS) {
     return LHS.getValue() == RHS.getValue() &&
            LHS.getPseudoValue() == RHS.getPseudoValue() &&
            LHS.getSize() == RHS.getSize() &&
            LHS.getOffset() == RHS.getOffset() &&

lib/CodeGen/MIRPrinter.cpp

 
 #include "llvm/CodeGen/MIRPrinter.h"
 #include "llvm/ADT/DenseMap.h"
 #include "llvm/ADT/None.h"
 #include "llvm/ADT/STLExtras.h"
 #include "llvm/ADT/SmallBitVector.h"
 #include "llvm/ADT/SmallPtrSet.h"
 #include "llvm/ADT/SmallVector.h"
-#include "llvm/ADT/StringExtras.h"
 #include "llvm/ADT/StringRef.h"
 #include "llvm/ADT/Twine.h"
 #include "llvm/CodeGen/GlobalISel/RegisterBank.h"
 #include "llvm/CodeGen/MIRYamlMapping.h"
 #include "llvm/CodeGen/MachineBasicBlock.h"
 #include "llvm/CodeGen/MachineConstantPool.h"
 #include "llvm/CodeGen/MachineFrameInfo.h"
 #include "llvm/CodeGen/MachineFunction.h"
MIPrinter(raw_ostream &OS, ModuleSlotTracker &MST,
             const DenseMap<const uint32_t *, unsigned> &RegisterMaskIds,
             const DenseMap<int, FrameIndexOperand> &StackObjectOperandMapping)
       : OS(OS), MST(MST), RegisterMaskIds(RegisterMaskIds),
         StackObjectOperandMapping(StackObjectOperandMapping) {}
 
   void print(const MachineBasicBlock &MBB);
 
   void print(const MachineInstr &MI);
-  void printIRValueReference(const Value &V);
   void printStackObjectReference(int FrameIndex);
   void print(const MachineInstr &MI, unsigned OpIdx,
              const TargetRegisterInfo *TRI, bool ShouldPrintRegisterTies,
              LLT TypeToPrint, bool PrintDef = true);
-  void print(const LLVMContext &Context, const TargetInstrInfo &TII,
-             const MachineMemOperand &Op);
-  void printSyncScope(const LLVMContext &Context, SyncScope::ID SSID);
 };
 
 } // end namespace llvm
 
 namespace llvm {
 namespace yaml {
 
 /// This struct serializes the LLVM IR module.
if (NeedComma)
       OS << ',';
     OS << " debug-location ";
     DL->printAsOperand(OS, MST);
   }
 
   if (!MI.memoperands_empty()) {
     OS << " :: ";
     const LLVMContext &Context = MF->getFunction().getContext();
+    const MachineFrameInfo &MFI = MF->getFrameInfo();
     bool NeedComma = false;
     for (const auto *Op : MI.memoperands()) {
       if (NeedComma)
         OS << ", ";
-      print(Context, *TII, *Op);
+      Op->print(OS, MST, SSNs, Context, &MFI, TII);
       NeedComma = true;
     }
   }
 }
 
-void MIPrinter::printIRValueReference(const Value &V) {
-  if (isa<GlobalValue>(V)) {
-    V.printAsOperand(OS, /*PrintType=*/false, MST);
-    return;
-  }
-  if (isa<Constant>(V)) {
-    // Machine memory operands can load/store to/from constant value pointers.
-    OS << '`';
-    V.printAsOperand(OS, /*PrintType=*/true, MST);
-    OS << '`';
-    return;
-  }
-  OS << "%ir.";
-  if (V.hasName()) {
-    printLLVMNameWithoutPrefix(OS, V.getName());
-    return;
-  }
-  MachineOperand::printIRSlotNumber(OS, MST.getLocalSlot(&V));
-}
-
 void MIPrinter::printStackObjectReference(int FrameIndex) {
   auto ObjectInfo = StackObjectOperandMapping.find(FrameIndex);
   assert(ObjectInfo != StackObjectOperandMapping.end() &&
          "Invalid frame index");
   const FrameIndexOperand &Operand = ObjectInfo->second;
   MachineOperand::printStackObjectReference(OS, Operand.ID, Operand.IsFixed,
                                             Operand.Name);
 }
if (RegMaskInfo != RegisterMaskIds.end())
       OS << StringRef(TRI->getRegMaskNames()[RegMaskInfo->second]).lower();
     else
       printCustomRegMask(Op.getRegMask(), OS, TRI);
     break;
   }
   }
 }
 
-static const char *getTargetMMOFlagName(const TargetInstrInfo &TII,
-                                        unsigned TMMOFlag) {
-  auto Flags = TII.getSerializableMachineMemOperandTargetFlags();
-  for (const auto &I : Flags) {
-    if (I.first == TMMOFlag) {
-      return I.second;
-    }
-  }
-  return nullptr;
-}
-
-void MIPrinter::print(const LLVMContext &Context, const TargetInstrInfo &TII,
-                      const MachineMemOperand &Op) {
-  OS << '(';
-  if (Op.isVolatile())
-    OS << "volatile ";
-  if (Op.isNonTemporal())
-    OS << "non-temporal ";
-  if (Op.isDereferenceable())
-    OS << "dereferenceable ";
-  if (Op.isInvariant())
-    OS << "invariant ";
-  if (Op.getFlags() & MachineMemOperand::MOTargetFlag1)
-    OS << '"' << getTargetMMOFlagName(TII, MachineMemOperand::MOTargetFlag1)
-       << "\" ";
-  if (Op.getFlags() & MachineMemOperand::MOTargetFlag2)
-    OS << '"' << getTargetMMOFlagName(TII, MachineMemOperand::MOTargetFlag2)
-       << "\" ";
-  if (Op.getFlags() & MachineMemOperand::MOTargetFlag3)
-    OS << '"' << getTargetMMOFlagName(TII, MachineMemOperand::MOTargetFlag3)
-       << "\" ";
-
-  assert((Op.isLoad() || Op.isStore()) && "machine memory operand must be a load or store (or both)");
-  if (Op.isLoad())
-    OS << "load ";
-  if (Op.isStore())
-    OS << "store ";
-
-  printSyncScope(Context, Op.getSyncScopeID());
-
-  if (Op.getOrdering() != AtomicOrdering::NotAtomic)
-    OS << toIRString(Op.getOrdering()) << ' ';
-  if (Op.getFailureOrdering() != AtomicOrdering::NotAtomic)
-    OS << toIRString(Op.getFailureOrdering()) << ' ';
-
-  OS << Op.getSize();
-  if (const Value *Val = Op.getValue()) {
-    OS << ((Op.isLoad() && Op.isStore()) ? " on "
-                                         : Op.isLoad() ? " from " : " into ");
-    printIRValueReference(*Val);
-  } else if (const PseudoSourceValue *PVal = Op.getPseudoValue()) {
-    OS << ((Op.isLoad() && Op.isStore()) ? " on "
-                                         : Op.isLoad() ? " from " : " into ");
-    assert(PVal && "Expected a pseudo source value");
-    switch (PVal->kind()) {
-    case PseudoSourceValue::Stack:
-      OS << "stack";
-      break;
-    case PseudoSourceValue::GOT:
-      OS << "got";
-      break;
-    case PseudoSourceValue::JumpTable:
-      OS << "jump-table";
-      break;
-    case PseudoSourceValue::ConstantPool:
-      OS << "constant-pool";
-      break;
-    case PseudoSourceValue::FixedStack:
-      printStackObjectReference(
-          cast<FixedStackPseudoSourceValue>(PVal)->getFrameIndex());
-      break;
-    case PseudoSourceValue::GlobalValueCallEntry:
-      OS << "call-entry ";
-      cast<GlobalValuePseudoSourceValue>(PVal)->getValue()->printAsOperand(
-          OS, /*PrintType=*/false, MST);
-      break;
-    case PseudoSourceValue::ExternalSymbolCallEntry:
-      OS << "call-entry &";
-      printLLVMNameWithoutPrefix(
-          OS, cast<ExternalSymbolPseudoSourceValue>(PVal)->getSymbol());
-      break;
-    case PseudoSourceValue::TargetCustom:
-      llvm_unreachable("TargetCustom pseudo source values are not supported");
-      break;
-    }
-  }
-  MachineOperand::printOperandOffset(OS, Op.getOffset());
-  if (Op.getBaseAlignment() != Op.getSize())
-    OS << ", align " << Op.getBaseAlignment();
-  auto AAInfo = Op.getAAInfo();
-  if (AAInfo.TBAA) {
-    OS << ", !tbaa ";
-    AAInfo.TBAA->printAsOperand(OS, MST);
-  }
-  if (AAInfo.Scope) {
-    OS << ", !alias.scope ";
-    AAInfo.Scope->printAsOperand(OS, MST);
-  }
-  if (AAInfo.NoAlias) {
-    OS << ", !noalias ";
-    AAInfo.NoAlias->printAsOperand(OS, MST);
-  }
-  if (Op.getRanges()) {
-    OS << ", !range ";
-    Op.getRanges()->printAsOperand(OS, MST);
-  }
-  if (unsigned AS = Op.getAddrSpace())
-    OS << ", addrspace " << AS;
-  OS << ')';
-}
-
-void MIPrinter::printSyncScope(const LLVMContext &Context, SyncScope::ID SSID) {
-  switch (SSID) {
-  case SyncScope::System: {
-    break;
-  }
-  default: {
-    if (SSNs.empty())
-      Context.getSyncScopeNames(SSNs);
-
-    OS << "syncscope(\"";
-    PrintEscapedString(SSNs[SSID], OS);
-    OS << "\") ";
-    break;
-  }
-  }
-}
-
 void llvm::printMIR(raw_ostream &OS, const Module &M) {
   yaml::Output Out(OS);
   Out << const_cast<Module &>(M);
 }
 
