[opaque pointer types] Pass value type to LoadInst creation.

This cleans up all LoadInst creation in LLVM to explicitly pass the
value type rather than deriving it from the pointer's element-type.

Differential Revision: https://reviews.llvm.org/D57172

llvm-svn: 352911

llvm/lib/Bitcode/Reader/BitcodeReader.cpp

@@ -4404,7 +4404,7 @@ Error BitcodeReader::parseFunctionBody(Function *F) {
       unsigned Align;
       if (Error Err = parseAlignmentValue(Record[OpNum], Align))
         return Err;
-      I = new LoadInst(Op, "", Record[OpNum+1], Align, Ordering, SSID);
+      I = new LoadInst(Ty, Op, "", Record[OpNum + 1], Align, Ordering, SSID);
 
       InstructionList.push_back(I);
       break;

llvm/lib/CodeGen/AtomicExpandPass.cpp

@@ -381,7 +381,7 @@ LoadInst *AtomicExpand::convertAtomicLoadToIntegerType(LoadInst *LI) {
                               Addr->getType()->getPointerAddressSpace());
   Value *NewAddr = Builder.CreateBitCast(Addr, PT);
 
-  auto *NewLI = Builder.CreateLoad(NewAddr);
+  auto *NewLI = Builder.CreateLoad(NewTy, NewAddr);
   NewLI->setAlignment(LI->getAlignment());
   NewLI->setVolatile(LI->isVolatile());
   NewLI->setAtomic(LI->getOrdering(), LI->getSyncScopeID());
@@ -1769,8 +1769,8 @@ bool AtomicExpand::expandAtomicOpToLibcall(
     // from call}
     Type *FinalResultTy = I->getType();
     Value *V = UndefValue::get(FinalResultTy);
-    Value *ExpectedOut =
+    Value *ExpectedOut = Builder.CreateAlignedLoad(
-        Builder.CreateAlignedLoad(AllocaCASExpected, AllocaAlignment);
+        CASExpected->getType(), AllocaCASExpected, AllocaAlignment);
     Builder.CreateLifetimeEnd(AllocaCASExpected_i8, SizeVal64);
     V = Builder.CreateInsertValue(V, ExpectedOut, 0);
     V = Builder.CreateInsertValue(V, Result, 1);
@@ -1780,7 +1780,8 @@ bool AtomicExpand::expandAtomicOpToLibcall(
     if (UseSizedLibcall)
       V = Builder.CreateBitOrPointerCast(Result, I->getType());
     else {
-      V = Builder.CreateAlignedLoad(AllocaResult, AllocaAlignment);
+      V = Builder.CreateAlignedLoad(I->getType(), AllocaResult,
+                                    AllocaAlignment);
       Builder.CreateLifetimeEnd(AllocaResult_i8, SizeVal64);
     }
     I->replaceAllUsesWith(V);

llvm/lib/CodeGen/GCRootLowering.cpp

@@ -213,7 +213,7 @@ bool LowerIntrinsics::DoLowering(Function &F, GCStrategy &S) {
       }
       case Intrinsic::gcread: {
         // Replace a read barrier with a simple load.
-        Value *Ld = new LoadInst(CI->getArgOperand(1), "", CI);
+        Value *Ld = new LoadInst(CI->getType(), CI->getArgOperand(1), "", CI);
         Ld->takeName(CI);
         CI->replaceAllUsesWith(Ld);
         CI->eraseFromParent();

llvm/lib/CodeGen/InterleavedLoadCombinePass.cpp

@@ -1218,7 +1218,7 @@ bool InterleavedLoadCombineImpl::combine(std::list<VectorInfo> &InterleavedLoad,
                                       "interleaved.wide.ptrcast");
 
   // Create the wide load and update the MemorySSA.
-  auto LI = Builder.CreateAlignedLoad(CI, InsertionPoint->getAlignment(),
+  auto LI = Builder.CreateAlignedLoad(ILTy, CI, InsertionPoint->getAlignment(),
                                       "interleaved.wide.load");
   auto MSSAU = MemorySSAUpdater(&MSSA);
   MemoryUse *MSSALoad = cast<MemoryUse>(MSSAU.createMemoryAccessBefore(

llvm/lib/CodeGen/PreISelIntrinsicLowering.cpp

@@ -44,7 +44,7 @@ static bool lowerLoadRelative(Function &F) {
     Value *OffsetPtr =
         B.CreateGEP(Int8Ty, CI->getArgOperand(0), CI->getArgOperand(1));
     Value *OffsetPtrI32 = B.CreateBitCast(OffsetPtr, Int32PtrTy);
-    Value *OffsetI32 = B.CreateAlignedLoad(OffsetPtrI32, 4);
+    Value *OffsetI32 = B.CreateAlignedLoad(Int32Ty, OffsetPtrI32, 4);
 
     Value *ResultPtr = B.CreateGEP(Int8Ty, CI->getArgOperand(0), OffsetI32);
 

llvm/lib/CodeGen/SafeStack.cpp

@@ -371,7 +371,7 @@ Value *SafeStack::getStackGuard(IRBuilder<> &IRB, Function &F) {
   if (!StackGuardVar)
     StackGuardVar =
         F.getParent()->getOrInsertGlobal("__stack_chk_guard", StackPtrTy);
-  return IRB.CreateLoad(StackGuardVar, "StackGuard");
+  return IRB.CreateLoad(StackPtrTy, StackGuardVar, "StackGuard");
 }
 
 void SafeStack::findInsts(Function &F,
@@ -452,7 +452,8 @@ SafeStack::createStackRestorePoints(IRBuilder<> &IRB, Function &F,
     ++NumUnsafeStackRestorePoints;
 
     IRB.SetInsertPoint(I->getNextNode());
-    Value *CurrentTop = DynamicTop ? IRB.CreateLoad(DynamicTop) : StaticTop;
+    Value *CurrentTop =
+        DynamicTop ? IRB.CreateLoad(StackPtrTy, DynamicTop) : StaticTop;
     IRB.CreateStore(CurrentTop, UnsafeStackPtr);
   }
 
@@ -461,7 +462,7 @@ SafeStack::createStackRestorePoints(IRBuilder<> &IRB, Function &F,
 
 void SafeStack::checkStackGuard(IRBuilder<> &IRB, Function &F, ReturnInst &RI,
                                 AllocaInst *StackGuardSlot, Value *StackGuard) {
-  Value *V = IRB.CreateLoad(StackGuardSlot);
+  Value *V = IRB.CreateLoad(StackPtrTy, StackGuardSlot);
   Value *Cmp = IRB.CreateICmpNE(StackGuard, V);
 
   auto SuccessProb = BranchProbabilityInfo::getBranchProbStackProtector(true);
@@ -659,7 +660,8 @@ void SafeStack::moveDynamicAllocasToUnsafeStack(
     uint64_t TySize = DL.getTypeAllocSize(Ty);
     Value *Size = IRB.CreateMul(ArraySize, ConstantInt::get(IntPtrTy, TySize));
 
-    Value *SP = IRB.CreatePtrToInt(IRB.CreateLoad(UnsafeStackPtr), IntPtrTy);
+    Value *SP = IRB.CreatePtrToInt(IRB.CreateLoad(StackPtrTy, UnsafeStackPtr),
+                                   IntPtrTy);
     SP = IRB.CreateSub(SP, Size);
 
     // Align the SP value to satisfy the AllocaInst, type and stack alignments.
@@ -697,7 +699,7 @@ void SafeStack::moveDynamicAllocasToUnsafeStack(
 
       if (II->getIntrinsicID() == Intrinsic::stacksave) {
         IRBuilder<> IRB(II);
-        Instruction *LI = IRB.CreateLoad(UnsafeStackPtr);
+        Instruction *LI = IRB.CreateLoad(StackPtrTy, UnsafeStackPtr);
         LI->takeName(II);
         II->replaceAllUsesWith(LI);
         II->eraseFromParent();
@@ -792,7 +794,7 @@ bool SafeStack::run() {
   // Load the current stack pointer (we'll also use it as a base pointer).
   // FIXME: use a dedicated register for it ?
   Instruction *BasePointer =
-      IRB.CreateLoad(UnsafeStackPtr, false, "unsafe_stack_ptr");
+      IRB.CreateLoad(StackPtrTy, UnsafeStackPtr, false, "unsafe_stack_ptr");
   assert(BasePointer->getType() == StackPtrTy);
 
   AllocaInst *StackGuardSlot = nullptr;

llvm/lib/CodeGen/ScalarizeMaskedMemIntrin.cpp

@@ -143,7 +143,7 @@ static void scalarizeMaskedLoad(CallInst *CI) {
 
   // Short-cut if the mask is all-true.
   if (isa<Constant>(Mask) && cast<Constant>(Mask)->isAllOnesValue()) {
-    Value *NewI = Builder.CreateAlignedLoad(Ptr, AlignVal);
+    Value *NewI = Builder.CreateAlignedLoad(VecType, Ptr, AlignVal);
     CI->replaceAllUsesWith(NewI);
     CI->eraseFromParent();
     return;
@@ -166,7 +166,7 @@ static void scalarizeMaskedLoad(CallInst *CI) {
         continue;
       Value *Gep =
           Builder.CreateInBoundsGEP(EltTy, FirstEltPtr, Builder.getInt32(Idx));
-      LoadInst *Load = Builder.CreateAlignedLoad(Gep, AlignVal);
+      LoadInst *Load = Builder.CreateAlignedLoad(EltTy, Gep, AlignVal);
       VResult =
           Builder.CreateInsertElement(VResult, Load, Builder.getInt32(Idx));
     }
@@ -198,7 +198,7 @@ static void scalarizeMaskedLoad(CallInst *CI) {
 
     Value *Gep =
         Builder.CreateInBoundsGEP(EltTy, FirstEltPtr, Builder.getInt32(Idx));
-    LoadInst *Load = Builder.CreateAlignedLoad(Gep, AlignVal);
+    LoadInst *Load = Builder.CreateAlignedLoad(EltTy, Gep, AlignVal);
     Value *NewVResult = Builder.CreateInsertElement(VResult, Load,
                                                     Builder.getInt32(Idx));
 
@@ -366,6 +366,7 @@ static void scalarizeMaskedGather(CallInst *CI) {
   Value *Src0 = CI->getArgOperand(3);
 
   VectorType *VecType = cast<VectorType>(CI->getType());
+  Type *EltTy = VecType->getElementType();
 
   IRBuilder<> Builder(CI->getContext());
   Instruction *InsertPt = CI;
@@ -387,7 +388,7 @@ static void scalarizeMaskedGather(CallInst *CI) {
       Value *Ptr = Builder.CreateExtractElement(Ptrs, Builder.getInt32(Idx),
                                                 "Ptr" + Twine(Idx));
       LoadInst *Load =
-          Builder.CreateAlignedLoad(Ptr, AlignVal, "Load" + Twine(Idx));
+          Builder.CreateAlignedLoad(EltTy, Ptr, AlignVal, "Load" + Twine(Idx));
       VResult = Builder.CreateInsertElement(
           VResult, Load, Builder.getInt32(Idx), "Res" + Twine(Idx));
     }
@@ -418,7 +419,7 @@ static void scalarizeMaskedGather(CallInst *CI) {
     Value *Ptr = Builder.CreateExtractElement(Ptrs, Builder.getInt32(Idx),
                                               "Ptr" + Twine(Idx));
     LoadInst *Load =
-        Builder.CreateAlignedLoad(Ptr, AlignVal, "Load" + Twine(Idx));
+        Builder.CreateAlignedLoad(EltTy, Ptr, AlignVal, "Load" + Twine(Idx));
     Value *NewVResult = Builder.CreateInsertElement(VResult, Load,
                                                     Builder.getInt32(Idx),
                                                     "Res" + Twine(Idx));

llvm/lib/CodeGen/ShadowStackGCLowering.cpp

@@ -312,7 +312,8 @@ bool ShadowStackGCLowering::runOnFunction(Function &F) {
   AtEntry.SetInsertPoint(IP->getParent(), IP);
 
   // Initialize the map pointer and load the current head of the shadow stack.
-  Instruction *CurrentHead = AtEntry.CreateLoad(Head, "gc_currhead");
+  Instruction *CurrentHead =
+      AtEntry.CreateLoad(StackEntryTy->getPointerTo(), Head, "gc_currhead");
   Instruction *EntryMapPtr = CreateGEP(Context, AtEntry, ConcreteStackEntryTy,
                                        StackEntry, 0, 1, "gc_frame.map");
   AtEntry.CreateStore(FrameMap, EntryMapPtr);
@@ -353,7 +354,8 @@ bool ShadowStackGCLowering::runOnFunction(Function &F) {
     Instruction *EntryNextPtr2 =
         CreateGEP(Context, *AtExit, ConcreteStackEntryTy, StackEntry, 0, 0,
                   "gc_frame.next");
-    Value *SavedHead = AtExit->CreateLoad(EntryNextPtr2, "gc_savedhead");
+    Value *SavedHead = AtExit->CreateLoad(StackEntryTy->getPointerTo(),
+                                          EntryNextPtr2, "gc_savedhead");
     AtExit->CreateStore(SavedHead, Head);
   }
 

llvm/lib/CodeGen/SjLjEHPrepare.cpp

@@ -189,14 +189,16 @@ Value *SjLjEHPrepare::setupFunctionContext(Function &F,
         Builder.CreateConstGEP2_32(FunctionContextTy, FuncCtx, 0, 2, "__data");
 
     // The exception values come back in context->__data[0].
+    Type *Int32Ty = Type::getInt32Ty(F.getContext());
     Value *ExceptionAddr = Builder.CreateConstGEP2_32(doubleUnderDataTy, FCData,
                                                       0, 0, "exception_gep");
-    Value *ExnVal = Builder.CreateLoad(ExceptionAddr, true, "exn_val");
+    Value *ExnVal = Builder.CreateLoad(Int32Ty, ExceptionAddr, true, "exn_val");
     ExnVal = Builder.CreateIntToPtr(ExnVal, Builder.getInt8PtrTy());
 
     Value *SelectorAddr = Builder.CreateConstGEP2_32(doubleUnderDataTy, FCData,
                                                      0, 1, "exn_selector_gep");
-    Value *SelVal = Builder.CreateLoad(SelectorAddr, true, "exn_selector_val");
+    Value *SelVal =
+        Builder.CreateLoad(Int32Ty, SelectorAddr, true, "exn_selector_val");
 
     substituteLPadValues(LPI, ExnVal, SelVal);
   }

llvm/lib/CodeGen/StackProtector.cpp

@@ -322,7 +322,7 @@ static Value *getStackGuard(const TargetLoweringBase *TLI, Module *M,
                             IRBuilder<> &B,
                             bool *SupportsSelectionDAGSP = nullptr) {
   if (Value *Guard = TLI->getIRStackGuard(B))
-    return B.CreateLoad(Guard, true, "StackGuard");
+    return B.CreateLoad(B.getInt8PtrTy(), Guard, true, "StackGuard");
 
   // Use SelectionDAG SSP handling, since there isn't an IR guard.
   //
@@ -417,7 +417,7 @@ bool StackProtector::InsertStackProtectors() {
       // Generate the function-based epilogue instrumentation.
       // The target provides a guard check function, generate a call to it.
       IRBuilder<> B(RI);
-      LoadInst *Guard = B.CreateLoad(AI, true, "Guard");
+      LoadInst *Guard = B.CreateLoad(B.getInt8PtrTy(), AI, true, "Guard");
       CallInst *Call = B.CreateCall(GuardCheck, {Guard});
       Call->setAttributes(GuardCheck->getAttributes());
       Call->setCallingConv(GuardCheck->getCallingConv());
@@ -472,7 +472,7 @@ bool StackProtector::InsertStackProtectors() {
       // Generate the stack protector instructions in the old basic block.
       IRBuilder<> B(BB);
       Value *Guard = getStackGuard(TLI, M, B);
-      LoadInst *LI2 = B.CreateLoad(AI, true);
+      LoadInst *LI2 = B.CreateLoad(B.getInt8PtrTy(), AI, true);
       Value *Cmp = B.CreateICmpEQ(Guard, LI2);
       auto SuccessProb =
           BranchProbabilityInfo::getBranchProbStackProtector(true);

llvm/lib/CodeGen/WasmEHPrepare.cpp

@@ -344,7 +344,8 @@ void WasmEHPrepare::prepareEHPad(BasicBlock *BB, bool NeedLSDA,
   PersCI->setDoesNotThrow();
 
   // Pseudocode: int selector = __wasm.landingpad_context.selector;
-  Instruction *Selector = IRB.CreateLoad(SelectorField, "selector");
+  Instruction *Selector =
+      IRB.CreateLoad(IRB.getInt32Ty(), SelectorField, "selector");
 
   // Replace the return value from wasm.get.ehselector() with the selector value
   // loaded from __wasm_lpad_context.selector.

llvm/lib/CodeGen/WinEHPrepare.cpp

@@ -1079,7 +1079,8 @@ AllocaInst *WinEHPrepare::insertPHILoads(PHINode *PN, Function &F) {
     SpillSlot = new AllocaInst(PN->getType(), DL->getAllocaAddrSpace(), nullptr,
                                Twine(PN->getName(), ".wineh.spillslot"),
                                &F.getEntryBlock().front());
-    Value *V = new LoadInst(SpillSlot, Twine(PN->getName(), ".wineh.reload"),
+    Value *V = new LoadInst(PN->getType(), SpillSlot,
+                            Twine(PN->getName(), ".wineh.reload"),
                             &*PHIBlock->getFirstInsertionPt());
     PN->replaceAllUsesWith(V);
     return SpillSlot;
@@ -1221,13 +1222,15 @@ void WinEHPrepare::replaceUseWithLoad(Value *V, Use &U, AllocaInst *&SpillSlot,
     Value *&Load = Loads[IncomingBlock];
     // Insert the load into the predecessor block
     if (!Load)
-      Load = new LoadInst(SpillSlot, Twine(V->getName(), ".wineh.reload"),
+      Load = new LoadInst(V->getType(), SpillSlot,
+                          Twine(V->getName(), ".wineh.reload"),
                           /*Volatile=*/false, IncomingBlock->getTerminator());
 
     U.set(Load);
   } else {
     // Reload right before the old use.
-    auto *Load = new LoadInst(SpillSlot, Twine(V->getName(), ".wineh.reload"),
+    auto *Load = new LoadInst(V->getType(), SpillSlot,
+                              Twine(V->getName(), ".wineh.reload"),
                               /*Volatile=*/false, UsingInst);
     U.set(Load);
   }

llvm/lib/ExecutionEngine/Orc/IndirectionUtils.cpp

@@ -235,14 +235,13 @@ void makeStub(Function &F, Value &ImplPointer) {
   assert(F.isDeclaration() && "Can't turn a definition into a stub.");
   assert(F.getParent() && "Function isn't in a module.");
   Module &M = *F.getParent();
-  FunctionType *FTy = F.getFunctionType();
   BasicBlock *EntryBlock = BasicBlock::Create(M.getContext(), "entry", &F);
   IRBuilder<> Builder(EntryBlock);
-  LoadInst *ImplAddr = Builder.CreateLoad(&ImplPointer);
+  LoadInst *ImplAddr = Builder.CreateLoad(F.getType(), &ImplPointer);
   std::vector<Value*> CallArgs;
   for (auto &A : F.args())
     CallArgs.push_back(&A);
-  CallInst *Call = Builder.CreateCall(FTy, ImplAddr, CallArgs);
+  CallInst *Call = Builder.CreateCall(F.getFunctionType(), ImplAddr, CallArgs);
   Call->setTailCall();
   Call->setAttributes(F.getAttributes());
   if (F.getReturnType()->isVoidTy())

llvm/lib/FuzzMutate/RandomIRBuilder.cpp

@@ -53,7 +53,8 @@ Value *RandomIRBuilder::newSource(BasicBlock &BB, ArrayRef<Instruction *> Insts,
       IP = ++I->getIterator();
       assert(IP != BB.end() && "guaranteed by the findPointer");
     }
-    auto *NewLoad = new LoadInst(Ptr, "L", &*IP);
+    auto *NewLoad = new LoadInst(
+        cast<PointerType>(Ptr->getType())->getElementType(), Ptr, "L", &*IP);
 
     // Only sample this load if it really matches the descriptor
     if (Pred.matches(Srcs, NewLoad))

llvm/lib/IR/AutoUpgrade.cpp

@@ -1162,16 +1162,16 @@ static Value *UpgradeMaskedStore(IRBuilder<> &Builder,
 static Value *UpgradeMaskedLoad(IRBuilder<> &Builder,
                                 Value *Ptr, Value *Passthru, Value *Mask,
                                 bool Aligned) {
+  Type *ValTy = Passthru->getType();
   // Cast the pointer to the right type.
-  Ptr = Builder.CreateBitCast(Ptr,
+  Ptr = Builder.CreateBitCast(Ptr, llvm::PointerType::getUnqual(ValTy));
-                             llvm::PointerType::getUnqual(Passthru->getType()));
   unsigned Align =
     Aligned ? cast<VectorType>(Passthru->getType())->getBitWidth() / 8 : 1;
 
   // If the mask is all ones just emit a regular store.
   if (const auto *C = dyn_cast<Constant>(Mask))
     if (C->isAllOnesValue())
-      return Builder.CreateAlignedLoad(Ptr, Align);
+      return Builder.CreateAlignedLoad(ValTy, Ptr, Align);
 
