[InferAddressSpaces] Support assumed addrspaces from addrspace predicates.

[hliao]

Summary

• CUDA cannot associate memory space with pointer types. Even though Clang could add extra attributes to specify the address space explicitly on a pointer type, it breaks the portability between Clang and NVCC.
• This change proposes to assume the address space from a pointer from the assumption built upon target-specific address space predicates, such as __isGlobal from CUDA. E.g.,

foo(float *p) {
  __builtin_assume(__isGlobal(p));
  // From there, we could assume p is a global pointer instead of a
  // generic one.
}

This makes the code portable without introducing the implementation-specific features.

Note that NVCC starts to support __builtin_assume from version 11.

[hliao]

Fix formatting.

[hliao]

Fix formatting again following clang-format.

[arsenm]

llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp

800–801

I disagree we need to do anything here, constant isn't a real address space. It would be cleaner if we didn't have it as a backend concept at all. We really just need global plus a slightly more contexts where memory can be marked as invariant. We also don't have a reasonable way we could implement this

llvm/lib/Transforms/Scalar/InferAddressSpaces.cpp

888

Does this really need the parent check? Doesn't isValidAssumeForContext understand the control flow?

[tra]

llvm/include/llvm/Analysis/AssumptionCache.h

109–110

Perhaps this should have a default of =nullptr and save us on spelling it out for most of the users who don't care about it.

llvm/lib/Transforms/Scalar/InferAddressSpaces.cpp

154

What do the values represent? It's worth a comment as it's not obvious.

Perhaps we should consolidate both maps into a struct or class. Plumbing each map separately through all the calls gets tedious.

202–203

I think this could've been a reference, too.

874–902

I can't say I'm happy about the way updateAddressSpace updates PredicateAS here, but delegates updates to InferredAddrSpace to the caller. I think both should be updated in one place -- either here, or in the callee.

[nikic]

Is it actually necessary to thread this through AssumptionCache, given how InferAddressSpaces is the only place that looks at these assumes?

[rampitec]

Is there anything to remove assume() call after address space is inferred? We do not need it anymore.

[hliao]

In D112041#3073560, @rampitec wrote:

Is there anything to remove assume() call after address space is inferred? We do not need it anymore.

along with a few other intrinsics, assume intrinsic is discarded in SDAG and GISel.

[rampitec]

In D112041#3073637, @hliao wrote:

In D112041#3073560, @rampitec wrote:

Is there anything to remove assume() call after address space is inferred? We do not need it anymore.

along with a few other intrinsics, assume intrinsic is discarded in SDAG and GISel.

We may want to discard these earlier for the sake of Value::hasOneUse(). These are really not needed after casts are inserted.

[hliao]

Revise following reviewers' feedback.

[hliao]

llvm/lib/Transforms/Scalar/InferAddressSpaces.cpp

202–203

yeah, but we should address that in another change.

874–902

the difference is that, here, we assume a use of pointer could be inferred with a new addrspace. The original is on a def of value.

[hliao]

Revise commit message.

[hliao]

In D112041#3073676, @rampitec wrote:

In D112041#3073637, @hliao wrote:

In D112041#3073560, @rampitec wrote:

Is there anything to remove assume() call after address space is inferred? We do not need it anymore.

along with a few other intrinsics, assume intrinsic is discarded in SDAG and GISel.

We may want to discard these earlier for the sake of Value::hasOneUse(). These are really not needed after casts are inserted.

That sounds reasonable but may face the limit due to the fact that we run addrspace inferring several times (if my memory is right, 3 times if the backend one is counted.) Among them, we expect pointer values are prompted from memory to register so that we could infer addrspace for them further. If we remove these assume intrinsic earlier, there would be the risk that later addrspace inferring may not be able to leverage those assumptions.
NVPTX runs that inferring just once. But after that, there would be too much optimizations at the IR level.

I checked the most hasOneUse() usage in IR passes. Most of them are not applied to pointer arithmetic. Only a few cases are applied to pointers but also have quite limited conditions. I would expect we may need to enhance them case by case if we found real cases where the extra use from assume intrinsic makes code quality worse.

[rampitec]

In D112041#3074418, @hliao wrote:

In D112041#3073676, @rampitec wrote:

In D112041#3073637, @hliao wrote:

In D112041#3073560, @rampitec wrote:

Is there anything to remove assume() call after address space is inferred? We do not need it anymore.

along with a few other intrinsics, assume intrinsic is discarded in SDAG and GISel.

We may want to discard these earlier for the sake of Value::hasOneUse(). These are really not needed after casts are inserted.

That sounds reasonable but may face the limit due to the fact that we run addrspace inferring several times (if my memory is right, 3 times if the backend one is counted.) Among them, we expect pointer values are prompted from memory to register so that we could infer addrspace for them further. If we remove these assume intrinsic earlier, there would be the risk that later addrspace inferring may not be able to leverage those assumptions.
NVPTX runs that inferring just once. But after that, there would be too much optimizations at the IR level.

I checked the most hasOneUse() usage in IR passes. Most of them are not applied to pointer arithmetic. Only a few cases are applied to pointers but also have quite limited conditions. I would expect we may need to enhance them case by case if we found real cases where the extra use from assume intrinsic makes code quality worse.

Thanks, that sounds reasonable. If we start seeing problems because of this we can always remove these later. Conceptually LGTM.

[tra]

llvm/lib/Transforms/Scalar/InferAddressSpaces.cpp

851–852

I.e. I'd move these lines into updateAddressSpace and change its return type to bool, as we no longer would need NewAS' value.

874–902

I'm not sure how it explains why the function returns values to update InferredAddrSpace outside of it, but updates PredicatedAS inside.

For consistency sake, I'd update InferredAddrSpace here as well and, possibly, combine both maps into a single structure as I've suggested above.

[hliao]

Sorry for the late reply.

llvm/lib/Transforms/Scalar/InferAddressSpaces.cpp

874–902

PredicatedAS holds only the deduced operand AS based on the assumption or other conditions. It may or may not help to infer its user's final AS. In case the user still cannot be inferred, PredicatedAS won't be used to change the final IR anyway. It's only an intermediate state during the inferring. However, I think it's a good idea to put relevant updates into updateAddressSpace. That seems a little bit cleaner.

[hliao]

Move the updating action into updateAddrSpace and return a boolean true if it's really updated.

[arsenm]

llvm/lib/Transforms/Scalar/InferAddressSpaces.cpp

240

The pass is already using UninitializedAddressSpace as a sentinal value; just use that instead of Optional<unsigned>?

957

*AS instead of getValue

[tra]

LGTM in general, modulo remaining nits.

llvm/lib/Transforms/Scalar/InferAddressSpaces.cpp

202–203

You're already changing function signatures.It would be reasonable to incorporate this cleanup, too.

[hliao]

Updated.

[nikic]

I'd still like an answer to:

In D112041#3073464, @nikic wrote:

Is it actually necessary to thread this through AssumptionCache, given how InferAddressSpaces is the only place that looks at these assumes?

I'd prefer not to infect AC with TTI if not necessary.

[hliao]

In D112041#3106936, @nikic wrote:

I'd still like an answer to:

In D112041#3073464, @nikic wrote:

Is it actually necessary to thread this through AssumptionCache, given how InferAddressSpaces is the only place that looks at these assumes?

I'd prefer not to infect AC with TTI if not necessary.

We need assume intrinsic to preserve those predicates through optimizations as infer-addrespace may run several times. With assume used the same way, we probably just duplicate the logic in AC if we want to track them inside this pass. That's not a good practice though.

[hliao]

Rebase and kindly ping for review.

[hliao]

llvm/lib/Transforms/Scalar/InferAddressSpaces.cpp

202–203

revised in the latest patch.

[hliao]

Rebase to the main branch.

[hliao]

Rebase and ping for further review.

[hliao]

Rebase and kindly PING for review.

clang/test/CodeGen/thinlto-distributed-newpm.ll

 ; CHECK-O: Running pass: IPSCCPPass
 ; CHECK-O: Running pass: CalledValuePropagationPass
 ; CHECK-O: Running pass: GlobalOptPass
 ; CHECK-O: Invalidating analysis: InnerAnalysisManagerProxy
 ; CHECK-O: Running analysis: InnerAnalysisManagerProxy
 ; CHECK-O: Running pass: PromotePass
 ; CHECK-O: Running analysis: DominatorTreeAnalysis on main
 ; CHECK-O: Running analysis: AssumptionAnalysis on main
+; CHECK-O: Running analysis: TargetIRAnalysis on main
 ; CHECK-O: Running pass: DeadArgumentEliminationPass
 ; CHECK-O: Running pass: InstCombinePass on main
 ; CHECK-O: Running analysis: TargetLibraryAnalysis on main
 ; CHECK-O: Running analysis: OptimizationRemarkEmitterAnalysis on main
-; CHECK-O: Running analysis: TargetIRAnalysis on main
 ; CHECK-O: Running analysis: AAManager on main
 ; CHECK-O: Running analysis: BasicAA on main
 ; CHECK-O: Running analysis: ScopedNoAliasAA on main
 ; CHECK-O: Running analysis: TypeBasedAA on main
 ; CHECK-O: Running analysis: OuterAnalysisManagerProxy
 ; CHECK-O: Running pass: SimplifyCFGPass on main
 ; CHECK-O: Running analysis: InnerAnalysisManagerProxy
 ; CHECK-O: Running analysis: LazyCallGraphAnalysis

llvm/include/llvm/Analysis/AssumptionCache.h

 #include "llvm/Pass.h"
 #include <memory>
 
 namespace llvm {
 
 class AssumeInst;
 class Function;
 class raw_ostream;
+class TargetTransformInfo;
 class Value;
 
 /// A cache of \@llvm.assume calls within a function.
 ///
 /// This cache provides fast lookup of assumptions within a function by caching
 /// them and amortizing the cost of scanning for them across all queries. Passes
 /// that create new assumptions are required to call registerAssumption() to
 /// register any new \@llvm.assume calls that they create. Deletions of
public:
   };
 
 private:
   /// The function for which this cache is handling assumptions.
   ///
   /// We track this to lazily populate our assumptions.
   Function &F;
 
+  TargetTransformInfo *TTI;
+
   /// Vector of weak value handles to calls of the \@llvm.assume
   /// intrinsic.
   SmallVector<ResultElem, 4> AssumeHandles;
 
   class AffectedValueCallbackVH final : public CallbackVH {
     AssumptionCache *AC;
 
     void deleted() override;
private:
   bool Scanned = false;
 
   /// Scan the function for assumptions and add them to the cache.
   void scanFunction();
 
 public:
   /// Construct an AssumptionCache from a function by scanning all of
   /// its instructions.
-  AssumptionCache(Function &F) : F(F) {}
+  AssumptionCache(Function &F, TargetTransformInfo *TTI = nullptr)
+      : F(F), TTI(TTI) {}
 
   /// This cache is designed to be self-updating and so it should never be
   /// invalidated.
   bool invalidate(Function &, const PreservedAnalyses &,
                   FunctionAnalysisManager::Invalidator &) {
     return false;
   }
 

 class AssumptionAnalysis : public AnalysisInfoMixin<AssumptionAnalysis> {
   friend AnalysisInfoMixin<AssumptionAnalysis>;
 
   static AnalysisKey Key;
 
 public:
   using Result = AssumptionCache;
 
-  AssumptionCache run(Function &F, FunctionAnalysisManager &) {
-    return AssumptionCache(F);
-  }
+  AssumptionCache run(Function &F, FunctionAnalysisManager &);
 };
 
 /// Printer pass for the \c AssumptionAnalysis results.
 class AssumptionPrinterPass : public PassInfoMixin<AssumptionPrinterPass> {
   raw_ostream &OS;
 
 public:
   explicit AssumptionPrinterPass(raw_ostream &OS) : OS(OS) {}

llvm/include/llvm/Analysis/TargetTransformInfo.h

 #include "llvm/IR/Operator.h"
 #include "llvm/IR/PassManager.h"
 #include "llvm/Pass.h"
 #include "llvm/Support/AtomicOrdering.h"
 #include "llvm/Support/BranchProbability.h"
 #include "llvm/Support/DataTypes.h"
 #include "llvm/Support/InstructionCost.h"
 #include <functional>
+#include <utility>
 
 namespace llvm {
 
 namespace Intrinsic {
 typedef unsigned ID;
 }
 
 class AssumptionCache;
public:
   bool isNoopAddrSpaceCast(unsigned FromAS, unsigned ToAS) const;
 
   /// Return true if globals in this address space can have initializers other
   /// than `undef`.
   bool canHaveNonUndefGlobalInitializerInAddressSpace(unsigned AS) const;
 
   unsigned getAssumedAddrSpace(const Value *V) const;
 
