[CodeGen] Refactor TLI/AtomicExpand interface to make LLSC explicit.

We used to have this magic "hasLoadLinkedStoreConditional()" callback,
which really meant two things:
- expand cmpxchg (to ll/sc).
- expand atomic loads using ll/sc (rather than cmpxchg).

Remove it, and, instead, introduce explicit callbacks:
- bool shouldExpandAtomicCmpXchgInIR(inst)
- AtomicExpansionKind shouldExpandAtomicLoadInIR(inst)

Differential Revision: http://reviews.llvm.org/D12557

llvm-svn: 247429

docs/Atomics.rst

@@ -446,7 +446,7 @@ It is often easiest for backends to use AtomicExpandPass to lower some of the
 atomic constructs. Here are some lowerings it can do:
 
 * cmpxchg -> loop with load-linked/store-conditional
-  by overriding ``hasLoadLinkedStoreConditional()``, ``emitLoadLinked()``,
+  by overriding ``shouldExpandAtomicCmpXchgInIR()``, ``emitLoadLinked()``,
   ``emitStoreConditional()``
 * large loads/stores -> ll-sc/cmpxchg
   by overriding ``shouldExpandAtomicStoreInIR()``/``shouldExpandAtomicLoadInIR()``

include/llvm/Target/TargetLowering.h

@@ -124,15 +124,14 @@ public:
                           // mask (ex: x86 blends).
   };
 
-  /// Enum that specifies what a AtomicRMWInst is expanded to, if at all. Exists
+  /// Enum that specifies what an atomic load/AtomicRMWInst is expanded
-  /// because different targets have different levels of support for these
+  /// to, if at all. Exists because different targets have different levels of
-  /// atomic RMW instructions, and also have different options w.r.t. what they
+  /// support for these atomic instructions, and also have different options
-  /// should expand to.
+  /// w.r.t. what they should expand to.
   enum class AtomicExpansionKind {
     None,      // Don't expand the instruction.
     LLSC,      // Expand the instruction into loadlinked/storeconditional; used
-               // by ARM/AArch64. Implies `hasLoadLinkedStoreConditional`
+               // by ARM/AArch64.
-               // returns true.
     CmpXChg,   // Expand the instruction into cmpxchg; used by at least X86.
   };
 
@@ -1027,10 +1026,6 @@ public:
   /// \name Helpers for atomic expansion.
   /// @{
 
-  /// True if AtomicExpandPass should use emitLoadLinked/emitStoreConditional
-  /// and expand AtomicCmpXchgInst.
-  virtual bool hasLoadLinkedStoreConditional() const { return false; }
-
   /// Perform a load-linked operation on Addr, returning a "Value *" with the
   /// corresponding pointee type. This may entail some non-trivial operations to
   /// truncate or reconstruct types that will be illegal in the backend. See
@@ -1111,12 +1106,20 @@ public:
   /// Returns true if arguments should be sign-extended in lib calls.
   virtual bool shouldSignExtendTypeInLibCall(EVT Type, bool IsSigned) const {
     return IsSigned;
- }
+  }
 
-  /// Returns true if the given (atomic) load should be expanded by the
+  /// Returns how the given (atomic) load should be expanded by the
-  /// IR-level AtomicExpand pass into a load-linked instruction
+  /// IR-level AtomicExpand pass.
-  /// (through emitLoadLinked()).
+  virtual AtomicExpansionKind shouldExpandAtomicLoadInIR(LoadInst *LI) const {
-  virtual bool shouldExpandAtomicLoadInIR(LoadInst *LI) const { return false; }
+    return AtomicExpansionKind::None;
+  }
+
+  /// Returns true if the given atomic cmpxchg should be expanded by the
+  /// IR-level AtomicExpand pass into a load-linked/store-conditional sequence
+  /// (through emitLoadLinked() and emitStoreConditional()).
+  virtual bool shouldExpandAtomicCmpXchgInIR(AtomicCmpXchgInst *AI) const {
+    return false;
+  }
 