   // Convert the mask from an integer type to a vector of i1.
   unsigned NumElts = Passthru->getType()->getVectorNumElements();
@@ -2199,7 +2199,7 @@ void llvm::UpgradeIntrinsicCall(CallInst *CI, Function *NewFn) {
       Type *VT = VectorType::get(EltTy, NumSrcElts);
       Value *Op = Builder.CreatePointerCast(CI->getArgOperand(0),
                                             PointerType::getUnqual(VT));
-      Value *Load = Builder.CreateAlignedLoad(Op, 1);
+      Value *Load = Builder.CreateAlignedLoad(VT, Op, 1);
       if (NumSrcElts == 2)
         Rep = Builder.CreateShuffleVector(Load, UndefValue::get(Load->getType()),
                                           { 0, 1, 0, 1 });
@@ -2945,7 +2945,7 @@ void llvm::UpgradeIntrinsicCall(CallInst *CI, Function *NewFn) {
       // Convert the type of the pointer to a pointer to the stored type.
       Value *BC =
           Builder.CreateBitCast(Ptr, PointerType::getUnqual(VTy), "cast");
-      LoadInst *LI = Builder.CreateAlignedLoad(BC, VTy->getBitWidth() / 8);
+      LoadInst *LI = Builder.CreateAlignedLoad(VTy, BC, VTy->getBitWidth() / 8);
       LI->setMetadata(M->getMDKindID("nontemporal"), Node);
       Rep = LI;
     } else if (IsX86 &&

llvm/lib/Target/AArch64/AArch64PromoteConstant.cpp

@@ -493,7 +493,8 @@ void AArch64PromoteConstant::insertDefinitions(Function &F,
   for (const auto &IPI : InsertPts) {
     // Create the load of the global variable.
     IRBuilder<> Builder(IPI.first);
-    LoadInst *LoadedCst = Builder.CreateLoad(&PromotedGV);
+    LoadInst *LoadedCst =
+        Builder.CreateLoad(PromotedGV.getValueType(), &PromotedGV);
     LLVM_DEBUG(dbgs() << "**********\n");
     LLVM_DEBUG(dbgs() << "New def: ");
     LLVM_DEBUG(LoadedCst->print(dbgs()));

llvm/lib/Target/AMDGPU/AMDGPUCodeGenPrepare.cpp

@@ -806,7 +806,7 @@ bool AMDGPUCodeGenPrepare::visitLoadInst(LoadInst &I) {
     Type *I32Ty = Builder.getInt32Ty();
     Type *PT = PointerType::get(I32Ty, I.getPointerAddressSpace());
     Value *BitCast= Builder.CreateBitCast(I.getPointerOperand(), PT);
-    LoadInst *WidenLoad = Builder.CreateLoad(BitCast);
+    LoadInst *WidenLoad = Builder.CreateLoad(I32Ty, BitCast);
     WidenLoad->copyMetadata(I);
 
     // If we have range metadata, we need to convert the type, and not make

llvm/lib/Target/AMDGPU/AMDGPULibCalls.cpp

@@ -1357,12 +1357,12 @@ bool AMDGPULibCalls::fold_sincos(CallInst *CI, IRBuilder<> &B,
   if (!isSin) { // CI->cos, UI->sin
     B.SetInsertPoint(&*ItOld);
     UI->replaceAllUsesWith(&*Call);
-    Instruction *Reload = B.CreateLoad(Alloc);
+    Instruction *Reload = B.CreateLoad(Alloc->getAllocatedType(), Alloc);
     CI->replaceAllUsesWith(Reload);
     UI->eraseFromParent();
     CI->eraseFromParent();
   } else { // CI->sin, UI->cos
-    Instruction *Reload = B.CreateLoad(Alloc);
+    Instruction *Reload = B.CreateLoad(Alloc->getAllocatedType(), Alloc);
     UI->replaceAllUsesWith(Reload);
     CI->replaceAllUsesWith(Call);
     UI->eraseFromParent();

llvm/lib/Target/AMDGPU/AMDGPULowerKernelArguments.cpp

@@ -132,6 +132,7 @@ bool AMDGPULowerKernelArguments::runOnFunction(Function &F) {
                                       KernArgBaseAlign);
 
     Value *ArgPtr;
+    Type *AdjustedArgTy;
     if (DoShiftOpt) { // FIXME: Handle aggregate types
       // Since we don't have sub-dword scalar loads, avoid doing an extload by
       // loading earlier than the argument address, and extracting the relevant
@@ -144,25 +145,25 @@ bool AMDGPULowerKernelArguments::runOnFunction(Function &F) {
         KernArgSegment,
         AlignDownOffset,
         Arg.getName() + ".kernarg.offset.align.down");
-      ArgPtr = Builder.CreateBitCast(ArgPtr,
+      AdjustedArgTy = Builder.getInt32Ty();
-                                     Builder.getInt32Ty()->getPointerTo(AS),
-                                     ArgPtr->getName() + ".cast");
     } else {
       ArgPtr = Builder.CreateConstInBoundsGEP1_64(
         KernArgSegment,
         EltOffset,
         Arg.getName() + ".kernarg.offset");
-      ArgPtr = Builder.CreateBitCast(ArgPtr, ArgTy->getPointerTo(AS),
+      AdjustedArgTy = ArgTy;
-                                     ArgPtr->getName() + ".cast");
     }
 
     if (IsV3 && Size >= 32) {
       V4Ty = VectorType::get(VT->getVectorElementType(), 4);
       // Use the hack that clang uses to avoid SelectionDAG ruining v3 loads
-      ArgPtr = Builder.CreateBitCast(ArgPtr, V4Ty->getPointerTo(AS));
+      AdjustedArgTy = V4Ty;
     }
 
-    LoadInst *Load = Builder.CreateAlignedLoad(ArgPtr, AdjustedAlign);
+    ArgPtr = Builder.CreateBitCast(ArgPtr, AdjustedArgTy->getPointerTo(AS),
+                                   ArgPtr->getName() + ".cast");
+    LoadInst *Load =
+        Builder.CreateAlignedLoad(AdjustedArgTy, ArgPtr, AdjustedAlign);
     Load->setMetadata(LLVMContext::MD_invariant_load, MDNode::get(Ctx, {}));
 
     MDBuilder MDB(Ctx);

llvm/lib/Target/AMDGPU/AMDGPUPromoteAlloca.cpp

@@ -245,10 +245,10 @@ AMDGPUPromoteAlloca::getLocalSizeYZ(IRBuilder<> &Builder) {
   // 32-bit and extract sequence is already present, and it is probably easier
   // to CSE this. The loads should be mergable later anyway.
   Value *GEPXY = Builder.CreateConstInBoundsGEP1_64(CastDispatchPtr, 1);
-  LoadInst *LoadXY = Builder.CreateAlignedLoad(GEPXY, 4);
+  LoadInst *LoadXY = Builder.CreateAlignedLoad(I32Ty, GEPXY, 4);
 
   Value *GEPZU = Builder.CreateConstInBoundsGEP1_64(CastDispatchPtr, 2);
-  LoadInst *LoadZU = Builder.CreateAlignedLoad(GEPZU, 4);
+  LoadInst *LoadZU = Builder.CreateAlignedLoad(I32Ty, GEPZU, 4);
 
   MDNode *MD = MDNode::get(Mod->getContext(), None);
   LoadXY->setMetadata(LLVMContext::MD_invariant_load, MD);
@@ -426,7 +426,7 @@ static bool tryPromoteAllocaToVector(AllocaInst *Alloca) {
       Value *Index = calculateVectorIndex(Ptr, GEPVectorIdx);
 
       Value *BitCast = Builder.CreateBitCast(Alloca, VecPtrTy);
-      Value *VecValue = Builder.CreateLoad(BitCast);
+      Value *VecValue = Builder.CreateLoad(VectorTy, BitCast);
       Value *ExtractElement = Builder.CreateExtractElement(VecValue, Index);
       Inst->replaceAllUsesWith(ExtractElement);
       Inst->eraseFromParent();
@@ -441,7 +441,7 @@ static bool tryPromoteAllocaToVector(AllocaInst *Alloca) {
       Value *Ptr = SI->getPointerOperand();
       Value *Index = calculateVectorIndex(Ptr, GEPVectorIdx);
       Value *BitCast = Builder.CreateBitCast(Alloca, VecPtrTy);
-      Value *VecValue = Builder.CreateLoad(BitCast);
+      Value *VecValue = Builder.CreateLoad(VectorTy, BitCast);
       Value *NewVecValue = Builder.CreateInsertElement(VecValue,
                                                        SI->getValueOperand(),
                                                        Index);

llvm/lib/Target/ARM/ARMParallelDSP.cpp

@@ -688,12 +688,12 @@ bool ARMParallelDSP::MatchSMLAD(Function &F) {
 }
 
 static LoadInst *CreateLoadIns(IRBuilder<NoFolder> &IRB, LoadInst &BaseLoad,
-                               const Type *LoadTy) {
+                               Type *LoadTy) {
   const unsigned AddrSpace = BaseLoad.getPointerAddressSpace();
 
   Value *VecPtr = IRB.CreateBitCast(BaseLoad.getPointerOperand(),
                                     LoadTy->getPointerTo(AddrSpace));
-  return IRB.CreateAlignedLoad(VecPtr, BaseLoad.getAlignment());
+  return IRB.CreateAlignedLoad(LoadTy, VecPtr, BaseLoad.getAlignment());
 }
 
 Instruction *ARMParallelDSP::CreateSMLADCall(LoadInst *VecLd0, LoadInst *VecLd1,
@@ -709,7 +709,7 @@ Instruction *ARMParallelDSP::CreateSMLADCall(LoadInst *VecLd0, LoadInst *VecLd1,
                               ++BasicBlock::iterator(InsertAfter));
 
   // Replace the reduction chain with an intrinsic call
-  const Type *Ty = IntegerType::get(M->getContext(), 32);
+  Type *Ty = IntegerType::get(M->getContext(), 32);
   LoadInst *NewLd0 = CreateLoadIns(Builder, VecLd0[0], Ty);
   LoadInst *NewLd1 = CreateLoadIns(Builder, VecLd1[0], Ty);
   Value* Args[] = { NewLd0, NewLd1, Acc };

llvm/lib/Target/NVPTX/NVPTXLowerArgs.cpp

@@ -169,7 +169,8 @@ void NVPTXLowerArgs::handleByValParam(Argument *Arg) {
   Value *ArgInParam = new AddrSpaceCastInst(
       Arg, PointerType::get(StructType, ADDRESS_SPACE_PARAM), Arg->getName(),
       FirstInst);
-  LoadInst *LI = new LoadInst(ArgInParam, Arg->getName(), FirstInst);
+  LoadInst *LI =
+      new LoadInst(StructType, ArgInParam, Arg->getName(), FirstInst);
   new StoreInst(LI, AllocA, FirstInst);
 }
 

llvm/lib/Target/WebAssembly/WebAssemblyLowerEmscriptenEHSjLj.cpp

@@ -445,7 +445,8 @@ Value *WebAssemblyLowerEmscriptenEHSjLj::wrapInvoke(CallOrInvoke *CI) {
 
   // Post-invoke
   // %__THREW__.val = __THREW__; __THREW__ = 0;
-  Value *Threw = IRB.CreateLoad(ThrewGV, ThrewGV->getName() + ".val");
+  Value *Threw =
+      IRB.CreateLoad(IRB.getInt32Ty(), ThrewGV, ThrewGV->getName() + ".val");
   IRB.CreateStore(IRB.getInt32(0), ThrewGV);
   return Threw;
 }
@@ -548,8 +549,8 @@ void WebAssemblyLowerEmscriptenEHSjLj::wrapTestSetjmp(
   BasicBlock *ElseBB1 = BasicBlock::Create(C, "if.else1", F);
   BasicBlock *EndBB1 = BasicBlock::Create(C, "if.end", F);
   Value *ThrewCmp = IRB.CreateICmpNE(Threw, IRB.getInt32(0));
-  Value *ThrewValue =
+  Value *ThrewValue = IRB.CreateLoad(IRB.getInt32Ty(), ThrewValueGV,
-      IRB.CreateLoad(ThrewValueGV, ThrewValueGV->getName() + ".val");
+                                     ThrewValueGV->getName() + ".val");
   Value *ThrewValueCmp = IRB.CreateICmpNE(ThrewValue, IRB.getInt32(0));
   Value *Cmp1 = IRB.CreateAnd(ThrewCmp, ThrewValueCmp, "cmp1");
   IRB.CreateCondBr(Cmp1, ThenBB1, ElseBB1);
@@ -561,8 +562,8 @@ void WebAssemblyLowerEmscriptenEHSjLj::wrapTestSetjmp(
   BasicBlock *EndBB2 = BasicBlock::Create(C, "if.end2", F);
   Value *ThrewInt = IRB.CreateIntToPtr(Threw, Type::getInt32PtrTy(C),
                                        Threw->getName() + ".i32p");
-  Value *LoadedThrew =
+  Value *LoadedThrew = IRB.CreateLoad(IRB.getInt32Ty(), ThrewInt,
-      IRB.CreateLoad(ThrewInt, ThrewInt->getName() + ".loaded");
+                                      ThrewInt->getName() + ".loaded");
   Value *ThenLabel = IRB.CreateCall(
       TestSetjmpF, {LoadedThrew, SetjmpTable, SetjmpTableSize}, "label");
   Value *Cmp2 = IRB.CreateICmpEQ(ThenLabel, IRB.getInt32(0));

llvm/lib/Target/X86/X86ISelLowering.cpp

@@ -25362,7 +25362,6 @@ X86TargetLowering::lowerIdempotentRMWIntoFencedLoad(AtomicRMWInst *AI) const {
   // We must restrict the ordering to avoid generating loads with Release or
   // ReleaseAcquire orderings.
   auto Order = AtomicCmpXchgInst::getStrongestFailureOrdering(AI->getOrdering());
-  auto Ptr = AI->getPointerOperand();
 
   // Before the load we need a fence. Here is an example lifted from
   // http://www.hpl.hp.com/techreports/2012/HPL-2012-68.pdf showing why a fence
@@ -25397,8 +25396,9 @@ X86TargetLowering::lowerIdempotentRMWIntoFencedLoad(AtomicRMWInst *AI) const {
   Builder.CreateCall(MFence, {});
 
   // Finally we can emit the atomic load.
-  LoadInst *Loaded = Builder.CreateAlignedLoad(Ptr,
+  LoadInst *Loaded =
-          AI->getType()->getPrimitiveSizeInBits());
+      Builder.CreateAlignedLoad(AI->getType(), AI->getPointerOperand(),
+                                AI->getType()->getPrimitiveSizeInBits());
   Loaded->setAtomic(Order, SSID);
   AI->replaceAllUsesWith(Loaded);
   AI->eraseFromParent();

llvm/lib/Target/X86/X86InterleavedAccess.cpp

@@ -193,7 +193,7 @@ void X86InterleavedAccessGroup::decompose(
 
   // Decompose the load instruction.
   LoadInst *LI = cast<LoadInst>(VecInst);
-  Type *VecBasePtrTy = SubVecTy->getPointerTo(LI->getPointerAddressSpace());
+  Type *VecBaseTy, *VecBasePtrTy;
   Value *VecBasePtr;
   unsigned int NumLoads = NumSubVectors;
   // In the case of stride 3 with a vector of 32 elements load the information
@@ -201,18 +201,21 @@ void X86InterleavedAccessGroup::decompose(
   // [0,1...,VF/2-1,VF/2+VF,VF/2+VF+1,...,2VF-1]
   unsigned VecLength = DL.getTypeSizeInBits(VecWidth);
   if (VecLength == 768 || VecLength == 1536) {
-    Type *VecTran =
+    VecBaseTy = VectorType::get(Type::getInt8Ty(LI->getContext()), 16);
-        VectorType::get(Type::getInt8Ty(LI->getContext()), 16)->getPointerTo();
+    VecBasePtrTy = VecBaseTy->getPointerTo(LI->getPointerAddressSpace());
-    VecBasePtr = Builder.CreateBitCast(LI->getPointerOperand(), VecTran);
-    NumLoads = NumSubVectors * (VecLength / 384);
-  } else
     VecBasePtr = Builder.CreateBitCast(LI->getPointerOperand(), VecBasePtrTy);
+    NumLoads = NumSubVectors * (VecLength / 384);
+  } else {
+    VecBaseTy = SubVecTy;
+    VecBasePtrTy = VecBaseTy->getPointerTo(LI->getPointerAddressSpace());
+    VecBasePtr = Builder.CreateBitCast(LI->getPointerOperand(), VecBasePtrTy);
+  }
   // Generate N loads of T type.
   for (unsigned i = 0; i < NumLoads; i++) {
     // TODO: Support inbounds GEP.
     Value *NewBasePtr = Builder.CreateGEP(VecBasePtr, Builder.getInt32(i));
     Instruction *NewLoad =
-        Builder.CreateAlignedLoad(NewBasePtr, LI->getAlignment());
+        Builder.CreateAlignedLoad(VecBaseTy, NewBasePtr, LI->getAlignment());
     DecomposedVectors.push_back(NewLoad);
   }
 }

llvm/lib/Target/X86/X86WinEHState.cpp

@@ -432,7 +432,7 @@ void WinEHStatePass::linkExceptionRegistration(IRBuilder<> &Builder,
   // Next = [fs:00]
   Constant *FSZero =
       Constant::getNullValue(LinkTy->getPointerTo()->getPointerTo(257));
-  Value *Next = Builder.CreateLoad(FSZero);
+  Value *Next = Builder.CreateLoad(LinkTy->getPointerTo(), FSZero);
   Builder.CreateStore(Next, Builder.CreateStructGEP(LinkTy, Link, 0));
   // [fs:00] = Link
   Builder.CreateStore(Link, FSZero);
@@ -447,8 +447,8 @@ void WinEHStatePass::unlinkExceptionRegistration(IRBuilder<> &Builder) {
   }
   Type *LinkTy = getEHLinkRegistrationType();
   // [fs:00] = Link->Next
-  Value *Next =
+  Value *Next = Builder.CreateLoad(LinkTy->getPointerTo(),
-      Builder.CreateLoad(Builder.CreateStructGEP(LinkTy, Link, 0));
+                                   Builder.CreateStructGEP(LinkTy, Link, 0));
   Constant *FSZero =
       Constant::getNullValue(LinkTy->getPointerTo()->getPointerTo(257));
   Builder.CreateStore(Next, FSZero);
@@ -783,7 +783,7 @@ void WinEHStatePass::addStateStores(Function &F, WinEHFuncInfo &FuncInfo) {
     if (InCleanup) {
       Value *StateField =
           Builder.CreateStructGEP(nullptr, RegNode, StateFieldIndex);
-      State = Builder.CreateLoad(StateField);
+      State = Builder.CreateLoad(Builder.getInt32Ty(), StateField);
     } else {
       State = Builder.getInt32(getStateForCallSite(BlockColors, FuncInfo, CS));
     }

llvm/lib/Transforms/Coroutines/CoroCleanup.cpp

@@ -49,7 +49,7 @@ static void lowerSubFn(IRBuilder<> &Builder, CoroSubFnInst *SubFn) {
   Builder.SetInsertPoint(SubFn);
   auto *FramePtr = Builder.CreateBitCast(FrameRaw, FramePtrTy);
   auto *Gep = Builder.CreateConstInBoundsGEP2_32(FrameTy, FramePtr, 0, Index);
-  auto *Load = Builder.CreateLoad(Gep);
+  auto *Load = Builder.CreateLoad(FrameTy->getElementType(Index), Gep);
 
   SubFn->replaceAllUsesWith(Load);
 }

llvm/lib/Transforms/Coroutines/CoroEarly.cpp

@@ -97,7 +97,7 @@ void Lowerer::lowerCoroDone(IntrinsicInst *II) {
   Builder.SetInsertPoint(II);
   auto *BCI = Builder.CreateBitCast(Operand, FramePtrTy);
   auto *Gep = Builder.CreateConstInBoundsGEP1_32(FrameTy, BCI, 0);
-  auto *Load = Builder.CreateLoad(Gep);
+  auto *Load = Builder.CreateLoad(FrameTy, Gep);
   auto *Cond = Builder.CreateICmpEQ(Load, NullPtr);
 
   II->replaceAllUsesWith(Cond);

llvm/lib/Transforms/Coroutines/CoroFrame.cpp

@@ -471,10 +471,10 @@ static Instruction *splitBeforeCatchSwitch(CatchSwitchInst *CatchSwitch) {
 static Instruction *insertSpills(SpillInfo &Spills, coro::Shape &Shape) {
   auto *CB = Shape.CoroBegin;
   IRBuilder<> Builder(CB->getNextNode());
-  PointerType *FramePtrTy = Shape.FrameTy->getPointerTo();
+  StructType *FrameTy = Shape.FrameTy;
+  PointerType *FramePtrTy = FrameTy->getPointerTo();
   auto *FramePtr =
       cast<Instruction>(Builder.CreateBitCast(CB, FramePtrTy, "FramePtr"));
-  Type *FrameTy = FramePtrTy->getElementType();
 
   Value *CurrentValue = nullptr;
   BasicBlock *CurrentBlock = nullptr;
@@ -501,7 +501,7 @@ static Instruction *insertSpills(SpillInfo &Spills, coro::Shape &Shape) {
                                                      Twine(".reload.addr"));
     return isa<AllocaInst>(CurrentValue)
                ? G
-               : Builder.CreateLoad(G,
+               : Builder.CreateLoad(FrameTy->getElementType(Index), G,
                                     CurrentValue->getName() + Twine(".reload"));
   };
 

llvm/lib/Transforms/Coroutines/CoroSplit.cpp

@@ -93,7 +93,7 @@ static BasicBlock *createResumeEntryBlock(Function &F, coro::Shape &Shape) {
   auto *FrameTy = Shape.FrameTy;
   auto *GepIndex = Builder.CreateConstInBoundsGEP2_32(
       FrameTy, FramePtr, 0, coro::Shape::IndexField, "index.addr");
-  auto *Index = Builder.CreateLoad(GepIndex, "index");
+  auto *Index = Builder.CreateLoad(Shape.getIndexType(), GepIndex, "index");
   auto *Switch =
       Builder.CreateSwitch(Index, UnreachBB, Shape.CoroSuspends.size());
   Shape.ResumeSwitch = Switch;
@@ -229,7 +229,8 @@ static void handleFinalSuspend(IRBuilder<> &Builder, Value *FramePtr,
     Builder.SetInsertPoint(OldSwitchBB->getTerminator());
     auto *GepIndex = Builder.CreateConstInBoundsGEP2_32(Shape.FrameTy, FramePtr,
                                                         0, 0, "ResumeFn.addr");
-    auto *Load = Builder.CreateLoad(GepIndex);
+    auto *Load = Builder.CreateLoad(
+        Shape.FrameTy->getElementType(coro::Shape::ResumeField), GepIndex);
     auto *NullPtr =
         ConstantPointerNull::get(cast<PointerType>(Load->getType()));
     auto *Cond = Builder.CreateICmpEQ(Load, NullPtr);

llvm/lib/Transforms/IPO/ArgumentPromotion.cpp

@@ -263,7 +263,8 @@ doPromotion(Function *F, SmallPtrSetImpl<Argument *> &ArgsToPromote,
           Value *Idx = GetElementPtrInst::Create(
               STy, *AI, Idxs, (*AI)->getName() + "." + Twine(i), Call);
           // TODO: Tell AA about the new values?
-          Args.push_back(new LoadInst(Idx, Idx->getName() + ".val", Call));
+          Args.push_back(new LoadInst(STy->getElementType(i), Idx,
+                                      Idx->getName() + ".val", Call));
           ArgAttrVec.push_back(AttributeSet());
         }
       } else if (!I->use_empty()) {
@@ -299,7 +300,8 @@ doPromotion(Function *F, SmallPtrSetImpl<Argument *> &ArgsToPromote,
           }
           // Since we're replacing a load make sure we take the alignment
           // of the previous load.
-          LoadInst *newLoad = new LoadInst(V, V->getName() + ".val", Call);
+          LoadInst *newLoad =
+              new LoadInst(OrigLoad->getType(), V, V->getName() + ".val", Call);
           newLoad->setAlignment(OrigLoad->getAlignment());
           // Transfer the AA info too.
           AAMDNodes AAInfo;

llvm/lib/Transforms/IPO/GlobalOpt.cpp

@@ -905,9 +905,10 @@ OptimizeGlobalAddressOfMalloc(GlobalVariable *GV, CallInst *CI, Type *AllocTy,
 
