Modify ModRefInfo values using static inline method abstractions [NFC].

[asbirlea]

Summary

The aim is to make ModRefInfo checks and changes more intuitive
and less error prone using inline methods that abstract the bit operations.

Ideally ModRefInfo would become an enum class, but that change will require
a wider set of changes into FunctionModRefBehavior.

[asbirlea]

clang-format.

[asbirlea]

Replace all bit-wise operations involving ModRefInfo with inline method wrappers.

There is nothing (as of this patch) preventing bit-wise operations from being added instead of using the wrappers.

[sanjoy]

Hi,

I have some minor comments inline.

I only skimmed through the parts that seemed to be mechanical transformations, but please let me know if there are specific places that should review more carefully.

include/llvm/Analysis/AliasAnalysis.h

111

Why static?

112–113

isModOrRef, isMod etc. (alternatively isModOrRefSet or isModSet) sound more idiomatic to me.

[edit: I also have another note later on the setXXX naming].

134

We should stick to the <verb><noun> format here too, so unionModRef and intersectModRef.

553

This does not seem NFC to me -- doesn't getModRefBehavior return a FunctionModRefBehavior with the extra bits set? Or is the semantics of enum such that ModRefInfo will strip the unnecessary bits?

lib/Analysis/AliasAnalysis.cpp

151

One tricky aspect here is that setModAndRef(XX) sounds a bit like "change XX in place and set that bit" when the semantics is more like getWithModAndRefSet, but that's a bit verbose. What do you think? I personally think the verbosity is worth it, but you're touching a lot of this code so I'll leave this to your intuition.

We should also at least mark these "setters" as WARN_UNUSED_VALUE since setModAndRef(XX); (or some other spelling) with no assignment of the result is certainly a bug.

452

Why do you need this change?

470

Same here -- this change seems unnecessary.

[asbirlea]

Address comments.

[asbirlea]

Thanks for the review, Sanjoy! Much appreciated!

include/llvm/Analysis/AliasAnalysis.h

111

My bad, removed.

112–113

Changing to isModSet makes sense, happy to make that change.
isMod is not the same, however. The goal is to check if it modifies, versus whether it *only* modifies.
Concretely, isModSet will check MRI & MRI_Mod, versus isMod should check MRI==MRI_Mod.

553

Should have researched this more before hand, yes, I assumed ModRefInfo would strip the unnecessary bits. While this happens to work, it's UB according to the standard.
Correct way is to mask out unnecessary bits before creating a ModRefInfo.
Replacing with a wrapper and adding comment to document this.

lib/Analysis/AliasAnalysis.cpp

151

I wouldn't really mind the verbosity, but it feels so much simpler to read setMod than getWithModSet. I made arguments const and added LLVM_NODISCARD.
Thanks a lot for the suggestions!

452

Not really needed in this patch. I reverse-created this patch from the one adding MRI_Must, where I need to check the alias result against multiple values. Undo-ing the change in this patch.

470

Same as above, change undone.

[sanjoy]

lgtm with some comments inline

include/llvm/Analysis/AliasAnalysis.h

221

UB if those bits are not stripped out.

I don't see how this is relevant here.

lib/Analysis/AliasAnalysis.cpp

267

Shouldn't this be isMod? Before it was checking that the return value of getArgModRefInfo was == MRI_Mod.

360

This change seems unnecessary.

377

This change seems unnecessary.

404

This change seems unnecessary.

lib/Analysis/AliasAnalysisEvaluator.cpp

35

This change seems unnecessary.

258

This change seems unnecessary.

283

This change seems unnecessary.

328

This change seems unnecessary.

[asbirlea]

Address comments.

[asbirlea]

Thank you for the comments and the review! I will land this tomorrow.

include/llvm/Analysis/AliasAnalysis.h

221

The comment was meant so possible future changes don't assume ModRefInfo(FMRB) will drop most significant bits. Removing and hoping that doesn't happen :).

lib/Analysis/AliasAnalysis.cpp

267

I made this change intentionally, here's why:

Before, it was checking == MRI_Mod and then resetting mask to MRI_ModRef. If result of getArgModRefInfo was already MRI_ModRef, it was left unchanged, but this doesn't stand out.