   /// Returns how the IR-level AtomicExpand pass should expand the given
   /// AtomicRMW, if at all. Default is to never expand.

lib/CodeGen/AtomicExpandPass.cpp

@@ -46,7 +46,7 @@ namespace {
   private:
     bool bracketInstWithFences(Instruction *I, AtomicOrdering Order,
                                bool IsStore, bool IsLoad);
-    bool expandAtomicLoad(LoadInst *LI);
+    bool tryExpandAtomicLoad(LoadInst *LI);
     bool expandAtomicLoadToLL(LoadInst *LI);
     bool expandAtomicLoadToCmpXchg(LoadInst *LI);
     bool expandAtomicStore(StoreInst *SI);
@@ -109,7 +109,7 @@ bool AtomicExpand::runOnFunction(Function &F) {
         FenceOrdering = RMWI->getOrdering();
         RMWI->setOrdering(Monotonic);
         IsStore = IsLoad = true;
-      } else if (CASI && !TLI->hasLoadLinkedStoreConditional() &&
+      } else if (CASI && !TLI->shouldExpandAtomicCmpXchgInIR(CASI) &&
                  (isAtLeastRelease(CASI->getSuccessOrdering()) ||
                   isAtLeastAcquire(CASI->getSuccessOrdering()))) {
         // If a compare and swap is lowered to LL/SC, we can do smarter fence
@@ -127,8 +127,8 @@ bool AtomicExpand::runOnFunction(Function &F) {
       }
     }
 
-    if (LI && TLI->shouldExpandAtomicLoadInIR(LI)) {
+    if (LI) {
-      MadeChange |= expandAtomicLoad(LI);
+      MadeChange |= tryExpandAtomicLoad(LI);
     } else if (SI && TLI->shouldExpandAtomicStoreInIR(SI)) {
       MadeChange |= expandAtomicStore(SI);
     } else if (RMWI) {
@@ -142,7 +142,7 @@ bool AtomicExpand::runOnFunction(Function &F) {
       } else {
         MadeChange |= tryExpandAtomicRMW(RMWI);
       }
-    } else if (CASI && TLI->hasLoadLinkedStoreConditional()) {
+    } else if (CASI && TLI->shouldExpandAtomicCmpXchgInIR(CASI)) {
       MadeChange |= expandAtomicCmpXchg(CASI);
     }
   }
@@ -170,11 +170,18 @@ bool AtomicExpand::bracketInstWithFences(Instruction *I, AtomicOrdering Order,
   return (LeadingFence || TrailingFence);
 }
 
-bool AtomicExpand::expandAtomicLoad(LoadInst *LI) {
+bool AtomicExpand::tryExpandAtomicLoad(LoadInst *LI) {
-  if (TLI->hasLoadLinkedStoreConditional())
+  switch (TLI->shouldExpandAtomicLoadInIR(LI)) {
+  case TargetLoweringBase::AtomicExpansionKind::None:
+    return false;
+  case TargetLoweringBase::AtomicExpansionKind::LLSC: {
     return expandAtomicLoadToLL(LI);
-  else
+  }
+  case TargetLoweringBase::AtomicExpansionKind::CmpXChg: {
     return expandAtomicLoadToCmpXchg(LI);
+  }
+  }
+  llvm_unreachable("Unhandled case in tryExpandAtomicLoad");
 }
 
 bool AtomicExpand::expandAtomicLoadToLL(LoadInst *LI) {
@@ -243,11 +250,6 @@ bool AtomicExpand::tryExpandAtomicRMW(AtomicRMWInst *AI) {
   case TargetLoweringBase::AtomicExpansionKind::None:
     return false;
   case TargetLoweringBase::AtomicExpansionKind::LLSC: {
-    assert(TLI->hasLoadLinkedStoreConditional() &&
-           "TargetLowering requested we expand AtomicRMW instruction into "
-           "load-linked/store-conditional combos, but such instructions aren't "
-           "supported");
-
     return expandAtomicRMWToLLSC(AI);
   }
   case TargetLoweringBase::AtomicExpansionKind::CmpXChg: {
@@ -503,11 +505,8 @@ bool AtomicExpand::isIdempotentRMW(AtomicRMWInst* RMWI) {
 }
 