       // Replace the cmp X, 0 with a use of the bool value.
       // Sink the load to where the compare was, if atomic rules allow us to.
-      Value *LV = new LoadInst(InitBool, InitBool->getName()+".val", false, 0,
+      Value *LV = new LoadInst(InitBool->getValueType(), InitBool,
+                               InitBool->getName() + ".val", false, 0,
                                LI->getOrdering(), LI->getSyncScopeID(),
-                               LI->isUnordered() ? (Instruction*)ICI : LI);
+                               LI->isUnordered() ? (Instruction *)ICI : LI);
       InitBoolUsed = true;
       switch (ICI->getPredicate()) {
       default: llvm_unreachable("Unknown ICmp Predicate!");
@@ -1040,7 +1041,8 @@ static void ReplaceUsesOfMallocWithGlobal(Instruction *Alloc,
     }
 
     // Insert a load from the global, and use it instead of the malloc.
-    Value *NL = new LoadInst(GV, GV->getName()+".val", InsertPt);
+    Value *NL =
+        new LoadInst(GV->getValueType(), GV, GV->getName() + ".val", InsertPt);
     U->replaceUsesOfWith(Alloc, NL);
   }
 }
@@ -1163,10 +1165,10 @@ static Value *GetHeapSROAValue(Value *V, unsigned FieldNo,
   if (LoadInst *LI = dyn_cast<LoadInst>(V)) {
     // This is a scalarized version of the load from the global.  Just create
     // a new Load of the scalarized global.
-    Result = new LoadInst(GetHeapSROAValue(LI->getOperand(0), FieldNo,
+    Value *V = GetHeapSROAValue(LI->getOperand(0), FieldNo,
-                                           InsertedScalarizedValues,
+                                InsertedScalarizedValues, PHIsToRewrite);
-                                           PHIsToRewrite),
+    Result = new LoadInst(V->getType()->getPointerElementType(), V,
-                          LI->getName()+".f"+Twine(FieldNo), LI);
+                          LI->getName() + ".f" + Twine(FieldNo), LI);
   } else {
     PHINode *PN = cast<PHINode>(V);
     // PN's type is pointer to struct.  Make a new PHI of pointer to struct
@@ -1356,7 +1358,9 @@ static GlobalVariable *PerformHeapAllocSRoA(GlobalVariable *GV, CallInst *CI,
   // Within the NullPtrBlock, we need to emit a comparison and branch for each
   // pointer, because some may be null while others are not.
   for (unsigned i = 0, e = FieldGlobals.size(); i != e; ++i) {
-    Value *GVVal = new LoadInst(FieldGlobals[i], "tmp", NullPtrBlock);
+    Value *GVVal =
+        new LoadInst(cast<GlobalVariable>(FieldGlobals[i])->getValueType(),
+                     FieldGlobals[i], "tmp", NullPtrBlock);
     Value *Cmp = new ICmpInst(*NullPtrBlock, ICmpInst::ICMP_NE, GVVal,
                               Constant::getNullValue(GVVal->getType()));
     BasicBlock *FreeBlock = BasicBlock::Create(Cmp->getContext(), "free_it",
@@ -1700,7 +1704,8 @@ static bool TryToShrinkGlobalToBoolean(GlobalVariable *GV, Constant *OtherVal) {
         if (LoadInst *LI = dyn_cast<LoadInst>(StoredVal)) {
           assert(LI->getOperand(0) == GV && "Not a copy!");
           // Insert a new load, to preserve the saved value.
-          StoreVal = new LoadInst(NewGV, LI->getName()+".b", false, 0,
+          StoreVal = new LoadInst(NewGV->getValueType(), NewGV,
+                                  LI->getName() + ".b", false, 0,
                                   LI->getOrdering(), LI->getSyncScopeID(), LI);
         } else {
           assert((isa<CastInst>(StoredVal) || isa<SelectInst>(StoredVal)) &&
@@ -1716,8 +1721,9 @@ static bool TryToShrinkGlobalToBoolean(GlobalVariable *GV, Constant *OtherVal) {
     } else {
       // Change the load into a load of bool then a select.
       LoadInst *LI = cast<LoadInst>(UI);
-      LoadInst *NLI = new LoadInst(NewGV, LI->getName()+".b", false, 0,
+      LoadInst *NLI =
-                                   LI->getOrdering(), LI->getSyncScopeID(), LI);
+          new LoadInst(NewGV->getValueType(), NewGV, LI->getName() + ".b",
+                       false, 0, LI->getOrdering(), LI->getSyncScopeID(), LI);
       Instruction *NSI;
       if (IsOneZero)
         NSI = new ZExtInst(NLI, LI->getType(), "", LI);

llvm/lib/Transforms/IPO/LowerTypeTests.cpp

@@ -618,7 +618,7 @@ Value *LowerTypeTestsModule::createBitSetTest(IRBuilder<> &B,
     }
 
     Value *ByteAddr = B.CreateGEP(Int8Ty, ByteArray, BitOffset);
-    Value *Byte = B.CreateLoad(ByteAddr);
+    Value *Byte = B.CreateLoad(Int8Ty, ByteAddr);
 
     Value *ByteAndMask =
         B.CreateAnd(Byte, ConstantExpr::getPtrToInt(TIL.BitMask, Int8Ty));

llvm/lib/Transforms/IPO/WholeProgramDevirt.cpp

@@ -1183,7 +1183,7 @@ void DevirtModule::applyVirtualConstProp(CallSiteInfo &CSInfo, StringRef FnName,
     Value *Addr =
         B.CreateGEP(Int8Ty, B.CreateBitCast(Call.VTable, Int8PtrTy), Byte);
     if (RetType->getBitWidth() == 1) {
-      Value *Bits = B.CreateLoad(Addr);
+      Value *Bits = B.CreateLoad(Int8Ty, Addr);
       Value *BitsAndBit = B.CreateAnd(Bits, Bit);
       auto IsBitSet = B.CreateICmpNE(BitsAndBit, ConstantInt::get(Int8Ty, 0));
       Call.replaceAndErase("virtual-const-prop-1-bit", FnName, RemarksEnabled,

llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp

@@ -171,7 +171,7 @@ Instruction *InstCombiner::SimplifyAnyMemTransfer(AnyMemTransferInst *MI) {
 
   Value *Src = Builder.CreateBitCast(MI->getArgOperand(1), NewSrcPtrTy);
   Value *Dest = Builder.CreateBitCast(MI->getArgOperand(0), NewDstPtrTy);
-  LoadInst *L = Builder.CreateLoad(Src);
+  LoadInst *L = Builder.CreateLoad(IntType, Src);
   // Alignment from the mem intrinsic will be better, so use it.
   L->setAlignment(CopySrcAlign);
   if (CopyMD)
@@ -1182,7 +1182,8 @@ static Value *simplifyMaskedLoad(const IntrinsicInst &II,
   if (maskIsAllOneOrUndef(II.getArgOperand(2))) {
     Value *LoadPtr = II.getArgOperand(0);
     unsigned Alignment = cast<ConstantInt>(II.getArgOperand(1))->getZExtValue();
-    return Builder.CreateAlignedLoad(LoadPtr, Alignment, "unmaskedload");
+    return Builder.CreateAlignedLoad(II.getType(), LoadPtr, Alignment,
+                                     "unmaskedload");
   }
 
   return nullptr;
@@ -1499,7 +1500,7 @@ static Value *simplifyNeonVld1(const IntrinsicInst &II,
 
   auto *BCastInst = Builder.CreateBitCast(II.getArgOperand(0),
                                           PointerType::get(II.getType(), 0));
-  return Builder.CreateAlignedLoad(BCastInst, Alignment);
+  return Builder.CreateAlignedLoad(II.getType(), BCastInst, Alignment);
 }
 
 // Returns true iff the 2 intrinsics have the same operands, limiting the
@@ -2300,7 +2301,7 @@ Instruction *InstCombiner::visitCallInst(CallInst &CI) {
                                    &DT) >= 16) {
       Value *Ptr = Builder.CreateBitCast(II->getArgOperand(0),
                                          PointerType::getUnqual(II->getType()));
-      return new LoadInst(Ptr);
+      return new LoadInst(II->getType(), Ptr);
     }
     break;
   case Intrinsic::ppc_vsx_lxvw4x:
@@ -2308,7 +2309,7 @@ Instruction *InstCombiner::visitCallInst(CallInst &CI) {
     // Turn PPC VSX loads into normal loads.
     Value *Ptr = Builder.CreateBitCast(II->getArgOperand(0),
                                        PointerType::getUnqual(II->getType()));
-    return new LoadInst(Ptr, Twine(""), false, 1);
+    return new LoadInst(II->getType(), Ptr, Twine(""), false, 1);
   }
   case Intrinsic::ppc_altivec_stvx:
   case Intrinsic::ppc_altivec_stvxl:
@@ -2336,7 +2337,7 @@ Instruction *InstCombiner::visitCallInst(CallInst &CI) {
                                   II->getType()->getVectorNumElements());
       Value *Ptr = Builder.CreateBitCast(II->getArgOperand(0),
                                          PointerType::getUnqual(VTy));
-      Value *Load = Builder.CreateLoad(Ptr);
+      Value *Load = Builder.CreateLoad(VTy, Ptr);
       return new FPExtInst(Load, II->getType());
     }
     break;
@@ -2346,7 +2347,7 @@ Instruction *InstCombiner::visitCallInst(CallInst &CI) {
                                    &DT) >= 32) {
       Value *Ptr = Builder.CreateBitCast(II->getArgOperand(0),
                                          PointerType::getUnqual(II->getType()));
-      return new LoadInst(Ptr);
+      return new LoadInst(II->getType(), Ptr);
     }
     break;
   case Intrinsic::ppc_qpx_qvstfs:

llvm/lib/Transforms/InstCombine/InstCombineLoadStoreAlloca.cpp

@@ -298,7 +298,7 @@ void PointerReplacer::replace(Instruction *I) {
   if (auto *LT = dyn_cast<LoadInst>(I)) {
     auto *V = getReplacement(LT->getPointerOperand());
     assert(V && "Operand not replaced");
-    auto *NewI = new LoadInst(V);
+    auto *NewI = new LoadInst(I->getType(), V);
     NewI->takeName(LT);
     IC.InsertNewInstWith(NewI, *LT);
     IC.replaceInstUsesWith(*LT, NewI);
@@ -465,7 +465,7 @@ static LoadInst *combineLoadToNewType(InstCombiner &IC, LoadInst &LI, Type *NewT
     NewPtr = IC.Builder.CreateBitCast(Ptr, NewTy->getPointerTo(AS));
 
   LoadInst *NewLoad = IC.Builder.CreateAlignedLoad(
-      NewPtr, LI.getAlignment(), LI.isVolatile(), LI.getName() + Suffix);
+      NewTy, NewPtr, LI.getAlignment(), LI.isVolatile(), LI.getName() + Suffix);
   NewLoad->setAtomic(LI.getOrdering(), LI.getSyncScopeID());
   MDBuilder MDB(NewLoad->getContext());
   for (const auto &MDPair : MD) {
@@ -724,7 +724,8 @@ static Instruction *unpackLoadToAggregate(InstCombiner &IC, LoadInst &LI) {
       auto *Ptr = IC.Builder.CreateInBoundsGEP(ST, Addr, makeArrayRef(Indices),
                                                Name + ".elt");
       auto EltAlign = MinAlign(Align, SL->getElementOffset(i));
-      auto *L = IC.Builder.CreateAlignedLoad(Ptr, EltAlign, Name + ".unpack");
+      auto *L = IC.Builder.CreateAlignedLoad(ST->getElementType(i), Ptr,
+                                             EltAlign, Name + ".unpack");
       // Propagate AA metadata. It'll still be valid on the narrowed load.
       AAMDNodes AAMD;
       LI.getAAMetadata(AAMD);
@@ -774,8 +775,8 @@ static Instruction *unpackLoadToAggregate(InstCombiner &IC, LoadInst &LI) {
       };
       auto *Ptr = IC.Builder.CreateInBoundsGEP(AT, Addr, makeArrayRef(Indices),
                                                Name + ".elt");
-      auto *L = IC.Builder.CreateAlignedLoad(Ptr, MinAlign(Align, Offset),
+      auto *L = IC.Builder.CreateAlignedLoad(
-                                             Name + ".unpack");
+          AT->getElementType(), Ptr, MinAlign(Align, Offset), Name + ".unpack");
       AAMDNodes AAMD;
       LI.getAAMetadata(AAMD);
       L->setAAMetadata(AAMD);
@@ -1065,10 +1066,12 @@ Instruction *InstCombiner::visitLoadInst(LoadInst &LI) {
       unsigned Align = LI.getAlignment();
       if (isSafeToLoadUnconditionally(SI->getOperand(1), Align, DL, SI) &&
           isSafeToLoadUnconditionally(SI->getOperand(2), Align, DL, SI)) {
-        LoadInst *V1 = Builder.CreateLoad(SI->getOperand(1),
+        LoadInst *V1 =
-                                          SI->getOperand(1)->getName()+".val");
+            Builder.CreateLoad(LI.getType(), SI->getOperand(1),
-        LoadInst *V2 = Builder.CreateLoad(SI->getOperand(2),
+                               SI->getOperand(1)->getName() + ".val");
-                                          SI->getOperand(2)->getName()+".val");
+        LoadInst *V2 =
+            Builder.CreateLoad(LI.getType(), SI->getOperand(2),
+                               SI->getOperand(2)->getName() + ".val");
         assert(LI.isUnordered() && "implied by above");
         V1->setAlignment(Align);
         V1->setAtomic(LI.getOrdering(), LI.getSyncScopeID());

llvm/lib/Transforms/InstCombine/InstCombinePHI.cpp

@@ -595,7 +595,8 @@ Instruction *InstCombiner::FoldPHIArgLoadIntoPHI(PHINode &PN) {
 
   Value *InVal = FirstLI->getOperand(0);
   NewPN->addIncoming(InVal, PN.getIncomingBlock(0));
-  LoadInst *NewLI = new LoadInst(NewPN, "", isVolatile, LoadAlignment);
+  LoadInst *NewLI =
+      new LoadInst(FirstLI->getType(), NewPN, "", isVolatile, LoadAlignment);
 
   unsigned KnownIDs[] = {
     LLVMContext::MD_tbaa,

llvm/lib/Transforms/InstCombine/InstructionCombining.cpp

@@ -2686,7 +2686,7 @@ Instruction *InstCombiner::visitExtractValueInst(ExtractValueInst &EV) {
       Builder.SetInsertPoint(L);
       Value *GEP = Builder.CreateInBoundsGEP(L->getType(),
                                              L->getPointerOperand(), Indices);
-      Instruction *NL = Builder.CreateLoad(GEP);
+      Instruction *NL = Builder.CreateLoad(EV.getType(), GEP);
       // Whatever aliasing information we had for the orignal load must also
       // hold for the smaller load, so propagate the annotations.
       AAMDNodes Nodes;

llvm/lib/Transforms/Instrumentation/AddressSanitizer.cpp

@@ -949,8 +949,9 @@ struct FunctionStackPoisoner : public InstVisitor<FunctionStackPoisoner> {
                                      DynamicAreaOffset);
     }
 
-    IRB.CreateCall(AsanAllocasUnpoisonFunc,
+    IRB.CreateCall(
-                   {IRB.CreateLoad(DynamicAllocaLayout), DynamicAreaPtr});
+        AsanAllocasUnpoisonFunc,
+        {IRB.CreateLoad(IntptrTy, DynamicAllocaLayout), DynamicAreaPtr});
   }
 
   // Unpoison dynamic allocas redzones.
@@ -1552,7 +1553,7 @@ void AddressSanitizer::instrumentAddress(Instruction *OrigIns,
   Value *ShadowPtr = memToShadow(AddrLong, IRB);
   Value *CmpVal = Constant::getNullValue(ShadowTy);
   Value *ShadowValue =
-      IRB.CreateLoad(IRB.CreateIntToPtr(ShadowPtr, ShadowPtrTy));
+      IRB.CreateLoad(ShadowTy, IRB.CreateIntToPtr(ShadowPtr, ShadowPtrTy));
 
   Value *Cmp = IRB.CreateICmpNE(ShadowValue, CmpVal);
   size_t Granularity = 1ULL << Mapping.Scale;
@@ -2444,7 +2445,7 @@ void AddressSanitizer::maybeInsertDynamicShadowAtFunctionEntry(Function &F) {
   } else {
     Value *GlobalDynamicAddress = F.getParent()->getOrInsertGlobal(
         kAsanShadowMemoryDynamicAddress, IntptrTy);
-    LocalDynamicShadow = IRB.CreateLoad(GlobalDynamicAddress);
+    LocalDynamicShadow = IRB.CreateLoad(IntptrTy, GlobalDynamicAddress);
   }
 }
 
@@ -2948,9 +2949,9 @@ void FunctionStackPoisoner::processStaticAllocas() {
     // void *LocalStackBase = (FakeStack) ? FakeStack : alloca(LocalStackSize);
     Constant *OptionDetectUseAfterReturn = F.getParent()->getOrInsertGlobal(
         kAsanOptionDetectUseAfterReturn, IRB.getInt32Ty());
-    Value *UseAfterReturnIsEnabled =
+    Value *UseAfterReturnIsEnabled = IRB.CreateICmpNE(
-        IRB.CreateICmpNE(IRB.CreateLoad(OptionDetectUseAfterReturn),
+        IRB.CreateLoad(IRB.getInt32Ty(), OptionDetectUseAfterReturn),
-                         Constant::getNullValue(IRB.getInt32Ty()));
+        Constant::getNullValue(IRB.getInt32Ty()));
     Instruction *Term =
         SplitBlockAndInsertIfThen(UseAfterReturnIsEnabled, InsBefore, false);
     IRBuilder<> IRBIf(Term);
@@ -3084,7 +3085,7 @@ void FunctionStackPoisoner::processStaticAllocas() {
             FakeStack,
             ConstantInt::get(IntptrTy, ClassSize - ASan.LongSize / 8));
         Value *SavedFlagPtr = IRBPoison.CreateLoad(
-            IRBPoison.CreateIntToPtr(SavedFlagPtrPtr, IntptrPtrTy));
+            IntptrTy, IRBPoison.CreateIntToPtr(SavedFlagPtrPtr, IntptrPtrTy));
         IRBPoison.CreateStore(
             Constant::getNullValue(IRBPoison.getInt8Ty()),
             IRBPoison.CreateIntToPtr(SavedFlagPtr, IRBPoison.getInt8PtrTy()));

llvm/lib/Transforms/Instrumentation/DataFlowSanitizer.cpp

@@ -1033,7 +1033,8 @@ Value *DFSanFunction::getShadow(Value *V) {
             DFS.ArgTLS ? &*F->getEntryBlock().begin()
                        : cast<Instruction>(ArgTLSPtr)->getNextNode();
         IRBuilder<> IRB(ArgTLSPos);
-        Shadow = IRB.CreateLoad(getArgTLS(A->getArgNo(), ArgTLSPos));
+        Shadow =
+            IRB.CreateLoad(DFS.ShadowTy, getArgTLS(A->getArgNo(), ArgTLSPos));
         break;
       }
       case DataFlowSanitizer::IA_Args: {
@@ -1183,7 +1184,7 @@ Value *DFSanFunction::loadShadow(Value *Addr, uint64_t Size, uint64_t Align,
     const auto i = AllocaShadowMap.find(AI);
     if (i != AllocaShadowMap.end()) {
       IRBuilder<> IRB(Pos);
-      return IRB.CreateLoad(i->second);
+      return IRB.CreateLoad(DFS.ShadowTy, i->second);
     }
   }
 