With this change, it's obvious that if the result of getArgModRefInfo is either MRI_ModRefor MRI_Mod, then the mask should be MRI_ModRef. The comment I added above also highlights this: that if CS2 mods the location, we're looking to further check if CS1 reads or writes, so mask should should be MRI_ModRef.

529

Undone this too.

[asbirlea]

Missed nits (stray comment, braces).

include/llvm/Analysis/AliasAnalysis.h

enum AliasResult {
   /// The two locations precisely alias each other.
   MustAlias,
 };
 
 /// Flags indicating whether a memory access modifies or references memory.
 ///
 /// This is no access at all, a modification, a reference, or both
 /// a modification and a reference. These are specifically structured such that
-/// they form a two bit matrix and bit-tests for 'mod' or 'ref' work with any
-/// of the possible values.
+/// they form a two bit matrix and bit-tests for 'mod' or 'ref'
+/// work with any of the possible values.
+
 enum ModRefInfo {
   /// The access neither references nor modifies the value stored in memory.
   MRI_NoModRef = 0,
-  /// The access references the value stored in memory.
+  /// The access may reference the value stored in memory.
   MRI_Ref = 1,
-  /// The access modifies the value stored in memory.
+  /// The access may modify the value stored in memory.
   MRI_Mod = 2,
-  /// The access both references and modifies the value stored in memory.
-  MRI_ModRef = MRI_Ref | MRI_Mod
+  /// The access may reference and may modify the value stored in memory.
+  MRI_ModRef = MRI_Ref | MRI_Mod,
 };
+static inline bool isNoModRef(ModRefInfo MRI) { return MRI == MRI_NoModRef; }
+static inline bool modOrRefSet(ModRefInfo MRI) { return MRI & MRI_ModRef; }
+static inline bool modAndRefSet(ModRefInfo MRI) {
+  return (MRI & MRI_ModRef) == MRI_ModRef;
+}
+static inline bool modSet(ModRefInfo MRI) { return MRI & MRI_Mod; }
+static inline bool refSet(ModRefInfo MRI) { return MRI & MRI_Ref; }
+
+static inline ModRefInfo setRef(ModRefInfo MRI) {
+  return ModRefInfo(MRI | MRI_Ref);
+}
+static inline ModRefInfo setMod(ModRefInfo MRI) {
+  return ModRefInfo(MRI | MRI_Mod);
+}
+static inline ModRefInfo setModAndRef(ModRefInfo MRI) {
+  return ModRefInfo(MRI | MRI_ModRef);
+}
+static inline ModRefInfo clearMod(ModRefInfo MRI) {
+  return ModRefInfo(MRI & MRI_Ref);
+}
+static inline ModRefInfo clearRef(ModRefInfo MRI) {
+  return ModRefInfo(MRI & MRI_Mod);
+}
+static inline ModRefInfo modRefUnion(ModRefInfo MRI1, ModRefInfo MRI2) {
+  return ModRefInfo(MRI1 | MRI2);
+}
+static inline ModRefInfo modRefIntersection(ModRefInfo MRI1, ModRefInfo MRI2) {
+  return ModRefInfo(MRI1 & MRI2);
+}
 
 /// The locations at which a function might access memory.
 ///
 /// These are primarily used in conjunction with the \c AccessKind bits to
 /// describe both the nature of access and the locations of access for a
 /// function call.
 enum FunctionModRefLocation {
   /// Base case is no access to memory.
enum FunctionModRefBehavior {
   FMRB_DoesNotReadMemory = FMRL_Anywhere | MRI_Mod,
 
   /// This indicates that the function could not be classified into one of the
   /// behaviors above.
   FMRB_UnknownModRefBehavior = FMRL_Anywhere | MRI_ModRef
 };
 
 class AAResults {
 public:
   // Make these results default constructable and movable. We have to spell
   // these out because MSVC won't synthesize them.
   AAResults(const TargetLibraryInfo &TLI) : TLI(TLI) {}
   AAResults(AAResults &&Arg);
   ~AAResults();
 