 void llvm::printMIR(raw_ostream &OS, const MachineFunction &MF) {
   MIRPrinter Printer(OS);
   Printer.print(MF);
 }

lib/CodeGen/MachineInstr.cpp

if (!SkipDebugLoc) {
     if (const DebugLoc &DL = getDebugLoc()) {
       if (!FirstOp)
         OS << ',';
       OS << " debug-location ";
       DL->printAsOperand(OS, MST);
     }
   }
 
-  bool HaveSemi = false;
   if (!memoperands_empty()) {
-    if (!HaveSemi) {
-      OS << ";";
-      HaveSemi = true;
+    SmallVector<StringRef, 0> SSNs;
+    const LLVMContext *Context = nullptr;
+    std::unique_ptr<LLVMContext> CtxPtr;
+    const MachineFrameInfo *MFI = nullptr;
+    if (const MachineFunction *MF = getMFIfAvailable(*this)) {
+      MFI = &MF->getFrameInfo();
+      Context = &MF->getFunction().getContext();
+    } else {
+      CtxPtr = llvm::make_unique<LLVMContext>();
+      Context = CtxPtr.get();
     }
 
-    OS << " mem:";
-    for (mmo_iterator i = memoperands_begin(), e = memoperands_end();
-         i != e; ++i) {
-      (*i)->print(OS, MST);
-      if (std::next(i) != e)
-        OS << " ";
+    OS << " :: ";
+    bool NeedComma = false;
+    for (const MachineMemOperand *Op : memoperands()) {
+      if (NeedComma)
+        OS << ", ";
+      Op->print(OS, MST, SSNs, *Context, MFI, TII);
+      NeedComma = true;
     }
   }
 
   if (SkipDebugLoc)
     return;
 
+  bool HaveSemi = false;
   // Print debug location information.
   if (isDebugValue() && getOperand(e - 2).isMetadata()) {
     if (!HaveSemi)
       OS << ";";
     auto *DV = cast<DILocalVariable>(getOperand(e - 2).getMetadata());
     OS << " line no:" <<  DV->getLine();
     if (auto *InlinedAt = debugLoc->getInlinedAt()) {
       DebugLoc InlinedAtDL(InlinedAt);

lib/CodeGen/MachineOperand.cpp

 //===- lib/CodeGen/MachineOperand.cpp -------------------------------------===//
 //
 //                     The LLVM Compiler Infrastructure
 //
 // This file is distributed under the University of Illinois Open Source
 // License. See LICENSE.TXT for details.
 //
 //===----------------------------------------------------------------------===//
 //
 /// \file Methods common to all machine operands.
 //
 //===----------------------------------------------------------------------===//
 
 #include "llvm/CodeGen/MachineOperand.h"
+#include "llvm/ADT/StringExtras.h"
 #include "llvm/Analysis/Loads.h"
 #include "llvm/CodeGen/MIRPrinter.h"
 #include "llvm/CodeGen/MachineFrameInfo.h"
 #include "llvm/CodeGen/MachineJumpTableInfo.h"
 #include "llvm/CodeGen/MachineRegisterInfo.h"
 #include "llvm/CodeGen/TargetInstrInfo.h"
 #include "llvm/CodeGen/TargetRegisterInfo.h"
 #include "llvm/IR/Constants.h"
if (const Function *F = BB.getParent()) {
     }
   }
   if (Slot)
     MachineOperand::printIRSlotNumber(OS, *Slot);
   else
     OS << "<unknown>";
 }
 
+static void printIRValueReference(raw_ostream &OS, const Value &V,
+                                  ModuleSlotTracker &MST) {
+  if (isa<GlobalValue>(V)) {
+    V.printAsOperand(OS, /*PrintType=*/false, MST);
+    return;
+  }
+  if (isa<Constant>(V)) {
+    // Machine memory operands can load/store to/from constant value pointers.
+    OS << '`';
+    V.printAsOperand(OS, /*PrintType=*/true, MST);
+    OS << '`';
+    return;
+  }
+  OS << "%ir.";
+  if (V.hasName()) {
+    printLLVMNameWithoutPrefix(OS, V.getName());
+    return;
+  }
+  MachineOperand::printIRSlotNumber(OS, MST.getLocalSlot(&V));
+}
+
+static void printSyncScope(raw_ostream &OS, const LLVMContext &Context,
+                           SyncScope::ID SSID,
+                           SmallVectorImpl<StringRef> &SSNs) {
+  switch (SSID) {
+  case SyncScope::System:
+    break;
+  default:
+    if (SSNs.empty())
+      Context.getSyncScopeNames(SSNs);
+
+    OS << "syncscope(\"";
+    PrintEscapedString(SSNs[SSID], OS);
+    OS << "\") ";
+    break;
+  }
+}
+
+static const char *getTargetMMOFlagName(const TargetInstrInfo &TII,
+                                        unsigned TMMOFlag) {
+  auto Flags = TII.getSerializableMachineMemOperandTargetFlags();
+  for (const auto &I : Flags) {
+    if (I.first == TMMOFlag) {
+      return I.second;
+    }
+  }
+  return nullptr;
+}
+
+static void printFrameIndex(raw_ostream& OS, int FrameIndex, bool IsFixed,
+                            const MachineFrameInfo *MFI) {
+  StringRef Name;
+  if (MFI) {
+    IsFixed = MFI->isFixedObjectIndex(FrameIndex);
+    if (const AllocaInst *Alloca = MFI->getObjectAllocation(FrameIndex))
+      if (Alloca->hasName())
+        Name = Alloca->getName();
+    if (IsFixed)
+      FrameIndex -= MFI->getObjectIndexBegin();
+  }
+  MachineOperand::printStackObjectReference(OS, FrameIndex, IsFixed, Name);
+}
+
 void MachineOperand::printSubRegIdx(raw_ostream &OS, uint64_t Index,
                                     const TargetRegisterInfo *TRI) {
   OS << "%subreg.";
   if (TRI)
     OS << TRI->getSubRegIndexName(Index);
   else
     OS << Index;
 }
void MachineOperand::print(raw_ostream &OS, ModuleSlotTracker &MST,
   case MachineOperand::MO_FPImmediate:
     getFPImm()->printAsOperand(OS, /*PrintType=*/true, MST);
     break;
   case MachineOperand::MO_MachineBasicBlock:
     OS << printMBBReference(*getMBB());
     break;
   case MachineOperand::MO_FrameIndex: {
     int FrameIndex = getIndex();
     bool IsFixed = false;
-    StringRef Name;
-    if (const MachineFunction *MF = getMFIfAvailable(*this)) {
-      const MachineFrameInfo &MFI = MF->getFrameInfo();
-      IsFixed = MFI.isFixedObjectIndex(FrameIndex);
-      if (const AllocaInst *Alloca = MFI.getObjectAllocation(FrameIndex))
-        if (Alloca->hasName())
-          Name = Alloca->getName();
-      if (IsFixed)
-        FrameIndex -= MFI.getObjectIndexBegin();
-    }
-    printStackObjectReference(OS, FrameIndex, IsFixed, Name);
+    const MachineFrameInfo *MFI = nullptr;
+    if (const MachineFunction *MF = getMFIfAvailable(*this))
+      MFI = &MF->getFrameInfo();
+    printFrameIndex(OS, FrameIndex, IsFixed, MFI);
     break;
   }
   case MachineOperand::MO_ConstantPoolIndex:
     OS << "%const." << getIndex();
     printOperandOffset(OS, getOffset());
     break;
   case MachineOperand::MO_TargetIndex: {
     OS << "target-index(";

 uint64_t MachineMemOperand::getAlignment() const {
   return MinAlign(getBaseAlignment(), getOffset());
 }
 
 void MachineMemOperand::print(raw_ostream &OS) const {
   ModuleSlotTracker DummyMST(nullptr);
   print(OS, DummyMST);
 }
+
 void MachineMemOperand::print(raw_ostream &OS, ModuleSlotTracker &MST) const {
-  assert((isLoad() || isStore()) && "SV has to be a load, store or both.");
+  SmallVector<StringRef, 0> SSNs;
+  LLVMContext Ctx;
+  print(OS, MST, SSNs, Ctx, nullptr, nullptr);
+}
 