@@ -1208,7 +1209,7 @@ Value *DFSanFunction::loadShadow(Value *Addr, uint64_t Size, uint64_t Align,
   case 0:
     return DFS.ZeroShadow;
   case 1: {
-    LoadInst *LI = new LoadInst(ShadowAddr, "", Pos);
+    LoadInst *LI = new LoadInst(DFS.ShadowTy, ShadowAddr, "", Pos);
     LI->setAlignment(ShadowAlign);
     return LI;
   }
@@ -1216,8 +1217,9 @@ Value *DFSanFunction::loadShadow(Value *Addr, uint64_t Size, uint64_t Align,
     IRBuilder<> IRB(Pos);
     Value *ShadowAddr1 = IRB.CreateGEP(DFS.ShadowTy, ShadowAddr,
                                        ConstantInt::get(DFS.IntptrTy, 1));
-    return combineShadows(IRB.CreateAlignedLoad(ShadowAddr, ShadowAlign),
+    return combineShadows(
-                          IRB.CreateAlignedLoad(ShadowAddr1, ShadowAlign), Pos);
+        IRB.CreateAlignedLoad(DFS.ShadowTy, ShadowAddr, ShadowAlign),
+        IRB.CreateAlignedLoad(DFS.ShadowTy, ShadowAddr1, ShadowAlign), Pos);
   }
   }
   if (!AvoidNewBlocks && Size % (64 / DFS.ShadowWidth) == 0) {
@@ -1236,7 +1238,8 @@ Value *DFSanFunction::loadShadow(Value *Addr, uint64_t Size, uint64_t Align,
     IRBuilder<> IRB(Pos);
     Value *WideAddr =
         IRB.CreateBitCast(ShadowAddr, Type::getInt64PtrTy(*DFS.Ctx));
-    Value *WideShadow = IRB.CreateAlignedLoad(WideAddr, ShadowAlign);
+    Value *WideShadow =
+        IRB.CreateAlignedLoad(IRB.getInt64Ty(), WideAddr, ShadowAlign);
     Value *TruncShadow = IRB.CreateTrunc(WideShadow, DFS.ShadowTy);
     Value *ShlShadow = IRB.CreateShl(WideShadow, DFS.ShadowWidth);
     Value *ShrShadow = IRB.CreateLShr(WideShadow, 64 - DFS.ShadowWidth);
@@ -1269,7 +1272,8 @@ Value *DFSanFunction::loadShadow(Value *Addr, uint64_t Size, uint64_t Align,
       IRBuilder<> NextIRB(NextBB);
       WideAddr = NextIRB.CreateGEP(Type::getInt64Ty(*DFS.Ctx), WideAddr,
                                    ConstantInt::get(DFS.IntptrTy, 1));
-      Value *NextWideShadow = NextIRB.CreateAlignedLoad(WideAddr, ShadowAlign);
+      Value *NextWideShadow = NextIRB.CreateAlignedLoad(NextIRB.getInt64Ty(),
+                                                        WideAddr, ShadowAlign);
       ShadowsEq = NextIRB.CreateICmpEQ(WideShadow, NextWideShadow);
       LastBr->setSuccessor(0, NextBB);
       LastBr = NextIRB.CreateCondBr(ShadowsEq, FallbackBB, FallbackBB);
@@ -1646,7 +1650,8 @@ void DFSanVisitor::visitCallSite(CallSite CS) {
         }
 
         if (!FT->getReturnType()->isVoidTy()) {
-          LoadInst *LabelLoad = IRB.CreateLoad(DFSF.LabelReturnAlloca);
+          LoadInst *LabelLoad =
+              IRB.CreateLoad(DFSF.DFS.ShadowTy, DFSF.LabelReturnAlloca);
           DFSF.setShadow(CustomCI, LabelLoad);
         }
 
@@ -1684,7 +1689,7 @@ void DFSanVisitor::visitCallSite(CallSite CS) {
 
     if (DFSF.DFS.getInstrumentedABI() == DataFlowSanitizer::IA_TLS) {
       IRBuilder<> NextIRB(Next);
-      LoadInst *LI = NextIRB.CreateLoad(DFSF.getRetvalTLS());
+      LoadInst *LI = NextIRB.CreateLoad(DFSF.DFS.ShadowTy, DFSF.getRetvalTLS());
       DFSF.SkipInsts.insert(LI);
       DFSF.setShadow(CS.getInstruction(), LI);
       DFSF.NonZeroChecks.push_back(LI);

llvm/lib/Transforms/Instrumentation/EfficiencySanitizer.cpp

@@ -790,7 +790,7 @@ bool EfficiencySanitizer::insertCounterUpdate(Instruction *I,
     ConstantExpr::getGetElementPtr(
         ArrayType::get(IRB.getInt64Ty(), getStructCounterSize(StructTy)),
         CounterArray, Indices);
-  Value *Load = IRB.CreateLoad(Counter);
+  Value *Load = IRB.CreateLoad(IRB.getInt64Ty(), Counter);
   IRB.CreateStore(IRB.CreateAdd(Load, ConstantInt::get(IRB.getInt64Ty(), 1)),
                   Counter);
   return true;
@@ -875,7 +875,8 @@ bool EfficiencySanitizer::instrumentFastpathWorkingSet(
   // memory access, if they are not already set.
   Value *ValueMask = ConstantInt::get(ShadowTy, 0x81); // 10000001B
 
-  Value *OldValue = IRB.CreateLoad(IRB.CreateIntToPtr(ShadowPtr, ShadowPtrTy));
+  Value *OldValue =
+      IRB.CreateLoad(ShadowTy, IRB.CreateIntToPtr(ShadowPtr, ShadowPtrTy));
   // The AND and CMP will be turned into a TEST instruction by the compiler.
   Value *Cmp = IRB.CreateICmpNE(IRB.CreateAnd(OldValue, ValueMask), ValueMask);
   Instruction *CmpTerm = SplitBlockAndInsertIfThen(Cmp, I, false);

llvm/lib/Transforms/Instrumentation/GCOVProfiling.cpp

@@ -817,7 +817,7 @@ bool GCOVProfiler::emitProfileArcs() {
 
           // Skip phis, landingpads.
           IRBuilder<> Builder(&*BB.getFirstInsertionPt());
-          Value *Count = Builder.CreateLoad(Phi);
+          Value *Count = Builder.CreateLoad(Builder.getInt64Ty(), Phi);
           Count = Builder.CreateAdd(Count, Builder.getInt64(1));
           Builder.CreateStore(Count, Phi);
 
@@ -828,7 +828,7 @@ bool GCOVProfiler::emitProfileArcs() {
             const unsigned Edge = It->second;
             Value *Counter =
                 Builder.CreateConstInBoundsGEP2_64(Counters, 0, Edge);
-            Value *Count = Builder.CreateLoad(Counter);
+            Value *Count = Builder.CreateLoad(Builder.getInt64Ty(), Counter);
             Count = Builder.CreateAdd(Count, Builder.getInt64(1));
             Builder.CreateStore(Count, Counter);
           }
@@ -1089,17 +1089,20 @@ Function *GCOVProfiler::insertCounterWriteout(
   auto *StartFileCallArgsPtr = Builder.CreateStructGEP(FileInfoPtr, 0);
   auto *StartFileCall = Builder.CreateCall(
       StartFile,
-      {Builder.CreateLoad(Builder.CreateStructGEP(StartFileCallArgsPtr, 0)),
+      {Builder.CreateLoad(StartFileCallArgsTy->getElementType(0),
-       Builder.CreateLoad(Builder.CreateStructGEP(StartFileCallArgsPtr, 1)),
+                          Builder.CreateStructGEP(StartFileCallArgsPtr, 0)),
-       Builder.CreateLoad(Builder.CreateStructGEP(StartFileCallArgsPtr, 2))});
+       Builder.CreateLoad(StartFileCallArgsTy->getElementType(1),
+                          Builder.CreateStructGEP(StartFileCallArgsPtr, 1)),
+       Builder.CreateLoad(StartFileCallArgsTy->getElementType(2),
+                          Builder.CreateStructGEP(StartFileCallArgsPtr, 2))});
   if (auto AK = TLI->getExtAttrForI32Param(false))
     StartFileCall->addParamAttr(2, AK);
-  auto *NumCounters =
+  auto *NumCounters = Builder.CreateLoad(
-      Builder.CreateLoad(Builder.CreateStructGEP(FileInfoPtr, 1));
+      FileInfoTy->getElementType(1), Builder.CreateStructGEP(FileInfoPtr, 1));
-  auto *EmitFunctionCallArgsArray =
+  auto *EmitFunctionCallArgsArray = Builder.CreateLoad(
-      Builder.CreateLoad(Builder.CreateStructGEP(FileInfoPtr, 2));
+      FileInfoTy->getElementType(2), Builder.CreateStructGEP(FileInfoPtr, 2));
-  auto *EmitArcsCallArgsArray =
+  auto *EmitArcsCallArgsArray = Builder.CreateLoad(
-      Builder.CreateLoad(Builder.CreateStructGEP(FileInfoPtr, 3));
+      FileInfoTy->getElementType(3), Builder.CreateStructGEP(FileInfoPtr, 3));
   auto *EnterCounterLoopCond =
       Builder.CreateICmpSLT(Builder.getInt32(0), NumCounters);
   Builder.CreateCondBr(EnterCounterLoopCond, CounterLoopHeader, FileLoopLatch);
@@ -1111,11 +1114,16 @@ Function *GCOVProfiler::insertCounterWriteout(
       Builder.CreateInBoundsGEP(EmitFunctionCallArgsArray, {JV});
   auto *EmitFunctionCall = Builder.CreateCall(
       EmitFunction,
-      {Builder.CreateLoad(Builder.CreateStructGEP(EmitFunctionCallArgsPtr, 0)),
+      {Builder.CreateLoad(EmitFunctionCallArgsTy->getElementType(0),
-       Builder.CreateLoad(Builder.CreateStructGEP(EmitFunctionCallArgsPtr, 1)),
+                          Builder.CreateStructGEP(EmitFunctionCallArgsPtr, 0)),
-       Builder.CreateLoad(Builder.CreateStructGEP(EmitFunctionCallArgsPtr, 2)),
+       Builder.CreateLoad(EmitFunctionCallArgsTy->getElementType(1),
-       Builder.CreateLoad(Builder.CreateStructGEP(EmitFunctionCallArgsPtr, 3)),
+                          Builder.CreateStructGEP(EmitFunctionCallArgsPtr, 1)),
+       Builder.CreateLoad(EmitFunctionCallArgsTy->getElementType(2),
+                          Builder.CreateStructGEP(EmitFunctionCallArgsPtr, 2)),
+       Builder.CreateLoad(EmitFunctionCallArgsTy->getElementType(3),
+                          Builder.CreateStructGEP(EmitFunctionCallArgsPtr, 3)),
        Builder.CreateLoad(
+           EmitFunctionCallArgsTy->getElementType(4),
            Builder.CreateStructGEP(EmitFunctionCallArgsPtr, 4))});
   if (auto AK = TLI->getExtAttrForI32Param(false)) {
     EmitFunctionCall->addParamAttr(0, AK);
@@ -1127,8 +1135,10 @@ Function *GCOVProfiler::insertCounterWriteout(
       Builder.CreateInBoundsGEP(EmitArcsCallArgsArray, {JV});
   auto *EmitArcsCall = Builder.CreateCall(
       EmitArcs,
-      {Builder.CreateLoad(Builder.CreateStructGEP(EmitArcsCallArgsPtr, 0)),
+      {Builder.CreateLoad(EmitArcsCallArgsTy->getElementType(0),
-       Builder.CreateLoad(Builder.CreateStructGEP(EmitArcsCallArgsPtr, 1))});
+                          Builder.CreateStructGEP(EmitArcsCallArgsPtr, 0)),
+       Builder.CreateLoad(EmitArcsCallArgsTy->getElementType(1),
+                          Builder.CreateStructGEP(EmitArcsCallArgsPtr, 1))});
   if (auto AK = TLI->getExtAttrForI32Param(false))
     EmitArcsCall->addParamAttr(0, AK);
   auto *NextJV = Builder.CreateAdd(JV, Builder.getInt32(1));

llvm/lib/Transforms/Instrumentation/HWAddressSanitizer.cpp

@@ -428,7 +428,7 @@ Value *HWAddressSanitizer::getDynamicShadowNonTls(IRBuilder<> &IRB) {
     Value *GlobalDynamicAddress =
         IRB.GetInsertBlock()->getParent()->getParent()->getOrInsertGlobal(
             kHwasanShadowMemoryDynamicAddress, Int8PtrTy);
-    return IRB.CreateLoad(GlobalDynamicAddress);
+    return IRB.CreateLoad(Int8PtrTy, GlobalDynamicAddress);
   }
 }
 
@@ -557,7 +557,7 @@ void HWAddressSanitizer::instrumentMemAccessInline(Value *Ptr, bool IsWrite,
                                   IRB.getInt8Ty());
   Value *AddrLong = untagPointer(IRB, PtrLong);
   Value *Shadow = memToShadow(AddrLong, IRB);
-  Value *MemTag = IRB.CreateLoad(Shadow);
+  Value *MemTag = IRB.CreateLoad(Int8Ty, Shadow);
   Value *TagMismatch = IRB.CreateICmpNE(PtrTag, MemTag);
 
   int matchAllTag = ClMatchAllTag.getNumOccurrences() > 0 ?
@@ -841,7 +841,7 @@ Value *HWAddressSanitizer::emitPrologue(IRBuilder<> &IRB,
   Value *SlotPtr = getHwasanThreadSlotPtr(IRB, IntptrTy);
   assert(SlotPtr);
 
-  Instruction *ThreadLong = IRB.CreateLoad(SlotPtr);
+  Instruction *ThreadLong = IRB.CreateLoad(IntptrTy, SlotPtr);
 
   Function *F = IRB.GetInsertBlock()->getParent();
   if (F->getFnAttribute("hwasan-abi").getValueAsString() == "interceptor") {
@@ -855,7 +855,7 @@ Value *HWAddressSanitizer::emitPrologue(IRBuilder<> &IRB,
     // FIXME: This should call a new runtime function with a custom calling
     // convention to avoid needing to spill all arguments here.
     IRB.CreateCall(HwasanThreadEnterFunc);
-    LoadInst *ReloadThreadLong = IRB.CreateLoad(SlotPtr);
+    LoadInst *ReloadThreadLong = IRB.CreateLoad(IntptrTy, SlotPtr);
 
     IRB.SetInsertPoint(&*Br->getSuccessor(0)->begin());
     PHINode *ThreadLongPhi = IRB.CreatePHI(IntptrTy, 2);

llvm/lib/Transforms/Instrumentation/InstrProfiling.cpp

@@ -195,6 +195,7 @@ public:
       // block.
       Value *LiveInValue = SSA.GetValueInMiddleOfBlock(ExitBlock);
       Value *Addr = cast<StoreInst>(Store)->getPointerOperand();
+      Type *Ty = LiveInValue->getType();
       IRBuilder<> Builder(InsertPos);
       if (AtomicCounterUpdatePromoted)
         // automic update currently can only be promoted across the current
@@ -202,7 +203,7 @@ public:
         Builder.CreateAtomicRMW(AtomicRMWInst::Add, Addr, LiveInValue,
                                 AtomicOrdering::SequentiallyConsistent);
       else {
-        LoadInst *OldVal = Builder.CreateLoad(Addr, "pgocount.promoted");
+        LoadInst *OldVal = Builder.CreateLoad(Ty, Addr, "pgocount.promoted");
         auto *NewVal = Builder.CreateAdd(OldVal, LiveInValue);
         auto *NewStore = Builder.CreateStore(NewVal, Addr);
 
@@ -603,7 +604,8 @@ void InstrProfiling::lowerIncrement(InstrProfIncrementInst *Inc) {
     Builder.CreateAtomicRMW(AtomicRMWInst::Add, Addr, Inc->getStep(),
                             AtomicOrdering::Monotonic);
   } else {
-    Value *Load = Builder.CreateLoad(Addr, "pgocount");
+    Value *IncStep = Inc->getStep();
+    Value *Load = Builder.CreateLoad(IncStep->getType(), Addr, "pgocount");
     auto *Count = Builder.CreateAdd(Load, Inc->getStep());
     auto *Store = Builder.CreateStore(Count, Addr);
     if (isCounterPromotionEnabled())
@@ -950,7 +952,7 @@ bool InstrProfiling::emitRuntimeHook() {
     User->setComdat(M->getOrInsertComdat(User->getName()));
 
   IRBuilder<> IRB(BasicBlock::Create(M->getContext(), "", User));
-  auto *Load = IRB.CreateLoad(Var);
+  auto *Load = IRB.CreateLoad(Int32Ty, Var);
   IRB.CreateRet(Load);
 
   // Mark the user variable as used so that it isn't stripped out.

llvm/lib/Transforms/Instrumentation/MemorySanitizer.cpp

@@ -1401,7 +1401,7 @@ struct MemorySanitizerVisitor : public InstVisitor<MemorySanitizerVisitor> {
             IRB.CreateAnd(OriginLong, ConstantInt::get(MS.IntptrTy, ~Mask));
       }
       OriginPtr =
-          IRB.CreateIntToPtr(OriginLong, PointerType::get(IRB.getInt32Ty(), 0));
+          IRB.CreateIntToPtr(OriginLong, PointerType::get(MS.OriginTy, 0));
     }
     return std::make_pair(ShadowPtr, OriginPtr);
   }
@@ -1618,8 +1618,8 @@ struct MemorySanitizerVisitor : public InstVisitor<MemorySanitizerVisitor> {
               // ParamTLS overflow.
               *ShadowPtr = getCleanShadow(V);
             } else {
-              *ShadowPtr =
+              *ShadowPtr = EntryIRB.CreateAlignedLoad(getShadowTy(&FArg), Base,
-                  EntryIRB.CreateAlignedLoad(Base, kShadowTLSAlignment);
+                                                      kShadowTLSAlignment);
             }
           }
           LLVM_DEBUG(dbgs()
@@ -1627,7 +1627,7 @@ struct MemorySanitizerVisitor : public InstVisitor<MemorySanitizerVisitor> {
           if (MS.TrackOrigins && !Overflow) {
             Value *OriginPtr =
                 getOriginPtrForArgument(&FArg, EntryIRB, ArgOffset);
-            setOrigin(A, EntryIRB.CreateLoad(OriginPtr));
+            setOrigin(A, EntryIRB.CreateLoad(MS.OriginTy, OriginPtr));
           } else {
             setOrigin(A, getCleanOrigin());
           }
@@ -1758,7 +1758,8 @@ struct MemorySanitizerVisitor : public InstVisitor<MemorySanitizerVisitor> {
     if (PropagateShadow) {
       std::tie(ShadowPtr, OriginPtr) =
           getShadowOriginPtr(Addr, IRB, ShadowTy, Alignment, /*isStore*/ false);
-      setShadow(&I, IRB.CreateAlignedLoad(ShadowPtr, Alignment, "_msld"));
+      setShadow(&I,
+                IRB.CreateAlignedLoad(ShadowTy, ShadowPtr, Alignment, "_msld"));
     } else {
       setShadow(&I, getCleanShadow(&I));
     }
@@ -1772,7 +1773,8 @@ struct MemorySanitizerVisitor : public InstVisitor<MemorySanitizerVisitor> {
     if (MS.TrackOrigins) {
       if (PropagateShadow) {
         unsigned OriginAlignment = std::max(kMinOriginAlignment, Alignment);
-        setOrigin(&I, IRB.CreateAlignedLoad(OriginPtr, OriginAlignment));
+        setOrigin(
+            &I, IRB.CreateAlignedLoad(MS.OriginTy, OriginPtr, OriginAlignment));
       } else {
         setOrigin(&I, getCleanOrigin());
       }
@@ -2452,7 +2454,8 @@ struct MemorySanitizerVisitor : public InstVisitor<MemorySanitizerVisitor> {
       unsigned Alignment = 1;
       std::tie(ShadowPtr, OriginPtr) =
           getShadowOriginPtr(Addr, IRB, ShadowTy, Alignment, /*isStore*/ false);
-      setShadow(&I, IRB.CreateAlignedLoad(ShadowPtr, Alignment, "_msld"));
+      setShadow(&I,
+                IRB.CreateAlignedLoad(ShadowTy, ShadowPtr, Alignment, "_msld"));
     } else {
       setShadow(&I, getCleanShadow(&I));
     }
@@ -2462,7 +2465,7 @@ struct MemorySanitizerVisitor : public InstVisitor<MemorySanitizerVisitor> {
 
     if (MS.TrackOrigins) {
       if (PropagateShadow)
-        setOrigin(&I, IRB.CreateLoad(OriginPtr));
+        setOrigin(&I, IRB.CreateLoad(MS.OriginTy, OriginPtr));
       else
         setOrigin(&I, getCleanOrigin());
     }
@@ -2845,9 +2848,9 @@ struct MemorySanitizerVisitor : public InstVisitor<MemorySanitizerVisitor> {
     if (ClCheckAccessAddress)
       insertShadowCheck(Addr, &I);
 
-    Value *Shadow = IRB.CreateAlignedLoad(ShadowPtr, Alignment, "_ldmxcsr");
+    Value *Shadow = IRB.CreateAlignedLoad(Ty, ShadowPtr, Alignment, "_ldmxcsr");
-    Value *Origin =
+    Value *Origin = MS.TrackOrigins ? IRB.CreateLoad(MS.OriginTy, OriginPtr)
-        MS.TrackOrigins ? IRB.CreateLoad(OriginPtr) : getCleanOrigin();
+                                    : getCleanOrigin();
     insertShadowCheck(Shadow, Origin, &I);
   }
 
@@ -2921,7 +2924,7 @@ struct MemorySanitizerVisitor : public InstVisitor<MemorySanitizerVisitor> {
 