   /// Register a specific AA result.
   template <typename AAResultT> void addAAResult(AAResultT &AAResult) {
public:
   ///
   /// For function calls, this delegates to the alias-analysis specific
   /// call-site mod-ref behavior queries. Otherwise it delegates to the specific
   /// helpers above.
   ModRefInfo getModRefInfo(const Instruction *I,
                            const Optional<MemoryLocation> &OptLoc) {
     if (OptLoc == None) {
       if (auto CS = ImmutableCallSite(I)) {
-        auto MRB = getModRefBehavior(CS);
-        if ((MRB & MRI_ModRef) == MRI_ModRef)
-          return MRI_ModRef;
-        if (MRB & MRI_Ref)
-          return MRI_Ref;
-        if (MRB & MRI_Mod)
-          return MRI_Mod;
-        return MRI_NoModRef;
+        return ModRefInfo(getModRefBehavior(CS));
       }
     }
 
     const MemoryLocation &Loc = OptLoc.getValueOr(MemoryLocation());
 
     switch (I->getOpcode()) {
     case Instruction::VAArg:  return getModRefInfo((const VAArgInst*)I, Loc);
     case Instruction::Load:   return getModRefInfo((const LoadInst*)I,  Loc);
public:
 
   /// Return information about whether two call sites may refer to the same set
   /// of memory locations. See the AA documentation for details:
   ///   http://llvm.org/docs/AliasAnalysis.html#ModRefInfo
   ModRefInfo getModRefInfo(ImmutableCallSite CS1, ImmutableCallSite CS2);
 
   /// \brief Return information about whether a particular call site modifies
   /// or reads the specified memory location \p MemLoc before instruction \p I
-  /// in a BasicBlock. A ordered basic block \p OBB can be used to speed up
+  /// in a BasicBlock. An ordered basic block \p OBB can be used to speed up
   /// instruction ordering queries inside the BasicBlock containing \p I.
   ModRefInfo callCapturesBefore(const Instruction *I,
                                 const MemoryLocation &MemLoc, DominatorTree *DT,
                                 OrderedBasicBlock *OBB = nullptr);
 
   /// \brief A convenience wrapper to synthesize a memory location.
   ModRefInfo callCapturesBefore(const Instruction *I, const Value *P,
                                 uint64_t Size, DominatorTree *DT,

lib/Analysis/AliasAnalysis.cpp

bool AAResults::pointsToConstantMemory(const MemoryLocation &Loc,
 
   return false;
 }
 
 ModRefInfo AAResults::getArgModRefInfo(ImmutableCallSite CS, unsigned ArgIdx) {
   ModRefInfo Result = MRI_ModRef;
 
   for (const auto &AA : AAs) {
-    Result = ModRefInfo(Result & AA->getArgModRefInfo(CS, ArgIdx));
+    Result = modRefIntersection(Result, AA->getArgModRefInfo(CS, ArgIdx));
 
     // Early-exit the moment we reach the bottom of the lattice.
-    if (Result == MRI_NoModRef)
+    if (isNoModRef(Result))
       return Result;
   }
 
   return Result;
 }
 
 ModRefInfo AAResults::getModRefInfo(Instruction *I, ImmutableCallSite Call) {
   // We may have two calls
   if (auto CS = ImmutableCallSite(I)) {
     // Check if the two calls modify the same memory
     return getModRefInfo(CS, Call);
   } else if (I->isFenceLike()) {
     // If this is a fence, just return MRI_ModRef.
     return MRI_ModRef;
   } else {
     // Otherwise, check if the call modifies or references the
     // location this memory access defines.  The best we can say
     // is that if the call references what this instruction
     // defines, it must be clobbered by this location.
     const MemoryLocation DefLoc = MemoryLocation::get(I);
-    if (getModRefInfo(Call, DefLoc) != MRI_NoModRef)
-      return MRI_ModRef;
+    ModRefInfo MR = getModRefInfo(Call, DefLoc);
+    if (modOrRefSet(MR))
+      return setModAndRef(MR);
   }
   return MRI_NoModRef;
 }
 