+  std::pair<const Value *, unsigned>
+  getPredicatedAddrSpace(const Value *V) const;
+
   /// Rewrite intrinsic call \p II such that \p OldV will be replaced with \p
   /// NewV, which has a different address space. This should happen for every
   /// operand index that collectFlatAddressOperands returned for the intrinsic.
   /// \returns nullptr if the intrinsic was not handled. Otherwise, returns the
   /// new value (which may be the original \p II with modified operands).
   Value *rewriteIntrinsicWithAddressSpace(IntrinsicInst *II, Value *OldV,
                                           Value *NewV) const;
 
public:
   virtual bool isAlwaysUniform(const Value *V) = 0;
   virtual unsigned getFlatAddressSpace() = 0;
   virtual bool collectFlatAddressOperands(SmallVectorImpl<int> &OpIndexes,
                                           Intrinsic::ID IID) const = 0;
   virtual bool isNoopAddrSpaceCast(unsigned FromAS, unsigned ToAS) const = 0;
   virtual bool
   canHaveNonUndefGlobalInitializerInAddressSpace(unsigned AS) const = 0;
   virtual unsigned getAssumedAddrSpace(const Value *V) const = 0;
+  virtual std::pair<const Value *, unsigned>
+  getPredicatedAddrSpace(const Value *V) const = 0;
   virtual Value *rewriteIntrinsicWithAddressSpace(IntrinsicInst *II,
                                                   Value *OldV,
                                                   Value *NewV) const = 0;
   virtual bool isLoweredToCall(const Function *F) = 0;
   virtual void getUnrollingPreferences(Loop *L, ScalarEvolution &,
                                        UnrollingPreferences &UP,
                                        OptimizationRemarkEmitter *ORE) = 0;
   virtual void getPeelingPreferences(Loop *L, ScalarEvolution &SE,
public:
   canHaveNonUndefGlobalInitializerInAddressSpace(unsigned AS) const override {
     return Impl.canHaveNonUndefGlobalInitializerInAddressSpace(AS);
   }
 
   unsigned getAssumedAddrSpace(const Value *V) const override {
     return Impl.getAssumedAddrSpace(V);
   }
 
+  std::pair<const Value *, unsigned>
+  getPredicatedAddrSpace(const Value *V) const override {
+    return Impl.getPredicatedAddrSpace(V);
+  }
+
   Value *rewriteIntrinsicWithAddressSpace(IntrinsicInst *II, Value *OldV,
                                           Value *NewV) const override {
     return Impl.rewriteIntrinsicWithAddressSpace(II, OldV, NewV);
   }
 
   bool isLoweredToCall(const Function *F) override {
     return Impl.isLoweredToCall(F);
   }

llvm/include/llvm/Analysis/TargetTransformInfoImpl.h

 #include "llvm/Analysis/VectorUtils.h"
 #include "llvm/IR/DataLayout.h"
 #include "llvm/IR/Function.h"
 #include "llvm/IR/GetElementPtrTypeIterator.h"
 #include "llvm/IR/IntrinsicInst.h"
 #include "llvm/IR/Operator.h"
 #include "llvm/IR/PatternMatch.h"
 #include "llvm/IR/Type.h"
+#include <utility>
 
 using namespace llvm::PatternMatch;
 
 namespace llvm {
 
 /// Base class for use as a mix-in that aids implementing
 /// a TargetTransformInfo-compatible class.
 class TargetTransformInfoImplBase {
public:
 
   bool isNoopAddrSpaceCast(unsigned, unsigned) const { return false; }
   bool canHaveNonUndefGlobalInitializerInAddressSpace(unsigned AS) const {
     return AS == 0;
   };
 
   unsigned getAssumedAddrSpace(const Value *V) const { return -1; }
 
+  std::pair<const Value *, unsigned>
+  getPredicatedAddrSpace(const Value *V) const {
+    return std::make_pair(nullptr, -1);
+  }
+
   Value *rewriteIntrinsicWithAddressSpace(IntrinsicInst *II, Value *OldV,
                                           Value *NewV) const {
     return nullptr;
   }
 
   bool isLoweredToCall(const Function *F) const {
     assert(F && "A concrete function must be provided to this routine.");
 

llvm/include/llvm/CodeGen/BasicTTIImpl.h

public:
   bool isNoopAddrSpaceCast(unsigned FromAS, unsigned ToAS) const {
     return getTLI()->getTargetMachine().isNoopAddrSpaceCast(FromAS, ToAS);
   }
 
   unsigned getAssumedAddrSpace(const Value *V) const {
     return getTLI()->getTargetMachine().getAssumedAddrSpace(V);
   }
 
+  std::pair<const Value *, unsigned>
+  getPredicatedAddrSpace(const Value *V) const {
+    return getTLI()->getTargetMachine().getPredicatedAddrSpace(V);
+  }
+
   Value *rewriteIntrinsicWithAddressSpace(IntrinsicInst *II, Value *OldV,
                                           Value *NewV) const {
     return nullptr;
   }
 
   bool isLegalAddImmediate(int64_t imm) {
     return getTLI()->isLegalAddImmediate(imm);
   }

llvm/include/llvm/Target/TargetMachine.h

 #include "llvm/IR/PassManager.h"
 #include "llvm/Pass.h"
 #include "llvm/Support/CodeGen.h"
 #include "llvm/Support/Error.h"
 #include "llvm/Support/PGOOptions.h"
 #include "llvm/Target/CGPassBuilderOption.h"
 #include "llvm/Target/TargetOptions.h"
 #include <string>
+#include <utility>
 
 namespace llvm {
 
 class AAManager;
 template <typename IRUnitT, typename AnalysisManagerT, typename... ExtraArgTs>
 class PassManager;
 using ModulePassManager = PassManager<Module>;
 
public:
   /// If the specified generic pointer could be assumed as a pointer to a
   /// specific address space, return that address space.
   ///
   /// Under offloading programming, the offloading target may be passed with
   /// values only prepared on the host side and could assume certain
   /// properties.
   virtual unsigned getAssumedAddrSpace(const Value *V) const { return -1; }
 
+  /// If the specified predicate checks whether a generic pointer falls within
+  /// a specified address space, return that generic pointer and the address
+  /// space being queried.
+  ///
+  /// Such predicates could be specified in @llvm.assume intrinsics for the
+  /// optimizer to assume that the given generic pointer always falls within
+  /// the address space based on that predicate.
+  virtual std::pair<const Value *, unsigned>
+  getPredicatedAddrSpace(const Value *V) const {
+    return std::make_pair(nullptr, -1);
+  }
+
   /// Get a \c TargetIRAnalysis appropriate for the target.
   ///
   /// This is used to construct the new pass manager's target IR analysis pass,
   /// set up appropriately for this target machine. Even the old pass manager
   /// uses this to answer queries about the IR.
   TargetIRAnalysis getTargetIRAnalysis();
 
   /// Return a TargetTransformInfo for a given function.

llvm/lib/Analysis/AssumptionCache.cpp

 //
 //===----------------------------------------------------------------------===//
 
 #include "llvm/Analysis/AssumeBundleQueries.h"
 #include "llvm/Analysis/AssumptionCache.h"
 #include "llvm/ADT/STLExtras.h"
 #include "llvm/ADT/SmallPtrSet.h"
 #include "llvm/ADT/SmallVector.h"
+#include "llvm/Analysis/TargetTransformInfo.h"
 #include "llvm/IR/BasicBlock.h"
 #include "llvm/IR/Function.h"
 #include "llvm/IR/InstrTypes.h"
 #include "llvm/IR/Instruction.h"
 #include "llvm/IR/Instructions.h"
 #include "llvm/IR/Intrinsics.h"
 #include "llvm/IR/PassManager.h"
 #include "llvm/IR/PatternMatch.h"
if (AVI != AffectedValues.end())
     return AVI->second;
 
   auto AVIP = AffectedValues.insert(
       {AffectedValueCallbackVH(V, this), SmallVector<ResultElem, 1>()});
   return AVIP.first->second;
 }
 
 static void
-findAffectedValues(CallBase *CI,
+findAffectedValues(CallBase *CI, TargetTransformInfo *TTI,
                    SmallVectorImpl<AssumptionCache::ResultElem> &Affected) {
   // Note: This code must be kept in-sync with the code in
   // computeKnownBitsFromAssume in ValueTracking.
 
   auto AddAffected = [&Affected](Value *V, unsigned Idx =
                                                AssumptionCache::ExprResultIdx) {
     if (isa<Argument>(V)) {
       Affected.push_back({V, Idx});
if (match(Cond, m_ICmp(Pred, m_Value(A), m_Value(B)))) {
     Value *X;
     // Handle (A + C1) u< C2, which is the canonical form of A > C3 && A < C4,
     // and recognized by LVI at least.
     if (Pred == ICmpInst::ICMP_ULT &&
         match(A, m_Add(m_Value(X), m_ConstantInt())) &&
         match(B, m_ConstantInt()))
       AddAffected(X);
   }
+
+  if (TTI) {
+    const Value *Ptr;
+    unsigned AS;
+    std::tie(Ptr, AS) = TTI->getPredicatedAddrSpace(Cond);
+    if (Ptr)
+      AddAffected(const_cast<Value *>(Ptr->stripInBoundsOffsets()));
+  }
 }
 
 void AssumptionCache::updateAffectedValues(AssumeInst *CI) {
   SmallVector<AssumptionCache::ResultElem, 16> Affected;
-  findAffectedValues(CI, Affected);
+  findAffectedValues(CI, TTI, Affected);
 
   for (auto &AV : Affected) {
     auto &AVV = getOrInsertAffectedValues(AV.Assume);
     if (llvm::none_of(AVV, [&](ResultElem &Elem) {
           return Elem.Assume == CI && Elem.Index == AV.Index;
         }))
       AVV.push_back({CI, AV.Index});
   }
 }
 
 void AssumptionCache::unregisterAssumption(AssumeInst *CI) {
   SmallVector<AssumptionCache::ResultElem, 16> Affected;
-  findAffectedValues(CI, Affected);
+  findAffectedValues(CI, TTI, Affected);
 
   for (auto &AV : Affected) {
     auto AVI = AffectedValues.find_as(AV.Assume);
     if (AVI == AffectedValues.end())
       continue;
     bool Found = false;
     bool HasNonnull = false;
     for (ResultElem &Elem : AVI->second) {
for (auto &VH : AssumeHandles) {
     assert(AssumptionSet.insert(VH).second &&
            "Cache contains multiple copies of a call!");
   }
 #endif
 
   updateAffectedValues(CI);
 }
 
+AssumptionCache AssumptionAnalysis::run(Function &F,
+                                        FunctionAnalysisManager &FAM) {
+  auto &TTI = FAM.getResult<TargetIRAnalysis>(F);
+  return AssumptionCache(F, &TTI);
+}
+
 AnalysisKey AssumptionAnalysis::Key;
 
 PreservedAnalyses AssumptionPrinterPass::run(Function &F,
                                              FunctionAnalysisManager &AM) {
   AssumptionCache &AC = AM.getResult<AssumptionAnalysis>(F);
 
   OS << "Cached assumptions for function: " << F.getName() << "\n";
   for (auto &VH : AC.assumptions())
AssumptionCache &AssumptionCacheTracker::getAssumptionCache(Function &F) {
   // We probe the function map twice to try and avoid creating a value handle
   // around the function in common cases. This makes insertion a bit slower,
   // but if we have to insert we're going to scan the whole function so that
   // shouldn't matter.
   auto I = AssumptionCaches.find_as(&F);
   if (I != AssumptionCaches.end())
     return *I->second;
 
+  auto *TTIWP = getAnalysisIfAvailable<TargetTransformInfoWrapperPass>();
+  auto *TTI = TTIWP ? &TTIWP->getTTI(F) : nullptr;
+
   // Ok, build a new cache by scanning the function, insert it and the value
   // handle into our map, and return the newly populated cache.
   auto IP = AssumptionCaches.insert(std::make_pair(
-      FunctionCallbackVH(&F, this), std::make_unique<AssumptionCache>(F)));
+      FunctionCallbackVH(&F, this), std::make_unique<AssumptionCache>(F, TTI)));
   assert(IP.second && "Scanning function already in the map?");
   return *IP.first->second;
 }
 
 AssumptionCache *AssumptionCacheTracker::lookupAssumptionCache(Function &F) {
   auto I = AssumptionCaches.find_as(&F);
   if (I != AssumptionCaches.end())
     return I->second.get();

llvm/lib/Analysis/TargetTransformInfo.cpp

bool TargetTransformInfo::canHaveNonUndefGlobalInitializerInAddressSpace(
     unsigned AS) const {
   return TTIImpl->canHaveNonUndefGlobalInitializerInAddressSpace(AS);
 }
 
 unsigned TargetTransformInfo::getAssumedAddrSpace(const Value *V) const {
   return TTIImpl->getAssumedAddrSpace(V);
 }
 