         Value *Origin = IRB.CreateSelect(
             IRB.CreateICmpNE(Acc, Constant::getNullValue(Acc->getType())),
-            getOrigin(PassThru), IRB.CreateLoad(OriginPtr));
+            getOrigin(PassThru), IRB.CreateLoad(MS.OriginTy, OriginPtr));
 
         setOrigin(&I, Origin);
       } else {
@@ -3284,12 +3287,13 @@ struct MemorySanitizerVisitor : public InstVisitor<MemorySanitizerVisitor> {
              "Could not find insertion point for retval shadow load");
     }
     IRBuilder<> IRBAfter(&*NextInsn);
-    Value *RetvalShadow =
+    Value *RetvalShadow = IRBAfter.CreateAlignedLoad(
-      IRBAfter.CreateAlignedLoad(getShadowPtrForRetval(&I, IRBAfter),
+        getShadowTy(&I), getShadowPtrForRetval(&I, IRBAfter),
-                                 kShadowTLSAlignment, "_msret");
+        kShadowTLSAlignment, "_msret");
     setShadow(&I, RetvalShadow);
     if (MS.TrackOrigins)
-      setOrigin(&I, IRBAfter.CreateLoad(getOriginPtrForRetval(IRBAfter)));
+      setOrigin(&I, IRBAfter.CreateLoad(MS.OriginTy,
+                                        getOriginPtrForRetval(IRBAfter)));
   }
 
   bool isAMustTailRetVal(Value *RetVal) {
@@ -3837,7 +3841,8 @@ struct VarArgAMD64Helper : public VarArgHelper {
       // If there is a va_start in this function, make a backup copy of
       // va_arg_tls somewhere in the function entry block.
       IRBuilder<> IRB(MSV.ActualFnStart->getFirstNonPHI());
-      VAArgOverflowSize = IRB.CreateLoad(MS.VAArgOverflowSizeTLS);
+      VAArgOverflowSize =
+          IRB.CreateLoad(IRB.getInt64Ty(), MS.VAArgOverflowSizeTLS);
       Value *CopySize =
         IRB.CreateAdd(ConstantInt::get(MS.IntptrTy, AMD64FpEndOffset),
                       VAArgOverflowSize);
@@ -3856,11 +3861,13 @@ struct VarArgAMD64Helper : public VarArgHelper {
       IRBuilder<> IRB(OrigInst->getNextNode());
       Value *VAListTag = OrigInst->getArgOperand(0);
 
+      Type *RegSaveAreaPtrTy = Type::getInt64PtrTy(*MS.C);
       Value *RegSaveAreaPtrPtr = IRB.CreateIntToPtr(
           IRB.CreateAdd(IRB.CreatePtrToInt(VAListTag, MS.IntptrTy),
                         ConstantInt::get(MS.IntptrTy, 16)),
-          PointerType::get(Type::getInt64PtrTy(*MS.C), 0));
+          PointerType::get(RegSaveAreaPtrTy, 0));
-      Value *RegSaveAreaPtr = IRB.CreateLoad(RegSaveAreaPtrPtr);
+      Value *RegSaveAreaPtr =
+          IRB.CreateLoad(RegSaveAreaPtrTy, RegSaveAreaPtrPtr);
       Value *RegSaveAreaShadowPtr, *RegSaveAreaOriginPtr;
       unsigned Alignment = 16;
       std::tie(RegSaveAreaShadowPtr, RegSaveAreaOriginPtr) =
@@ -3871,11 +3878,13 @@ struct VarArgAMD64Helper : public VarArgHelper {
       if (MS.TrackOrigins)
         IRB.CreateMemCpy(RegSaveAreaOriginPtr, Alignment, VAArgTLSOriginCopy,
                          Alignment, AMD64FpEndOffset);
+      Type *OverflowArgAreaPtrTy = Type::getInt64PtrTy(*MS.C);
       Value *OverflowArgAreaPtrPtr = IRB.CreateIntToPtr(
           IRB.CreateAdd(IRB.CreatePtrToInt(VAListTag, MS.IntptrTy),
                         ConstantInt::get(MS.IntptrTy, 8)),
-          PointerType::get(Type::getInt64PtrTy(*MS.C), 0));
+          PointerType::get(OverflowArgAreaPtrTy, 0));
-      Value *OverflowArgAreaPtr = IRB.CreateLoad(OverflowArgAreaPtrPtr);
+      Value *OverflowArgAreaPtr =
+          IRB.CreateLoad(OverflowArgAreaPtrTy, OverflowArgAreaPtrPtr);
       Value *OverflowArgAreaShadowPtr, *OverflowArgAreaOriginPtr;
       std::tie(OverflowArgAreaShadowPtr, OverflowArgAreaOriginPtr) =
           MSV.getShadowOriginPtr(OverflowArgAreaPtr, IRB, IRB.getInt8Ty(),
@@ -3977,7 +3986,7 @@ struct VarArgMIPS64Helper : public VarArgHelper {
     assert(!VAArgSize && !VAArgTLSCopy &&
            "finalizeInstrumentation called twice");
     IRBuilder<> IRB(MSV.ActualFnStart->getFirstNonPHI());
-    VAArgSize = IRB.CreateLoad(MS.VAArgOverflowSizeTLS);
+    VAArgSize = IRB.CreateLoad(IRB.getInt64Ty(), MS.VAArgOverflowSizeTLS);
     Value *CopySize = IRB.CreateAdd(ConstantInt::get(MS.IntptrTy, 0),
                                     VAArgSize);
 
@@ -3994,10 +4003,12 @@ struct VarArgMIPS64Helper : public VarArgHelper {
       CallInst *OrigInst = VAStartInstrumentationList[i];
       IRBuilder<> IRB(OrigInst->getNextNode());
       Value *VAListTag = OrigInst->getArgOperand(0);
+      Type *RegSaveAreaPtrTy = Type::getInt64PtrTy(*MS.C);
       Value *RegSaveAreaPtrPtr =
           IRB.CreateIntToPtr(IRB.CreatePtrToInt(VAListTag, MS.IntptrTy),
-                             PointerType::get(Type::getInt64PtrTy(*MS.C), 0));
+                             PointerType::get(RegSaveAreaPtrTy, 0));
-      Value *RegSaveAreaPtr = IRB.CreateLoad(RegSaveAreaPtrPtr);
+      Value *RegSaveAreaPtr =
+          IRB.CreateLoad(RegSaveAreaPtrTy, RegSaveAreaPtrPtr);
       Value *RegSaveAreaShadowPtr, *RegSaveAreaOriginPtr;
       unsigned Alignment = 8;
       std::tie(RegSaveAreaShadowPtr, RegSaveAreaOriginPtr) =
@@ -4147,7 +4158,7 @@ struct VarArgAArch64Helper : public VarArgHelper {
         IRB.CreateAdd(IRB.CreatePtrToInt(VAListTag, MS.IntptrTy),
                       ConstantInt::get(MS.IntptrTy, offset)),
         Type::getInt64PtrTy(*MS.C));
-    return IRB.CreateLoad(SaveAreaPtrPtr);
+    return IRB.CreateLoad(Type::getInt64Ty(*MS.C), SaveAreaPtrPtr);
   }
 
   // Retrieve a va_list field of 'int' size.
@@ -4157,7 +4168,7 @@ struct VarArgAArch64Helper : public VarArgHelper {
         IRB.CreateAdd(IRB.CreatePtrToInt(VAListTag, MS.IntptrTy),
                       ConstantInt::get(MS.IntptrTy, offset)),
         Type::getInt32PtrTy(*MS.C));
-    Value *SaveArea32 = IRB.CreateLoad(SaveAreaPtr);
+    Value *SaveArea32 = IRB.CreateLoad(IRB.getInt32Ty(), SaveAreaPtr);
     return IRB.CreateSExt(SaveArea32, MS.IntptrTy);
   }
 
@@ -4168,7 +4179,8 @@ struct VarArgAArch64Helper : public VarArgHelper {
       // If there is a va_start in this function, make a backup copy of
       // va_arg_tls somewhere in the function entry block.
       IRBuilder<> IRB(MSV.ActualFnStart->getFirstNonPHI());
-      VAArgOverflowSize = IRB.CreateLoad(MS.VAArgOverflowSizeTLS);
+      VAArgOverflowSize =
+          IRB.CreateLoad(IRB.getInt64Ty(), MS.VAArgOverflowSizeTLS);
       Value *CopySize =
         IRB.CreateAdd(ConstantInt::get(MS.IntptrTy, AArch64VAEndOffset),
                       VAArgOverflowSize);
@@ -4411,7 +4423,7 @@ struct VarArgPowerPC64Helper : public VarArgHelper {
     assert(!VAArgSize && !VAArgTLSCopy &&
            "finalizeInstrumentation called twice");
     IRBuilder<> IRB(MSV.ActualFnStart->getFirstNonPHI());
-    VAArgSize = IRB.CreateLoad(MS.VAArgOverflowSizeTLS);
+    VAArgSize = IRB.CreateLoad(IRB.getInt64Ty(), MS.VAArgOverflowSizeTLS);
     Value *CopySize = IRB.CreateAdd(ConstantInt::get(MS.IntptrTy, 0),
                                     VAArgSize);
 
@@ -4428,10 +4440,12 @@ struct VarArgPowerPC64Helper : public VarArgHelper {
       CallInst *OrigInst = VAStartInstrumentationList[i];
       IRBuilder<> IRB(OrigInst->getNextNode());
       Value *VAListTag = OrigInst->getArgOperand(0);
+      Type *RegSaveAreaPtrTy = Type::getInt64PtrTy(*MS.C);
       Value *RegSaveAreaPtrPtr =
           IRB.CreateIntToPtr(IRB.CreatePtrToInt(VAListTag, MS.IntptrTy),
-                             PointerType::get(Type::getInt64PtrTy(*MS.C), 0));
+                             PointerType::get(RegSaveAreaPtrTy, 0));
-      Value *RegSaveAreaPtr = IRB.CreateLoad(RegSaveAreaPtrPtr);
+      Value *RegSaveAreaPtr =
+          IRB.CreateLoad(RegSaveAreaPtrTy, RegSaveAreaPtrPtr);
       Value *RegSaveAreaShadowPtr, *RegSaveAreaOriginPtr;
       unsigned Alignment = 8;
       std::tie(RegSaveAreaShadowPtr, RegSaveAreaOriginPtr) =

llvm/lib/Transforms/Instrumentation/SanitizerCoverage.cpp

@@ -822,7 +822,7 @@ void SanitizerCoverageModule::InjectCoverageAtBlock(Function &F, BasicBlock &BB,
     auto CounterPtr = IRB.CreateGEP(
         Function8bitCounterArray,
         {ConstantInt::get(IntptrTy, 0), ConstantInt::get(IntptrTy, Idx)});
-    auto Load = IRB.CreateLoad(CounterPtr);
+    auto Load = IRB.CreateLoad(Int8Ty, CounterPtr);
     auto Inc = IRB.CreateAdd(Load, ConstantInt::get(Int8Ty, 1));
     auto Store = IRB.CreateStore(Inc, CounterPtr);
     SetNoSanitizeMetadata(Load);
@@ -835,7 +835,7 @@ void SanitizerCoverageModule::InjectCoverageAtBlock(Function &F, BasicBlock &BB,
     auto FrameAddrPtr =
         IRB.CreateCall(GetFrameAddr, {Constant::getNullValue(Int32Ty)});
     auto FrameAddrInt = IRB.CreatePtrToInt(FrameAddrPtr, IntptrTy);
-    auto LowestStack = IRB.CreateLoad(SanCovLowestStack);
+    auto LowestStack = IRB.CreateLoad(IntptrTy, SanCovLowestStack);
     auto IsStackLower = IRB.CreateICmpULT(FrameAddrInt, LowestStack);
     auto ThenTerm = SplitBlockAndInsertIfThen(IsStackLower, &*IP, false);
     IRBuilder<> ThenIRB(ThenTerm);

llvm/lib/Transforms/Scalar/GVN.cpp

@@ -1234,10 +1234,10 @@ bool GVN::PerformLoadPRE(LoadInst *LI, AvailValInBlkVect &ValuesPerBlock,
     BasicBlock *UnavailablePred = PredLoad.first;
     Value *LoadPtr = PredLoad.second;
 
-    auto *NewLoad = new LoadInst(LoadPtr, LI->getName()+".pre",
+    auto *NewLoad =
-                                 LI->isVolatile(), LI->getAlignment(),
+        new LoadInst(LI->getType(), LoadPtr, LI->getName() + ".pre",
-                                 LI->getOrdering(), LI->getSyncScopeID(),
+                     LI->isVolatile(), LI->getAlignment(), LI->getOrdering(),
-                                 UnavailablePred->getTerminator());
+                     LI->getSyncScopeID(), UnavailablePred->getTerminator());
     NewLoad->setDebugLoc(LI->getDebugLoc());
 
     // Transfer the old load's AA tags to the new load.

llvm/lib/Transforms/Scalar/JumpThreading.cpp

@@ -1445,11 +1445,11 @@ bool JumpThreadingPass::SimplifyPartiallyRedundantLoad(LoadInst *LoadI) {
   if (UnavailablePred) {
     assert(UnavailablePred->getTerminator()->getNumSuccessors() == 1 &&
            "Can't handle critical edge here!");
-    LoadInst *NewVal =
+    LoadInst *NewVal = new LoadInst(
-        new LoadInst(LoadedPtr->DoPHITranslation(LoadBB, UnavailablePred),
+        LoadI->getType(), LoadedPtr->DoPHITranslation(LoadBB, UnavailablePred),
-                     LoadI->getName() + ".pr", false, LoadI->getAlignment(),
+        LoadI->getName() + ".pr", false, LoadI->getAlignment(),
-                     LoadI->getOrdering(), LoadI->getSyncScopeID(),
+        LoadI->getOrdering(), LoadI->getSyncScopeID(),
-                     UnavailablePred->getTerminator());
+        UnavailablePred->getTerminator());
     NewVal->setDebugLoc(LoadI->getDebugLoc());
     if (AATags)
       NewVal->setAAMetadata(AATags);

llvm/lib/Transforms/Scalar/LICM.cpp

@@ -1947,7 +1947,8 @@ bool llvm::promoteLoopAccessesToScalars(
   // Set up the preheader to have a definition of the value.  It is the live-out
   // value from the preheader that uses in the loop will use.
   LoadInst *PreheaderLoad = new LoadInst(
-      SomePtr, SomePtr->getName() + ".promoted", Preheader->getTerminator());
+      SomePtr->getType()->getPointerElementType(), SomePtr,
+      SomePtr->getName() + ".promoted", Preheader->getTerminator());
   if (SawUnorderedAtomic)
     PreheaderLoad->setOrdering(AtomicOrdering::Unordered);
   PreheaderLoad->setAlignment(Alignment);

llvm/lib/Transforms/Scalar/LoopLoadElimination.cpp

@@ -427,9 +427,9 @@ public:
     auto *PH = L->getLoopPreheader();
     Value *InitialPtr = SEE.expandCodeFor(PtrSCEV->getStart(), Ptr->getType(),
                                           PH->getTerminator());
-    Value *Initial =
+    Value *Initial = new LoadInst(
-        new LoadInst(InitialPtr, "load_initial", /* isVolatile */ false,
+        Cand.Load->getType(), InitialPtr, "load_initial",
-                     Cand.Load->getAlignment(), PH->getTerminator());
+        /* isVolatile */ false, Cand.Load->getAlignment(), PH->getTerminator());
 
     PHINode *PHI = PHINode::Create(Initial->getType(), 2, "store_forwarded",
                                    &L->getHeader()->front());

llvm/lib/Transforms/Scalar/LowerAtomic.cpp

@@ -26,7 +26,7 @@ static bool LowerAtomicCmpXchgInst(AtomicCmpXchgInst *CXI) {
   Value *Cmp = CXI->getCompareOperand();
   Value *Val = CXI->getNewValOperand();
 
-  LoadInst *Orig = Builder.CreateLoad(Ptr);
+  LoadInst *Orig = Builder.CreateLoad(Val->getType(), Ptr);
   Value *Equal = Builder.CreateICmpEQ(Orig, Cmp);
   Value *Res = Builder.CreateSelect(Equal, Val, Orig);
   Builder.CreateStore(Res, Ptr);
@@ -44,7 +44,7 @@ static bool LowerAtomicRMWInst(AtomicRMWInst *RMWI) {
   Value *Ptr = RMWI->getPointerOperand();
   Value *Val = RMWI->getValOperand();
 
-  LoadInst *Orig = Builder.CreateLoad(Ptr);
+  LoadInst *Orig = Builder.CreateLoad(Val->getType(), Ptr);
   Value *Res = nullptr;
 
   switch (RMWI->getOperation()) {

llvm/lib/Transforms/Scalar/RewriteStatepointsForGC.cpp

@@ -1636,7 +1636,7 @@ makeStatepointExplicit(DominatorTree &DT, CallSite CS,
 // for sanity checking.
 static void
 insertRelocationStores(iterator_range<Value::user_iterator> GCRelocs,
-                       DenseMap<Value *, Value *> &AllocaMap,
+                       DenseMap<Value *, AllocaInst *> &AllocaMap,
                        DenseSet<Value *> &VisitedLiveValues) {
   for (User *U : GCRelocs) {
     GCRelocateInst *Relocate = dyn_cast<GCRelocateInst>(U);
@@ -1671,7 +1671,7 @@ insertRelocationStores(iterator_range<Value::user_iterator> GCRelocs,
 // "insertRelocationStores" but works for rematerialized values.
 static void insertRematerializationStores(
     const RematerializedValueMapTy &RematerializedValues,
-    DenseMap<Value *, Value *> &AllocaMap,
+    DenseMap<Value *, AllocaInst *> &AllocaMap,
     DenseSet<Value *> &VisitedLiveValues) {
   for (auto RematerializedValuePair: RematerializedValues) {
     Instruction *RematerializedValue = RematerializedValuePair.first;
@@ -1704,7 +1704,7 @@ static void relocationViaAlloca(
 #endif
 
   // TODO-PERF: change data structures, reserve
-  DenseMap<Value *, Value *> AllocaMap;
+  DenseMap<Value *, AllocaInst *> AllocaMap;
   SmallVector<AllocaInst *, 200> PromotableAllocas;
   // Used later to chack that we have enough allocas to store all values
   std::size_t NumRematerializedValues = 0;
@@ -1774,7 +1774,7 @@ static void relocationViaAlloca(
       SmallVector<AllocaInst *, 64> ToClobber;
       for (auto Pair : AllocaMap) {
         Value *Def = Pair.first;
-        AllocaInst *Alloca = cast<AllocaInst>(Pair.second);
+        AllocaInst *Alloca = Pair.second;
 
         // This value was relocated
         if (VisitedLiveValues.count(Def)) {
@@ -1806,7 +1806,7 @@ static void relocationViaAlloca(
   // Update use with load allocas and add store for gc_relocated.
   for (auto Pair : AllocaMap) {
     Value *Def = Pair.first;
-    Value *Alloca = Pair.second;
+    AllocaInst *Alloca = Pair.second;
 
     // We pre-record the uses of allocas so that we dont have to worry about
     // later update that changes the user information..
@@ -1834,13 +1834,15 @@ static void relocationViaAlloca(
         PHINode *Phi = cast<PHINode>(Use);
         for (unsigned i = 0; i < Phi->getNumIncomingValues(); i++) {
           if (Def == Phi->getIncomingValue(i)) {
-            LoadInst *Load = new LoadInst(
+            LoadInst *Load =
-                Alloca, "", Phi->getIncomingBlock(i)->getTerminator());
+                new LoadInst(Alloca->getAllocatedType(), Alloca, "",
+                             Phi->getIncomingBlock(i)->getTerminator());
             Phi->setIncomingValue(i, Load);
           }
         }
       } else {
-        LoadInst *Load = new LoadInst(Alloca, "", Use);
+        LoadInst *Load =
+            new LoadInst(Alloca->getAllocatedType(), Alloca, "", Use);
         Use->replaceUsesOfWith(Def, Load);
       }
     }

llvm/lib/Transforms/Scalar/SROA.cpp

@@ -1231,15 +1231,14 @@ static bool isSafePHIToSpeculate(PHINode &PN) {
 static void speculatePHINodeLoads(PHINode &PN) {
   LLVM_DEBUG(dbgs() << "    original: " << PN << "\n");
 
-  Type *LoadTy = cast<PointerType>(PN.getType())->getElementType();
+  LoadInst *SomeLoad = cast<LoadInst>(PN.user_back());
+  Type *LoadTy = SomeLoad->getType();
   IRBuilderTy PHIBuilder(&PN);
   PHINode *NewPN = PHIBuilder.CreatePHI(LoadTy, PN.getNumIncomingValues(),
                                         PN.getName() + ".sroa.speculated");
 
   // Get the AA tags and alignment to use from one of the loads.  It doesn't
   // matter which one we get and if any differ.
-  LoadInst *SomeLoad = cast<LoadInst>(PN.user_back());
-
   AAMDNodes AATags;
   SomeLoad->getAAMetadata(AATags);
   unsigned Align = SomeLoad->getAlignment();
@@ -1270,7 +1269,8 @@ static void speculatePHINodeLoads(PHINode &PN) {
     IRBuilderTy PredBuilder(TI);
 
     LoadInst *Load = PredBuilder.CreateLoad(
-        InVal, (PN.getName() + ".sroa.speculate.load." + Pred->getName()));
+        LoadTy, InVal,
+        (PN.getName() + ".sroa.speculate.load." + Pred->getName()));
     ++NumLoadsSpeculated;
     Load->setAlignment(Align);
     if (AATags)
@@ -1330,10 +1330,10 @@ static void speculateSelectInstLoads(SelectInst &SI) {
     assert(LI->isSimple() && "We only speculate simple loads");
 
     IRB.SetInsertPoint(LI);
-    LoadInst *TL =
+    LoadInst *TL = IRB.CreateLoad(LI->getType(), TV,
-        IRB.CreateLoad(TV, LI->getName() + ".sroa.speculate.load.true");
+                                  LI->getName() + ".sroa.speculate.load.true");
-    LoadInst *FL =
+    LoadInst *FL = IRB.CreateLoad(LI->getType(), FV,
-        IRB.CreateLoad(FV, LI->getName() + ".sroa.speculate.load.false");
+                                  LI->getName() + ".sroa.speculate.load.false");
     NumLoadsSpeculated += 2;
 
     // Transfer alignment and AA info if present.
@@ -2410,14 +2410,16 @@ private:
     unsigned EndIndex = getIndex(NewEndOffset);
     assert(EndIndex > BeginIndex && "Empty vector!");
 
-    Value *V = IRB.CreateAlignedLoad(&NewAI, NewAI.getAlignment(), "load");
+    Value *V = IRB.CreateAlignedLoad(NewAI.getAllocatedType(), &NewAI,
+                                     NewAI.getAlignment(), "load");
     return extractVector(IRB, V, BeginIndex, EndIndex, "vec");
   }
 