 bool AtomicExpand::simplifyIdempotentRMW(AtomicRMWInst* RMWI) {
-  if (auto ResultingLoad = TLI->lowerIdempotentRMWIntoFencedLoad(RMWI)) {
+  if (auto ResultingLoad = TLI->lowerIdempotentRMWIntoFencedLoad(RMWI))
-    if (TLI->shouldExpandAtomicLoadInIR(ResultingLoad))
+    return tryExpandAtomicLoad(ResultingLoad);
-      expandAtomicLoad(ResultingLoad);
-    return true;
-  }
   return false;
 }
 

lib/Target/AArch64/AArch64ISelLowering.cpp

@@ -9492,19 +9492,21 @@ bool AArch64TargetLowering::shouldExpandAtomicStoreInIR(StoreInst *SI) const {
 // Loads and stores less than 128-bits are already atomic; ones above that
 // are doomed anyway, so defer to the default libcall and blame the OS when
 // things go wrong.
-bool AArch64TargetLowering::shouldExpandAtomicLoadInIR(LoadInst *LI) const {
+TargetLowering::AtomicExpansionKind
+AArch64TargetLowering::shouldExpandAtomicLoadInIR(LoadInst *LI) const {
   unsigned Size = LI->getType()->getPrimitiveSizeInBits();
-  return Size == 128;
+  return Size == 128 ? AtomicExpansionKind::LLSC : AtomicExpansionKind::None;
 }
 
 // For the real atomic operations, we have ldxr/stxr up to 128 bits,
-TargetLoweringBase::AtomicExpansionKind
+TargetLowering::AtomicExpansionKind
 AArch64TargetLowering::shouldExpandAtomicRMWInIR(AtomicRMWInst *AI) const {
   unsigned Size = AI->getType()->getPrimitiveSizeInBits();
   return Size <= 128 ? AtomicExpansionKind::LLSC : AtomicExpansionKind::None;
 }
 
-bool AArch64TargetLowering::hasLoadLinkedStoreConditional() const {
+bool AArch64TargetLowering::shouldExpandAtomicCmpXchgInIR(
+    AtomicCmpXchgInst *AI) const {
   return true;
 }
 

lib/Target/AArch64/AArch64ISelLowering.h

@@ -343,17 +343,19 @@ public:
   bool shouldConvertConstantLoadToIntImm(const APInt &Imm,
                                          Type *Ty) const override;
 
-  bool hasLoadLinkedStoreConditional() const override;
   Value *emitLoadLinked(IRBuilder<> &Builder, Value *Addr,
                         AtomicOrdering Ord) const override;
   Value *emitStoreConditional(IRBuilder<> &Builder, Value *Val,
                               Value *Addr, AtomicOrdering Ord) const override;
 
-  bool shouldExpandAtomicLoadInIR(LoadInst *LI) const override;
+  TargetLoweringBase::AtomicExpansionKind
+  shouldExpandAtomicLoadInIR(LoadInst *LI) const override;
   bool shouldExpandAtomicStoreInIR(StoreInst *SI) const override;
   TargetLoweringBase::AtomicExpansionKind
   shouldExpandAtomicRMWInIR(AtomicRMWInst *AI) const override;
 