+std::pair<const Value *, unsigned>
+TargetTransformInfo::getPredicatedAddrSpace(const Value *V) const {
+  return TTIImpl->getPredicatedAddrSpace(V);
+}
+
 Value *TargetTransformInfo::rewriteIntrinsicWithAddressSpace(
     IntrinsicInst *II, Value *OldV, Value *NewV) const {
   return TTIImpl->rewriteIntrinsicWithAddressSpace(II, OldV, NewV);
 }
 
 bool TargetTransformInfo::isLoweredToCall(const Function *F) const {
   return TTIImpl->isLoweredToCall(F);
 }

llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.h

 //===----------------------------------------------------------------------===//
 
 #ifndef LLVM_LIB_TARGET_AMDGPU_AMDGPUTARGETMACHINE_H
 #define LLVM_LIB_TARGET_AMDGPU_AMDGPUTARGETMACHINE_H
 
 #include "GCNSubtarget.h"
 #include "llvm/CodeGen/TargetPassConfig.h"
 #include "llvm/Target/TargetMachine.h"
+#include <utility>
 
 namespace llvm {
 
 class ScheduleDAGMILive;
 
 //===----------------------------------------------------------------------===//
 // AMDGPU Target Machine (R600+)
 //===----------------------------------------------------------------------===//
public:
   void registerDefaultAliasAnalyses(AAManager &) override;
 
   /// Get the integer value of a null pointer in the given address space.
   static int64_t getNullPointerValue(unsigned AddrSpace);
 
   bool isNoopAddrSpaceCast(unsigned SrcAS, unsigned DestAS) const override;
 
   unsigned getAssumedAddrSpace(const Value *V) const override;
+
+  std::pair<const Value *, unsigned>
+  getPredicatedAddrSpace(const Value *V) const override;
 };
 
 //===----------------------------------------------------------------------===//
 // GCN Target Machine (SI+)
 //===----------------------------------------------------------------------===//
 
 class GCNTargetMachine final : public AMDGPUTargetMachine {
 private:

llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp

 #include "llvm/CodeGen/GlobalISel/InstructionSelect.h"
 #include "llvm/CodeGen/GlobalISel/Legalizer.h"
 #include "llvm/CodeGen/GlobalISel/Localizer.h"
 #include "llvm/CodeGen/GlobalISel/RegBankSelect.h"
 #include "llvm/CodeGen/MIRParser/MIParser.h"
 #include "llvm/CodeGen/Passes.h"
 #include "llvm/CodeGen/RegAllocRegistry.h"
 #include "llvm/CodeGen/TargetPassConfig.h"
+#include "llvm/IR/IntrinsicsAMDGPU.h"
 #include "llvm/IR/LegacyPassManager.h"
 #include "llvm/IR/PassManager.h"
+#include "llvm/IR/PatternMatch.h"
 #include "llvm/InitializePasses.h"
 #include "llvm/MC/TargetRegistry.h"
 #include "llvm/Passes/PassBuilder.h"
 #include "llvm/Transforms/IPO.h"
 #include "llvm/Transforms/IPO/AlwaysInliner.h"
 #include "llvm/Transforms/IPO/GlobalDCE.h"
 #include "llvm/Transforms/IPO/Internalize.h"
 #include "llvm/Transforms/IPO/PassManagerBuilder.h"
if (Ptr->getType()->getPointerAddressSpace() != AMDGPUAS::CONSTANT_ADDRESS)
     return AMDGPUAS::UNKNOWN_ADDRESS_SPACE;
   // For a generic pointer loaded from the constant memory, it could be assumed
   // as a global pointer since the constant memory is only populated on the
   // host side. As implied by the offload programming model, only global
   // pointers could be referenced on the host side.
   return AMDGPUAS::GLOBAL_ADDRESS;
 }
 
+std::pair<const Value *, unsigned>
+AMDGPUTargetMachine::getPredicatedAddrSpace(const Value *V) const {
+  if (auto *II = dyn_cast<IntrinsicInst>(V)) {
+    switch (II->getIntrinsicID()) {
+    case Intrinsic::amdgcn_is_shared:
+      return std::make_pair(II->getArgOperand(0), AMDGPUAS::LOCAL_ADDRESS);
+    case Intrinsic::amdgcn_is_private:
+      return std::make_pair(II->getArgOperand(0), AMDGPUAS::PRIVATE_ADDRESS);
+    default:
+      break;
+    }
+    return std::make_pair(nullptr, -1);
+  }
+  // Check the global pointer predication based on
+  // (!is_share(p) && !is_private(p)). Note that logic 'and' is commutative and
+  // the order of 'is_shared' and 'is_private' is not significant.
+  Value *Ptr;
+  if (match(
+          const_cast<Value *>(V),
+          m_c_And(m_Not(m_Intrinsic<Intrinsic::amdgcn_is_shared>(m_Value(Ptr))),
+                  m_Not(m_Intrinsic<Intrinsic::amdgcn_is_private>(
+                      m_Deferred(Ptr))))))
+    return std::make_pair(Ptr, AMDGPUAS::GLOBAL_ADDRESS);
+
+  return std::make_pair(nullptr, -1);
+}
+
 //===----------------------------------------------------------------------===//
 // GCN Target Machine (SI+)
 //===----------------------------------------------------------------------===//
 
 GCNTargetMachine::GCNTargetMachine(const Target &T, const Triple &TT,
                                    StringRef CPU, StringRef FS,
                                    TargetOptions Options,
                                    Optional<Reloc::Model> RM,

llvm/lib/Target/NVPTX/NVPTXTargetMachine.h

 //===----------------------------------------------------------------------===//
 
 #ifndef LLVM_LIB_TARGET_NVPTX_NVPTXTARGETMACHINE_H
 #define LLVM_LIB_TARGET_NVPTX_NVPTXTARGETMACHINE_H
 
 #include "ManagedStringPool.h"
 #include "NVPTXSubtarget.h"
 #include "llvm/Target/TargetMachine.h"
+#include <utility>
 
 namespace llvm {
 
 /// NVPTXTargetMachine
 ///
 class NVPTXTargetMachine : public LLVMTargetMachine {
   bool is64bit;
   // Use 32-bit pointers for accessing const/local/short AS.
public:
   void adjustPassManager(PassManagerBuilder &) override;
   void registerPassBuilderCallbacks(PassBuilder &PB) override;
 
   TargetTransformInfo getTargetTransformInfo(const Function &F) override;
 
   bool isMachineVerifierClean() const override {
     return false;
   }
+
+  std::pair<const Value *, unsigned>
+  getPredicatedAddrSpace(const Value *V) const override;
 }; // NVPTXTargetMachine.
 
 class NVPTXTargetMachine32 : public NVPTXTargetMachine {
   virtual void anchor();
 public:
   NVPTXTargetMachine32(const Target &T, const Triple &TT, StringRef CPU,
                        StringRef FS, const TargetOptions &Options,
                        Optional<Reloc::Model> RM, Optional<CodeModel::Model> CM,

llvm/lib/Target/NVPTX/NVPTXTargetMachine.cpp

 #include "NVPTXTargetObjectFile.h"
 #include "NVPTXTargetTransformInfo.h"
 #include "TargetInfo/NVPTXTargetInfo.h"
 #include "llvm/ADT/STLExtras.h"
 #include "llvm/ADT/Triple.h"
 #include "llvm/Analysis/TargetTransformInfo.h"
 #include "llvm/CodeGen/Passes.h"
 #include "llvm/CodeGen/TargetPassConfig.h"
+#include "llvm/IR/IntrinsicsNVPTX.h"
 #include "llvm/IR/LegacyPassManager.h"
 #include "llvm/MC/TargetRegistry.h"
 #include "llvm/Pass.h"
 #include "llvm/Passes/PassBuilder.h"
 #include "llvm/Support/CommandLine.h"
 #include "llvm/Target/TargetMachine.h"
 #include "llvm/Target/TargetOptions.h"
 #include "llvm/Transforms/IPO/PassManagerBuilder.h"
PB.registerPipelineStartEPCallback(
       });
 }
 
 TargetTransformInfo
 NVPTXTargetMachine::getTargetTransformInfo(const Function &F) {
   return TargetTransformInfo(NVPTXTTIImpl(this, F));
 }
 
+std::pair<const Value *, unsigned>
+NVPTXTargetMachine::getPredicatedAddrSpace(const Value *V) const {
+  if (auto *II = dyn_cast<IntrinsicInst>(V)) {
+    switch (II->getIntrinsicID()) {
+    case Intrinsic::nvvm_isspacep_const:
+      return std::make_pair(II->getArgOperand(0), llvm::ADDRESS_SPACE_CONST);
+    case Intrinsic::nvvm_isspacep_global:
+      return std::make_pair(II->getArgOperand(0), llvm::ADDRESS_SPACE_GLOBAL);
+    case Intrinsic::nvvm_isspacep_local:
+      return std::make_pair(II->getArgOperand(0), llvm::ADDRESS_SPACE_LOCAL);
+    case Intrinsic::nvvm_isspacep_shared:
+      return std::make_pair(II->getArgOperand(0), llvm::ADDRESS_SPACE_SHARED);
+    default:
+      break;
+    }
+  }
+  return std::make_pair(nullptr, -1);
+}
+
 void NVPTXPassConfig::addEarlyCSEOrGVNPass() {
   if (getOptLevel() == CodeGenOpt::Aggressive)
     addPass(createGVNPass());
   else
     addPass(createEarlyCSEPass());
 }
 
 void NVPTXPassConfig::addAddressSpaceInferencePasses() {

llvm/lib/Transforms/Scalar/InferAddressSpaces.cpp

 #include "llvm/Transforms/Scalar/InferAddressSpaces.h"
 #include "llvm/ADT/ArrayRef.h"
 #include "llvm/ADT/DenseMap.h"
 #include "llvm/ADT/DenseSet.h"
 #include "llvm/ADT/None.h"
 #include "llvm/ADT/Optional.h"
 #include "llvm/ADT/SetVector.h"
 #include "llvm/ADT/SmallVector.h"
+#include "llvm/Analysis/AssumptionCache.h"
 #include "llvm/Analysis/TargetTransformInfo.h"
+#include "llvm/Analysis/ValueTracking.h"
 #include "llvm/IR/BasicBlock.h"
 #include "llvm/IR/Constant.h"
 #include "llvm/IR/Constants.h"
+#include "llvm/IR/Dominators.h"
 #include "llvm/IR/Function.h"
 #include "llvm/IR/IRBuilder.h"
 #include "llvm/IR/InstIterator.h"
 #include "llvm/IR/Instruction.h"
 #include "llvm/IR/Instructions.h"
 #include "llvm/IR/IntrinsicInst.h"
 #include "llvm/IR/Intrinsics.h"
 #include "llvm/IR/LLVMContext.h"
 #include "llvm/IR/Operator.h"
 #include "llvm/IR/PassManager.h"
 #include "llvm/IR/Type.h"
 #include "llvm/IR/Use.h"
 #include "llvm/IR/User.h"
 #include "llvm/IR/Value.h"
 #include "llvm/IR/ValueHandle.h"
+#include "llvm/InitializePasses.h"
 #include "llvm/Pass.h"
 #include "llvm/Support/Casting.h"
 #include "llvm/Support/CommandLine.h"
 #include "llvm/Support/Compiler.h"
 #include "llvm/Support/Debug.h"
 #include "llvm/Support/ErrorHandling.h"
 #include "llvm/Support/raw_ostream.h"
 #include "llvm/Transforms/Scalar.h"
cl::desc("The default address space is assumed as the flat address space. "
              "This is mainly for test purpose."));
 
 static const unsigned UninitializedAddressSpace =
     std::numeric_limits<unsigned>::max();
 
 namespace {
 
 using ValueToAddrSpaceMapTy = DenseMap<const Value *, unsigned>;
+// Different from ValueToAddrSpaceMapTy, where a new addrspace is inferred on
+// the *def* of a value, PredicatedAddrSpaceMapTy is map where a new
+// addrspace is inferred on the *use* of a pointer. This map is introduced to
+// infer addrspace from the addrspace predicate assumption built from assume
+// intrinsic. In that scenario, only specific uses (under valid assumption
+// context) could be inferred with a new addrspace.
+using PredicatedAddrSpaceMapTy =
+    DenseMap<std::pair<const Value *, const Value *>, unsigned>;
 using PostorderStackTy = llvm::SmallVector<PointerIntPair<Value *, 1, bool>, 4>;
 
 class InferAddressSpaces : public FunctionPass {
   unsigned FlatAddrSpace = 0;
 
 public:
   static char ID;
 