+void MachineMemOperand::print(raw_ostream &OS, ModuleSlotTracker &MST,
+                              SmallVectorImpl<StringRef> &SSNs,
+                              const LLVMContext &Context,
+                              const MachineFrameInfo *MFI,
+                              const TargetInstrInfo *TII) const {
+  OS << '(';
   if (isVolatile())
-    OS << "Volatile ";
+    OS << "volatile ";
+  if (isNonTemporal())
+    OS << "non-temporal ";
+  if (isDereferenceable())
+    OS << "dereferenceable ";
+  if (isInvariant())
+    OS << "invariant ";
+  if (getFlags() & MachineMemOperand::MOTargetFlag1)
+    OS << '"' << getTargetMMOFlagName(*TII, MachineMemOperand::MOTargetFlag1)
+       << "\" ";
+  if (getFlags() & MachineMemOperand::MOTargetFlag2)
+    OS << '"' << getTargetMMOFlagName(*TII, MachineMemOperand::MOTargetFlag2)
+       << "\" ";
+  if (getFlags() & MachineMemOperand::MOTargetFlag3)
+    OS << '"' << getTargetMMOFlagName(*TII, MachineMemOperand::MOTargetFlag3)
+       << "\" ";
 
+  assert((isLoad() || isStore()) &&
+         "machine memory operand must be a load or store (or both)");
   if (isLoad())
-    OS << "LD";
+    OS << "load ";
   if (isStore())
-    OS << "ST";
-  OS << getSize();
-
-  // Print the address information.
-  OS << "[";
-  if (const Value *V = getValue())
-    V->printAsOperand(OS, /*PrintType=*/false, MST);
-  else if (const PseudoSourceValue *PSV = getPseudoValue())
-    PSV->printCustom(OS);
-  else
-    OS << "<unknown>";
+    OS << "store ";
 
-  unsigned AS = getAddrSpace();
-  if (AS != 0)
-    OS << "(addrspace=" << AS << ')';
-
-  // If the alignment of the memory reference itself differs from the alignment
-  // of the base pointer, print the base alignment explicitly, next to the base
-  // pointer.
-  if (getBaseAlignment() != getAlignment())
-    OS << "(align=" << getBaseAlignment() << ")";
-
-  if (getOffset() != 0)
-    OS << "+" << getOffset();
-  OS << "]";
-
-  // Print the alignment of the reference.
-  if (getBaseAlignment() != getAlignment() || getBaseAlignment() != getSize())
-    OS << "(align=" << getAlignment() << ")";
-
-  // Print TBAA info.
-  if (const MDNode *TBAAInfo = getAAInfo().TBAA) {
-    OS << "(tbaa=";
-    if (TBAAInfo->getNumOperands() > 0)
-      TBAAInfo->getOperand(0)->printAsOperand(OS, MST);
-    else
-      OS << "<unknown>";
-    OS << ")";
-  }
+  printSyncScope(OS, Context, getSyncScopeID(), SSNs);
 
-  // Print AA scope info.
-  if (const MDNode *ScopeInfo = getAAInfo().Scope) {
-    OS << "(alias.scope=";
-    if (ScopeInfo->getNumOperands() > 0)
-      for (unsigned i = 0, ie = ScopeInfo->getNumOperands(); i != ie; ++i) {
-        ScopeInfo->getOperand(i)->printAsOperand(OS, MST);
-        if (i != ie - 1)
-          OS << ",";
-      }
-    else
-      OS << "<unknown>";
-    OS << ")";
-  }
+  if (getOrdering() != AtomicOrdering::NotAtomic)
+    OS << toIRString(getOrdering()) << ' ';
+  if (getFailureOrdering() != AtomicOrdering::NotAtomic)
+    OS << toIRString(getFailureOrdering()) << ' ';
 
-  // Print AA noalias scope info.
-  if (const MDNode *NoAliasInfo = getAAInfo().NoAlias) {
-    OS << "(noalias=";
-    if (NoAliasInfo->getNumOperands() > 0)
-      for (unsigned i = 0, ie = NoAliasInfo->getNumOperands(); i != ie; ++i) {
-        NoAliasInfo->getOperand(i)->printAsOperand(OS, MST);
-        if (i != ie - 1)
-          OS << ",";
-      }
-    else
-      OS << "<unknown>";
-    OS << ")";
+  OS << getSize();
+  if (const Value *Val = getValue()) {
+    OS << ((isLoad() && isStore()) ? " on " : isLoad() ? " from " : " into ");
+    printIRValueReference(OS, *Val, MST);
+  } else if (const PseudoSourceValue *PVal = getPseudoValue()) {
+    OS << ((isLoad() && isStore()) ? " on " : isLoad() ? " from " : " into ");
+    assert(PVal && "Expected a pseudo source value");
+    switch (PVal->kind()) {
+    case PseudoSourceValue::Stack:
+      OS << "stack";
+      break;
+    case PseudoSourceValue::GOT:
+      OS << "got";
+      break;
+    case PseudoSourceValue::JumpTable:
+      OS << "jump-table";
+      break;
+    case PseudoSourceValue::ConstantPool:
+      OS << "constant-pool";
+      break;
+    case PseudoSourceValue::FixedStack: {
+      int FrameIndex = cast<FixedStackPseudoSourceValue>(PVal)->getFrameIndex();
+      bool IsFixed = true;
+      printFrameIndex(OS, FrameIndex, IsFixed, MFI);
+      break;
+    }
+    case PseudoSourceValue::GlobalValueCallEntry:
+      OS << "call-entry ";
+      cast<GlobalValuePseudoSourceValue>(PVal)->getValue()->printAsOperand(
+          OS, /*PrintType=*/false, MST);
+      break;
+    case PseudoSourceValue::ExternalSymbolCallEntry:
+      OS << "call-entry &";
+      printLLVMNameWithoutPrefix(
+          OS, cast<ExternalSymbolPseudoSourceValue>(PVal)->getSymbol());
+      break;
+    case PseudoSourceValue::TargetCustom:
+      llvm_unreachable("TargetCustom pseudo source values are not supported");
+      break;
     }
-
-  if (const MDNode *Ranges = getRanges()) {
-    unsigned NumRanges = Ranges->getNumOperands();
-    if (NumRanges != 0) {
-      OS << "(ranges=";
-
-      for (unsigned I = 0; I != NumRanges; ++I) {
-        Ranges->getOperand(I)->printAsOperand(OS, MST);
-        if (I != NumRanges - 1)
-          OS << ',';
-      }
+  MachineOperand::printOperandOffset(OS, getOffset());
+  if (getBaseAlignment() != getSize())
+    OS << ", align " << getBaseAlignment();
+  auto AAInfo = getAAInfo();
+  if (AAInfo.TBAA) {
+    OS << ", !tbaa ";
+    AAInfo.TBAA->printAsOperand(OS, MST);
+  }
+  if (AAInfo.Scope) {
+    OS << ", !alias.scope ";
+    AAInfo.Scope->printAsOperand(OS, MST);
+  }
+  if (AAInfo.NoAlias) {
+    OS << ", !noalias ";
+    AAInfo.NoAlias->printAsOperand(OS, MST);
+  }
+  if (getRanges()) {
+    OS << ", !range ";
+    getRanges()->printAsOperand(OS, MST);
+  }
+  // FIXME: Implement addrspace printing/parsing in MIR.
+  // For now, print this even though parsing it is not available in MIR.
+  if (unsigned AS = getAddrSpace())
+    OS << ", addrspace " << AS;
 