 ModRefInfo AAResults::getModRefInfo(ImmutableCallSite CS,
                                     const MemoryLocation &Loc) {
   ModRefInfo Result = MRI_ModRef;
 
   for (const auto &AA : AAs) {
-    Result = ModRefInfo(Result & AA->getModRefInfo(CS, Loc));
+    Result = modRefIntersection(Result, AA->getModRefInfo(CS, Loc));
 
     // Early-exit the moment we reach the bottom of the lattice.
-    if (Result == MRI_NoModRef)
+    if (isNoModRef(Result))
       return Result;
   }
 
   // Try to refine the mod-ref info further using other API entry points to the
   // aggregate set of AA results.
   auto MRB = getModRefBehavior(CS);
   if (MRB == FMRB_DoesNotAccessMemory ||
       MRB == FMRB_OnlyAccessesInaccessibleMem)
     return MRI_NoModRef;
 
   if (onlyReadsMemory(MRB))
-    Result = ModRefInfo(Result & MRI_Ref);
+    Result = clearMod(Result);
   else if (doesNotReadMemory(MRB))
-    Result = ModRefInfo(Result & MRI_Mod);
+    Result = clearRef(Result);
 
   if (onlyAccessesArgPointees(MRB) || onlyAccessesInaccessibleOrArgMem(MRB)) {
     bool DoesAlias = false;
     ModRefInfo AllArgsMask = MRI_NoModRef;
     if (doesAccessArgPointees(MRB)) {
       for (auto AI = CS.arg_begin(), AE = CS.arg_end(); AI != AE; ++AI) {
         const Value *Arg = *AI;
         if (!Arg->getType()->isPointerTy())
           continue;
         unsigned ArgIdx = std::distance(CS.arg_begin(), AI);
         MemoryLocation ArgLoc = MemoryLocation::getForArgument(CS, ArgIdx, TLI);
         AliasResult ArgAlias = alias(ArgLoc, Loc);
         if (ArgAlias != NoAlias) {
           ModRefInfo ArgMask = getArgModRefInfo(CS, ArgIdx);
           DoesAlias = true;
-          AllArgsMask = ModRefInfo(AllArgsMask | ArgMask);
+          AllArgsMask = modRefUnion(AllArgsMask, ArgMask);
         }
       }
     }
+    // Return MRI_NoModRef if no alias found with any argument.
     if (!DoesAlias)
       return MRI_NoModRef;
-    Result = ModRefInfo(Result & AllArgsMask);
+    // Logical & between other AA analyses and argument analysis.
+    Result = modRefIntersection(Result, AllArgsMask);
   }
 
   // If Loc is a constant memory location, the call definitely could not
   // modify the memory location.
-  if ((Result & MRI_Mod) &&
-      pointsToConstantMemory(Loc, /*OrLocal*/ false))
-    Result = ModRefInfo(Result & ~MRI_Mod);
+  if (modSet(Result) && pointsToConstantMemory(Loc, /*OrLocal*/ false))
+    Result = clearMod(Result);
 
   return Result;
 }
 
 ModRefInfo AAResults::getModRefInfo(ImmutableCallSite CS1,
                                     ImmutableCallSite CS2) {
   ModRefInfo Result = MRI_ModRef;
 
   for (const auto &AA : AAs) {
-    Result = ModRefInfo(Result & AA->getModRefInfo(CS1, CS2));
+    Result = modRefIntersection(Result, AA->getModRefInfo(CS1, CS2));
 
     // Early-exit the moment we reach the bottom of the lattice.
-    if (Result == MRI_NoModRef)
+    if (isNoModRef(Result))
       return Result;
   }
 
   // Try to refine the mod-ref info further using other API entry points to the
   // aggregate set of AA results.
 