   InferAddressSpaces() :
     FunctionPass(ID), FlatAddrSpace(UninitializedAddressSpace) {}
   InferAddressSpaces(unsigned AS) : FunctionPass(ID), FlatAddrSpace(AS) {}
 
   void getAnalysisUsage(AnalysisUsage &AU) const override {
     AU.setPreservesCFG();
+    AU.addPreserved<DominatorTreeWrapperPass>();
+    AU.addRequired<AssumptionCacheTracker>();
     AU.addRequired<TargetTransformInfoWrapperPass>();
   }
 
   bool runOnFunction(Function &F) override;
 };
 
 class InferAddressSpacesImpl {
+  AssumptionCache &AC;
+  DominatorTree *DT = nullptr;
   const TargetTransformInfo *TTI = nullptr;
   const DataLayout *DL = nullptr;
 
   /// Target specific address space which uses of should be replaced if
   /// possible.
   unsigned FlatAddrSpace = 0;
 
-  // Returns the new address space of V if updated; otherwise, returns None.
-  Optional<unsigned>
-  updateAddressSpace(const Value &V,
-                     const ValueToAddrSpaceMapTy &InferredAddrSpace) const;
+  // Try to update the address space of V. If V is updated, returns true and
+  // false otherwise.
+  bool updateAddressSpace(const Value &V,
+                          ValueToAddrSpaceMapTy &InferredAddrSpace,
+                          PredicatedAddrSpaceMapTy &PredicatedAS) const;
 
   // Tries to infer the specific address space of each address expression in
   // Postorder.
   void inferAddressSpaces(ArrayRef<WeakTrackingVH> Postorder,
-                          ValueToAddrSpaceMapTy *InferredAddrSpace) const;
+                          ValueToAddrSpaceMapTy &InferredAddrSpace,
+                          PredicatedAddrSpaceMapTy &PredicatedAS) const;
 
   bool isSafeToCastConstAddrSpace(Constant *C, unsigned NewAS) const;
 
   Value *cloneInstructionWithNewAddressSpace(
       Instruction *I, unsigned NewAddrSpace,
       const ValueToValueMapTy &ValueWithNewAddrSpace,
+      const PredicatedAddrSpaceMapTy &PredicatedAS,
       SmallVectorImpl<const Use *> *UndefUsesToFix) const;
 
   // Changes the flat address expressions in function F to point to specific
   // address spaces if InferredAddrSpace says so. Postorder is the postorder of
   // all flat expressions in the use-def graph of function F.
   bool rewriteWithNewAddressSpaces(
       const TargetTransformInfo &TTI, ArrayRef<WeakTrackingVH> Postorder,
-      const ValueToAddrSpaceMapTy &InferredAddrSpace, Function *F) const;
+      const ValueToAddrSpaceMapTy &InferredAddrSpace,
+      const PredicatedAddrSpaceMapTy &PredicatedAS, Function *F) const;
 
   void appendsFlatAddressExpressionToPostorderStack(
       Value *V, PostorderStackTy &PostorderStack,
       DenseSet<Value *> &Visited) const;
 
   bool rewriteIntrinsicOperands(IntrinsicInst *II,
                                 Value *OldV, Value *NewV) const;
   void collectRewritableIntrinsicOperands(IntrinsicInst *II,
                                           PostorderStackTy &PostorderStack,
                                           DenseSet<Value *> &Visited) const;
 
   std::vector<WeakTrackingVH> collectFlatAddressExpressions(Function &F) const;
 
   Value *cloneValueWithNewAddressSpace(
       Value *V, unsigned NewAddrSpace,
       const ValueToValueMapTy &ValueWithNewAddrSpace,
+      const PredicatedAddrSpaceMapTy &PredicatedAS,
       SmallVectorImpl<const Use *> *UndefUsesToFix) const;
   unsigned joinAddressSpaces(unsigned AS1, unsigned AS2) const;
 
+  unsigned getPredicatedAddrSpace(const Value &V, Value *Opnd) const;
+
 public:
-  InferAddressSpacesImpl(const TargetTransformInfo *TTI, unsigned FlatAddrSpace)
-      : TTI(TTI), FlatAddrSpace(FlatAddrSpace) {}
+  InferAddressSpacesImpl(AssumptionCache &AC, DominatorTree *DT,
+                         const TargetTransformInfo *TTI, unsigned FlatAddrSpace)
+      : AC(AC), DT(DT), TTI(TTI), FlatAddrSpace(FlatAddrSpace) {}
   bool run(Function &F);
 };
 
 } // end anonymous namespace
 
 char InferAddressSpaces::ID = 0;
 
 namespace llvm {
 
 void initializeInferAddressSpacesPass(PassRegistry &);
 
 } // end namespace llvm
 
-INITIALIZE_PASS(InferAddressSpaces, DEBUG_TYPE, "Infer address spaces",
+INITIALIZE_PASS_BEGIN(InferAddressSpaces, DEBUG_TYPE, "Infer address spaces",
+                      false, false)
+INITIALIZE_PASS_DEPENDENCY(AssumptionCacheTracker)
+INITIALIZE_PASS_DEPENDENCY(TargetTransformInfoWrapperPass)
+INITIALIZE_PASS_END(InferAddressSpaces, DEBUG_TYPE, "Infer address spaces",
                     false, false)
 
 // Check whether that's no-op pointer bicast using a pair of
 // `ptrtoint`/`inttoptr` due to the missing no-op pointer bitcast over
 // different address spaces.
 static bool isNoopPtrIntCastPair(const Operator *I2P, const DataLayout &DL,
                                  const TargetTransformInfo *TTI) {
   assert(I2P->getOpcode() == Instruction::IntToPtr);
   auto *P2I = dyn_cast<Operator>(I2P->getOperand(0));

 }
 
 // A helper function for cloneInstructionWithNewAddressSpace. Returns the clone
 // of OperandUse.get() in the new address space. If the clone is not ready yet,
 // returns an undef in the new address space as a placeholder.
 static Value *operandWithNewAddressSpaceOrCreateUndef(
     const Use &OperandUse, unsigned NewAddrSpace,
     const ValueToValueMapTy &ValueWithNewAddrSpace,
+    const PredicatedAddrSpaceMapTy &PredicatedAS,
     SmallVectorImpl<const Use *> *UndefUsesToFix) {
   Value *Operand = OperandUse.get();
 
   Type *NewPtrTy = PointerType::getWithSamePointeeType(
       cast<PointerType>(Operand->getType()), NewAddrSpace);
 
   if (Constant *C = dyn_cast<Constant>(Operand))
     return ConstantExpr::getAddrSpaceCast(C, NewPtrTy);
 
   if (Value *NewOperand = ValueWithNewAddrSpace.lookup(Operand))
     return NewOperand;
 
+  Instruction *Inst = cast<Instruction>(OperandUse.getUser());
+  auto I = PredicatedAS.find(std::make_pair(Inst, Operand));
+  if (I != PredicatedAS.end()) {
+    // Insert an addrspacecast on that operand before the user.
+    unsigned NewAS = I->second;
+    Type *NewPtrTy = PointerType::getWithSamePointeeType(
+        cast<PointerType>(Operand->getType()), NewAS);
+    auto *NewI = new AddrSpaceCastInst(Operand, NewPtrTy);
+    NewI->insertBefore(Inst);
+    return NewI;
+  }
+
   UndefUsesToFix->push_back(&OperandUse);
   return UndefValue::get(NewPtrTy);
 }
 
 // Returns a clone of `I` with its operands converted to those specified in
 // ValueWithNewAddrSpace. Due to potential cycles in the data flow graph, an
 // operand whose address space needs to be modified might not exist in
 // ValueWithNewAddrSpace. In that case, uses undef as a placeholder operand and
 // adds that operand use to UndefUsesToFix so that caller can fix them later.
 //
 // Note that we do not necessarily clone `I`, e.g., if it is an addrspacecast
 // from a pointer whose type already matches. Therefore, this function returns a
 // Value* instead of an Instruction*.
 //
 // This may also return nullptr in the case the instruction could not be
 // rewritten.
 Value *InferAddressSpacesImpl::cloneInstructionWithNewAddressSpace(
     Instruction *I, unsigned NewAddrSpace,
     const ValueToValueMapTy &ValueWithNewAddrSpace,
+    const PredicatedAddrSpaceMapTy &PredicatedAS,
     SmallVectorImpl<const Use *> *UndefUsesToFix) const {
   Type *NewPtrType = PointerType::getWithSamePointeeType(
       cast<PointerType>(I->getType()), NewAddrSpace);
 
   if (I->getOpcode() == Instruction::AddrSpaceCast) {
     Value *Src = I->getOperand(0);
     // Because `I` is flat, the source address space must be specific.
     // Therefore, the inferred address space must be the source space, according
     // to our algorithm.
     assert(Src->getType()->getPointerAddressSpace() == NewAddrSpace);
     if (Src->getType() != NewPtrType)
       return new BitCastInst(Src, NewPtrType);
     return Src;
   }
 
   if (IntrinsicInst *II = dyn_cast<IntrinsicInst>(I)) {
     // Technically the intrinsic ID is a pointer typed argument, so specially
     // handle calls early.
     assert(II->getIntrinsicID() == Intrinsic::ptrmask);
     Value *NewPtr = operandWithNewAddressSpaceOrCreateUndef(
         II->getArgOperandUse(0), NewAddrSpace, ValueWithNewAddrSpace,
-        UndefUsesToFix);
+        PredicatedAS, UndefUsesToFix);
     Value *Rewrite =
         TTI->rewriteIntrinsicWithAddressSpace(II, II->getArgOperand(0), NewPtr);
     if (Rewrite) {
       assert(Rewrite != II && "cannot modify this pointer operation in place");
       return Rewrite;
     }
 
     return nullptr;
Value *InferAddressSpacesImpl::cloneInstructionWithNewAddressSpace(
 
   // Computes the converted pointer operands.
   SmallVector<Value *, 4> NewPointerOperands;
   for (const Use &OperandUse : I->operands()) {
     if (!OperandUse.get()->getType()->isPointerTy())
       NewPointerOperands.push_back(nullptr);
     else
       NewPointerOperands.push_back(operandWithNewAddressSpaceOrCreateUndef(
-                                     OperandUse, NewAddrSpace, ValueWithNewAddrSpace, UndefUsesToFix));
+          OperandUse, NewAddrSpace, ValueWithNewAddrSpace, PredicatedAS,
+          UndefUsesToFix));
   }
 
   switch (I->getOpcode()) {
   case Instruction::BitCast:
     return new BitCastInst(NewPointerOperands[0], NewPtrType);
   case Instruction::PHI: {
     assert(I->getType()->isPointerTy());
     PHINode *PHI = cast<PHINode>(I);

 // Returns a clone of the value `V`, with its operands replaced as specified in
 // ValueWithNewAddrSpace. This function is called on every flat address
 // expression whose address space needs to be modified, in postorder.
 //
 // See cloneInstructionWithNewAddressSpace for the meaning of UndefUsesToFix.
 Value *InferAddressSpacesImpl::cloneValueWithNewAddressSpace(
     Value *V, unsigned NewAddrSpace,
     const ValueToValueMapTy &ValueWithNewAddrSpace,
+    const PredicatedAddrSpaceMapTy &PredicatedAS,
     SmallVectorImpl<const Use *> *UndefUsesToFix) const {
   // All values in Postorder are flat address expressions.
   assert(V->getType()->getPointerAddressSpace() == FlatAddrSpace &&
          isAddressExpression(*V, *DL, TTI));
 
   if (Instruction *I = dyn_cast<Instruction>(V)) {
     Value *NewV = cloneInstructionWithNewAddressSpace(
-      I, NewAddrSpace, ValueWithNewAddrSpace, UndefUsesToFix);
+        I, NewAddrSpace, ValueWithNewAddrSpace, PredicatedAS, UndefUsesToFix);
     if (Instruction *NewI = dyn_cast_or_null<Instruction>(NewV)) {
       if (NewI->getParent() == nullptr) {
         NewI->insertBefore(I);
         NewI->takeName(I);
       }
     }
     return NewV;
   }
bool InferAddressSpacesImpl::run(Function &F) {
   }
 
   // Collects all flat address expressions in postorder.
   std::vector<WeakTrackingVH> Postorder = collectFlatAddressExpressions(F);
 
   // Runs a data-flow analysis to refine the address spaces of every expression
   // in Postorder.
   ValueToAddrSpaceMapTy InferredAddrSpace;
-  inferAddressSpaces(Postorder, &InferredAddrSpace);
+  PredicatedAddrSpaceMapTy PredicatedAS;
+  inferAddressSpaces(Postorder, InferredAddrSpace, PredicatedAS);
 