   Value *rewriteIntegerLoad(LoadInst &LI) {
     assert(IntTy && "We cannot insert an integer to the alloca");
     assert(!LI.isVolatile());
-    Value *V = IRB.CreateAlignedLoad(&NewAI, NewAI.getAlignment(), "load");
+    Value *V = IRB.CreateAlignedLoad(NewAI.getAllocatedType(), &NewAI,
+                                     NewAI.getAlignment(), "load");
     V = convertValue(DL, IRB, V, IntTy);
     assert(NewBeginOffset >= NewAllocaBeginOffset && "Out of bounds offset");
     uint64_t Offset = NewBeginOffset - NewAllocaBeginOffset;
@@ -2461,7 +2463,8 @@ private:
                (canConvertValue(DL, NewAllocaTy, TargetTy) ||
                 (IsLoadPastEnd && NewAllocaTy->isIntegerTy() &&
                  TargetTy->isIntegerTy()))) {
-      LoadInst *NewLI = IRB.CreateAlignedLoad(&NewAI, NewAI.getAlignment(),
+      LoadInst *NewLI = IRB.CreateAlignedLoad(NewAI.getAllocatedType(), &NewAI,
+                                              NewAI.getAlignment(),
                                               LI.isVolatile(), LI.getName());
       if (AATags)
         NewLI->setAAMetadata(AATags);
@@ -2497,9 +2500,9 @@ private:
           }
     } else {
       Type *LTy = TargetTy->getPointerTo(AS);
-      LoadInst *NewLI = IRB.CreateAlignedLoad(getNewAllocaSlicePtr(IRB, LTy),
+      LoadInst *NewLI = IRB.CreateAlignedLoad(
-                                              getSliceAlign(TargetTy),
+          TargetTy, getNewAllocaSlicePtr(IRB, LTy), getSliceAlign(TargetTy),
-                                              LI.isVolatile(), LI.getName());
+          LI.isVolatile(), LI.getName());
       if (AATags)
         NewLI->setAAMetadata(AATags);
       if (LI.isVolatile())
@@ -2525,8 +2528,8 @@ private:
       // basis for the new value. This allows us to replace the uses of LI with
       // the computed value, and then replace the placeholder with LI, leaving
       // LI only used for this computation.
-      Value *Placeholder =
+      Value *Placeholder = new LoadInst(
-          new LoadInst(UndefValue::get(LI.getType()->getPointerTo(AS)));
+          LI.getType(), UndefValue::get(LI.getType()->getPointerTo(AS)));
       V = insertInteger(DL, IRB, Placeholder, V, NewBeginOffset - BeginOffset,
                         "insert");
       LI.replaceAllUsesWith(V);
@@ -2557,7 +2560,8 @@ private:
         V = convertValue(DL, IRB, V, SliceTy);
 
       // Mix in the existing elements.
-      Value *Old = IRB.CreateAlignedLoad(&NewAI, NewAI.getAlignment(), "load");
+      Value *Old = IRB.CreateAlignedLoad(NewAI.getAllocatedType(), &NewAI,
+                                         NewAI.getAlignment(), "load");
       V = insertVector(IRB, Old, V, BeginIndex, "vec");
     }
     StoreInst *Store = IRB.CreateAlignedStore(V, &NewAI, NewAI.getAlignment());
@@ -2573,8 +2577,8 @@ private:
     assert(IntTy && "We cannot extract an integer from the alloca");
     assert(!SI.isVolatile());
     if (DL.getTypeSizeInBits(V->getType()) != IntTy->getBitWidth()) {
-      Value *Old =
+      Value *Old = IRB.CreateAlignedLoad(NewAI.getAllocatedType(), &NewAI,
-          IRB.CreateAlignedLoad(&NewAI, NewAI.getAlignment(), "oldload");
+                                         NewAI.getAlignment(), "oldload");
       Old = convertValue(DL, IRB, Old, IntTy);
       assert(BeginOffset >= NewAllocaBeginOffset && "Out of bounds offset");
       uint64_t Offset = BeginOffset - NewAllocaBeginOffset;
@@ -2766,8 +2770,8 @@ private:
       if (NumElements > 1)
         Splat = getVectorSplat(Splat, NumElements);
 
-      Value *Old =
+      Value *Old = IRB.CreateAlignedLoad(NewAI.getAllocatedType(), &NewAI,
-          IRB.CreateAlignedLoad(&NewAI, NewAI.getAlignment(), "oldload");
+                                         NewAI.getAlignment(), "oldload");
       V = insertVector(IRB, Old, Splat, BeginIndex, "vec");
     } else if (IntTy) {
       // If this is a memset on an alloca where we can widen stores, insert the
@@ -2779,8 +2783,8 @@ private:
 
       if (IntTy && (BeginOffset != NewAllocaBeginOffset ||
                     EndOffset != NewAllocaBeginOffset)) {
-        Value *Old =
+        Value *Old = IRB.CreateAlignedLoad(NewAI.getAllocatedType(), &NewAI,
-            IRB.CreateAlignedLoad(&NewAI, NewAI.getAlignment(), "oldload");
+                                           NewAI.getAlignment(), "oldload");
         Old = convertValue(DL, IRB, Old, IntTy);
         uint64_t Offset = NewBeginOffset - NewAllocaBeginOffset;
         V = insertInteger(DL, IRB, Old, V, Offset, "insert");
@@ -2940,18 +2944,18 @@ private:
 
     // Reset the other pointer type to match the register type we're going to
     // use, but using the address space of the original other pointer.
+    Type *OtherTy;
     if (VecTy && !IsWholeAlloca) {
       if (NumElements == 1)
-        OtherPtrTy = VecTy->getElementType();
+        OtherTy = VecTy->getElementType();
       else
-        OtherPtrTy = VectorType::get(VecTy->getElementType(), NumElements);
+        OtherTy = VectorType::get(VecTy->getElementType(), NumElements);
-
-      OtherPtrTy = OtherPtrTy->getPointerTo(OtherAS);
     } else if (IntTy && !IsWholeAlloca) {
-      OtherPtrTy = SubIntTy->getPointerTo(OtherAS);
+      OtherTy = SubIntTy;
     } else {
-      OtherPtrTy = NewAllocaTy->getPointerTo(OtherAS);
+      OtherTy = NewAllocaTy;
     }
+    OtherPtrTy = OtherTy->getPointerTo(OtherAS);
 
     Value *SrcPtr = getAdjustedPtr(IRB, DL, OtherPtr, OtherOffset, OtherPtrTy,
                                    OtherPtr->getName() + ".");
@@ -2965,28 +2969,30 @@ private:
 
     Value *Src;
     if (VecTy && !IsWholeAlloca && !IsDest) {
-      Src = IRB.CreateAlignedLoad(&NewAI, NewAI.getAlignment(), "load");
+      Src = IRB.CreateAlignedLoad(NewAI.getAllocatedType(), &NewAI,
+                                  NewAI.getAlignment(), "load");
       Src = extractVector(IRB, Src, BeginIndex, EndIndex, "vec");
     } else if (IntTy && !IsWholeAlloca && !IsDest) {
-      Src = IRB.CreateAlignedLoad(&NewAI, NewAI.getAlignment(), "load");
+      Src = IRB.CreateAlignedLoad(NewAI.getAllocatedType(), &NewAI,
+                                  NewAI.getAlignment(), "load");
       Src = convertValue(DL, IRB, Src, IntTy);
       uint64_t Offset = NewBeginOffset - NewAllocaBeginOffset;
       Src = extractInteger(DL, IRB, Src, SubIntTy, Offset, "extract");
     } else {
-      LoadInst *Load = IRB.CreateAlignedLoad(SrcPtr, SrcAlign, II.isVolatile(),
+      LoadInst *Load = IRB.CreateAlignedLoad(OtherTy, SrcPtr, SrcAlign,
-                                             "copyload");
+                                             II.isVolatile(), "copyload");
       if (AATags)
         Load->setAAMetadata(AATags);
       Src = Load;
     }
 
     if (VecTy && !IsWholeAlloca && IsDest) {
-      Value *Old =
+      Value *Old = IRB.CreateAlignedLoad(NewAI.getAllocatedType(), &NewAI,
-          IRB.CreateAlignedLoad(&NewAI, NewAI.getAlignment(), "oldload");
+                                         NewAI.getAlignment(), "oldload");
       Src = insertVector(IRB, Old, Src, BeginIndex, "vec");
     } else if (IntTy && !IsWholeAlloca && IsDest) {
-      Value *Old =
+      Value *Old = IRB.CreateAlignedLoad(NewAI.getAllocatedType(), &NewAI,
-          IRB.CreateAlignedLoad(&NewAI, NewAI.getAlignment(), "oldload");
+                                         NewAI.getAlignment(), "oldload");
       Old = convertValue(DL, IRB, Old, IntTy);
       uint64_t Offset = NewBeginOffset - NewAllocaBeginOffset;
       Src = insertInteger(DL, IRB, Old, Src, Offset, "insert");
@@ -3293,7 +3299,7 @@ private:
       // Load the single value and insert it using the indices.
       Value *GEP =
           IRB.CreateInBoundsGEP(nullptr, Ptr, GEPIndices, Name + ".gep");
-      LoadInst *Load = IRB.CreateAlignedLoad(GEP, Align, Name + ".load");
+      LoadInst *Load = IRB.CreateAlignedLoad(Ty, GEP, Align, Name + ".load");
       if (AATags)
         Load->setAAMetadata(AATags);
       Agg = IRB.CreateInsertValue(Agg, Load, Indices, Name + ".insert");
@@ -3787,6 +3793,7 @@ bool SROA::presplitLoadsAndStores(AllocaInst &AI, AllocaSlices &AS) {
       auto AS = LI->getPointerAddressSpace();
       auto *PartPtrTy = PartTy->getPointerTo(AS);
       LoadInst *PLoad = IRB.CreateAlignedLoad(
+          PartTy,
           getAdjustedPtr(IRB, DL, BasePtr,
                          APInt(DL.getIndexSizeInBits(AS), PartOffset),
                          PartPtrTy, BasePtr->getName() + "."),
@@ -3928,6 +3935,7 @@ bool SROA::presplitLoadsAndStores(AllocaInst &AI, AllocaSlices &AS) {
         IRB.SetInsertPoint(LI);
         auto AS = LI->getPointerAddressSpace();
         PLoad = IRB.CreateAlignedLoad(
+            PartTy,
             getAdjustedPtr(IRB, DL, LoadBasePtr,
                            APInt(DL.getIndexSizeInBits(AS), PartOffset),
                            LoadPartPtrTy, LoadBasePtr->getName() + "."),

llvm/lib/Transforms/Scalar/Scalarizer.cpp

@@ -743,7 +743,8 @@ bool ScalarizerVisitor::visitLoadInst(LoadInst &LI) {
   Res.resize(NumElems);
 
   for (unsigned I = 0; I < NumElems; ++I)
-    Res[I] = Builder.CreateAlignedLoad(Ptr[I], Layout.getElemAlign(I),
+    Res[I] = Builder.CreateAlignedLoad(Layout.VecTy->getElementType(), Ptr[I],
+                                       Layout.getElemAlign(I),
                                        LI.getName() + ".i" + Twine(I));
   gather(&LI, Res);
   return true;

llvm/lib/Transforms/Utils/CodeExtractor.cpp

@@ -844,7 +844,8 @@ Function *CodeExtractor::constructFunction(const ValueSet &inputs,
       Instruction *TI = newFunction->begin()->getTerminator();
       GetElementPtrInst *GEP = GetElementPtrInst::Create(
           StructTy, &*AI, Idx, "gep_" + inputs[i]->getName(), TI);
-      RewriteVal = new LoadInst(GEP, "loadgep_" + inputs[i]->getName(), TI);
+      RewriteVal = new LoadInst(StructTy->getElementType(i), GEP,
+                                "loadgep_" + inputs[i]->getName(), TI);
     } else
       RewriteVal = &*AI++;
 
@@ -1054,7 +1055,8 @@ CallInst *CodeExtractor::emitCallAndSwitchStatement(Function *newFunction,
     } else {
       Output = ReloadOutputs[i];
     }
-    LoadInst *load = new LoadInst(Output, outputs[i]->getName()+".reload");
+    LoadInst *load = new LoadInst(outputs[i]->getType(), Output,
+                                  outputs[i]->getName() + ".reload");
     Reloads.push_back(load);
     codeReplacer->getInstList().push_back(load);
     std::vector<User *> Users(outputs[i]->user_begin(), outputs[i]->user_end());

llvm/lib/Transforms/Utils/DemoteRegToStack.cpp

@@ -72,7 +72,8 @@ AllocaInst *llvm::DemoteRegToStack(Instruction &I, bool VolatileLoads,
           Value *&V = Loads[PN->getIncomingBlock(i)];
           if (!V) {
             // Insert the load into the predecessor block
-            V = new LoadInst(Slot, I.getName()+".reload", VolatileLoads,
+            V = new LoadInst(I.getType(), Slot, I.getName() + ".reload",
+                             VolatileLoads,
                              PN->getIncomingBlock(i)->getTerminator());
           }
           PN->setIncomingValue(i, V);
@@ -80,7 +81,8 @@ AllocaInst *llvm::DemoteRegToStack(Instruction &I, bool VolatileLoads,
 
     } else {
       // If this is a normal instruction, just insert a load.
-      Value *V = new LoadInst(Slot, I.getName()+".reload", VolatileLoads, U);
+      Value *V = new LoadInst(I.getType(), Slot, I.getName() + ".reload",
+                              VolatileLoads, U);
       U->replaceUsesOfWith(&I, V);
     }
   }
@@ -141,7 +143,8 @@ AllocaInst *llvm::DemotePHIToStack(PHINode *P, Instruction *AllocaPoint) {
   for (; isa<PHINode>(InsertPt) || InsertPt->isEHPad(); ++InsertPt)
     /* empty */;   // Don't insert before PHI nodes or landingpad instrs.
 
-  Value *V = new LoadInst(Slot, P->getName() + ".reload", &*InsertPt);
+  Value *V =
+      new LoadInst(P->getType(), Slot, P->getName() + ".reload", &*InsertPt);
   P->replaceAllUsesWith(V);
 
   // Delete PHI.

llvm/lib/Transforms/Utils/LowerMemIntrinsics.cpp

@@ -72,7 +72,7 @@ void llvm::createMemCpyLoopKnownSize(Instruction *InsertBefore, Value *SrcAddr,
     // Loop Body
     Value *SrcGEP =
         LoopBuilder.CreateInBoundsGEP(LoopOpType, SrcAddr, LoopIndex);
-    Value *Load = LoopBuilder.CreateLoad(SrcGEP, SrcIsVolatile);
+    Value *Load = LoopBuilder.CreateLoad(LoopOpType, SrcGEP, SrcIsVolatile);
     Value *DstGEP =
         LoopBuilder.CreateInBoundsGEP(LoopOpType, DstAddr, LoopIndex);
     LoopBuilder.CreateStore(Load, DstGEP, DstIsVolatile);
@@ -114,7 +114,7 @@ void llvm::createMemCpyLoopKnownSize(Instruction *InsertBefore, Value *SrcAddr,
                              : RBuilder.CreateBitCast(SrcAddr, SrcPtrType);
       Value *SrcGEP = RBuilder.CreateInBoundsGEP(
           OpTy, CastedSrc, ConstantInt::get(TypeOfCopyLen, GepIndex));
-      Value *Load = RBuilder.CreateLoad(SrcGEP, SrcIsVolatile);
+      Value *Load = RBuilder.CreateLoad(OpTy, SrcGEP, SrcIsVolatile);
 
       // Cast destination to operand type and store.
       PointerType *DstPtrType = PointerType::get(OpTy, DstAS);
@@ -181,7 +181,7 @@ void llvm::createMemCpyLoopUnknownSize(Instruction *InsertBefore,
   LoopIndex->addIncoming(ConstantInt::get(CopyLenType, 0U), PreLoopBB);
 
   Value *SrcGEP = LoopBuilder.CreateInBoundsGEP(LoopOpType, SrcAddr, LoopIndex);
-  Value *Load = LoopBuilder.CreateLoad(SrcGEP, SrcIsVolatile);
+  Value *Load = LoopBuilder.CreateLoad(LoopOpType, SrcGEP, SrcIsVolatile);
   Value *DstGEP = LoopBuilder.CreateInBoundsGEP(LoopOpType, DstAddr, LoopIndex);
   LoopBuilder.CreateStore(Load, DstGEP, DstIsVolatile);
 
@@ -234,7 +234,7 @@ void llvm::createMemCpyLoopUnknownSize(Instruction *InsertBefore,
     Value *FullOffset = ResBuilder.CreateAdd(RuntimeBytesCopied, ResidualIndex);
     Value *SrcGEP =
         ResBuilder.CreateInBoundsGEP(Int8Type, SrcAsInt8, FullOffset);
-    Value *Load = ResBuilder.CreateLoad(SrcGEP, SrcIsVolatile);
+    Value *Load = ResBuilder.CreateLoad(Int8Type, SrcGEP, SrcIsVolatile);
     Value *DstGEP =
         ResBuilder.CreateInBoundsGEP(Int8Type, DstAsInt8, FullOffset);
     ResBuilder.CreateStore(Load, DstGEP, DstIsVolatile);
@@ -292,6 +292,8 @@ static void createMemMoveLoop(Instruction *InsertBefore,
   BasicBlock *OrigBB = InsertBefore->getParent();
   Function *F = OrigBB->getParent();
 
+  Type *EltTy = cast<PointerType>(SrcAddr->getType())->getElementType();
+
   // Create the a comparison of src and dst, based on which we jump to either
   // the forward-copy part of the function (if src >= dst) or the backwards-copy
   // part (if src < dst).
@@ -330,7 +332,7 @@ static void createMemMoveLoop(Instruction *InsertBefore,
   Value *IndexPtr = LoopBuilder.CreateSub(
       LoopPhi, ConstantInt::get(TypeOfCopyLen, 1), "index_ptr");
   Value *Element = LoopBuilder.CreateLoad(
-      LoopBuilder.CreateInBoundsGEP(SrcAddr, IndexPtr), "element");
+      EltTy, LoopBuilder.CreateInBoundsGEP(SrcAddr, IndexPtr), "element");
   LoopBuilder.CreateStore(Element,
                           LoopBuilder.CreateInBoundsGEP(DstAddr, IndexPtr));
   LoopBuilder.CreateCondBr(
@@ -347,7 +349,7 @@ static void createMemMoveLoop(Instruction *InsertBefore,
   IRBuilder<> FwdLoopBuilder(FwdLoopBB);
   PHINode *FwdCopyPhi = FwdLoopBuilder.CreatePHI(TypeOfCopyLen, 0, "index_ptr");
   Value *FwdElement = FwdLoopBuilder.CreateLoad(
-      FwdLoopBuilder.CreateInBoundsGEP(SrcAddr, FwdCopyPhi), "element");
+      EltTy, FwdLoopBuilder.CreateInBoundsGEP(SrcAddr, FwdCopyPhi), "element");
   FwdLoopBuilder.CreateStore(
       FwdElement, FwdLoopBuilder.CreateInBoundsGEP(DstAddr, FwdCopyPhi));
   Value *FwdIndexPtr = FwdLoopBuilder.CreateAdd(

llvm/lib/Transforms/Utils/SimplifyCFG.cpp

@@ -5090,7 +5090,9 @@ Value *SwitchLookupTable::BuildLookup(Value *Index, IRBuilder<> &Builder) {
     Value *GEPIndices[] = {Builder.getInt32(0), Index};
     Value *GEP = Builder.CreateInBoundsGEP(Array->getValueType(), Array,
                                            GEPIndices, "switch.gep");
-    return Builder.CreateLoad(GEP, "switch.load");
+    return Builder.CreateLoad(
+        cast<ArrayType>(Array->getValueType())->getElementType(), GEP,
+        "switch.load");
   }
   }
   llvm_unreachable("Unknown lookup table kind!");

llvm/lib/Transforms/Utils/SimplifyLibCalls.cpp

@@ -333,11 +333,12 @@ Value *LibCallSimplifier::optimizeStrCmp(CallInst *CI, IRBuilder<> &B) {
     return ConstantInt::get(CI->getType(), Str1.compare(Str2));
 
   if (HasStr1 && Str1.empty()) // strcmp("", x) -> -*x
-    return B.CreateNeg(
+    return B.CreateNeg(B.CreateZExt(
-        B.CreateZExt(B.CreateLoad(Str2P, "strcmpload"), CI->getType()));
+        B.CreateLoad(B.getInt8Ty(), Str2P, "strcmpload"), CI->getType()));
 
   if (HasStr2 && Str2.empty()) // strcmp(x,"") -> *x
-    return B.CreateZExt(B.CreateLoad(Str1P, "strcmpload"), CI->getType());
+    return B.CreateZExt(B.CreateLoad(B.getInt8Ty(), Str1P, "strcmpload"),
+                        CI->getType());
 
   // strcmp(P, "x") -> memcmp(P, "x", 2)
   uint64_t Len1 = GetStringLength(Str1P);
@@ -397,11 +398,12 @@ Value *LibCallSimplifier::optimizeStrNCmp(CallInst *CI, IRBuilder<> &B) {
   }
 
   if (HasStr1 && Str1.empty()) // strncmp("", x, n) -> -*x
-    return B.CreateNeg(
+    return B.CreateNeg(B.CreateZExt(
-        B.CreateZExt(B.CreateLoad(Str2P, "strcmpload"), CI->getType()));
+        B.CreateLoad(B.getInt8Ty(), Str2P, "strcmpload"), CI->getType()));
 
   if (HasStr2 && Str2.empty()) // strncmp(x, "", n) -> *x
-    return B.CreateZExt(B.CreateLoad(Str1P, "strcmpload"), CI->getType());
+    return B.CreateZExt(B.CreateLoad(B.getInt8Ty(), Str1P, "strcmpload"),
+                        CI->getType());
 
   uint64_t Len1 = GetStringLength(Str1P);
   uint64_t Len2 = GetStringLength(Str2P);
@@ -590,7 +592,8 @@ Value *LibCallSimplifier::optimizeStringLength(CallInst *CI, IRBuilder<> &B,
   // strlen(x) != 0 --> *x != 0
   // strlen(x) == 0 --> *x == 0
   if (isOnlyUsedInZeroEqualityComparison(CI))
-    return B.CreateZExt(B.CreateLoad(Src, "strlenfirst"), CI->getType());
+    return B.CreateZExt(B.CreateLoad(B.getIntNTy(CharSize), Src, "strlenfirst"),
+                        CI->getType());
 
   return nullptr;
 }
@@ -844,10 +847,12 @@ Value *LibCallSimplifier::optimizeMemCmp(CallInst *CI, IRBuilder<> &B) {
 