-      OS << ')';
-    }
-  }
-
-  if (isNonTemporal())
-    OS << "(nontemporal)";
-  if (isDereferenceable())
-    OS << "(dereferenceable)";
-  if (isInvariant())
-    OS << "(invariant)";
-  if (getFlags() & MOTargetFlag1)
-    OS << "(flag1)";
-  if (getFlags() & MOTargetFlag2)
-    OS << "(flag2)";
-  if (getFlags() & MOTargetFlag3)
-    OS << "(flag3)";
-}

lib/CodeGen/SelectionDAG/SelectionDAGDumper.cpp

 #include "llvm/CodeGen/TargetSubtargetInfo.h"
 #include "llvm/CodeGen/ValueTypes.h"
 #include "llvm/IR/BasicBlock.h"
 #include "llvm/IR/Constants.h"
 #include "llvm/IR/DebugInfoMetadata.h"
 #include "llvm/IR/DebugLoc.h"
 #include "llvm/IR/Function.h"
 #include "llvm/IR/Intrinsics.h"
+#include "llvm/IR/ModuleSlotTracker.h"
 #include "llvm/IR/Value.h"
 #include "llvm/Support/Casting.h"
 #include "llvm/Support/CommandLine.h"
 #include "llvm/Support/Compiler.h"
 #include "llvm/Support/Debug.h"
 #include "llvm/Support/ErrorHandling.h"
 #include "llvm/Support/Printable.h"
 #include "llvm/Support/raw_ostream.h"

 #ifndef NDEBUG
     OS << 't' << Node.PersistentId;
 #else
     OS << (const void*)&Node;
 #endif
   });
 }
 
+// Print the MMO with more information from the SelectionDAG.
+static void printMemOperand(raw_ostream &OS, const MachineMemOperand &MMO,
+                            const SelectionDAG *G) {
+  const MachineFunction &MF = G->getMachineFunction();
+  const Function &F = MF.getFunction();
+  const MachineFrameInfo &MFI = MF.getFrameInfo();
+  const TargetInstrInfo *TII = G->getSubtarget().getInstrInfo();
+  ModuleSlotTracker MST(F.getParent());
+  MST.incorporateFunction(F);
+  SmallVector<StringRef, 0> SSNs;
+  MMO.print(OS, MST, SSNs, *G->getContext(), &MFI, TII);
+}
+
 #if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
 LLVM_DUMP_METHOD void SDNode::dump() const { dump(nullptr); }
 
 LLVM_DUMP_METHOD void SDNode::dump(const SelectionDAG *G) const {
   print(dbgs(), G);
   dbgs() << '\n';
 }
 #endif
if (getFlags().hasVectorReduction())
     OS << " vector-reduction";
 
   if (const MachineSDNode *MN = dyn_cast<MachineSDNode>(this)) {
     if (!MN->memoperands_empty()) {
       OS << "<";
       OS << "Mem:";
       for (MachineSDNode::mmo_iterator i = MN->memoperands_begin(),
            e = MN->memoperands_end(); i != e; ++i) {
-        OS << **i;
+        printMemOperand(OS, **i, G);
         if (std::next(i) != e)
           OS << " ";
       }
       OS << ">";
     }
   } else if (const ShuffleVectorSDNode *SVN =
                dyn_cast<ShuffleVectorSDNode>(this)) {
     OS << "<";
if (const MachineSDNode *MN = dyn_cast<MachineSDNode>(this)) {
     if (MD->getMD())
       OS << "<" << MD->getMD() << ">";
     else
       OS << "<null>";
   } else if (const VTSDNode *N = dyn_cast<VTSDNode>(this)) {
     OS << ":" << N->getVT().getEVTString();
   }
   else if (const LoadSDNode *LD = dyn_cast<LoadSDNode>(this)) {
-    OS << "<" << *LD->getMemOperand();
+    OS << "<";
+
+    printMemOperand(OS, *LD->getMemOperand(), G);
 
     bool doExt = true;
     switch (LD->getExtensionType()) {
     default: doExt = false; break;
     case ISD::EXTLOAD:  OS << ", anyext"; break;
     case ISD::SEXTLOAD: OS << ", sext"; break;
     case ISD::ZEXTLOAD: OS << ", zext"; break;
     }
     if (doExt)
       OS << " from " << LD->getMemoryVT().getEVTString();
 
     const char *AM = getIndexedModeName(LD->getAddressingMode());
     if (*AM)
       OS << ", " << AM;
 
     OS << ">";
   } else if (const StoreSDNode *ST = dyn_cast<StoreSDNode>(this)) {
-    OS << "<" << *ST->getMemOperand();
+    OS << "<";
+    printMemOperand(OS, *ST->getMemOperand(), G);
 
     if (ST->isTruncatingStore())
       OS << ", trunc to " << ST->getMemoryVT().getEVTString();
 
     const char *AM = getIndexedModeName(ST->getAddressingMode());
     if (*AM)
       OS << ", " << AM;
 
     OS << ">";
   } else if (const MemSDNode* M = dyn_cast<MemSDNode>(this)) {
-    OS << "<" << *M->getMemOperand() << ">";
+    OS << "<";
+    printMemOperand(OS, *M->getMemOperand(), G);
+    OS << ">";
   } else if (const BlockAddressSDNode *BA =
                dyn_cast<BlockAddressSDNode>(this)) {
     int64_t offset = BA->getOffset();
     OS << "<";
     BA->getBlockAddress()->getFunction()->printAsOperand(OS, false);
     OS << ", ";
     BA->getBlockAddress()->getBasicBlock()->printAsOperand(OS, false);
     OS << ">";

test/CodeGen/AArch64/GlobalISel/arm64-fallback.ll

 ; FALLBACK-WITH-REPORT-OUT: fmov d0, #1.0
 define [1 x double] @constant() {
   ret [1 x double] [double 1.0]
 }
 
   ; The key problem here is that we may fail to create an MBB referenced by a
   ; PHI. If so, we cannot complete the G_PHI and mustn't try or bad things
   ; happen.
-; FALLBACK-WITH-REPORT-ERR: remark: <unknown>:0:0: cannot select: G_STORE %6:gpr(s32), %2:gpr(p0); mem:ST4[%addr] (in function: pending_phis)
+; FALLBACK-WITH-REPORT-ERR: remark: <unknown>:0:0: cannot select: G_STORE %6:gpr(s32), %2:gpr(p0) :: (store seq_cst 4 into %ir.addr) (in function: pending_phis)
 ; FALLBACK-WITH-REPORT-ERR: warning: Instruction selection used fallback path for pending_phis
 ; FALLBACK-WITH-REPORT-OUT-LABEL: pending_phis:
 define i32 @pending_phis(i1 %tst, i32 %val, i32* %addr) {
   br i1 %tst, label %true, label %false
 
 end:
   %res = phi i32 [%val, %true], [42, %false]
   ret i32 %res
 
 true:
   store atomic i32 42, i32* %addr seq_cst, align 4
   br label %end
 
 false:
   br label %end
 
 }
 
-; FALLBACK-WITH-REPORT-ERR: remark: <unknown>:0:0: unable to legalize instruction: %0:_(s24) = G_LOAD %1:_(p0); mem:LD3[undef](align=1) (in function: odd_type_load)
+; FALLBACK-WITH-REPORT-ERR: remark: <unknown>:0:0: unable to legalize instruction: %0:_(s24) = G_LOAD %1:_(p0) :: (load 3 from `i24* undef`, align 1) (in function: odd_type_load)
 ; FALLBACK-WITH-REPORT-ERR: warning: Instruction selection used fallback path for odd_type_load
 ; FALLBACK-WITH-REPORT-OUT-LABEL: odd_type_load
 define i32 @odd_type_load() {
 entry:
   %ld = load i24, i24* undef, align 1
   %cst = zext i24 %ld to i32
   ret i32 %cst
 }
 
   ; General legalizer inability to handle types whose size wasn't a power of 2.
-; FALLBACK-WITH-REPORT-ERR: remark: <unknown>:0:0: unable to legalize instruction: G_STORE %1:_(s42), %0:_(p0); mem:ST6[%addr](align=8) (in function: odd_type)
+; FALLBACK-WITH-REPORT-ERR: remark: <unknown>:0:0: unable to legalize instruction: G_STORE %1:_(s42), %0:_(p0) :: (store 6 into %ir.addr, align 8) (in function: odd_type)
 ; FALLBACK-WITH-REPORT-ERR: warning: Instruction selection used fallback path for odd_type
 ; FALLBACK-WITH-REPORT-OUT-LABEL: odd_type:
 define void @odd_type(i42* %addr) {
   %val42 = load i42, i42* %addr
   store i42 %val42, i42* %addr
   ret void
 }
 