   // If CS1 or CS2 are readnone, they don't interact.
   auto CS1B = getModRefBehavior(CS1);
   if (CS1B == FMRB_DoesNotAccessMemory)
     return MRI_NoModRef;
 
   auto CS2B = getModRefBehavior(CS2);
   if (CS2B == FMRB_DoesNotAccessMemory)
     return MRI_NoModRef;
 
   // If they both only read from memory, there is no dependence.
   if (onlyReadsMemory(CS1B) && onlyReadsMemory(CS2B))
     return MRI_NoModRef;
 
   // If CS1 only reads memory, the only dependence on CS2 can be
   // from CS1 reading memory written by CS2.
   if (onlyReadsMemory(CS1B))
-    Result = ModRefInfo(Result & MRI_Ref);
+    Result = clearMod(Result);
   else if (doesNotReadMemory(CS1B))
-    Result = ModRefInfo(Result & MRI_Mod);
+    Result = clearRef(Result);
 
   // If CS2 only access memory through arguments, accumulate the mod/ref
   // information from CS1's references to the memory referenced by
   // CS2's arguments.
   if (onlyAccessesArgPointees(CS2B)) {
     ModRefInfo R = MRI_NoModRef;
     if (doesAccessArgPointees(CS2B)) {
       for (auto I = CS2.arg_begin(), E = CS2.arg_end(); I != E; ++I) {
         const Value *Arg = *I;
         if (!Arg->getType()->isPointerTy())
           continue;
         unsigned CS2ArgIdx = std::distance(CS2.arg_begin(), I);
         auto CS2ArgLoc = MemoryLocation::getForArgument(CS2, CS2ArgIdx, TLI);
 
-        // ArgMask indicates what CS2 might do to CS2ArgLoc, and the dependence
-        // of CS1 on that location is the inverse.
-        ModRefInfo ArgMask = getArgModRefInfo(CS2, CS2ArgIdx);
-        if (ArgMask == MRI_Mod)
+        // ArgModRefCS2 indicates what CS2 might do to CS2ArgLoc, and the
+        // dependence of CS1 on that location is the inverse:
+        // - If CS2 modifies location, dependence exists if CS1 reads or writes.
+        // - If CS2 only reads location, dependence exists if CS1 writes.
+        ModRefInfo ArgModRefCS2 = getArgModRefInfo(CS2, CS2ArgIdx);
+        ModRefInfo ArgMask;
+        if (modSet(ArgModRefCS2))
           ArgMask = MRI_ModRef;
-        else if (ArgMask == MRI_Ref)
+        else if (refSet(ArgModRefCS2))
           ArgMask = MRI_Mod;
 
-        ArgMask = ModRefInfo(ArgMask & getModRefInfo(CS1, CS2ArgLoc));
+        // ModRefCS1 indicates what CS1 might do to CS2ArgLoc, and we use
+        // above ArgMask to update dependence info.
+        ModRefInfo ModRefCS1 = getModRefInfo(CS1, CS2ArgLoc);
+        ArgMask = modRefIntersection(ArgMask, ModRefCS1);
 
-        R = ModRefInfo((R | ArgMask) & Result);
-        if (R == Result)
+        R = modRefIntersection(modRefUnion(R, ArgMask), Result);
+        if (R == Result) {
           break;
         }
       }
+    }
     return R;
   }
 
   // If CS1 only accesses memory through arguments, check if CS2 references
   // any of the memory referenced by CS1's arguments. If not, return NoModRef.
   if (onlyAccessesArgPointees(CS1B)) {
     ModRefInfo R = MRI_NoModRef;
     if (doesAccessArgPointees(CS1B)) {
       for (auto I = CS1.arg_begin(), E = CS1.arg_end(); I != E; ++I) {
         const Value *Arg = *I;
         if (!Arg->getType()->isPointerTy())
           continue;
         unsigned CS1ArgIdx = std::distance(CS1.arg_begin(), I);
         auto CS1ArgLoc = MemoryLocation::getForArgument(CS1, CS1ArgIdx, TLI);
 