   // Changes the address spaces of the flat address expressions who are inferred
   // to point to a specific address space.
-  return rewriteWithNewAddressSpaces(*TTI, Postorder, InferredAddrSpace, &F);
+  return rewriteWithNewAddressSpaces(*TTI, Postorder, InferredAddrSpace,
+                                     PredicatedAS, &F);
 }
 
 // Constants need to be tracked through RAUW to handle cases with nested
 // constant expressions, so wrap values in WeakTrackingVH.
 void InferAddressSpacesImpl::inferAddressSpaces(
     ArrayRef<WeakTrackingVH> Postorder,
-    ValueToAddrSpaceMapTy *InferredAddrSpace) const {
+    ValueToAddrSpaceMapTy &InferredAddrSpace,
+    PredicatedAddrSpaceMapTy &PredicatedAS) const {
   SetVector<Value *> Worklist(Postorder.begin(), Postorder.end());
   // Initially, all expressions are in the uninitialized address space.
   for (Value *V : Postorder)
-    (*InferredAddrSpace)[V] = UninitializedAddressSpace;
+    InferredAddrSpace[V] = UninitializedAddressSpace;
 
   while (!Worklist.empty()) {
     Value *V = Worklist.pop_back_val();
 
-    // Tries to update the address space of the stack top according to the
+    // Try to update the address space of the stack top according to the
     // address spaces of its operands.
-    LLVM_DEBUG(dbgs() << "Updating the address space of\n  " << *V << '\n');
-    Optional<unsigned> NewAS = updateAddressSpace(*V, *InferredAddrSpace);
-    if (!NewAS.hasValue())
+    if (!updateAddressSpace(*V, InferredAddrSpace, PredicatedAS))
       continue;
-    // If any updates are made, grabs its users to the worklist because
-    // their address spaces can also be possibly updated.
-    LLVM_DEBUG(dbgs() << "  to " << NewAS.getValue() << '\n');
-    (*InferredAddrSpace)[V] = NewAS.getValue();
 
     for (Value *User : V->users()) {
       // Skip if User is already in the worklist.
       if (Worklist.count(User))
         continue;
 
-      auto Pos = InferredAddrSpace->find(User);
+      auto Pos = InferredAddrSpace.find(User);
       // Our algorithm only updates the address spaces of flat address
       // expressions, which are those in InferredAddrSpace.
-      if (Pos == InferredAddrSpace->end())
+      if (Pos == InferredAddrSpace.end())
         continue;
 
       // Function updateAddressSpace moves the address space down a lattice
       // path. Therefore, nothing to do if User is already inferred as flat (the
       // bottom element in the lattice).
       if (Pos->second == FlatAddrSpace)
         continue;
 
       Worklist.insert(User);
     }
   }
 }
 
-Optional<unsigned> InferAddressSpacesImpl::updateAddressSpace(
-    const Value &V, const ValueToAddrSpaceMapTy &InferredAddrSpace) const {
+unsigned InferAddressSpacesImpl::getPredicatedAddrSpace(const Value &V,
+                                                        Value *Opnd) const {
+  const Instruction *I = dyn_cast<Instruction>(&V);
+  if (!I)
+    return UninitializedAddressSpace;
+
+  Opnd = Opnd->stripInBoundsOffsets();
+  for (auto &AssumeVH : AC.assumptionsFor(Opnd)) {
+    if (!AssumeVH)
+      continue;
+    CallInst *CI = cast<CallInst>(AssumeVH);
+    if (!isValidAssumeForContext(CI, I, DT))
+      continue;
+
+    const Value *Ptr;
+    unsigned AS;
+    std::tie(Ptr, AS) = TTI->getPredicatedAddrSpace(CI->getArgOperand(0));
+    if (Ptr)
+      return AS;
+  }
+
+  return UninitializedAddressSpace;
+}
+
+bool InferAddressSpacesImpl::updateAddressSpace(
+    const Value &V, ValueToAddrSpaceMapTy &InferredAddrSpace,
+    PredicatedAddrSpaceMapTy &PredicatedAS) const {
   assert(InferredAddrSpace.count(&V));
 
+  LLVM_DEBUG(dbgs() << "Updating the address space of\n  " << V << '\n');
+
   // The new inferred address space equals the join of the address spaces
   // of all its pointer operands.
   unsigned NewAS = UninitializedAddressSpace;
 
   const Operator &Op = cast<Operator>(V);
   if (Op.getOpcode() == Instruction::Select) {
     Value *Src0 = Op.getOperand(1);
     Value *Src1 = Op.getOperand(2);
if (Op.getOpcode() == Instruction::Select) {
     auto *C0 = dyn_cast<Constant>(Src0);
     auto *C1 = dyn_cast<Constant>(Src1);
 
     // If one of the inputs is a constant, we may be able to do a constant
     // addrspacecast of it. Defer inferring the address space until the input
     // address space is known.
     if ((C1 && Src0AS == UninitializedAddressSpace) ||
         (C0 && Src1AS == UninitializedAddressSpace))
-      return None;
+      return false;
 
     if (C0 && isSafeToCastConstAddrSpace(C0, Src1AS))
       NewAS = Src1AS;
     else if (C1 && isSafeToCastConstAddrSpace(C1, Src0AS))
       NewAS = Src0AS;
     else
       NewAS = joinAddressSpaces(Src0AS, Src1AS);
   } else {
     unsigned AS = TTI->getAssumedAddrSpace(&V);
     if (AS != UninitializedAddressSpace) {
       // Use the assumed address space directly.
       NewAS = AS;
     } else {
       // Otherwise, infer the address space from its pointer operands.
       for (Value *PtrOperand : getPointerOperands(V, *DL, TTI)) {
         auto I = InferredAddrSpace.find(PtrOperand);
-        unsigned OperandAS =
-            I != InferredAddrSpace.end()
-                ? I->second
-                : PtrOperand->getType()->getPointerAddressSpace();
+        unsigned OperandAS;
+        if (I == InferredAddrSpace.end()) {
+          OperandAS = PtrOperand->getType()->getPointerAddressSpace();
+          if (OperandAS == FlatAddrSpace) {
+            // Check AC for assumption dominating V.
+            unsigned AS = getPredicatedAddrSpace(V, PtrOperand);
+            if (AS != UninitializedAddressSpace) {
+              LLVM_DEBUG(dbgs()
+                         << "  deduce operand AS from the predicate addrspace "
+                         << AS << '\n');
+              OperandAS = AS;
+              // Record this use with the predicated AS.
+              PredicatedAS[std::make_pair(&V, PtrOperand)] = OperandAS;
+            }
+          }
+        } else
+          OperandAS = I->second;
 
         // join(flat, *) = flat. So we can break if NewAS is already flat.
         NewAS = joinAddressSpaces(NewAS, OperandAS);
         if (NewAS == FlatAddrSpace)
           break;
       }
     }
   }
 
   unsigned OldAS = InferredAddrSpace.lookup(&V);
   assert(OldAS != FlatAddrSpace);
   if (OldAS == NewAS)
-    return None;
-  return NewAS;
+    return false;
+
+  // If any updates are made, grabs its users to the worklist because
+  // their address spaces can also be possibly updated.
+  LLVM_DEBUG(dbgs() << "  to " << NewAS << '\n');
+  InferredAddrSpace[&V] = NewAS;
+  return true;
 }
 
 /// \p returns true if \p U is the pointer operand of a memory instruction with
 /// a single pointer operand that can have its address space changed by simply
 /// mutating the use to a new value. If the memory instruction is volatile,
 /// return true only if the target allows the memory instruction to be volatile
 /// in the new address space.
 static bool isSimplePointerUseValidToReplace(const TargetTransformInfo &TTI,
static Value::use_iterator skipToNextUser(Value::use_iterator I,
   while (I != End && I->getUser() == CurUser)
     ++I;
 
   return I;
 }
 
 bool InferAddressSpacesImpl::rewriteWithNewAddressSpaces(
     const TargetTransformInfo &TTI, ArrayRef<WeakTrackingVH> Postorder,
-    const ValueToAddrSpaceMapTy &InferredAddrSpace, Function *F) const {
+    const ValueToAddrSpaceMapTy &InferredAddrSpace,
+    const PredicatedAddrSpaceMapTy &PredicatedAS, Function *F) const {
   // For each address expression to be modified, creates a clone of it with its
   // pointer operands converted to the new address space. Since the pointer
   // operands are converted, the clone is naturally in the new address space by
   // construction.
   ValueToValueMapTy ValueWithNewAddrSpace;
   SmallVector<const Use *, 32> UndefUsesToFix;
   for (Value* V : Postorder) {
     unsigned NewAddrSpace = InferredAddrSpace.lookup(V);
 
     // In some degenerate cases (e.g. invalid IR in unreachable code), we may
     // not even infer the value to have its original address space.
     if (NewAddrSpace == UninitializedAddressSpace)
       continue;
 
     if (V->getType()->getPointerAddressSpace() != NewAddrSpace) {
-      Value *New = cloneValueWithNewAddressSpace(
-          V, NewAddrSpace, ValueWithNewAddrSpace, &UndefUsesToFix);
+      Value *New =
+          cloneValueWithNewAddressSpace(V, NewAddrSpace, ValueWithNewAddrSpace,
+                                        PredicatedAS, &UndefUsesToFix);
       if (New)
         ValueWithNewAddrSpace[V] = New;
     }
   }
 
   if (ValueWithNewAddrSpace.empty())
     return false;
 
bool InferAddressSpacesImpl::rewriteWithNewAddressSpaces(
 
   return true;
 }
 
 bool InferAddressSpaces::runOnFunction(Function &F) {
   if (skipFunction(F))
     return false;
 
+  auto *DTWP = getAnalysisIfAvailable<DominatorTreeWrapperPass>();
+  DominatorTree *DT = DTWP ? &DTWP->getDomTree() : nullptr;
   return InferAddressSpacesImpl(
+             getAnalysis<AssumptionCacheTracker>().getAssumptionCache(F), DT,
              &getAnalysis<TargetTransformInfoWrapperPass>().getTTI(F),
              FlatAddrSpace)
       .run(F);
 }
 
 FunctionPass *llvm::createInferAddressSpacesPass(unsigned AddressSpace) {
   return new InferAddressSpaces(AddressSpace);
 }
 
 InferAddressSpacesPass::InferAddressSpacesPass()
     : FlatAddrSpace(UninitializedAddressSpace) {}
 InferAddressSpacesPass::InferAddressSpacesPass(unsigned AddressSpace)
     : FlatAddrSpace(AddressSpace) {}
 
 PreservedAnalyses InferAddressSpacesPass::run(Function &F,
                                               FunctionAnalysisManager &AM) {
   bool Changed =
-      InferAddressSpacesImpl(&AM.getResult<TargetIRAnalysis>(F), FlatAddrSpace)
+      InferAddressSpacesImpl(AM.getResult<AssumptionAnalysis>(F),
+                             AM.getCachedResult<DominatorTreeAnalysis>(F),
+                             &AM.getResult<TargetIRAnalysis>(F), FlatAddrSpace)
           .run(F);
   if (Changed) {
     PreservedAnalyses PA;
     PA.preserveSet<CFGAnalyses>();
+    PA.preserve<DominatorTreeAnalysis>();
     return PA;
   }
   return PreservedAnalyses::all();
 }

llvm/test/Other/loop-pm-invalidation.ll

 ; RUN:     -passes='loop(no-op-loop,loop-deletion),invalidate<scalar-evolution>,loop(no-op-loop)' \
 ; RUN:     | FileCheck %s --check-prefix=CHECK-SCEV-INV-AFTER-DELETE
 
 define void @no_loops() {
 ; CHECK-LOOP-INV: Running pass: LoopSimplifyPass
 ; CHECK-LOOP-INV-NEXT: Running analysis: LoopAnalysis
 ; CHECK-LOOP-INV-NEXT: Running analysis: DominatorTreeAnalysis
 ; CHECK-LOOP-INV-NEXT: Running analysis: AssumptionAnalysis
+; CHECK-LOOP-INV-NEXT: Running analysis: TargetIRAnalysis
 ; CHECK-LOOP-INV-NEXT: Running pass: LCSSAPass
 ; CHECK-LOOP-INV-NEXT: Running pass: InvalidateAnalysisPass<{{.*}}LoopAnalysis
 ; CHECK-LOOP-INV-NEXT: Invalidating analysis: LoopAnalysis
 ; CHECK-LOOP-INV-NEXT: Running pass: LoopSimplifyPass
 ; CHECK-LOOP-INV-NEXT: Running analysis: LoopAnalysis
 ; CHECK-LOOP-INV-NEXT: Running pass: LCSSAPass
 ;
 ; CHECK-SCEV-INV: Running pass: LoopSimplifyPass
 ; CHECK-SCEV-INV-NEXT: Running analysis: LoopAnalysis
 ; CHECK-SCEV-INV-NEXT: Running analysis: DominatorTreeAnalysis
 ; CHECK-SCEV-INV-NEXT: Running analysis: AssumptionAnalysis
+; CHECK-SCEV-INV-NEXT: Running analysis: TargetIRAnalysis
 ; CHECK-SCEV-INV-NEXT: Running pass: LCSSAPass
 ; CHECK-SCEV-INV-NEXT: Running pass: InvalidateAnalysisPass<{{.*}}ScalarEvolutionAnalysis
 ; CHECK-SCEV-INV-NEXT: Running pass: LoopSimplifyPass
 ; CHECK-SCEV-INV-NEXT: Running pass: LCSSAPass
 