+  bool shouldExpandAtomicCmpXchgInIR(AtomicCmpXchgInst *AI) const override;
+
   bool useLoadStackGuardNode() const override;
   TargetLoweringBase::LegalizeTypeAction
   getPreferredVectorAction(EVT VT) const override;

lib/Target/ARM/ARMISelLowering.cpp

@@ -11430,8 +11430,6 @@ bool ARMTargetLowering::shouldConvertConstantLoadToIntImm(const APInt &Imm,
   return true;
 }
 
-bool ARMTargetLowering::hasLoadLinkedStoreConditional() const { return true; }
-
 Instruction* ARMTargetLowering::makeDMB(IRBuilder<> &Builder,
                                         ARM_MB::MemBOpt Domain) const {
   Module *M = Builder.GetInsertBlock()->getParent()->getParent();
@@ -11527,14 +11525,16 @@ bool ARMTargetLowering::shouldExpandAtomicStoreInIR(StoreInst *SI) const {
 // FIXME: ldrd and strd are atomic if the CPU has LPAE (e.g. A15 has that
 // guarantee, see DDI0406C ARM architecture reference manual,
 // sections A8.8.72-74 LDRD)
-bool ARMTargetLowering::shouldExpandAtomicLoadInIR(LoadInst *LI) const {
+TargetLowering::AtomicExpansionKind
+ARMTargetLowering::shouldExpandAtomicLoadInIR(LoadInst *LI) const {
   unsigned Size = LI->getType()->getPrimitiveSizeInBits();
-  return (Size == 64) && !Subtarget->isMClass();
+  return ((Size == 64) && !Subtarget->isMClass()) ? AtomicExpansionKind::LLSC
+                                                  : AtomicExpansionKind::None;
 }
 
 // For the real atomic operations, we have ldrex/strex up to 32 bits,
 // and up to 64 bits on the non-M profiles
-TargetLoweringBase::AtomicExpansionKind
+TargetLowering::AtomicExpansionKind
 ARMTargetLowering::shouldExpandAtomicRMWInIR(AtomicRMWInst *AI) const {
   unsigned Size = AI->getType()->getPrimitiveSizeInBits();
   return (Size <= (Subtarget->isMClass() ? 32U : 64U))
@@ -11542,6 +11542,11 @@ ARMTargetLowering::shouldExpandAtomicRMWInIR(AtomicRMWInst *AI) const {
              : AtomicExpansionKind::None;
 }
 
+bool ARMTargetLowering::shouldExpandAtomicCmpXchgInIR(
+    AtomicCmpXchgInst *AI) const {
+  return true;
+}
+
 // This has so far only been implemented for MachO.
 bool ARMTargetLowering::useLoadStackGuardNode() const {
   return Subtarget->isTargetMachO();

lib/Target/ARM/ARMISelLowering.h

@@ -415,7 +415,6 @@ namespace llvm {
     bool functionArgumentNeedsConsecutiveRegisters(
         Type *Ty, CallingConv::ID CallConv, bool isVarArg) const override;
 
-    bool hasLoadLinkedStoreConditional() const override;
     Instruction *makeDMB(IRBuilder<> &Builder, ARM_MB::MemBOpt Domain) const;
     Value *emitLoadLinked(IRBuilder<> &Builder, Value *Addr,
                           AtomicOrdering Ord) const override;
@@ -436,10 +435,12 @@ namespace llvm {
     bool lowerInterleavedStore(StoreInst *SI, ShuffleVectorInst *SVI,
                                unsigned Factor) const override;
 
-    bool shouldExpandAtomicLoadInIR(LoadInst *LI) const override;
+    TargetLoweringBase::AtomicExpansionKind
+    shouldExpandAtomicLoadInIR(LoadInst *LI) const override;
     bool shouldExpandAtomicStoreInIR(StoreInst *SI) const override;
     TargetLoweringBase::AtomicExpansionKind
     shouldExpandAtomicRMWInIR(AtomicRMWInst *AI) const override;
+    bool shouldExpandAtomicCmpXchgInIR(AtomicCmpXchgInst *AI) const override;
 
     bool useLoadStackGuardNode() const override;
 

lib/Target/Hexagon/HexagonISelLowering.cpp

@@ -2499,9 +2499,12 @@ Value *HexagonTargetLowering::emitStoreConditional(IRBuilder<> &Builder,
   return Ext;
 }
 