-        // ArgMask indicates what CS1 might do to CS1ArgLoc; if CS1 might Mod
-        // CS1ArgLoc, then we care about either a Mod or a Ref by CS2. If CS1
-        // might Ref, then we care only about a Mod by CS2.
-        ModRefInfo ArgMask = getArgModRefInfo(CS1, CS1ArgIdx);
-        ModRefInfo ArgR = getModRefInfo(CS2, CS1ArgLoc);
-        if (((ArgMask & MRI_Mod) != MRI_NoModRef &&
-             (ArgR & MRI_ModRef) != MRI_NoModRef) ||
-            ((ArgMask & MRI_Ref) != MRI_NoModRef &&
-             (ArgR & MRI_Mod) != MRI_NoModRef))
-          R = ModRefInfo((R | ArgMask) & Result);
+        // ArgModRefCS1 indicates what CS1 might do to CS1ArgLoc; if CS1 might
+        // Mod CS1ArgLoc, then we care about either a Mod or a Ref by CS2. If
+        // CS1 might Ref, then we care only about a Mod by CS2.
+        ModRefInfo ArgModRefCS1 = getArgModRefInfo(CS1, CS1ArgIdx);
+        ModRefInfo ModRefCS2 = getModRefInfo(CS2, CS1ArgLoc);
+        if ((modSet(ArgModRefCS1) && modOrRefSet(ModRefCS2)) ||
+            (refSet(ArgModRefCS1) && modSet(ModRefCS2))) {
+          R = modRefIntersection(modRefUnion(R, ArgModRefCS1), Result);
+        }
 
-        if (R == Result)
+        if (R == Result) {
           break;
         }
       }
+    }
     return R;
   }
 
   return Result;
 }
 
 FunctionModRefBehavior AAResults::getModRefBehavior(ImmutableCallSite CS) {
   FunctionModRefBehavior Result = FMRB_UnknownModRefBehavior;

 ModRefInfo AAResults::getModRefInfo(const LoadInst *L,
                                     const MemoryLocation &Loc) {
   // Be conservative in the face of atomic.
   if (isStrongerThan(L->getOrdering(), AtomicOrdering::Unordered))
     return MRI_ModRef;
 
   // If the load address doesn't alias the given address, it doesn't read
   // or write the specified memory.
-  if (Loc.Ptr && !alias(MemoryLocation::get(L), Loc))
+  if (Loc.Ptr) {
+    AliasResult AR = alias(MemoryLocation::get(L), Loc);
+    if (AR == NoAlias)
       return MRI_NoModRef;
-
+    if (AR == MustAlias)
+      return MRI_Ref;
+  }
   // Otherwise, a load just reads.
   return MRI_Ref;
 }
 
 ModRefInfo AAResults::getModRefInfo(const StoreInst *S,
                                     const MemoryLocation &Loc) {
   // Be conservative in the face of atomic.
   if (isStrongerThan(S->getOrdering(), AtomicOrdering::Unordered))
     return MRI_ModRef;
 
   if (Loc.Ptr) {
+    AliasResult AR = alias(MemoryLocation::get(S), Loc);
     // If the store address cannot alias the pointer in question, then the
     // specified memory cannot be modified by the store.
-    if (!alias(MemoryLocation::get(S), Loc))
+    if (AR == NoAlias)
       return MRI_NoModRef;
 
     // If the pointer is a pointer to constant memory, then it could not have
     // been modified by this store.
     if (pointsToConstantMemory(Loc))
       return MRI_NoModRef;
   }
 
   // Otherwise, a store just writes.
   return MRI_Mod;
 }
 
 ModRefInfo AAResults::getModRefInfo(const FenceInst *S, const MemoryLocation &Loc) {
   // If we know that the location is a constant memory location, the fence
   // cannot modify this location.
   if (Loc.Ptr && pointsToConstantMemory(Loc))
     return MRI_Ref;
   return MRI_ModRef;
 }
 
 ModRefInfo AAResults::getModRefInfo(const VAArgInst *V,
                                     const MemoryLocation &Loc) {
   if (Loc.Ptr) {
+    AliasResult AR = alias(MemoryLocation::get(V), Loc);
     // If the va_arg address cannot alias the pointer in question, then the
     // specified memory cannot be accessed by the va_arg.
-    if (!alias(MemoryLocation::get(V), Loc))
+    if (AR == NoAlias)
       return MRI_NoModRef;
 
     // If the pointer is a pointer to constant memory, then it could not have
     // been modified by this va_arg.
     if (pointsToConstantMemory(Loc))
       return MRI_NoModRef;
   }
 