-; FALLBACK-WITH-REPORT-ERR: remark: <unknown>:0:0: unable to legalize instruction: G_STORE %1:_(<7 x s32>), %0:_(p0); mem:ST28[%addr](align=32) (in function: odd_vector)
+; FALLBACK-WITH-REPORT-ERR: remark: <unknown>:0:0: unable to legalize instruction: G_STORE %1:_(<7 x s32>), %0:_(p0) :: (store 28 into %ir.addr, align 32) (in function: odd_vector)
 ; FALLBACK-WITH-REPORT-ERR: warning: Instruction selection used fallback path for odd_vector
 ; FALLBACK-WITH-REPORT-OUT-LABEL: odd_vector:
 define void @odd_vector(<7 x i32>* %addr) {
   %vec = load <7 x i32>, <7 x i32>* %addr
   store <7 x i32> %vec, <7 x i32>* %addr
   ret void
 }
 
   ; AArch64 was asserting instead of returning an invalid mapping for unknown
   ; sizes.
 ; FALLBACK-WITH-REPORT-ERR: remark: <unknown>:0:0: unable to translate instruction: ret: '  ret i128 undef' (in function: sequence_sizes)
 ; FALLBACK-WITH-REPORT-ERR: warning: Instruction selection used fallback path for sequence_sizes
 ; FALLBACK-WITH-REPORT-LABEL: sequence_sizes:
 define i128 @sequence_sizes([8 x i8] %in) {
   ret i128 undef
 }
 
 ; Just to make sure we don't accidentally emit a normal load/store.
-; FALLBACK-WITH-REPORT-ERR: remark: <unknown>:0:0: cannot select: %2:gpr(s64) = G_LOAD %0:gpr(p0); mem:LD8[%addr] (in function: atomic_ops)
+; FALLBACK-WITH-REPORT-ERR: remark: <unknown>:0:0: cannot select: %2:gpr(s64) = G_LOAD %0:gpr(p0) :: (load seq_cst 8 from %ir.addr)  (in function: atomic_ops)
 ; FALLBACK-WITH-REPORT-ERR: warning: Instruction selection used fallback path for atomic_ops
 ; FALLBACK-WITH-REPORT-LABEL: atomic_ops:
 define i64 @atomic_ops(i64* %addr) {
   store atomic i64 0, i64* %addr unordered, align 8
   %res = load atomic i64, i64* %addr seq_cst, align 8
   ret i64 %res
 }
 
block:
   store i16* %dummy, i16** undef
   ret void
 
 end:
   %vec = load <2 x i16*>, <2 x i16*>* undef
   br label %block
 }
 
-; FALLBACK-WITH-REPORT-ERR: remark: <unknown>:0:0: unable to legalize instruction: G_STORE %0:_(<2 x p0>), %4:_(p0); mem:ST16[undef] (in function: vector_of_pointers_insertelement)
+; FALLBACK-WITH-REPORT-ERR: remark: <unknown>:0:0: unable to legalize instruction: G_STORE %0:_(<2 x p0>), %4:_(p0) :: (store 16 into `<2 x i16*>* undef`) (in function: vector_of_pointers_insertelement)
 ; FALLBACK-WITH-REPORT-ERR: warning: Instruction selection used fallback path for vector_of_pointers_insertelement
 ; FALLBACK-WITH-REPORT-OUT-LABEL: vector_of_pointers_insertelement:
 define void @vector_of_pointers_insertelement() {
   br label %end
 
 block:
   %dummy = insertelement <2 x i16*> %vec, i16* null, i32 0
   store <2 x i16*> %dummy, <2 x i16*>* undef
   ret void
 
 end:
   %vec = load <2 x i16*>, <2 x i16*>* undef
   br label %block
 }
 
-; FALLBACK-WITH-REPORT-ERR: remark: <unknown>:0:0: unable to legalize instruction: G_STORE %1:_(s96), %3:_(p0); mem:ST12[undef](align=4) (in function: nonpow2_insertvalue_narrowing)
+; FALLBACK-WITH-REPORT-ERR: remark: <unknown>:0:0: unable to legalize instruction: G_STORE %1:_(s96), %3:_(p0) :: (store 12 into `%struct96* undef`, align 4) (in function: nonpow2_insertvalue_narrowing)
 ; FALLBACK-WITH-REPORT-ERR: warning: Instruction selection used fallback path for nonpow2_insertvalue_narrowing
 ; FALLBACK-WITH-REPORT-OUT-LABEL: nonpow2_insertvalue_narrowing:
 %struct96 = type { float, float, float }
 define void @nonpow2_insertvalue_narrowing(float %a) {
   %dummy = insertvalue %struct96 undef, float %a, 0
   store %struct96 %dummy, %struct96* undef
   ret void
 }
 
-; FALLBACK-WITH-REPORT-ERR remark: <unknown>:0:0: unable to legalize instruction: G_STORE %3, %4; mem:ST12[undef](align=16) (in function: nonpow2_add_narrowing)
+; FALLBACK-WITH-REPORT-ERR remark: <unknown>:0:0: unable to legalize instruction: G_STORE %3, %4 :: (store 12 into `i96* undef`, align 16) (in function: nonpow2_add_narrowing)
 ; FALLBACK-WITH-REPORT-ERR: warning: Instruction selection used fallback path for nonpow2_add_narrowing
 ; FALLBACK-WITH-REPORT-OUT-LABEL: nonpow2_add_narrowing:
 define void @nonpow2_add_narrowing() {
   %a = add i128 undef, undef
   %b = trunc i128 %a to i96
   %dummy = add i96 %b, %b
   store i96 %dummy, i96* undef
   ret void
 }
 
-; FALLBACK-WITH-REPORT-ERR remark: <unknown>:0:0: unable to legalize instruction: G_STORE %3, %4; mem:ST12[undef](align=16) (in function: nonpow2_add_narrowing)
+; FALLBACK-WITH-REPORT-ERR remark: <unknown>:0:0: unable to legalize instruction: G_STORE %3, %4 :: (store 12 into `i96* undef`, align 16) (in function: nonpow2_add_narrowing)
 ; FALLBACK-WITH-REPORT-ERR: warning: Instruction selection used fallback path for nonpow2_or_narrowing
 ; FALLBACK-WITH-REPORT-OUT-LABEL: nonpow2_or_narrowing:
 define void @nonpow2_or_narrowing() {
   %a = add i128 undef, undef
   %b = trunc i128 %a to i96
   %dummy = or i96 %b, %b
   store i96 %dummy, i96* undef
   ret void
 }
 
-; FALLBACK-WITH-REPORT-ERR remark: <unknown>:0:0: unable to legalize instruction: G_STORE %0, %1; mem:ST12[undef](align=16) (in function: nonpow2_load_narrowing)
+; FALLBACK-WITH-REPORT-ERR remark: <unknown>:0:0: unable to legalize instruction: G_STORE %0, %1 :: (store 12 into `i96* undef`, align 16) (in function: nonpow2_load_narrowing)
 ; FALLBACK-WITH-REPORT-ERR: warning: Instruction selection used fallback path for nonpow2_load_narrowing
 ; FALLBACK-WITH-REPORT-OUT-LABEL: nonpow2_load_narrowing:
 define void @nonpow2_load_narrowing() {
   %dummy = load i96, i96* undef
   store i96 %dummy, i96* undef
   ret void
 }
 
-; FALLBACK-WITH-REPORT-ERR: remark: <unknown>:0:0: unable to legalize instruction: G_STORE %3:_(s96), %0:_(p0); mem:ST12[%c](align=16) (in function: nonpow2_store_narrowing
+; FALLBACK-WITH-REPORT-ERR: remark: <unknown>:0:0: unable to legalize instruction: G_STORE %3:_(s96), %0:_(p0) :: (store 12 into %ir.c, align 16) (in function: nonpow2_store_narrowing
 ; FALLBACK-WITH-REPORT-ERR: warning: Instruction selection used fallback path for nonpow2_store_narrowing
 ; FALLBACK-WITH-REPORT-OUT-LABEL: nonpow2_store_narrowing:
 define void @nonpow2_store_narrowing(i96* %c) {
   %a = add i128 undef, undef
   %b = trunc i128 %a to i96
   store i96 %b, i96* %c
   ret void
 }
 
-; FALLBACK-WITH-REPORT-ERR: remark: <unknown>:0:0: unable to legalize instruction: G_STORE %0:_(s96), %1:_(p0); mem:ST12[undef](align=16) (in function: nonpow2_constant_narrowing)
+; FALLBACK-WITH-REPORT-ERR: remark: <unknown>:0:0: unable to legalize instruction: G_STORE %0:_(s96), %1:_(p0) :: (store 12 into `i96* undef`, align 16) (in function: nonpow2_constant_narrowing)
 ; FALLBACK-WITH-REPORT-ERR: warning: Instruction selection used fallback path for nonpow2_constant_narrowing
 ; FALLBACK-WITH-REPORT-OUT-LABEL: nonpow2_constant_narrowing:
 define void @nonpow2_constant_narrowing() {
   store i96 0, i96* undef
   ret void
 }
 