 entry:
   ret void
 }
 
 define void @one_loop(i1* %ptr) {
 ; CHECK-LOOP-INV: Running pass: LoopSimplifyPass
 ; CHECK-LOOP-INV-NEXT: Running analysis: LoopAnalysis
 ; CHECK-LOOP-INV-NEXT: Running analysis: DominatorTreeAnalysis
 ; CHECK-LOOP-INV-NEXT: Running analysis: AssumptionAnalysis
+; CHECK-LOOP-INV-NEXT: Running analysis: TargetIRAnalysis
 ; CHECK-LOOP-INV-NEXT: Running pass: LCSSAPass
 ; CHECK-LOOP-INV-NEXT: Running analysis: AAManager
 ; CHECK-LOOP-INV-NEXT: Running analysis: TargetLibraryAnalysis
 ; CHECK-LOOP-INV-NEXT: Running analysis: ScalarEvolutionAnalysis
-; CHECK-LOOP-INV-NEXT: Running analysis: TargetIRAnalysis
 ; CHECK-LOOP-INV-NEXT: Running analysis: InnerAnalysisManagerProxy<{{.*}}Loop
 ; CHECK-LOOP-INV-NEXT: Running pass: NoOpLoopPass
 ; CHECK-LOOP-INV-NEXT: Running pass: InvalidateAnalysisPass<{{.*}}LoopAnalysis
 ; CHECK-LOOP-INV-NEXT: Clearing all analysis results for: <possibly invalidated loop>
 ; CHECK-LOOP-INV-NEXT: Invalidating analysis: LoopAnalysis
 ; CHECK-LOOP-INV-NEXT: Invalidating analysis: ScalarEvolutionAnalysis
 ; CHECK-LOOP-INV-NEXT: Invalidating analysis: InnerAnalysisManagerProxy<{{.*}}Loop
 ; CHECK-LOOP-INV-NEXT: Running pass: LoopSimplifyPass
 ; CHECK-LOOP-INV-NEXT: Running analysis: LoopAnalysis
 ; CHECK-LOOP-INV-NEXT: Running pass: LCSSAPass
 ; CHECK-LOOP-INV-NEXT: Running analysis: ScalarEvolutionAnalysis
 ; CHECK-LOOP-INV-NEXT: Running analysis: InnerAnalysisManagerProxy<{{.*}}Loop
 ; CHECK-LOOP-INV-NEXT: Running pass: NoOpLoopPass
 ;
 ; CHECK-SCEV-INV-NEXT: Running pass: LoopSimplifyPass
 ; CHECK-SCEV-INV-NEXT: Running analysis: LoopAnalysis
 ; CHECK-SCEV-INV-NEXT: Running analysis: DominatorTreeAnalysis
 ; CHECK-SCEV-INV-NEXT: Running analysis: AssumptionAnalysis
+; CHECK-SCEV-INV-NEXT: Running analysis: TargetIRAnalysis
 ; CHECK-SCEV-INV-NEXT: Running pass: LCSSAPass
 ; CHECK-SCEV-INV-NEXT: Running analysis: AAManager
 ; CHECK-SCEV-INV-NEXT: Running analysis: TargetLibraryAnalysis
 ; CHECK-SCEV-INV-NEXT: Running analysis: ScalarEvolutionAnalysis
-; CHECK-SCEV-INV-NEXT: Running analysis: TargetIRAnalysis
 ; CHECK-SCEV-INV-NEXT: Running analysis: InnerAnalysisManagerProxy<{{.*}}Loop
 ; CHECK-SCEV-INV-NEXT: Running pass: NoOpLoopPass
 ; CHECK-SCEV-INV-NEXT: Running pass: InvalidateAnalysisPass<{{.*}}ScalarEvolutionAnalysis
 ; CHECK-SCEV-INV-NEXT: Clearing all analysis results for: <possibly invalidated loop>
 ; CHECK-SCEV-INV-NEXT: Invalidating analysis: ScalarEvolutionAnalysis
 ; CHECK-SCEV-INV-NEXT: Invalidating analysis: InnerAnalysisManagerProxy<{{.*}}Loop
 ; CHECK-SCEV-INV-NEXT: Running pass: LoopSimplifyPass
 ; CHECK-SCEV-INV-NEXT: Running pass: LCSSAPass
exit:
   ret void
 }
 
 define void @nested_loops(i1* %ptr) {
 ; CHECK-LOOP-INV: Running pass: LoopSimplifyPass
 ; CHECK-LOOP-INV-NEXT: Running analysis: LoopAnalysis
 ; CHECK-LOOP-INV-NEXT: Running analysis: DominatorTreeAnalysis
 ; CHECK-LOOP-INV-NEXT: Running analysis: AssumptionAnalysis
+; CHECK-LOOP-INV-NEXT: Running analysis: TargetIRAnalysis
 ; CHECK-LOOP-INV-NEXT: Running pass: LCSSAPass
 ; CHECK-LOOP-INV-NEXT: Running analysis: AAManager
 ; CHECK-LOOP-INV-NEXT: Running analysis: TargetLibraryAnalysis
 ; CHECK-LOOP-INV-NEXT: Running analysis: ScalarEvolutionAnalysis
-; CHECK-LOOP-INV-NEXT: Running analysis: TargetIRAnalysis
 ; CHECK-LOOP-INV-NEXT: Running analysis: InnerAnalysisManagerProxy<{{.*}}Loop
 ; CHECK-LOOP-INV-NEXT: Running pass: NoOpLoopPass
 ; CHECK-LOOP-INV-NEXT: Running pass: NoOpLoopPass
 ; CHECK-LOOP-INV-NEXT: Running pass: InvalidateAnalysisPass<{{.*}}LoopAnalysis
 ; CHECK-LOOP-INV-NEXT: Clearing all analysis results for: <possibly invalidated loop>
 ; CHECK-LOOP-INV-NEXT: Clearing all analysis results for: <possibly invalidated loop>
 ; CHECK-LOOP-INV-NEXT: Invalidating analysis: LoopAnalysis
 ; CHECK-LOOP-INV-NEXT: Invalidating analysis: ScalarEvolutionAnalysis
 ; CHECK-LOOP-INV-NEXT: Invalidating analysis: InnerAnalysisManagerProxy<{{.*}}Loop
 ; CHECK-LOOP-INV-NEXT: Running pass: LoopSimplifyPass
 ; CHECK-LOOP-INV-NEXT: Running analysis: LoopAnalysis
 ; CHECK-LOOP-INV-NEXT: Running pass: LCSSAPass
 ; CHECK-LOOP-INV-NEXT: Running analysis: ScalarEvolutionAnalysis
 ; CHECK-LOOP-INV-NEXT: Running analysis: InnerAnalysisManagerProxy<{{.*}}Loop
 ; CHECK-LOOP-INV-NEXT: Running pass: NoOpLoopPass
 ; CHECK-LOOP-INV-NEXT: Running pass: NoOpLoopPass
 ;
 ; CHECK-SCEV-INV: Running pass: LoopSimplifyPass
 ; CHECK-SCEV-INV-NEXT: Running analysis: LoopAnalysis
 ; CHECK-SCEV-INV-NEXT: Running analysis: DominatorTreeAnalysis
 ; CHECK-SCEV-INV-NEXT: Running analysis: AssumptionAnalysis
+; CHECK-SCEV-INV-NEXT: Running analysis: TargetIRAnalysis
 ; CHECK-SCEV-INV-NEXT: Running pass: LCSSAPass
 ; CHECK-SCEV-INV-NEXT: Running analysis: AAManager
 ; CHECK-SCEV-INV-NEXT: Running analysis: TargetLibraryAnalysis
 ; CHECK-SCEV-INV-NEXT: Running analysis: ScalarEvolutionAnalysis
-; CHECK-SCEV-INV-NEXT: Running analysis: TargetIRAnalysis
 ; CHECK-SCEV-INV-NEXT: Running analysis: InnerAnalysisManagerProxy<{{.*}}Loop
 ; CHECK-SCEV-INV-NEXT: Running pass: NoOpLoopPass
 ; CHECK-SCEV-INV-NEXT: Running pass: NoOpLoopPass
 ; CHECK-SCEV-INV-NEXT: Running pass: InvalidateAnalysisPass<{{.*}}ScalarEvolutionAnalysis
 ; CHECK-SCEV-INV-NEXT: Clearing all analysis results for: <possibly invalidated loop>
 ; CHECK-SCEV-INV-NEXT: Clearing all analysis results for: <possibly invalidated loop>
 ; CHECK-SCEV-INV-NEXT: Invalidating analysis: ScalarEvolutionAnalysis
 ; CHECK-SCEV-INV-NEXT: Invalidating analysis: InnerAnalysisManagerProxy<{{.*}}Loop
exit:
   ret void
 }
 