 }
 
 ModRefInfo AAResults::getModRefInfo(const AtomicCmpXchgInst *CX,
                                     const MemoryLocation &Loc) {
   // Acquire/Release cmpxchg has properties that matter for arbitrary addresses.
   if (isStrongerThanMonotonic(CX->getSuccessOrdering()))
     return MRI_ModRef;
 
-  // If the cmpxchg address does not alias the location, it does not access it.
-  if (Loc.Ptr && !alias(MemoryLocation::get(CX), Loc))
+  if (Loc.Ptr) {
+    AliasResult AR = alias(MemoryLocation::get(CX), Loc);
+    // If the cmpxchg address does not alias the location, it does not access
+    // it.
+    if (AR == NoAlias)
       return MRI_NoModRef;
+  }
 
   return MRI_ModRef;
 }
 
 ModRefInfo AAResults::getModRefInfo(const AtomicRMWInst *RMW,
                                     const MemoryLocation &Loc) {
   // Acquire/Release atomicrmw has properties that matter for arbitrary addresses.
   if (isStrongerThanMonotonic(RMW->getOrdering()))
     return MRI_ModRef;
 
-  // If the atomicrmw address does not alias the location, it does not access it.
-  if (Loc.Ptr && !alias(MemoryLocation::get(RMW), Loc))
+  if (Loc.Ptr) {
+    AliasResult AR = alias(MemoryLocation::get(RMW), Loc);
+    // If the atomicrmw address does not alias the location, it does not access
+    // it.
+    if (AR == NoAlias)
       return MRI_NoModRef;
+  }
 
   return MRI_ModRef;
 }
 
 /// \brief Return information about whether a particular call site modifies
 /// or reads the specified memory location \p MemLoc before instruction \p I
-/// in a BasicBlock. A ordered basic block \p OBB can be used to speed up
+/// in a BasicBlock. An ordered basic block \p OBB can be used to speed up
 /// instruction-ordering queries inside the BasicBlock containing \p I.
 /// FIXME: this is really just shoring-up a deficiency in alias analysis.
 /// BasicAA isn't willing to spend linear time determining whether an alloca
 /// was captured before or after this particular call, while we are. However,
 /// with a smarter AA in place, this test is just wasting compile time.
 ModRefInfo AAResults::callCapturesBefore(const Instruction *I,
                                          const MemoryLocation &MemLoc,
                                          DominatorTree *DT,
ModRefInfo AAResults::callCapturesBefore(const Instruction *I,
   if (PointerMayBeCapturedBefore(Object, /* ReturnCaptures */ true,
                                  /* StoreCaptures */ true, I, DT,
                                  /* include Object */ true,
                                  /* OrderedBasicBlock */ OBB))
     return MRI_ModRef;
 
   unsigned ArgNo = 0;
   ModRefInfo R = MRI_NoModRef;
+  // Set flag only if no May found and all operands processed.
   for (auto CI = CS.data_operands_begin(), CE = CS.data_operands_end();
        CI != CE; ++CI, ++ArgNo) {
     // Only look at the no-capture or byval pointer arguments.  If this
     // pointer were passed to arguments that were neither of these, then it
     // couldn't be no-capture.
     if (!(*CI)->getType()->isPointerTy() ||
         (!CS.doesNotCapture(ArgNo) &&
          ArgNo < CS.getNumArgOperands() && !CS.isByValArgument(ArgNo)))
       continue;
 
+    AliasResult AR = alias(MemoryLocation(*CI), MemoryLocation(Object));
     // If this is a no-capture pointer argument, see if we can tell that it
     // is impossible to alias the pointer we're checking.  If not, we have to
     // assume that the call could touch the pointer, even though it doesn't
     // escape.
-    if (isNoAlias(MemoryLocation(*CI), MemoryLocation(Object)))
+    if (AR == NoAlias)
       continue;
     if (CS.doesNotAccessMemory(ArgNo))
       continue;
     if (CS.onlyReadsMemory(ArgNo)) {
       R = MRI_Ref;
       continue;
     }
     return MRI_ModRef;
bool AAResults::canInstructionRangeModRef(const Instruction &I1,
                                           const ModRefInfo Mode) {
   assert(I1.getParent() == I2.getParent() &&
          "Instructions not in same basic block!");
   BasicBlock::const_iterator I = I1.getIterator();
   BasicBlock::const_iterator E = I2.getIterator();
   ++E;  // Convert from inclusive to exclusive range.
 