   // memcmp(S1,S2,1) -> *(unsigned char*)LHS - *(unsigned char*)RHS
   if (Len == 1) {
-    Value *LHSV = B.CreateZExt(B.CreateLoad(castToCStr(LHS, B), "lhsc"),
+    Value *LHSV =
-                               CI->getType(), "lhsv");
+        B.CreateZExt(B.CreateLoad(B.getInt8Ty(), castToCStr(LHS, B), "lhsc"),
-    Value *RHSV = B.CreateZExt(B.CreateLoad(castToCStr(RHS, B), "rhsc"),
+                     CI->getType(), "lhsv");
-                               CI->getType(), "rhsv");
+    Value *RHSV =
+        B.CreateZExt(B.CreateLoad(B.getInt8Ty(), castToCStr(RHS, B), "rhsc"),
+                     CI->getType(), "rhsv");
     return B.CreateSub(LHSV, RHSV, "chardiff");
   }
 
@@ -877,12 +882,12 @@ Value *LibCallSimplifier::optimizeMemCmp(CallInst *CI, IRBuilder<> &B) {
       if (!LHSV) {
         Type *LHSPtrTy =
             IntType->getPointerTo(LHS->getType()->getPointerAddressSpace());
-        LHSV = B.CreateLoad(B.CreateBitCast(LHS, LHSPtrTy), "lhsv");
+        LHSV = B.CreateLoad(IntType, B.CreateBitCast(LHS, LHSPtrTy), "lhsv");
       }
       if (!RHSV) {
         Type *RHSPtrTy =
             IntType->getPointerTo(RHS->getType()->getPointerAddressSpace());
-        RHSV = B.CreateLoad(B.CreateBitCast(RHS, RHSPtrTy), "rhsv");
+        RHSV = B.CreateLoad(IntType, B.CreateBitCast(RHS, RHSPtrTy), "rhsv");
       }
       return B.CreateZExt(B.CreateICmpNE(LHSV, RHSV), CI->getType(), "memcmp");
     }
@@ -2286,7 +2291,8 @@ Value *LibCallSimplifier::optimizeFWrite(CallInst *CI, IRBuilder<> &B) {
     // If this is writing one byte, turn it into fputc.
     // This optimisation is only valid, if the return value is unused.
     if (Bytes == 1 && CI->use_empty()) { // fwrite(S,1,1,F) -> fputc(S[0],F)
-      Value *Char = B.CreateLoad(castToCStr(CI->getArgOperand(0), B), "char");
+      Value *Char = B.CreateLoad(B.getInt8Ty(),
+                                 castToCStr(CI->getArgOperand(0), B), "char");
       Value *NewCI = emitFPutC(Char, CI->getArgOperand(3), B, TLI);
       return NewCI ? ConstantInt::get(CI->getType(), 1) : nullptr;
     }

llvm/lib/Transforms/Utils/VNCoercion.cpp

@@ -386,12 +386,12 @@ Value *getLoadValueForLoad(LoadInst *SrcVal, unsigned Offset, Type *LoadTy,
     // memdep queries will find the new load.  We can't easily remove the old
     // load completely because it is already in the value numbering table.
     IRBuilder<> Builder(SrcVal->getParent(), ++BasicBlock::iterator(SrcVal));
-    Type *DestPTy = IntegerType::get(LoadTy->getContext(), NewLoadSize * 8);
+    Type *DestTy = IntegerType::get(LoadTy->getContext(), NewLoadSize * 8);
-    DestPTy =
+    Type *DestPTy =
-        PointerType::get(DestPTy, PtrVal->getType()->getPointerAddressSpace());
+        PointerType::get(DestTy, PtrVal->getType()->getPointerAddressSpace());
     Builder.SetCurrentDebugLocation(SrcVal->getDebugLoc());
     PtrVal = Builder.CreateBitCast(PtrVal, DestPTy);
-    LoadInst *NewLoad = Builder.CreateLoad(PtrVal);
+    LoadInst *NewLoad = Builder.CreateLoad(DestTy, PtrVal);
     NewLoad->takeName(SrcVal);
     NewLoad->setAlignment(SrcVal->getAlignment());
 

llvm/lib/Transforms/Vectorize/LoadStoreVectorizer.cpp

@@ -1181,7 +1181,7 @@ bool Vectorizer::vectorizeLoadChain(
 
   Value *Bitcast =
       Builder.CreateBitCast(L0->getPointerOperand(), VecTy->getPointerTo(AS));
-  LoadInst *LI = Builder.CreateAlignedLoad(Bitcast, Alignment);
+  LoadInst *LI = Builder.CreateAlignedLoad(VecTy, Bitcast, Alignment);
   propagateMetadata(LI, Chain);
 
   if (VecLoadTy) {

llvm/lib/Transforms/Vectorize/LoopVectorize.cpp

@@ -2100,8 +2100,8 @@ void InnerLoopVectorizer::vectorizeInterleaveGroup(Instruction *Instr,
                                      GroupMask, UndefVec, "wide.masked.vec");
       }
       else
-        NewLoad = Builder.CreateAlignedLoad(NewPtrs[Part], 
+        NewLoad = Builder.CreateAlignedLoad(VecTy, NewPtrs[Part],
-          Group->getAlignment(), "wide.vec");
+                                            Group->getAlignment(), "wide.vec");
       Group->addMetadata(NewLoad);
       NewLoads.push_back(NewLoad);
     }
@@ -2312,7 +2312,8 @@ void InnerLoopVectorizer::vectorizeMemoryInstruction(Instruction *Instr,
                                          UndefValue::get(DataTy),
                                          "wide.masked.load");
       else
-        NewLI = Builder.CreateAlignedLoad(VecPtr, Alignment, "wide.load");
+        NewLI =
+            Builder.CreateAlignedLoad(DataTy, VecPtr, Alignment, "wide.load");
 
       // Add metadata to the load, but setVectorValue to the reverse shuffle.
       addMetadata(NewLI, LI);

llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp

@@ -3136,7 +3136,7 @@ Value *BoUpSLP::vectorizeTree(TreeEntry *E) {
         Builder.SetInsertPoint(LI);
         PointerType *PtrTy = PointerType::get(VecTy, LI->getPointerAddressSpace());
         Value *Ptr = Builder.CreateBitCast(LI->getOperand(0), PtrTy);
-        LoadInst *V = Builder.CreateAlignedLoad(Ptr, LI->getAlignment());
+        LoadInst *V = Builder.CreateAlignedLoad(VecTy, Ptr, LI->getAlignment());
         Value *NewV = propagateMetadata(V, E->Scalars);
         if (!E->ReorderIndices.empty()) {
           OrdersType Mask;
@@ -3341,7 +3341,7 @@ Value *BoUpSLP::vectorizeTree(TreeEntry *E) {
         ExternalUses.push_back(ExternalUser(PO, cast<User>(VecPtr), 0));
 
       unsigned Alignment = LI->getAlignment();
-      LI = Builder.CreateLoad(VecPtr);
+      LI = Builder.CreateLoad(VecTy, VecPtr);
       if (!Alignment) {
         Alignment = DL->getABITypeAlignment(ScalarLoadTy);
       }

llvm/test/CodeGen/AMDGPU/lower-kernargs.ll

@@ -306,20 +306,18 @@ define amdgpu_kernel void @kern_v3i32(<3 x i32> %arg0) {
 ; HSA-LABEL: @kern_v3i32(
 ; HSA-NEXT:    [[KERN_V3I32_KERNARG_SEGMENT:%.*]] = call nonnull align 16 dereferenceable(16) i8 addrspace(4)* @llvm.amdgcn.kernarg.segment.ptr()
 ; HSA-NEXT:    [[ARG0_KERNARG_OFFSET:%.*]] = getelementptr inbounds i8, i8 addrspace(4)* [[KERN_V3I32_KERNARG_SEGMENT]], i64 0
-; HSA-NEXT:    [[ARG0_KERNARG_OFFSET_CAST:%.*]] = bitcast i8 addrspace(4)* [[ARG0_KERNARG_OFFSET]] to <3 x i32> addrspace(4)*
+; HSA-NEXT:    [[ARG0_KERNARG_OFFSET_CAST:%.*]] = bitcast i8 addrspace(4)* [[ARG0_KERNARG_OFFSET]] to <4 x i32> addrspace(4)*
-; HSA-NEXT:    [[TMP1:%.*]] = bitcast <3 x i32> addrspace(4)* [[ARG0_KERNARG_OFFSET_CAST]] to <4 x i32> addrspace(4)*
+; HSA-NEXT:    [[TMP:%.*]] = load <4 x i32>, <4 x i32> addrspace(4)* [[ARG0_KERNARG_OFFSET_CAST]], align 16, !invariant.load !0
-; HSA-NEXT:    [[TMP2:%.*]] = load <4 x i32>, <4 x i32> addrspace(4)* [[TMP1]], align 16, !invariant.load !0
+; HSA-NEXT:    [[ARG0_LOAD:%.*]] = shufflevector <4 x i32> [[TMP]], <4 x i32> undef, <3 x i32> <i32 0, i32 1, i32 2>
-; HSA-NEXT:    [[ARG0_LOAD:%.*]] = shufflevector <4 x i32> [[TMP2]], <4 x i32> undef, <3 x i32> <i32 0, i32 1, i32 2>
 ; HSA-NEXT:    store <3 x i32> [[ARG0_LOAD]], <3 x i32> addrspace(1)* undef, align 4
 ; HSA-NEXT:    ret void
 ;
 ; MESA-LABEL: @kern_v3i32(
 ; MESA-NEXT:    [[KERN_V3I32_KERNARG_SEGMENT:%.*]] = call nonnull align 16 dereferenceable(52) i8 addrspace(4)* @llvm.amdgcn.kernarg.segment.ptr()
 ; MESA-NEXT:    [[ARG0_KERNARG_OFFSET:%.*]] = getelementptr inbounds i8, i8 addrspace(4)* [[KERN_V3I32_KERNARG_SEGMENT]], i64 36
-; MESA-NEXT:    [[ARG0_KERNARG_OFFSET_CAST:%.*]] = bitcast i8 addrspace(4)* [[ARG0_KERNARG_OFFSET]] to <3 x i32> addrspace(4)*
+; MESA-NEXT:    [[ARG0_KERNARG_OFFSET_CAST:%.*]] = bitcast i8 addrspace(4)* [[ARG0_KERNARG_OFFSET]] to <4 x i32> addrspace(4)*
-; MESA-NEXT:    [[TMP1:%.*]] = bitcast <3 x i32> addrspace(4)* [[ARG0_KERNARG_OFFSET_CAST]] to <4 x i32> addrspace(4)*
+; MESA-NEXT:    [[TMP:%.*]] = load <4 x i32>, <4 x i32> addrspace(4)* [[ARG0_KERNARG_OFFSET_CAST]], align 4, !invariant.load !0
-; MESA-NEXT:    [[TMP2:%.*]] = load <4 x i32>, <4 x i32> addrspace(4)* [[TMP1]], align 4, !invariant.load !0
+; MESA-NEXT:    [[ARG0_LOAD:%.*]] = shufflevector <4 x i32> [[TMP]], <4 x i32> undef, <3 x i32> <i32 0, i32 1, i32 2>
-; MESA-NEXT:    [[ARG0_LOAD:%.*]] = shufflevector <4 x i32> [[TMP2]], <4 x i32> undef, <3 x i32> <i32 0, i32 1, i32 2>
 ; MESA-NEXT:    store <3 x i32> [[ARG0_LOAD]], <3 x i32> addrspace(1)* undef, align 4
 ; MESA-NEXT:    ret void
 ;
@@ -397,10 +395,9 @@ define amdgpu_kernel void @kern_i32_v3i32(i32 %arg0, <3 x i32> %arg1) {
 ; HSA-NEXT:    [[ARG0_KERNARG_OFFSET_CAST:%.*]] = bitcast i8 addrspace(4)* [[ARG0_KERNARG_OFFSET]] to i32 addrspace(4)*
 ; HSA-NEXT:    [[ARG0_LOAD:%.*]] = load i32, i32 addrspace(4)* [[ARG0_KERNARG_OFFSET_CAST]], align 16, !invariant.load !0
 ; HSA-NEXT:    [[ARG1_KERNARG_OFFSET:%.*]] = getelementptr inbounds i8, i8 addrspace(4)* [[KERN_I32_V3I32_KERNARG_SEGMENT]], i64 16
-; HSA-NEXT:    [[ARG1_KERNARG_OFFSET_CAST:%.*]] = bitcast i8 addrspace(4)* [[ARG1_KERNARG_OFFSET]] to <3 x i32> addrspace(4)*
+; HSA-NEXT:    [[ARG1_KERNARG_OFFSET_CAST:%.*]] = bitcast i8 addrspace(4)* [[ARG1_KERNARG_OFFSET]] to <4 x i32> addrspace(4)*
-; HSA-NEXT:    [[TMP1:%.*]] = bitcast <3 x i32> addrspace(4)* [[ARG1_KERNARG_OFFSET_CAST]] to <4 x i32> addrspace(4)*
+; HSA-NEXT:    [[TMP:%.*]] = load <4 x i32>, <4 x i32> addrspace(4)* [[ARG1_KERNARG_OFFSET_CAST]], align 16, !invariant.load !0
-; HSA-NEXT:    [[TMP2:%.*]] = load <4 x i32>, <4 x i32> addrspace(4)* [[TMP1]], align 16, !invariant.load !0
+; HSA-NEXT:    [[ARG1_LOAD:%.*]] = shufflevector <4 x i32> [[TMP]], <4 x i32> undef, <3 x i32> <i32 0, i32 1, i32 2>
-; HSA-NEXT:    [[ARG1_LOAD:%.*]] = shufflevector <4 x i32> [[TMP2]], <4 x i32> undef, <3 x i32> <i32 0, i32 1, i32 2>
 ; HSA-NEXT:    store i32 [[ARG0_LOAD]], i32 addrspace(1)* undef
 ; HSA-NEXT:    store <3 x i32> [[ARG1_LOAD]], <3 x i32> addrspace(1)* undef, align 4
 ; HSA-NEXT:    ret void
@@ -411,10 +408,9 @@ define amdgpu_kernel void @kern_i32_v3i32(i32 %arg0, <3 x i32> %arg1) {
 ; MESA-NEXT:    [[ARG0_KERNARG_OFFSET_CAST:%.*]] = bitcast i8 addrspace(4)* [[ARG0_KERNARG_OFFSET]] to i32 addrspace(4)*
 ; MESA-NEXT:    [[ARG0_LOAD:%.*]] = load i32, i32 addrspace(4)* [[ARG0_KERNARG_OFFSET_CAST]], align 4, !invariant.load !0
 ; MESA-NEXT:    [[ARG1_KERNARG_OFFSET:%.*]] = getelementptr inbounds i8, i8 addrspace(4)* [[KERN_I32_V3I32_KERNARG_SEGMENT]], i64 52
-; MESA-NEXT:    [[ARG1_KERNARG_OFFSET_CAST:%.*]] = bitcast i8 addrspace(4)* [[ARG1_KERNARG_OFFSET]] to <3 x i32> addrspace(4)*
+; MESA-NEXT:    [[ARG1_KERNARG_OFFSET_CAST:%.*]] = bitcast i8 addrspace(4)* [[ARG1_KERNARG_OFFSET]] to <4 x i32> addrspace(4)*
-; MESA-NEXT:    [[TMP1:%.*]] = bitcast <3 x i32> addrspace(4)* [[ARG1_KERNARG_OFFSET_CAST]] to <4 x i32> addrspace(4)*
+; MESA-NEXT:    [[TMP:%.*]] = load <4 x i32>, <4 x i32> addrspace(4)* [[ARG1_KERNARG_OFFSET_CAST]], align 4, !invariant.load !0
-; MESA-NEXT:    [[TMP2:%.*]] = load <4 x i32>, <4 x i32> addrspace(4)* [[TMP1]], align 4, !invariant.load !0
+; MESA-NEXT:    [[ARG1_LOAD:%.*]] = shufflevector <4 x i32> [[TMP]], <4 x i32> undef, <3 x i32> <i32 0, i32 1, i32 2>
-; MESA-NEXT:    [[ARG1_LOAD:%.*]] = shufflevector <4 x i32> [[TMP2]], <4 x i32> undef, <3 x i32> <i32 0, i32 1, i32 2>
 ; MESA-NEXT:    store i32 [[ARG0_LOAD]], i32 addrspace(1)* undef
 ; MESA-NEXT:    store <3 x i32> [[ARG1_LOAD]], <3 x i32> addrspace(1)* undef, align 4
 ; MESA-NEXT:    ret void

llvm/tools/bugpoint/Miscompilation.cpp

@@ -879,7 +879,8 @@ CleanupAndPrepareModules(BugDriver &BD, std::unique_ptr<Module> Test,
               BasicBlock::Create(F->getContext(), "lookupfp", FuncWrapper);
 
           // Check to see if we already looked up the value.
-          Value *CachedVal = new LoadInst(Cache, "fpcache", EntryBB);
+          Value *CachedVal =
+              new LoadInst(F->getType(), Cache, "fpcache", EntryBB);
           Value *IsNull = new ICmpInst(*EntryBB, ICmpInst::ICMP_EQ, CachedVal,
                                        NullPtr, "isNull");
           BranchInst::Create(LookupBB, DoCallBB, IsNull, EntryBB);

llvm/unittests/Analysis/AliasAnalysisTest.cpp

@@ -174,7 +174,7 @@ TEST_F(AliasAnalysisTest, getModRefInfo) {
   auto *Addr = ConstantPointerNull::get(PtrType);
 
   auto *Store1 = new StoreInst(Value, Addr, BB);
-  auto *Load1 = new LoadInst(Addr, "load", BB);
+  auto *Load1 = new LoadInst(IntType, Addr, "load", BB);
   auto *Add1 = BinaryOperator::CreateAdd(Value, Value, "add", BB);
   auto *VAArg1 = new VAArgInst(Addr, PtrType, "vaarg", BB);
   auto *CmpXChg1 = new AtomicCmpXchgInst(

llvm/unittests/Analysis/MemorySSATest.cpp

@@ -92,7 +92,7 @@ TEST_F(MemorySSATest, CreateALoad) {
   MemorySSAUpdater Updater(&MSSA);
   // Add the load
   B.SetInsertPoint(Merge);
-  LoadInst *LoadInst = B.CreateLoad(PointerArg);
+  LoadInst *LoadInst = B.CreateLoad(B.getInt8Ty(), PointerArg);
 
   // MemoryPHI should already exist.
   MemoryPhi *MP = MSSA.getMemoryAccess(Merge);
@@ -138,7 +138,7 @@ TEST_F(MemorySSATest, CreateLoadsAndStoreUpdater) {
 
   // Add the load
   B.SetInsertPoint(Merge, Merge->begin());
-  LoadInst *FirstLoad = B.CreateLoad(PointerArg);
+  LoadInst *FirstLoad = B.CreateLoad(B.getInt8Ty(), PointerArg);
 
   // MemoryPHI should not already exist.
   MemoryPhi *MP = MSSA.getMemoryAccess(Merge);
@@ -162,7 +162,7 @@ TEST_F(MemorySSATest, CreateLoadsAndStoreUpdater) {
   // We don't touch existing loads, so we need to create a new one to get a phi
   // Add the second load
   B.SetInsertPoint(Merge, Merge->begin());
-  LoadInst *SecondLoad = B.CreateLoad(PointerArg);
+  LoadInst *SecondLoad = B.CreateLoad(B.getInt8Ty(), PointerArg);
 
   // MemoryPHI should not already exist.
   MP = MSSA.getMemoryAccess(Merge);
@@ -228,7 +228,7 @@ TEST_F(MemorySSATest, CreateALoadUpdater) {
 
   // Add the load
   B.SetInsertPoint(Merge, Merge->begin());
-  LoadInst *LoadInst = B.CreateLoad(PointerArg);
+  LoadInst *LoadInst = B.CreateLoad(B.getInt8Ty(), PointerArg);
 
   // MemoryPHI should not already exist.
   MemoryPhi *MP = MSSA.getMemoryAccess(Merge);
@@ -262,7 +262,7 @@ TEST_F(MemorySSATest, SinkLoad) {
 