 define void @dead_loop() {
 ; CHECK-LOOP-INV: Running pass: LoopSimplifyPass
 ; CHECK-LOOP-INV-NEXT: Running analysis: LoopAnalysis
 ; CHECK-LOOP-INV-NEXT: Running analysis: DominatorTreeAnalysis
 ; CHECK-LOOP-INV-NEXT: Running analysis: AssumptionAnalysis
+; CHECK-LOOP-INV-NEXT: Running analysis: TargetIRAnalysis
 ; CHECK-LOOP-INV-NEXT: Running pass: LCSSAPass
 ; CHECK-LOOP-INV-NEXT: Running analysis: AAManager
 ; CHECK-LOOP-INV-NEXT: Running analysis: TargetLibraryAnalysis
 ; CHECK-LOOP-INV-NEXT: Running analysis: ScalarEvolutionAnalysis
-; CHECK-LOOP-INV-NEXT: Running analysis: TargetIRAnalysis
 ; CHECK-LOOP-INV-NEXT: Running analysis: InnerAnalysisManagerProxy<{{.*}}Loop
 ; CHECK-LOOP-INV-NEXT: Running pass: NoOpLoopPass
 ; CHECK-LOOP-INV-NEXT: Running pass: InvalidateAnalysisPass<{{.*}}LoopAnalysis
 ; CHECK-LOOP-INV-NEXT: Clearing all analysis results for: <possibly invalidated loop>
 ; CHECK-LOOP-INV-NEXT: Invalidating analysis: LoopAnalysis
 ; CHECK-LOOP-INV-NEXT: Invalidating analysis: ScalarEvolutionAnalysis
 ; CHECK-LOOP-INV-NEXT: Invalidating analysis: InnerAnalysisManagerProxy<{{.*}}Loop
 ; CHECK-LOOP-INV-NEXT: Running pass: LoopSimplifyPass
 ; CHECK-LOOP-INV-NEXT: Running analysis: LoopAnalysis
 ; CHECK-LOOP-INV-NEXT: Running pass: LCSSAPass
 ; CHECK-LOOP-INV-NEXT: Running analysis: ScalarEvolutionAnalysis
 ; CHECK-LOOP-INV-NEXT: Running analysis: InnerAnalysisManagerProxy<{{.*}}Loop
 ; CHECK-LOOP-INV-NEXT: Running pass: NoOpLoopPass
 ;
 ; CHECK-SCEV-INV: Running pass: LoopSimplifyPass
 ; CHECK-SCEV-INV-NEXT: Running analysis: LoopAnalysis
 ; CHECK-SCEV-INV-NEXT: Running analysis: DominatorTreeAnalysis
 ; CHECK-SCEV-INV-NEXT: Running analysis: AssumptionAnalysis
+; CHECK-SCEV-INV-NEXT: Running analysis: TargetIRAnalysis
 ; CHECK-SCEV-INV-NEXT: Running pass: LCSSAPass
 ; CHECK-SCEV-INV-NEXT: Running analysis: AAManager
 ; CHECK-SCEV-INV-NEXT: Running analysis: TargetLibraryAnalysis
 ; CHECK-SCEV-INV-NEXT: Running analysis: ScalarEvolutionAnalysis
-; CHECK-SCEV-INV-NEXT: Running analysis: TargetIRAnalysis
 ; CHECK-SCEV-INV-NEXT: Running analysis: InnerAnalysisManagerProxy<{{.*}}Loop
 ; CHECK-SCEV-INV-NEXT: Running pass: NoOpLoopPass
 ; CHECK-SCEV-INV-NEXT: Running pass: InvalidateAnalysisPass<{{.*}}ScalarEvolutionAnalysis
 ; CHECK-SCEV-INV-NEXT: Clearing all analysis results for: <possibly invalidated loop>
 ; CHECK-SCEV-INV-NEXT: Invalidating analysis: ScalarEvolutionAnalysis
 ; CHECK-SCEV-INV-NEXT: Invalidating analysis: InnerAnalysisManagerProxy<{{.*}}Loop
 ; CHECK-SCEV-INV-NEXT: Running pass: LoopSimplifyPass
 ; CHECK-SCEV-INV-NEXT: Running pass: LCSSAPass
 ; CHECK-SCEV-INV-NEXT: Running analysis: ScalarEvolutionAnalysis
 ; CHECK-SCEV-INV-NEXT: Running analysis: InnerAnalysisManagerProxy<{{.*}}Loop
 ; CHECK-SCEV-INV-NEXT: Running pass: NoOpLoopPass
 ;
 ; CHECK-SCEV-INV-AFTER-DELETE-LABEL: Running pass: LoopSimplifyPass on dead_loop
 ; CHECK-SCEV-INV-AFTER-DELETE-NEXT: Running analysis: LoopAnalysis
 ; CHECK-SCEV-INV-AFTER-DELETE-NEXT: Running analysis: DominatorTreeAnalysis
 ; CHECK-SCEV-INV-AFTER-DELETE-NEXT: Running analysis: AssumptionAnalysis
+; CHECK-SCEV-INV-AFTER-DELETE-NEXT: Running analysis: TargetIRAnalysis
 ; CHECK-SCEV-INV-AFTER-DELETE-NEXT: Running pass: LCSSAPass
 ; CHECK-SCEV-INV-AFTER-DELETE-NEXT: Running analysis: AAManager
 ; CHECK-SCEV-INV-AFTER-DELETE-NEXT: Running analysis: TargetLibraryAnalysis
 ; CHECK-SCEV-INV-AFTER-DELETE-NEXT: Running analysis: ScalarEvolutionAnalysis
-; CHECK-SCEV-INV-AFTER-DELETE-NEXT: Running analysis: TargetIRAnalysis
 ; CHECK-SCEV-INV-AFTER-DELETE-NEXT: Running analysis: InnerAnalysisManagerProxy<{{.*}}Loop
 ; CHECK-SCEV-INV-AFTER-DELETE-NEXT: Running pass: NoOpLoopPass
 ; CHECK-SCEV-INV-AFTER-DELETE-NEXT: Running pass: LoopDeletionPass
 ; CHECK-SCEV-INV-AFTER-DELETE-NEXT: Clearing all analysis results for:
 ; CHECK-SCEV-INV-AFTER-DELETE-NEXT: Running pass: InvalidateAnalysisPass<{{.*}}ScalarEvolutionAnalysis
 ; CHECK-SCEV-INV-AFTER-DELETE-NEXT: Invalidating analysis: ScalarEvolutionAnalysis
 ; CHECK-SCEV-INV-AFTER-DELETE-NEXT: Invalidating analysis: InnerAnalysisManagerProxy<{{.*}}Loop
 ; CHECK-SCEV-INV-AFTER-DELETE-NEXT: Running pass: LoopSimplifyPass

llvm/test/Other/new-pass-manager.ll

 ; RUN:     -passes='loop(repeat<3>(no-op-loop))' %s 2>&1 \
 ; RUN:     | FileCheck %s --check-prefix=CHECK-REPEAT-LOOP-PASS
 ; CHECK-REPEAT-LOOP-PASS: Running analysis: InnerAnalysisManagerProxy<{{.*}}>
 ; CHECK-REPEAT-LOOP-PASS-NEXT: Running analysis: PreservedCFGCheckerAnalysis on foo
 ; CHECK-REPEAT-LOOP-PASS-NEXT: Running pass: LoopSimplify
 ; CHECK-REPEAT-LOOP-PASS-NEXT: Running analysis: LoopAnalysis
 ; CHECK-REPEAT-LOOP-PASS-NEXT: Running analysis: DominatorTreeAnalysis
 ; CHECK-REPEAT-LOOP-PASS-NEXT: Running analysis: AssumptionAnalysis
+; CHECK-REPEAT-LOOP-PASS-NEXT: Running analysis: TargetIRAnalysis
 ; CHECK-REPEAT-LOOP-PASS-NEXT: Invalidating analysis: PreservedCFGCheckerAnalysis on foo
 ; CHECK-REPEAT-LOOP-PASS-NEXT: Running pass: LCSSAPass
 ; CHECK-REPEAT-LOOP-PASS-NEXT: Running analysis: PreservedCFGCheckerAnalysis on foo
 ; CHECK-REPEAT-LOOP-PASS-NEXT: Running analysis: AAManager
 ; CHECK-REPEAT-LOOP-PASS-NEXT: Running analysis: TargetLibraryAnalysis
 ; CHECK-REPEAT-LOOP-PASS-NEXT: Running analysis: BasicAA
 ; CHECK-REPEAT-LOOP-PASS-NEXT: Running analysis: ScopedNoAliasAA
 ; CHECK-REPEAT-LOOP-PASS-NEXT: Running analysis: TypeBasedAA
 ; CHECK-REPEAT-LOOP-PASS-NEXT: Running analysis: OuterAnalysisManagerProxy
 ; CHECK-REPEAT-LOOP-PASS-NEXT: Running analysis: ScalarEvolutionAnalysis
-; CHECK-REPEAT-LOOP-PASS-NEXT: Running analysis: TargetIRAnalysis
 ; CHECK-REPEAT-LOOP-PASS-NEXT: Running analysis: InnerAnalysisManagerProxy<{{.*}}>
 ; CHECK-REPEAT-LOOP-PASS-NEXT: Running pass: RepeatedPass
 ; CHECK-REPEAT-LOOP-PASS-NEXT: Running pass: NoOpLoopPass
 ; CHECK-REPEAT-LOOP-PASS-NEXT: Running pass: NoOpLoopPass
 ; CHECK-REPEAT-LOOP-PASS-NEXT: Running pass: NoOpLoopPass
 
 define void @foo(i1 %x, i8* %p1, i8* %p2) {
 entry:

llvm/test/Other/new-pm-lto-defaults.ll

 ; CHECK-O-NEXT: Running analysis: LazyCallGraphAnalysis
 ; CHECK-O1-NEXT: Running analysis: TargetLibraryAnalysis
 ; CHECK-O-NEXT: Running analysis: FunctionAnalysisManagerCGSCCProxy
 ; CHECK-O-NEXT: Running analysis: OuterAnalysisManagerProxy<{{.*}}LazyCallGraph{{.*}}>
 ; CHECK-O-NEXT: Running pass: PostOrderFunctionAttrsPass
 ; CHECK-O-NEXT: Running analysis: AAManager
 ; CHECK-O-NEXT: Running analysis: BasicAA
 ; CHECK-O1-NEXT: Running analysis: AssumptionAnalysis on foo
+; CHECK-O1-NEXT: Running analysis: TargetIRAnalysis
 ; CHECK-O1-NEXT: Running analysis: DominatorTreeAnalysis
 ; CHECK-O-NEXT: Running analysis: ScopedNoAliasAA
 ; CHECK-O-NEXT: Running analysis: TypeBasedAA
 ; CHECK-O-NEXT: Running analysis: OuterAnalysisManagerProxy
 ; CHECK-O-NEXT: Running pass: ReversePostOrderFunctionAttrsPass
 ; CHECK-O-NEXT: Running analysis: CallGraphAnalysis
 ; CHECK-O-NEXT: Running pass: GlobalSplitPass
 ; CHECK-O-NEXT: Running pass: WholeProgramDevirtPass

llvm/test/Other/new-pm-thinlto-prelink-pgo-defaults.ll

 ; CHECK-O-NEXT: Running analysis: OuterAnalysisManagerProxy<{{.*}}LazyCallGraph::SCC{{.*}}>
 ; CHECK-O-NEXT: Running pass: DevirtSCCRepeatedPass
 ; CHECK-O-NEXT: Running pass: InlinerPass
 ; CHECK-O-NEXT: Running pass: InlinerPass
 ; CHECK-O-NEXT: Running pass: PostOrderFunctionAttrsPass
 ; CHECK-O-NEXT: Running analysis: AAManager
 ; CHECK-O-NEXT: Running analysis: BasicAA
 ; CHECK-O-NEXT: Running analysis: AssumptionAnalysis
+; CHECK-O-NEXT: Running analysis: TargetIRAnalysis
 ; CHECK-O-NEXT: Running analysis: DominatorTreeAnalysis
 ; CHECK-O-NEXT: Running analysis: ScopedNoAliasAA
 ; CHECK-O-NEXT: Running analysis: TypeBasedAA
 ; CHECK-O-NEXT: Running analysis: OuterAnalysisManagerProxy
 ; CHECK-O3-NEXT: Running pass: ArgumentPromotionPass
-; CHECK-O3-NEXT: Running analysis: TargetIRAnalysis
 ; CHECK-O2-NEXT: Running pass: OpenMPOptCGSCCPass
 ; CHECK-O3-NEXT: Running pass: OpenMPOptCGSCCPass
 ; CHECK-O-NEXT: Running pass: SROA
 ; CHECK-O-NEXT: Running pass: EarlyCSEPass
-; CHECK-O1-NEXT: Running analysis: TargetIRAnalysis on foo
-; CHECK-O2-NEXT: Running analysis: TargetIRAnalysis on foo
-; CHECK-Os-NEXT: Running analysis: TargetIRAnalysis on foo
-; CHECK-Oz-NEXT: Running analysis: TargetIRAnalysis on foo
 ; CHECK-O-NEXT: Running analysis: MemorySSAAnalysis
 ; CHECK-O23SZ-NEXT: Running pass: SpeculativeExecutionPass
 ; CHECK-O23SZ-NEXT: Running pass: JumpThreadingPass
 ; CHECK-O23SZ-NEXT: Running analysis: LazyValueAnalysis
 ; CHECK-O23SZ-NEXT: Running pass: CorrelatedValuePropagationPass
 ; CHECK-O23SZ-NEXT: Invalidating analysis: LazyValueAnalysis
 ; CHECK-O-NEXT: Running pass: SimplifyCFGPass
 ; CHECK-O3-NEXT: Running pass: AggressiveInstCombinePass