 ; Currently can't handle vector lengths that aren't an exact multiple of
 ; natively supported vector lengths. Test that the fall-back works for those.
-; FALLBACK-WITH-REPORT-ERR-G_IMPLICIT_DEF-LEGALIZABLE: (FIXME: this is what is expected once we can legalize non-pow-of-2 G_IMPLICIT_DEF) remark: <unknown>:0:0: unable to legalize instruction: %1:_(<7 x s64>) = G_ADD %0, %0; (in function: nonpow2_vector_add_fewerelements
+; FALLBACK-WITH-REPORT-ERR-G_IMPLICIT_DEF-LEGALIZABLE: (FIXME: this is what is expected once we can legalize non-pow-of-2 G_IMPLICIT_DEF) remark: <unknown>:0:0: unable to legalize instruction: %1:_(<7 x s64>) = G_ADD %0, %0 (in function: nonpow2_vector_add_fewerelements
 ; FALLBACK-WITH-REPORT-ERR: remark: <unknown>:0:0: unable to legalize instruction: %2:_(s64) = G_EXTRACT_VECTOR_ELT %1:_(<7 x s64>), %3:_(s64) (in function: nonpow2_vector_add_fewerelements)
 ; FALLBACK-WITH-REPORT-ERR: warning: Instruction selection used fallback path for nonpow2_vector_add_fewerelements
 ; FALLBACK-WITH-REPORT-OUT-LABEL: nonpow2_vector_add_fewerelements:
 define void @nonpow2_vector_add_fewerelements() {
   %dummy = add <7 x i64> undef, undef
   %ex = extractelement <7 x i64> %dummy, i64 0
   store i64 %ex, i64* undef
   ret void
 }

test/CodeGen/AArch64/aarch64-stp-cluster.ll

entry:
   %arrayidx3 = getelementptr inbounds float, float* %P, i64 4
   store float %v, float* %arrayidx3
   ret void
 }
 
 ; CHECK: ********** MI Scheduling **********
 ; CHECK-LABEL: stp_volatile:%bb.0
 ; CHECK-NOT: Cluster ld/st
-; CHECK:SU(2):   STRXui %1:gpr64, %0:gpr64common, 3; mem:Volatile
-; CHECK:SU(3):   STRXui %1:gpr64, %0:gpr64common, 2; mem:Volatile
-; CHECK:SU(4):   STRXui %1:gpr64, %0:gpr64common, 1; mem:Volatile
-; CHECK:SU(5):   STRXui %1:gpr64, %0:gpr64common, 4; mem:Volatile
+; CHECK:SU(2):   STRXui %1:gpr64, %0:gpr64common, 3 :: (volatile
+; CHECK:SU(3):   STRXui %1:gpr64, %0:gpr64common, 2 :: (volatile
+; CHECK:SU(4):   STRXui %1:gpr64, %0:gpr64common, 1 :: (volatile
+; CHECK:SU(5):   STRXui %1:gpr64, %0:gpr64common, 4 :: (volatile
 define i64 @stp_volatile(i64* nocapture %P, i64 %v) {
 entry:
   %arrayidx = getelementptr inbounds i64, i64* %P, i64 3
   store volatile i64 %v, i64* %arrayidx
   %arrayidx1 = getelementptr inbounds i64, i64* %P, i64 2
   store volatile i64 %v, i64* %arrayidx1
   %arrayidx2 = getelementptr inbounds i64, i64* %P, i64 1
   store volatile i64 %v, i64* %arrayidx2
   %arrayidx3 = getelementptr inbounds i64, i64* %P, i64 4
   store volatile i64 %v, i64* %arrayidx3
   ret i64 %v
 }
 

test/CodeGen/AArch64/arm64-misched-memdep-bug.ll

 ; REQUIRES: asserts
 ; RUN: llc < %s -mtriple=arm64-linux-gnu -mcpu=cortex-a57 -enable-misched -verify-misched -debug-only=machine-scheduler -o - 2>&1 > /dev/null | FileCheck %s
 ;
 ; Test for bug in misched memory dependency calculation.
 ;
 ; CHECK: ********** MI Scheduling **********
 ; CHECK: misched_bug:%bb.0 entry
-; CHECK: SU(2):   %2:gpr32 = LDRWui %0:gpr64common, 1; mem:LD4[%ptr1_plus1]
+; CHECK: SU(2):   %2:gpr32 = LDRWui %0:gpr64common, 1 :: (load 4 from %ir.ptr1_plus1)
 ; CHECK:   Successors:
 ; CHECK-NEXT:    SU(5): Data Latency=4 Reg=%2
 ; CHECK-NEXT:    SU(4): Ord  Latency=0
-; CHECK: SU(3):   STRWui $wzr, %0:gpr64common, 0; mem:ST4[%ptr1]
+; CHECK: SU(3):   STRWui $wzr, %0:gpr64common, 0 :: (store 4 into %ir.ptr1)
 ; CHECK:   Successors:
 ; CHECK: SU(4): Ord  Latency=0
-; CHECK: SU(4):   STRWui $wzr, %1:gpr64common, 0; mem:ST4[%ptr2]
+; CHECK: SU(4):   STRWui $wzr, %1:gpr64common, 0 :: (store 4 into %ir.ptr2)
 ; CHECK: SU(5):   $w0 = COPY %2
 ; CHECK: ** ScheduleDAGMI::schedule picking next node
 define i32 @misched_bug(i32* %ptr1, i32* %ptr2) {
 entry:
   %ptr1_plus1 = getelementptr inbounds i32, i32* %ptr1, i64 1
   %val1 = load i32, i32* %ptr1_plus1, align 4
   store i32 0, i32* %ptr1, align 4
   store i32 0, i32* %ptr2, align 4
   ret i32 %val1
 }

test/CodeGen/AMDGPU/extload-align.ll

 ; RUN: llc -debug-only=machine-scheduler -march=amdgcn -mtriple=amdgcn---amdgiz -verify-machineinstrs %s -o - 2>&1| FileCheck -check-prefix=SI-NOHSA -check-prefix=FUNC -check-prefix=DEBUG %s
 target datalayout = "A5"
 ; REQUIRES: asserts
 
 ; Verify that the extload generated from %eval has the default
 ; alignment size (2) corresponding to the underlying memory size (i16)
 ; size and not 4 corresponding to the sign-extended size (i32).
 
 ; DEBUG: {{^}}# Machine code for function extload_align:
-; DEBUG: mem:LD2[<unknown>(addrspace=5)]
+; DEBUG: (load 2, addrspace 5)
 ; DEBUG: {{^}}# End machine code for function extload_align.
 
 define amdgpu_kernel void @extload_align(i32 addrspace(5)* %out, i32 %index) #0 {
   %v0 = alloca [4 x i16], addrspace(5)
   %a1 = getelementptr inbounds [4 x i16], [4 x i16] addrspace(5)* %v0, i32 0, i32 0
   %a2 = getelementptr inbounds [4 x i16], [4 x i16] addrspace(5)* %v0, i32 0, i32 1
   store i16 0, i16 addrspace(5)* %a1
   store i16 1, i16 addrspace(5)* %a2
   %a = getelementptr inbounds [4 x i16], [4 x i16] addrspace(5)* %v0, i32 0, i32 %index
   %val = load i16, i16 addrspace(5)* %a
   %eval = sext i16 %val to i32
   store i32 %eval, i32 addrspace(5)* %out
   ret void
 }

test/CodeGen/ARM/2012-06-12-SchedMemLatency.ll

 ; RUN: llc < %s -o /dev/null "-mtriple=thumbv7-apple-ios" -debug-only=post-RA-sched 2> %t
 ; RUN: FileCheck %s < %t
 ; REQUIRES: asserts
 ; Make sure that mayalias store-load dependencies have one cycle
 ; latency regardless of whether they are barriers or not.
 