-bool HexagonTargetLowering::shouldExpandAtomicLoadInIR(LoadInst *LI) const {
+TargetLowering::AtomicExpansionKind
+HexagonTargetLowering::shouldExpandAtomicLoadInIR(LoadInst *LI) const {
   // Do not expand loads and stores that don't exceed 64 bits.
-  return LI->getType()->getPrimitiveSizeInBits() > 64;
+  return LI->getType()->getPrimitiveSizeInBits() > 64
+             ? AtomicExpansionKind::LLSC
+             : AtomicExpansionKind::None;
 }
 
 bool HexagonTargetLowering::shouldExpandAtomicStoreInIR(StoreInst *SI) const {

lib/Target/Hexagon/HexagonISelLowering.h

@@ -209,14 +209,11 @@ bool isPositiveHalfWord(SDNode *N);
     bool isLegalICmpImmediate(int64_t Imm) const override;
 
     // Handling of atomic RMW instructions.
-    bool hasLoadLinkedStoreConditional() const override {
-      return true;
-    }
     Value *emitLoadLinked(IRBuilder<> &Builder, Value *Addr,
         AtomicOrdering Ord) const override;
     Value *emitStoreConditional(IRBuilder<> &Builder, Value *Val,
         Value *Addr, AtomicOrdering Ord) const override;
-    bool shouldExpandAtomicLoadInIR(LoadInst *LI) const override;
+    AtomicExpansionKind shouldExpandAtomicLoadInIR(LoadInst *LI) const override;
     bool shouldExpandAtomicStoreInIR(StoreInst *SI) const override;
     AtomicExpansionKind
     shouldExpandAtomicRMWInIR(AtomicRMWInst *AI) const override {

lib/Target/X86/X86ISelLowering.cpp

@@ -18406,12 +18406,14 @@ bool X86TargetLowering::shouldExpandAtomicStoreInIR(StoreInst *SI) const {
 
 // Note: this turns large loads into lock cmpxchg8b/16b.
 // FIXME: On 32 bits x86, fild/movq might be faster than lock cmpxchg8b.
-bool X86TargetLowering::shouldExpandAtomicLoadInIR(LoadInst *LI) const {
+TargetLowering::AtomicExpansionKind
+X86TargetLowering::shouldExpandAtomicLoadInIR(LoadInst *LI) const {
   auto PTy = cast<PointerType>(LI->getPointerOperand()->getType());
-  return needsCmpXchgNb(PTy->getElementType());
+  return needsCmpXchgNb(PTy->getElementType()) ? AtomicExpansionKind::CmpXChg
+                                               : AtomicExpansionKind::None;
 }
 
-TargetLoweringBase::AtomicExpansionKind
+TargetLowering::AtomicExpansionKind
 X86TargetLowering::shouldExpandAtomicRMWInIR(AtomicRMWInst *AI) const {
   unsigned NativeWidth = Subtarget->is64Bit() ? 64 : 32;
   Type *MemType = AI->getType();

lib/Target/X86/X86ISelLowering.h

@@ -1053,7 +1053,8 @@ namespace llvm {
 
     const MCPhysReg *getScratchRegisters(CallingConv::ID CC) const override;
 
-    bool shouldExpandAtomicLoadInIR(LoadInst *SI) const override;
+    TargetLoweringBase::AtomicExpansionKind
+    shouldExpandAtomicLoadInIR(LoadInst *SI) const override;
     bool shouldExpandAtomicStoreInIR(StoreInst *SI) const override;
     TargetLoweringBase::AtomicExpansionKind
     shouldExpandAtomicRMWInIR(AtomicRMWInst *AI) const override;