llvm/test/Transforms/InferAddressSpaces/AMDGPU/builtin-assumed-addrspace.ll

+; RUN: opt -S -mtriple=amdgcn-amd-amdhsa -infer-address-spaces -o - %s | FileCheck %s
+
+; CHECK-LABEL: @f0
+; CHECK: addrspacecast float* {{%.*}} to float addrspace(3)*
+; CHECK: getelementptr inbounds float, float addrspace(3)*
+; CHECK: load float, float addrspace(3)*
+define float @f0(float* %p) {
+entry:
+  %0 = bitcast float* %p to i8*
+  %1 = call i1 @llvm.amdgcn.is.shared(i8* %0)
+  tail call void @llvm.assume(i1 %1)
+  %2 = tail call i32 @llvm.amdgcn.workitem.id.x()
+  %idxprom = zext i32 %2 to i64
+  %arrayidx = getelementptr inbounds float, float* %p, i64 %idxprom
+  %3 = load float, float* %arrayidx, align 4
+  ret float %3
+}
+
+; CHECK-LABEL: @f1
+; CHECK: addrspacecast float* {{%.*}} to float addrspace(5)*
+; CHECK: getelementptr inbounds float, float addrspace(5)*
+; CHECK: load float, float addrspace(5)*
+define float @f1(float* %p) {
+entry:
+  %0 = bitcast float* %p to i8*
+  %1 = call i1 @llvm.amdgcn.is.private(i8* %0)
+  tail call void @llvm.assume(i1 %1)
+  %2 = tail call i32 @llvm.amdgcn.workitem.id.x()
+  %idxprom = zext i32 %2 to i64
+  %arrayidx = getelementptr inbounds float, float* %p, i64 %idxprom
+  %3 = load float, float* %arrayidx, align 4
+  ret float %3
+}
+
+; CHECK-LABEL: @f2
+; CHECK: addrspacecast float* {{%.*}} to float addrspace(1)*
+; CHECK: getelementptr inbounds float, float addrspace(1)*
+; CHECK: load float, float addrspace(1)*
+define float @f2(float* %p) {
+entry:
+  %0 = bitcast float* %p to i8*
+  %1 = call i1 @llvm.amdgcn.is.private(i8* %0)
+  %2 = xor i1 %1, -1
+  %3 = call i1 @llvm.amdgcn.is.shared(i8* %0)
+  %4 = xor i1 %3, -1
+  %5 = and i1 %2, %4
+  tail call void @llvm.assume(i1 %5)
+  %6 = tail call i32 @llvm.amdgcn.workitem.id.x()
+  %idxprom = zext i32 %6 to i64
+  %arrayidx = getelementptr inbounds float, float* %p, i64 %idxprom
+  %7 = load float, float* %arrayidx, align 4
+  ret float %7
+}
+
+; CHECK-LABEL: @g0
+; CHECK: if.then:
+; CHECK: addrspacecast float* {{%.*}} to float addrspace(3)*
+; CHECK: getelementptr inbounds float, float addrspace(3)*
+; CHECK: load float, float addrspace(3)*
+; CHECK: if.end:
+; CHECK: getelementptr inbounds float, float*
+; CHECK: load float, float*
+define float @g0(i32 %c, float* %p) {
+entry:
+  %tobool.not = icmp eq i32 %c, 0
+  br i1 %tobool.not, label %if.end, label %if.then
+
+if.then:
+  %0 = bitcast float* %p to i8*
+  %1 = call i1 @llvm.amdgcn.is.shared(i8* %0)
+  tail call void @llvm.assume(i1 %1)
+  %2 = tail call i32 @llvm.amdgcn.workitem.id.x()
+  %idxprom = zext i32 %2 to i64
+  %arrayidx = getelementptr inbounds float, float* %p, i64 %idxprom
+  %3 = load float, float* %arrayidx, align 4
+  %add = fadd float %3, 0.
+  br label %if.end
+
+if.end:
+  %s = phi float [ %add, %if.then ], [ 0., %entry ]
+  %4 = tail call i32 @llvm.amdgcn.workitem.id.y()
+  %idxprom2 = zext i32 %4 to i64
+  %arrayidx2 = getelementptr inbounds float, float* %p, i64 %idxprom2
+  %5 = load float, float* %arrayidx2, align 4
+  %add2 = fadd float %s, %5
+  ret float %add2
+}
+
+declare void @llvm.assume(i1)
+declare i1 @llvm.amdgcn.is.shared(i8* nocapture)
+declare i1 @llvm.amdgcn.is.private(i8* nocapture)
+declare i32 @llvm.amdgcn.workitem.id.x()
+declare i32 @llvm.amdgcn.workitem.id.y()

llvm/test/Transforms/InferAddressSpaces/NVPTX/builtin-assumed-addrspace.ll

+; RUN: opt -S -mtriple=nvptx64-nvidia-cuda -infer-address-spaces -o - %s | FileCheck %s
+
+; CHECK-LABEL: @f0
+; CHECK: addrspacecast float* {{%.*}} to float addrspace(4)*
+; CHECK: getelementptr inbounds float, float addrspace(4)*
+; CHECK: load float, float addrspace(4)*
+define float @f0(float* %p) {
+entry:
+  %0 = bitcast float* %p to i8*
+  %1 = call i1 @llvm.nvvm.isspacep.const(i8* %0)
+  tail call void @llvm.assume(i1 %1)
+  %2 = tail call i32 @llvm.nvvm.read.ptx.sreg.tid.x()
+  %idxprom = zext i32 %2 to i64
+  %arrayidx = getelementptr inbounds float, float* %p, i64 %idxprom
+  %3 = load float, float* %arrayidx, align 4
+  ret float %3
+}
+
+; CHECK-LABEL: @f1
+; CHECK: addrspacecast float* {{%.*}} to float addrspace(1)*
+; CHECK: getelementptr inbounds float, float addrspace(1)*
+; CHECK: load float, float addrspace(1)*
+define float @f1(float* %p) {
+entry:
+  %0 = bitcast float* %p to i8*
+  %1 = call i1 @llvm.nvvm.isspacep.global(i8* %0)
+  tail call void @llvm.assume(i1 %1)
+  %2 = tail call i32 @llvm.nvvm.read.ptx.sreg.tid.x()
+  %idxprom = zext i32 %2 to i64
+  %arrayidx = getelementptr inbounds float, float* %p, i64 %idxprom
+  %3 = load float, float* %arrayidx, align 4
+  ret float %3
+}
+
+; CHECK-LABEL: @f2
+; CHECK: addrspacecast float* {{%.*}} to float addrspace(5)*
+; CHECK: getelementptr inbounds float, float addrspace(5)*
+; CHECK: load float, float addrspace(5)*
+define float @f2(float* %p) {
+entry:
+  %0 = bitcast float* %p to i8*
+  %1 = call i1 @llvm.nvvm.isspacep.local(i8* %0)
+  tail call void @llvm.assume(i1 %1)
+  %2 = tail call i32 @llvm.nvvm.read.ptx.sreg.tid.x()
+  %idxprom = zext i32 %2 to i64
+  %arrayidx = getelementptr inbounds float, float* %p, i64 %idxprom
+  %3 = load float, float* %arrayidx, align 4
+  ret float %3
+}
+
+; CHECK-LABEL: @f3
+; CHECK: addrspacecast float* {{%.*}} to float addrspace(3)*
+; CHECK: getelementptr inbounds float, float addrspace(3)*
+; CHECK: load float, float addrspace(3)*
+define float @f3(float* %p) {
+entry:
+  %0 = bitcast float* %p to i8*
+  %1 = call i1 @llvm.nvvm.isspacep.shared(i8* %0)
+  tail call void @llvm.assume(i1 %1)
+  %2 = tail call i32 @llvm.nvvm.read.ptx.sreg.tid.x()
+  %idxprom = zext i32 %2 to i64
+  %arrayidx = getelementptr inbounds float, float* %p, i64 %idxprom
+  %3 = load float, float* %arrayidx, align 4
+  ret float %3
+}
+
+; CHECK-LABEL: @g0
+; CHECK: if.then:
+; CHECK: addrspacecast float* {{%.*}} to float addrspace(3)*
+; CHECK: getelementptr inbounds float, float addrspace(3)*
+; CHECK: load float, float addrspace(3)*
+; CHECK: if.end:
+; CHECK: getelementptr inbounds float, float*
+; CHECK: load float, float*
+define float @g0(i32 %c, float* %p) {
+entry:
+  %tobool.not = icmp eq i32 %c, 0
+  br i1 %tobool.not, label %if.end, label %if.then
+
+if.then:
+  %0 = bitcast float* %p to i8*
+  %1 = call i1 @llvm.nvvm.isspacep.shared(i8* %0)
+  tail call void @llvm.assume(i1 %1)
+  %2 = tail call i32 @llvm.nvvm.read.ptx.sreg.tid.x()
+  %idxprom = zext i32 %2 to i64
+  %arrayidx = getelementptr inbounds float, float* %p, i64 %idxprom
+  %3 = load float, float* %arrayidx, align 4
+  %add = fadd float %3, 0.
+  br label %if.end
+
+if.end:
+  %s = phi float [ %add, %if.then ], [ 0., %entry ]
+  %4 = tail call i32 @llvm.nvvm.read.ptx.sreg.tid.y()
+  %idxprom2 = zext i32 %4 to i64
+  %arrayidx2 = getelementptr inbounds float, float* %p, i64 %idxprom2
+  %5 = load float, float* %arrayidx2, align 4
+  %add2 = fadd float %s, %5
+  ret float %add2
+}
+
+declare void @llvm.assume(i1)
+declare i1 @llvm.nvvm.isspacep.const(i8*)
+declare i1 @llvm.nvvm.isspacep.global(i8*)
+declare i1 @llvm.nvvm.isspacep.local(i8*)
+declare i1 @llvm.nvvm.isspacep.shared(i8*)
+declare i32 @llvm.nvvm.read.ptx.sreg.tid.x()
+declare i32 @llvm.nvvm.read.ptx.sreg.tid.y()

llvm/test/Transforms/LoopRotate/pr35210.ll

 ;RUN: opt %s -aa-pipeline= -passes='adce,loop(loop-rotate),adce' -S -verify-cfg-preserved=0 -debug-pass-manager -debug-only=loop-rotate 2>&1 | FileCheck %s
 ;RUN: opt %s -aa-pipeline= -passes='adce,loop-mssa(loop-rotate),adce' -S -verify-cfg-preserved=0 -debug-pass-manager -debug-only=loop-rotate -verify-memoryssa 2>&1 | FileCheck %s --check-prefix=MSSA
 ;REQUIRES: asserts
 
 ; This test is to make sure we invalidate the post dominator pass after loop rotate simplifies the loop latch.
 ; The adce passes are here to make sure post dominator analysis is required.
 
 ; CHECK: Running pass: ADCEPass on f
 ; CHECK-NEXT: Running analysis: PostDominatorTreeAnalysis on f
 ; CHECK-NEXT: Running pass: LoopSimplifyPass on f
 ; CHECK-NEXT: Running analysis: LoopAnalysis on f
 ; CHECK-NEXT: Running analysis: DominatorTreeAnalysis on f
 ; CHECK-NEXT: Running analysis: AssumptionAnalysis on f
+; CHECK-NEXT: Running analysis: TargetIRAnalysis on f
 ; CHECK-NEXT: Running pass: LCSSAPass on f
 ; CHECK-NEXT: Running analysis: AAManager on f
 ; CHECK-NEXT: Running analysis: TargetLibraryAnalysis on f
 ; CHECK-NEXT: Running analysis: ScalarEvolutionAnalysis on f
-; CHECK-NEXT: Running analysis: TargetIRAnalysis on f
 ; CHECK-NEXT: Running analysis: InnerAnalysisManagerProxy{{.*}} on f
 ; CHECK-NEXT: Running pass: LoopRotatePass on Loop at depth 1 containing: %bb<header><exiting>,%bb4<latch>
 ; CHECK-NEXT: Folding loop latch bb4 into bb
 ; CHECK-NEXT: Invalidating analysis: PostDominatorTreeAnalysis on f
 ; CHECK-NEXT: Running pass: ADCEPass on f
 ; CHECK-NEXT: Running analysis: PostDominatorTreeAnalysis on f
 
 ; MSSA: Running pass: ADCEPass on f
 ; MSSA-NEXT: Running analysis: PostDominatorTreeAnalysis on f
 ; MSSA-NEXT: Running pass: LoopSimplifyPass on f
 ; MSSA-NEXT: Running analysis: LoopAnalysis on f
 ; MSSA-NEXT: Running analysis: DominatorTreeAnalysis on f
 ; MSSA-NEXT: Running analysis: AssumptionAnalysis on f
+; MSSA-NEXT: Running analysis: TargetIRAnalysis on f
 ; MSSA-NEXT: Running pass: LCSSAPass on f
 ; MSSA-NEXT: Running analysis: MemorySSAAnalysis on f
 ; MSSA-NEXT: Running analysis: AAManager on f
 ; MSSA-NEXT: Running analysis: TargetLibraryAnalysis on f
 ; MSSA-NEXT: Running analysis: ScalarEvolutionAnalysis on f
-; MSSA-NEXT: Running analysis: TargetIRAnalysis on f
 ; MSSA-NEXT: Running analysis: InnerAnalysisManagerProxy{{.*}} on f
 ; MSSA-NEXT: Running pass: LoopRotatePass on Loop at depth 1 containing: %bb<header><exiting>,%bb4<latch>
 ; MSSA-NEXT: Folding loop latch bb4 into bb
 ; MSSA-NEXT: Invalidating analysis: PostDominatorTreeAnalysis on f
 ; MSSA-NEXT: Running pass: ADCEPass on f
 ; MSSA-NEXT: Running analysis: PostDominatorTreeAnalysis on f
 
 ; CHECK-LABEL: define i8 @f() {

llvm/unittests/Analysis/AssumeBundleQueriesTest.cpp

std::unique_ptr<Module> Mod = parseAssemblyString(
       "ret void\n}\n",
       Err, C);
   if (!Mod)
     Err.print("AssumeQueryAPI", errs());
   Function *F = Mod->getFunction("test");
   BasicBlock::iterator First = F->begin()->begin();
   BasicBlock::iterator Second = F->begin()->begin();
   Second++;
-  AssumptionCacheTracker ACT;
-  AssumptionCache &AC = ACT.getAssumptionCache(*F);
+  AssumptionCache AC(*F);
   auto AR = AC.assumptionsFor(F->getArg(3));
   ASSERT_EQ(AR.size(), 0u);
   AR = AC.assumptionsFor(F->getArg(1));
   ASSERT_EQ(AR.size(), 1u);
   ASSERT_EQ(AR[0].Index, 0u);
   ASSERT_EQ(AR[0].Assume, &*Second);
   AR = AC.assumptionsFor(F->getArg(2));
   ASSERT_EQ(AR.size(), 1u);