 ; CHECK: ** List Scheduling
-; CHECK: SU(2){{.*}}STR{{.*}}Volatile
+; CHECK: SU(2){{.*}}STR{{.*}}(volatile
 ; CHECK-NOT: SU({{.*}}): Ord
 ; CHECK: SU(3): Ord Latency=1
 ; CHECK-NOT: SU({{.*}}): Ord
-; CHECK: SU(3){{.*}}LDR{{.*}}Volatile
+; CHECK: SU(3){{.*}}LDR{{.*}}(volatile
 ; CHECK-NOT: SU({{.*}}): Ord
 ; CHECK: SU(2): Ord Latency=1
 ; CHECK-NOT: SU({{.*}}): Ord
 ; CHECK: Successors:
 ; CHECK: ** List Scheduling
 ; CHECK: SU(2){{.*}}STR{{.*}}
 ; CHECK-NOT: SU({{.*}}): Ord
 ; CHECK: SU(3): Ord Latency=1

test/CodeGen/ARM/ldrd-memoper.ll

 ; RUN: llc %s -o /dev/null -mtriple=thumbv7-apple-ios -mcpu=cortex-a8 -debug-only=arm-ldst-opt 2> %t
 ; RUN: FileCheck %s < %t
 ; REQUIRES: asserts
 ; PR8113: ARMLoadStoreOptimizer must preserve memoperands.
 
 @b = external global i64*
 
-; CHECK: Formed {{.*}} t2LDRD{{.*}} mem:LD4[%0] LD4[%0+4]
+; CHECK: Formed {{.*}} t2LDRD{{.*}} (load 4 from %ir.0), (load 4 from %ir.0 + 4)
 define i64 @t(i64 %a) nounwind readonly {
 entry:
   %0 = load i64*, i64** @b, align 4
   %1 = load i64, i64* %0, align 4
   %2 = mul i64 %1, %a
   ret i64 %2
 }

test/CodeGen/ARM/misched-int-basic-thumb2.mir

--- |
   }
 #
 # CHECK:       ********** MI Scheduling **********
 # CHECK:       SU(2):   %2:rgpr = t2MOVi32imm @g1
 # CHECK_A9:    Latency    : 2
 # CHECK_SWIFT: Latency    : 2
 # CHECK_R52:   Latency    : 2
 #
-# CHECK:       SU(3):   %3:rgpr = t2LDRi12 %2:rgpr, 0, 14, $noreg; mem:LD4[@g1](dereferenceable)
+# CHECK:       SU(3):   %3:rgpr = t2LDRi12 %2:rgpr, 0, 14, $noreg :: (dereferenceable load 4 from @g1)
 # CHECK_A9:    Latency    : 1
 # CHECK_SWIFT: Latency    : 3
 # CHECK_R52:   Latency    : 4
 #
 # CHECK :      SU(6):   %6 = t2ADDrr %3:rgpr, %3:rgpr, 14, $noreg, $noreg
 # CHECK_A9:    Latency    : 1
 # CHECK_SWIFT: Latency    : 1
 # CHECK_R52:   Latency    : 3
 
 # CHECK:       SU(7):   %7:rgpr = t2SDIV %6:rgpr, %5:rgpr, 14, $noreg
 # CHECK_A9:    Latency    : 0
 # CHECK_SWIFT: Latency    : 14
 # CHECK_R52:   Latency    : 8
 
-# CHECK:       SU(8):   t2STRi12 %7:rgpr, %2:rgpr, 0, 14, $noreg; mem:ST4[@g1]
+# CHECK:       SU(8):   t2STRi12 %7:rgpr, %2:rgpr, 0, 14, $noreg :: (store 4 into @g1)
 # CHECK_A9:    Latency    : 1
 # CHECK_SWIFT: Latency    : 0
 # CHECK_R52:   Latency    : 4
 #
 # CHECK:       SU(9):   %8:rgpr = t2SMULBB %1:rgpr, %1:rgpr, 14, $noreg
 # CHECK_A9:    Latency    : 2
 # CHECK_SWIFT: Latency    : 4
 # CHECK_R52:   Latency    : 4

test/CodeGen/ARM/single-issue-r52.mir

define <8 x i8> @foo(i8* %A) {
     %tmp3 = extractvalue %struct.__neon_int8x8x4_t %tmp1, 1
     %tmp4 = add <8 x i8> %tmp2, %tmp3
     ret <8 x i8> %tmp4
   }
   declare %struct.__neon_int8x8x4_t @llvm.arm.neon.vld4.v8i8.p0i8(i8*, i32)
 
 # CHECK: ********** MI Scheduling **********
 # CHECK: ScheduleDAGMILive::schedule starting
-# CHECK: SU(1):   %1:qqpr = VLD4d8Pseudo %0:gpr, 8, 14, $noreg; mem:LD32[%A](align=8)
+# CHECK: SU(1):   %1:qqpr = VLD4d8Pseudo %0:gpr, 8, 14, $noreg :: (load 32 from %ir.A, align 8)
 # CHECK: Latency            : 8
 # CHECK: Single Issue       : true;
 # CHECK: SU(2):   %4:dpr = VADDv8i8 %1.dsub_0:qqpr, %1.dsub_1:qqpr, 14, $noreg
 # CHECK: Latency            : 5
 # CHECK: Single Issue       : false;
 # CHECK: SU(3):   %5:gpr, %6:gpr = VMOVRRD %4:dpr, 14, $noreg
 # CHECK: Latency            : 4
 # CHECK: Single Issue       : false;

test/CodeGen/Hexagon/post-inc-aa-metadata.ll

 ; RUN: llc -march=hexagon -debug-only=isel < %s 2>&1 | FileCheck %s
 ; REQUIRES: asserts
 
 ; Check that the generated post-increment load has TBAA information.
 ; CHECK-LABEL: Machine code for function fred:
-; CHECK: = V6_vL32b_pi %{{[0-9]+}}{{[^,]*}}, 64; mem:LD64[{{.*}}](tbaa=
+; CHECK: = V6_vL32b_pi %{{[0-9]+}}{{[^,]*}}, 64 :: (load 64{{.*}}!tbaa
 
 target triple = "hexagon"
 
 ; Function Attrs: norecurse nounwind
 define void @fred(<16 x i32>* nocapture %p, <16 x i32>* nocapture readonly %q, i32 %n) local_unnamed_addr #0 {
 entry:
   %tobool2 = icmp eq i32 %n, 0
   br i1 %tobool2, label %while.end, label %while.body

test/CodeGen/PowerPC/byval-agg-info.ll

 ; RUN: llc -verify-machineinstrs < %s -print-after=prologepilog >%t 2>&1 && FileCheck <%t %s
 target datalayout = "E-m:e-i64:64-n32:64"
 target triple = "powerpc64-unknown-linux-gnu"
 
 %struct.anon = type { i32, i32 }
 
 declare void @foo(%struct.anon* %v)
 define void @test(i32 %a, i32 %b, %struct.anon* byval nocapture %v) {
 entry:
   call void @foo(%struct.anon* %v)
   ret void
 }
 
 ; Make sure that the MMO on the store has no offset from the byval
-; variable itself (we used to have mem:ST8[%v+64]).
-; CHECK: STD killed renamable $x5, 176, $x1; mem:ST8[%v](align=16)
+; variable itself (we used to have (store 8 into %ir.v + 64)).
+; CHECK: STD killed renamable $x5, 176, $x1 :: (store 8 into %ir.v, align 16)
 

test/CodeGen/PowerPC/combine_loads_from_build_pair.ll

 ; RUN: llc -verify-machineinstrs -O0 -mcpu=g4 -mtriple=powerpc-apple-darwin8 < %s -debug -stop-after=machineverifier 2>&1 | FileCheck %s
 
 ; REQUIRES: asserts
 
 define i64 @func1(i64 %p1, i64 %p2, i64 %p3, i64 %p4, { i64, i8* } %struct) {
 ; Verify that we get a combine on the build_pair, creating a LD8 load somewhere
 ; between "Initial selection DAG" and "Optimized lowered selection DAG".
 ; The target is big-endian, and stack grows towards higher addresses,
 ; so we expect the LD8 to load from the address used in the original HIBITS
 ; load.
 ; CHECK-LABEL: Initial selection DAG:
-; CHECK-DAG:     [[LOBITS:t[0-9]+]]: i32,ch = load<LD4[FixedStack-2]>
-; CHECK-DAG:     [[HIBITS:t[0-9]+]]: i32,ch = load<LD4[FixedStack-1]>
+; CHECK-DAG:     [[LOBITS:t[0-9]+]]: i32,ch = load<(load 4 from %fixed-stack.1)>
+; CHECK-DAG:     [[HIBITS:t[0-9]+]]: i32,ch = load<(load 4 from %fixed-stack.2)>
 ; CHECK: Combining: t{{[0-9]+}}: i64 = build_pair [[LOBITS]], [[HIBITS]]
 ; CHECK-NEXT: Creating new node
-; CHECK-SAME: load<LD8[FixedStack-1]
+; CHECK-SAME: load<(load 8 from %fixed-stack.2, align 4)>
 ; CHECK-NEXT: into
-; CHECK-SAME: load<LD8[FixedStack-1]
+; CHECK-SAME: load<(load 8 from %fixed-stack.2, align 4)>
 ; CHECK-LABEL: Optimized lowered selection DAG:
   %result = extractvalue {i64, i8* } %struct, 0
   ret i64 %result
 }
 

test/CodeGen/PowerPC/unal-vec-negarith.ll

 ; RUN: llc -verify-machineinstrs -debug-only=isel <%s >%t 2>&1 && FileCheck <%t %s
 ; REQUIRES: asserts
 
 target datalayout = "E-m:e-i64:64-n32:64"
 target triple = "powerpc64-unknown-linux-gnu"
 
 define <16 x i8> @test_l_v16i8(<16 x i8>* %p) #0 {
 entry:
   %r = load <16 x i8>, <16 x i8>* %p, align 1
   ret <16 x i8> %r
 