   // Load in left block
   B.SetInsertPoint(Left, Left->begin());
-  LoadInst *LoadInst1 = B.CreateLoad(PointerArg);
+  LoadInst *LoadInst1 = B.CreateLoad(B.getInt8Ty(), PointerArg);
   // Store in merge block
   B.SetInsertPoint(Merge, Merge->begin());
   B.CreateStore(B.getInt8(16), PointerArg);
@@ -310,7 +310,7 @@ TEST_F(MemorySSATest, MoveAStore) {
   BranchInst::Create(Merge, Left);
   BranchInst::Create(Merge, Right);
   B.SetInsertPoint(Merge);
-  B.CreateLoad(PointerArg);
+  B.CreateLoad(B.getInt8Ty(), PointerArg);
   setupAnalyses();
   MemorySSA &MSSA = *Analyses->MSSA;
   MemorySSAUpdater Updater(&MSSA);
@@ -346,7 +346,7 @@ TEST_F(MemorySSATest, MoveAStoreUpdater) {
   BranchInst::Create(Merge, Left);
   BranchInst::Create(Merge, Right);
   B.SetInsertPoint(Merge);
-  auto *MergeLoad = B.CreateLoad(PointerArg);
+  auto *MergeLoad = B.CreateLoad(B.getInt8Ty(), PointerArg);
   setupAnalyses();
   MemorySSA &MSSA = *Analyses->MSSA;
   MemorySSAUpdater Updater(&MSSA);
@@ -392,7 +392,7 @@ TEST_F(MemorySSATest, MoveAStoreUpdaterMove) {
   BranchInst::Create(Merge, Left);
   BranchInst::Create(Merge, Right);
   B.SetInsertPoint(Merge);
-  auto *MergeLoad = B.CreateLoad(PointerArg);
+  auto *MergeLoad = B.CreateLoad(B.getInt8Ty(), PointerArg);
   setupAnalyses();
   MemorySSA &MSSA = *Analyses->MSSA;
   MemorySSAUpdater Updater(&MSSA);
@@ -436,7 +436,7 @@ TEST_F(MemorySSATest, MoveAStoreAllAround) {
   BranchInst::Create(Merge, Left);
   BranchInst::Create(Merge, Right);
   B.SetInsertPoint(Merge);
-  auto *MergeLoad = B.CreateLoad(PointerArg);
+  auto *MergeLoad = B.CreateLoad(B.getInt8Ty(), PointerArg);
   setupAnalyses();
   MemorySSA &MSSA = *Analyses->MSSA;
   MemorySSAUpdater Updater(&MSSA);
@@ -490,7 +490,7 @@ TEST_F(MemorySSATest, RemoveAPhi) {
   BranchInst::Create(Merge, Left);
   BranchInst::Create(Merge, Right);
   B.SetInsertPoint(Merge);
-  LoadInst *LoadInst = B.CreateLoad(PointerArg);
+  LoadInst *LoadInst = B.CreateLoad(B.getInt8Ty(), PointerArg);
 
   setupAnalyses();
   MemorySSA &MSSA = *Analyses->MSSA;
@@ -535,7 +535,7 @@ TEST_F(MemorySSATest, RemoveMemoryAccess) {
   BranchInst::Create(Merge, Left);
   BranchInst::Create(Merge, Right);
   B.SetInsertPoint(Merge);
-  LoadInst *LoadInst = B.CreateLoad(PointerArg);
+  LoadInst *LoadInst = B.CreateLoad(B.getInt8Ty(), PointerArg);
 
   setupAnalyses();
   MemorySSA &MSSA = *Analyses->MSSA;
@@ -631,7 +631,7 @@ TEST_F(MemorySSATest, TestStoreAndLoad) {
   Type *Int8 = Type::getInt8Ty(C);
   Value *Alloca = B.CreateAlloca(Int8, ConstantInt::get(Int8, 1), "A");
   Instruction *SI = B.CreateStore(ConstantInt::get(Int8, 0), Alloca);
-  Instruction *LI = B.CreateLoad(Alloca);
+  Instruction *LI = B.CreateLoad(Int8, Alloca);
 
   setupAnalyses();
   MemorySSA &MSSA = *Analyses->MSSA;
@@ -715,12 +715,12 @@ TEST_F(MemorySSATest, PartialWalkerCacheWithPhis) {
   B.SetInsertPoint(IfThen);
   Instruction *FirstStore = B.CreateStore(Zero, AllocA);
   B.CreateStore(Zero, AllocB);
-  Instruction *ALoad0 = B.CreateLoad(AllocA, "");
+  Instruction *ALoad0 = B.CreateLoad(Int8, AllocA, "");
   Instruction *BStore = B.CreateStore(Zero, AllocB);
   // Due to use optimization/etc. we make a store to A, which is removed after
   // we build MSSA. This helps keep the test case simple-ish.
   Instruction *KillStore = B.CreateStore(Zero, AllocA);
-  Instruction *ALoad = B.CreateLoad(AllocA, "");
+  Instruction *ALoad = B.CreateLoad(Int8, AllocA, "");
   B.CreateBr(IfEnd);
 
   B.SetInsertPoint(IfEnd);
@@ -771,7 +771,7 @@ TEST_F(MemorySSATest, WalkerInvariantLoadOpt) {
   Value *AllocA = B.CreateAlloca(Int8, One, "");
 
   Instruction *Store = B.CreateStore(One, AllocA);
-  Instruction *Load = B.CreateLoad(AllocA);
+  Instruction *Load = B.CreateLoad(Int8, AllocA);
 
   setupAnalyses();
   MemorySSA &MSSA = *Analyses->MSSA;
@@ -800,7 +800,7 @@ TEST_F(MemorySSATest, WalkerReopt) {
   Instruction *SIA = B.CreateStore(ConstantInt::get(Int8, 0), AllocaA);
   Value *AllocaB = B.CreateAlloca(Int8, ConstantInt::get(Int8, 1), "B");
   Instruction *SIB = B.CreateStore(ConstantInt::get(Int8, 0), AllocaB);
-  Instruction *LIA = B.CreateLoad(AllocaA);
+  Instruction *LIA = B.CreateLoad(Int8, AllocaA);
 
   setupAnalyses();
   MemorySSA &MSSA = *Analyses->MSSA;
@@ -834,11 +834,11 @@ TEST_F(MemorySSATest, MoveAboveMemoryDef) {
 
   StoreInst *StoreA0 = B.CreateStore(ConstantInt::get(Int8, 0), A);
   StoreInst *StoreB = B.CreateStore(ConstantInt::get(Int8, 0), B_);
-  LoadInst *LoadB = B.CreateLoad(B_);
+  LoadInst *LoadB = B.CreateLoad(Int8, B_);
   StoreInst *StoreA1 = B.CreateStore(ConstantInt::get(Int8, 4), A);
   StoreInst *StoreC = B.CreateStore(ConstantInt::get(Int8, 4), C);
   StoreInst *StoreA2 = B.CreateStore(ConstantInt::get(Int8, 4), A);
-  LoadInst *LoadC = B.CreateLoad(C);
+  LoadInst *LoadC = B.CreateLoad(Int8, C);
 
   setupAnalyses();
   MemorySSA &MSSA = *Analyses->MSSA;
@@ -902,7 +902,7 @@ TEST_F(MemorySSATest, Irreducible) {
   MemorySSA &MSSA = *Analyses->MSSA;
   MemorySSAUpdater Updater(&MSSA);
   // Create the load memory acccess
-  LoadInst *LoadInst = B.CreateLoad(FirstArg);
+  LoadInst *LoadInst = B.CreateLoad(B.getInt8Ty(), FirstArg);
   MemoryUse *LoadAccess = cast<MemoryUse>(Updater.createMemoryAccessInBB(
       LoadInst, nullptr, AfterLoopBB, MemorySSA::Beginning));
   Updater.insertUse(LoadAccess);
@@ -1010,15 +1010,15 @@ TEST_F(MemorySSATest, TestLoadMustAlias) {
 
   B.CreateStore(ConstantInt::get(Int8, 1), AllocaB);
   // Check load from LOE
-  LoadInst *LA1 = B.CreateLoad(AllocaA, "");
+  LoadInst *LA1 = B.CreateLoad(Int8, AllocaA, "");
   // Check load alias cached for second load
-  LoadInst *LA2 = B.CreateLoad(AllocaA, "");
+  LoadInst *LA2 = B.CreateLoad(Int8, AllocaA, "");
 
   B.CreateStore(ConstantInt::get(Int8, 1), AllocaA);
   // Check load from store/def
-  LoadInst *LA3 = B.CreateLoad(AllocaA, "");
+  LoadInst *LA3 = B.CreateLoad(Int8, AllocaA, "");
   // Check load alias cached for second load
-  LoadInst *LA4 = B.CreateLoad(AllocaA, "");
+  LoadInst *LA4 = B.CreateLoad(Int8, AllocaA, "");
 
   setupAnalyses();
   MemorySSA &MSSA = *Analyses->MSSA;
@@ -1103,13 +1103,13 @@ TEST_F(MemorySSATest, TestLoadMayAlias) {
   Argument *PointerA = &*ArgIt;
   Argument *PointerB = &*(++ArgIt);
   B.CreateStore(ConstantInt::get(Int8, 1), PointerB);
-  LoadInst *LA1 = B.CreateLoad(PointerA, "");
+  LoadInst *LA1 = B.CreateLoad(Int8, PointerA, "");
   B.CreateStore(ConstantInt::get(Int8, 0), PointerA);
-  LoadInst *LB1 = B.CreateLoad(PointerB, "");
+  LoadInst *LB1 = B.CreateLoad(Int8, PointerB, "");
   B.CreateStore(ConstantInt::get(Int8, 0), PointerA);
-  LoadInst *LA2 = B.CreateLoad(PointerA, "");
+  LoadInst *LA2 = B.CreateLoad(Int8, PointerA, "");
   B.CreateStore(ConstantInt::get(Int8, 0), PointerB);
-  LoadInst *LB2 = B.CreateLoad(PointerB, "");
+  LoadInst *LB2 = B.CreateLoad(Int8, PointerB, "");
 
   setupAnalyses();
   MemorySSA &MSSA = *Analyses->MSSA;

llvm/unittests/Analysis/OrderedInstructionsTest.cpp

@@ -44,13 +44,13 @@ TEST(OrderedInstructionsTest, DominanceTest) {
   BasicBlock *BBX = BasicBlock::Create(Ctx, "bbx", F);
   B.SetInsertPoint(BBX);
   Argument *PointerArg = &*F->arg_begin();
-  LoadInst *LoadInstX = B.CreateLoad(PointerArg);
+  LoadInst *LoadInstX = B.CreateLoad(B.getInt8Ty(), PointerArg);
-  LoadInst *LoadInstY = B.CreateLoad(PointerArg);
+  LoadInst *LoadInstY = B.CreateLoad(B.getInt8Ty(), PointerArg);
 
   // Create BBY with 1 load.
   BasicBlock *BBY = BasicBlock::Create(Ctx, "bby", F);
   B.SetInsertPoint(BBY);
-  LoadInst *LoadInstZ = B.CreateLoad(PointerArg);
+  LoadInst *LoadInstZ = B.CreateLoad(B.getInt8Ty(), PointerArg);
   B.CreateRet(LoadInstZ);
   std::unique_ptr<DominatorTree> DT(new DominatorTree(*F));
   OrderedInstructions OI(&*DT);

llvm/unittests/Analysis/PhiValuesTest.cpp

@@ -37,10 +37,10 @@ TEST(PhiValuesTest, SimplePhi) {
   BranchInst::Create(Then, If);
   BranchInst::Create(Then, Else);
 
-  Value *Val1 = new LoadInst(UndefValue::get(I32PtrTy), "val1", Entry);
+  Value *Val1 = new LoadInst(I32Ty, UndefValue::get(I32PtrTy), "val1", Entry);
-  Value *Val2 = new LoadInst(UndefValue::get(I32PtrTy), "val2", Entry);
+  Value *Val2 = new LoadInst(I32Ty, UndefValue::get(I32PtrTy), "val2", Entry);
-  Value *Val3 = new LoadInst(UndefValue::get(I32PtrTy), "val3", Entry);
+  Value *Val3 = new LoadInst(I32Ty, UndefValue::get(I32PtrTy), "val3", Entry);
-  Value *Val4 = new LoadInst(UndefValue::get(I32PtrTy), "val4", Entry);
+  Value *Val4 = new LoadInst(I32Ty, UndefValue::get(I32PtrTy), "val4", Entry);
 
   PHINode *Phi1 = PHINode::Create(I32Ty, 2, "phi1", Then);
   Phi1->addIncoming(Val1, If);
@@ -110,10 +110,10 @@ TEST(PhiValuesTest, DependentPhi) {
   BranchInst::Create(End, If2);
   BranchInst::Create(End, Else2);
 
-  Value *Val1 = new LoadInst(UndefValue::get(I32PtrTy), "val1", Entry);
+  Value *Val1 = new LoadInst(I32Ty, UndefValue::get(I32PtrTy), "val1", Entry);
-  Value *Val2 = new LoadInst(UndefValue::get(I32PtrTy), "val2", Entry);
+  Value *Val2 = new LoadInst(I32Ty, UndefValue::get(I32PtrTy), "val2", Entry);
-  Value *Val3 = new LoadInst(UndefValue::get(I32PtrTy), "val3", Entry);
+  Value *Val3 = new LoadInst(I32Ty, UndefValue::get(I32PtrTy), "val3", Entry);
-  Value *Val4 = new LoadInst(UndefValue::get(I32PtrTy), "val4", Entry);
+  Value *Val4 = new LoadInst(I32Ty, UndefValue::get(I32PtrTy), "val4", Entry);
 
   PHINode *Phi1 = PHINode::Create(I32Ty, 2, "phi1", Then);
   Phi1->addIncoming(Val1, If1);

llvm/unittests/Analysis/ScalarEvolutionTest.cpp

@@ -407,9 +407,11 @@ TEST_F(ScalarEvolutionsTest, CompareValueComplexity) {
 
   const int ValueDepth = 10;
   for (int i = 0; i < ValueDepth; i++) {
-    X = new LoadInst(new IntToPtrInst(X, IntPtrPtrTy, "", EntryBB), "",
+    X = new LoadInst(IntPtrTy, new IntToPtrInst(X, IntPtrPtrTy, "", EntryBB),
+                     "",
                      /*isVolatile*/ false, EntryBB);
-    Y = new LoadInst(new IntToPtrInst(Y, IntPtrPtrTy, "", EntryBB), "",
+    Y = new LoadInst(IntPtrTy, new IntToPtrInst(Y, IntPtrPtrTy, "", EntryBB),
+                     "",
                      /*isVolatile*/ false, EntryBB);
   }
 

llvm/unittests/Analysis/SparsePropagation.cpp

@@ -381,7 +381,7 @@ TEST_F(SparsePropagationTest, FunctionDefined) {
   BasicBlock *Else = BasicBlock::Create(Context, "else", F);
   F->arg_begin()->setName("cond");
   Builder.SetInsertPoint(If);
-  LoadInst *Cond = Builder.CreateLoad(F->arg_begin());
+  LoadInst *Cond = Builder.CreateLoad(Type::getInt1Ty(Context), F->arg_begin());
   Builder.CreateCondBr(Cond, Then, Else);
   Builder.SetInsertPoint(Then);
   Builder.CreateRet(Builder.getInt64(1));
@@ -421,7 +421,7 @@ TEST_F(SparsePropagationTest, FunctionOverDefined) {
   BasicBlock *Else = BasicBlock::Create(Context, "else", F);
   F->arg_begin()->setName("cond");
   Builder.SetInsertPoint(If);
-  LoadInst *Cond = Builder.CreateLoad(F->arg_begin());
+  LoadInst *Cond = Builder.CreateLoad(Type::getInt1Ty(Context), F->arg_begin());
   Builder.CreateCondBr(Cond, Then, Else);
   Builder.SetInsertPoint(Then);
   Builder.CreateRet(Builder.getInt64(0));

llvm/unittests/ExecutionEngine/MCJIT/MCJITMultipleModuleTest.cpp

@@ -211,10 +211,10 @@ TEST_F(MCJITMultipleModuleTest, two_module_global_variables_case) {
   GVB = insertGlobalInt32(B.get(), "GVB", initialNum);
   FA = startFunction(A.get(),
                      FunctionType::get(Builder.getInt32Ty(), {}, false), "FA");
-  endFunctionWithRet(FA, Builder.CreateLoad(GVA));
+  endFunctionWithRet(FA, Builder.CreateLoad(Builder.getInt32Ty(), GVA));
   FB = startFunction(B.get(),
                      FunctionType::get(Builder.getInt32Ty(), {}, false), "FB");
-  endFunctionWithRet(FB, Builder.CreateLoad(GVB));
+  endFunctionWithRet(FB, Builder.CreateLoad(Builder.getInt32Ty(), GVB));
 
   GVC = insertGlobalInt32(B.get(), "GVC", initialNum);
   GVC->setLinkage(GlobalValue::InternalLinkage);

llvm/unittests/ExecutionEngine/MCJIT/MCJITTest.cpp

@@ -101,7 +101,7 @@ TEST_F(MCJITTest, return_global) {
   Function *ReturnGlobal =
       startFunction(M.get(), FunctionType::get(Builder.getInt32Ty(), {}, false),
                     "ReturnGlobal");
-  Value *ReadGlobal = Builder.CreateLoad(GV);
+  Value *ReadGlobal = Builder.CreateLoad(Builder.getInt32Ty(), GV);
   endFunctionWithRet(ReturnGlobal, ReadGlobal);
 
   createJIT(std::move(M));

llvm/unittests/IR/IRBuilderTest.cpp

@@ -54,7 +54,7 @@ TEST_F(IRBuilderTest, Intrinsics) {
   CallInst *Call;
   IntrinsicInst *II;
 
-  V = Builder.CreateLoad(GV);
+  V = Builder.CreateLoad(GV->getValueType(), GV);
   I = cast<Instruction>(Builder.CreateFAdd(V, V));
   I->setHasNoInfs(true);
   I->setHasNoNaNs(false);
@@ -207,7 +207,7 @@ TEST_F(IRBuilderTest, FastMathFlags) {
   Value *F, *FC;
   Instruction *FDiv, *FAdd, *FCmp, *FCall;
 
-  F = Builder.CreateLoad(GV);
+  F = Builder.CreateLoad(GV->getValueType(), GV);
   F = Builder.CreateFAdd(F, F);
 
   EXPECT_FALSE(Builder.getFastMathFlags().any());
@@ -394,7 +394,7 @@ TEST_F(IRBuilderTest, WrapFlags) {
   // Test instructions.
   GlobalVariable *G = new GlobalVariable(*M, Builder.getInt32Ty(), true,
                                          GlobalValue::ExternalLinkage, nullptr);
-  Value *V = Builder.CreateLoad(G);
+  Value *V = Builder.CreateLoad(G->getValueType(), G);
   EXPECT_TRUE(
       cast<BinaryOperator>(Builder.CreateNSWAdd(V, V))->hasNoSignedWrap());
   EXPECT_TRUE(
@@ -461,7 +461,7 @@ TEST_F(IRBuilderTest, RAIIHelpersTest) {
   EXPECT_FALSE(Builder.getFastMathFlags().allowReciprocal());
   EXPECT_EQ(FPMathA, Builder.getDefaultFPMathTag());
 
-  Value *F = Builder.CreateLoad(GV);
+  Value *F = Builder.CreateLoad(GV->getValueType(), GV);
 
   {
     IRBuilder<>::InsertPointGuard Guard(Builder);

llvm/unittests/IR/PatternMatch.cpp

@@ -397,7 +397,7 @@ TEST_F(PatternMatchTest, LoadStoreOps) {
   //  store i32 42, i32* %0
 
   Value *Alloca = IRB.CreateAlloca(IRB.getInt32Ty());
-  Value *LoadInst = IRB.CreateLoad(Alloca);
+  Value *LoadInst = IRB.CreateLoad(IRB.getInt32Ty(), Alloca);
   Value *FourtyTwo = IRB.getInt32(42);
   Value *StoreInst = IRB.CreateStore(FourtyTwo, Alloca);
   Value *MatchLoad, *MatchStoreVal, *MatchStorePointer;

llvm/unittests/Linker/LinkModulesTest.cpp

@@ -83,7 +83,7 @@ TEST_F(LinkModuleTest, BlockAddress) {
   GEPIndices.push_back(&*F->arg_begin());
 
   Value *GEP = Builder.CreateGEP(AT, GV, GEPIndices, "switch.gep");
-  Value *Load = Builder.CreateLoad(GEP, "switch.load");
+  Value *Load = Builder.CreateLoad(AT->getElementType(), GEP, "switch.load");
 
   Builder.CreateRet(Load);
 

llvm/unittests/Transforms/Scalar/LoopPassManagerTest.cpp

@@ -908,7 +908,8 @@ TEST_F(LoopPassManagerTest, LoopChildInsertion) {
   ASSERT_THAT(BBI, F.end());
   auto CreateCondBr = [&](BasicBlock *TrueBB, BasicBlock *FalseBB,
                           const char *Name, BasicBlock *BB) {
-    auto *Cond = new LoadInst(&Ptr, Name, /*isVolatile*/ true, BB);
+    auto *Cond = new LoadInst(Type::getInt1Ty(Context), &Ptr, Name,
+                              /*isVolatile*/ true, BB);
     BranchInst::Create(TrueBB, FalseBB, Cond, BB);
   };
 
@@ -1110,7 +1111,8 @@ TEST_F(LoopPassManagerTest, LoopPeerInsertion) {
   ASSERT_THAT(BBI, F.end());
   auto CreateCondBr = [&](BasicBlock *TrueBB, BasicBlock *FalseBB,
                           const char *Name, BasicBlock *BB) {
-    auto *Cond = new LoadInst(&Ptr, Name, /*isVolatile*/ true, BB);
+    auto *Cond = new LoadInst(Type::getInt1Ty(Context), &Ptr, Name,
+                              /*isVolatile*/ true, BB);
     BranchInst::Create(TrueBB, FalseBB, Cond, BB);
   };
 
@@ -1503,8 +1505,9 @@ TEST_F(LoopPassManagerTest, LoopDeletion) {
             auto *NewLoop03BB =
                 BasicBlock::Create(Context, "loop.0.3", &F, &Loop0LatchBB);
             BranchInst::Create(NewLoop03BB, NewLoop03PHBB);
-            auto *Cond = new LoadInst(&Ptr, "cond.0.3", /*isVolatile*/ true,
+            auto *Cond =
-                                      NewLoop03BB);
+                new LoadInst(Type::getInt1Ty(Context), &Ptr, "cond.0.3",
+                             /*isVolatile*/ true, NewLoop03BB);
             BranchInst::Create(&Loop0LatchBB, NewLoop03BB, Cond, NewLoop03BB);
             Loop02PHBB.getTerminator()->replaceUsesOfWith(&Loop0LatchBB,
                                                           NewLoop03PHBB);

llvm/unittests/Transforms/Utils/FunctionComparatorTest.cpp

@@ -31,7 +31,7 @@ struct TestFunction {
     BB = BasicBlock::Create(Ctx, "", F);
     B.SetInsertPoint(BB);
     Argument *PointerArg = &*F->arg_begin();
-    LoadInst *LoadInst = B.CreateLoad(PointerArg);
+    LoadInst *LoadInst = B.CreateLoad(T, PointerArg);
     C = B.getInt8(addVal);
     I = cast<Instruction>(B.CreateAdd(LoadInst, C));
     B.CreateRet(I);