-; CHECK-NOT: v4i32,ch = llvm.ppc.altivec.lvx{{.*}}<LD31[%p+4294967281](align=1)>
-; CHECK:     v4i32,ch = llvm.ppc.altivec.lvx{{.*}}<LD31[%p+-15](align=1)>
+; CHECK-NOT: v4i32,ch = llvm.ppc.altivec.lvx{{.*}}<(load 31 from %ir.p + 4294967281, align 1)>
+; CHECK:     v4i32,ch = llvm.ppc.altivec.lvx{{.*}}<(load 31 from %ir.p - 15, align 1)>
 }
 
 attributes #0 = { nounwind "target-cpu"="pwr7" }
 

test/CodeGen/X86/GlobalISel/x86_64-fallback.ll

 ; RUN: llc -O0 -mtriple=x86_64-linux-gnu -global-isel -global-isel-abort=2 -pass-remarks-missed='gisel*' -verify-machineinstrs %s -o %t.out 2> %t.err
 ; RUN: FileCheck %s --check-prefix=FALLBACK-WITH-REPORT-OUT < %t.out
 ; RUN: FileCheck %s --check-prefix=FALLBACK-WITH-REPORT-ERR < %t.err
 ; This file checks that the fallback path to selection dag works.
 ; The test is fragile in the sense that it must be updated to expose
 ; something that fails with global-isel.
 ; When we cannot produce a test case anymore, that means we can remove
 ; the fallback path.
 
 ; Check that we fallback on invoke translation failures.
-; FALLBACK-WITH-REPORT-ERR: remark: <unknown>:0:0: unable to legalize instruction: G_STORE %1:_(s80), %0:_(p0); mem:ST10[%ptr](align=16) (in function: test_x86_fp80_dump)
+; FALLBACK-WITH-REPORT-ERR: remark: <unknown>:0:0: unable to legalize instruction: G_STORE %1:_(s80), %0:_(p0) :: (store 10 into %ir.ptr, align 16) (in function: test_x86_fp80_dump)
 ; FALLBACK-WITH-REPORT-ERR: warning: Instruction selection used fallback path for test_x86_fp80_dump
 ; FALLBACK-WITH-REPORT-OUT-LABEL: test_x86_fp80_dump:
 define void @test_x86_fp80_dump(x86_fp80* %ptr){
   store x86_fp80 0xK4002A000000000000000, x86_fp80* %ptr, align 16
   ret void
 }
 
 ; Check that we fallback on byVal argument

test/CodeGen/X86/merge-store-partially-alias-loads.ll

 ; X86-NEXT: movb [[HI1]], 3([[BASEREG]])
 ; X86-NEXT: retq
 
 ; DBGDAG-LABEL: Optimized legalized selection DAG: %bb.0 'merge_store_partial_overlap_load:'
 ; DBGDAG: [[ENTRYTOKEN:t[0-9]+]]: ch = EntryToken
 ; DBGDAG-DAG: [[BASEPTR:t[0-9]+]]: i64,ch = CopyFromReg [[ENTRYTOKEN]],
 ; DBGDAG-DAG: [[ADDPTR:t[0-9]+]]: i64 = add {{(nuw )?}}[[BASEPTR]], Constant:i64<2>
 
-; DBGDAG-DAG: [[LD2:t[0-9]+]]: i16,ch = load<LD2[%tmp81](align=1)> [[ENTRYTOKEN]], [[BASEPTR]], undef:i64
-; DBGDAG-DAG: [[LD1:t[0-9]+]]: i8,ch = load<LD1[%tmp12]> [[ENTRYTOKEN]], [[ADDPTR]], undef:i64
+; DBGDAG-DAG: [[LD2:t[0-9]+]]: i16,ch = load<(load 2 from %ir.tmp81, align 1)> [[ENTRYTOKEN]], [[BASEPTR]], undef:i64
+; DBGDAG-DAG: [[LD1:t[0-9]+]]: i8,ch = load<(load 1 from %ir.tmp12)> [[ENTRYTOKEN]], [[ADDPTR]], undef:i64
 
-; DBGDAG-DAG: [[ST1:t[0-9]+]]: ch = store<ST1[%tmp14]> [[ENTRYTOKEN]], [[LD1]], t{{[0-9]+}}, undef:i64
+; DBGDAG-DAG: [[ST1:t[0-9]+]]: ch = store<(store 1 into %ir.tmp14)> [[ENTRYTOKEN]], [[LD1]], t{{[0-9]+}}, undef:i64
 ; DBGDAG-DAG: [[LOADTOKEN:t[0-9]+]]: ch = TokenFactor [[LD2]]:1, [[LD1]]:1
-; DBGDAG-DAG: [[ST2:t[0-9]+]]: ch = store<ST2[%tmp10](align=1)> [[LOADTOKEN]], [[LD2]], t{{[0-9]+}}, undef:i64
+; DBGDAG-DAG: [[ST2:t[0-9]+]]: ch = store<(store 2 into %ir.tmp10, align 1)> [[LOADTOKEN]], [[LD2]], t{{[0-9]+}}, undef:i64
 
 ; DBGDAG: X86ISD::RET_FLAG t{{[0-9]+}},
 
 ; DBGDAG-LABEL: Instruction selection begins
 define void @merge_store_partial_overlap_load([4 x i8]* %tmp) {
   %tmp8 = getelementptr inbounds [4 x i8], [4 x i8]* %tmp, i32 0, i8 0
   %tmp10 = getelementptr inbounds [4 x i8], [4 x i8]* %tmp, i32 0, i8 1
   %tmp12 = getelementptr inbounds [4 x i8], [4 x i8]* %tmp, i32 0, i8 2

test/CodeGen/X86/stack-protector-weight.ll

 
 ; DARWIN-IR: # Machine code for function test_branch_weights:
 ; DARWIN-IR: successors: %bb.[[SUCCESS:[0-9]+]](0x7fffffff), %bb.[[FAILURE:[0-9]+]]
 ; DARWIN-IR: bb.[[SUCCESS]]{{[0-9a-zA-Z_.]+}}:
 ; DARWIN-IR: bb.[[FAILURE]]{{[0-9a-zA-Z_.]+}}:
 ; DARWIN-IR: CALL64pcrel32 @__stack_chk_fail
 
 ; MSVC-SELDAG: # Machine code for function test_branch_weights:
-; MSVC-SELDAG: mem:Volatile LD4[@__security_cookie]
-; MSVC-SELDAG: ST4[FixedStack0]
-; MSVC-SELDAG: LD4[FixedStack0]
+; MSVC-SELDAG: :: (volatile load 4 from @__security_cookie)
+; MSVC-SELDAG: (store 4 into stack)
+; MSVC-SELDAG: (volatile load 4 from %stack.0.StackGuardSlot)
 ; MSVC-SELDAG: CALLpcrel32 @__security_check_cookie
 
 ; MSVC always uses selection DAG now.
 ; MSVC-IR: # Machine code for function test_branch_weights:
-; MSVC-IR: mem:Volatile LD4[@__security_cookie]
-; MSVC-IR: ST4[FixedStack0]
-; MSVC-IR: LD4[FixedStack0]
+; MSVC-IR: :: (volatile load 4 from @__security_cookie)
+; MSVC-IR: (store 4 into stack)
+; MSVC-IR: (volatile load 4 from %stack.0.StackGuardSlot)
 ; MSVC-IR: CALLpcrel32 @__security_check_cookie
 
 define i32 @test_branch_weights(i32 %n) #0 {
 entry:
   %a = alloca [128 x i32], align 16
   %0 = bitcast [128 x i32]* %a to i8*
   call void @llvm.lifetime.start.p0i8(i64 512, i8* %0)
   %arraydecay = getelementptr inbounds [128 x i32], [128 x i32]* %a, i64 0, i64 0