   for (; I != E; ++I) // Check every instruction in range
-    if (getModRefInfo(&*I, Loc) & Mode)
+    if (modRefIntersection(getModRefInfo(&*I, Loc), Mode))
       return true;
   return false;
 }
 
 // Provide a definition for the root virtual destructor.
 AAResults::Concept::~Concept() = default;
 
 // Provide a definition for the static object used to identify passes.

lib/Analysis/AliasAnalysisEvaluator.cpp

 static cl::opt<bool> PrintAll("print-all-alias-modref-info", cl::ReallyHidden);
 
 static cl::opt<bool> PrintNoAlias("print-no-aliases", cl::ReallyHidden);
 static cl::opt<bool> PrintMayAlias("print-may-aliases", cl::ReallyHidden);
 static cl::opt<bool> PrintPartialAlias("print-partial-aliases", cl::ReallyHidden);
 static cl::opt<bool> PrintMustAlias("print-must-aliases", cl::ReallyHidden);
 
 static cl::opt<bool> PrintNoModRef("print-no-modref", cl::ReallyHidden);
-static cl::opt<bool> PrintMod("print-mod", cl::ReallyHidden);
 static cl::opt<bool> PrintRef("print-ref", cl::ReallyHidden);
+static cl::opt<bool> PrintMod("print-mod", cl::ReallyHidden);
 static cl::opt<bool> PrintModRef("print-modref", cl::ReallyHidden);
 
 static cl::opt<bool> EvalAAMD("evaluate-aa-metadata", cl::ReallyHidden);
 
 static void PrintResults(const char *Msg, bool P, const Value *V1,
                          const Value *V2, const Module *M) {
   if (PrintAll || P) {
     std::string o1, o2;
for (auto Pointer : Pointers) {
       if (ElTy->isSized()) Size = DL.getTypeStoreSize(ElTy);
 
       switch (AA.getModRefInfo(C, Pointer, Size)) {
       case MRI_NoModRef:
         PrintModRefResults("NoModRef", PrintNoModRef, I, Pointer,
                            F.getParent());
         ++NoModRefCount;
         break;
-      case MRI_Mod:
-        PrintModRefResults("Just Mod", PrintMod, I, Pointer, F.getParent());
-        ++ModCount;
-        break;
       case MRI_Ref:
         PrintModRefResults("Just Ref", PrintRef, I, Pointer, F.getParent());
         ++RefCount;
         break;
+      case MRI_Mod:
+        PrintModRefResults("Just Mod", PrintMod, I, Pointer, F.getParent());
+        ++ModCount;
+        break;
       case MRI_ModRef:
         PrintModRefResults("Both ModRef", PrintModRef, I, Pointer,
                            F.getParent());
         ++ModRefCount;
         break;
       }
     }
   }
 
   // Mod/ref alias analysis: compare all pairs of calls
   for (auto C = CallSites.begin(), Ce = CallSites.end(); C != Ce; ++C) {
     for (auto D = CallSites.begin(); D != Ce; ++D) {
       if (D == C)
         continue;
       switch (AA.getModRefInfo(*C, *D)) {
       case MRI_NoModRef:
         PrintModRefResults("NoModRef", PrintNoModRef, *C, *D, F.getParent());
         ++NoModRefCount;
         break;
-      case MRI_Mod:
-        PrintModRefResults("Just Mod", PrintMod, *C, *D, F.getParent());
-        ++ModCount;
-        break;
       case MRI_Ref:
         PrintModRefResults("Just Ref", PrintRef, *C, *D, F.getParent());
         ++RefCount;
         break;
+      case MRI_Mod:
+        PrintModRefResults("Just Mod", PrintMod, *C, *D, F.getParent());
+        ++ModCount;
+        break;
       case MRI_ModRef:
         PrintModRefResults("Both ModRef", PrintModRef, *C, *D, F.getParent());
         ++ModRefCount;
         break;
       }
     }
   }
 }
if (AliasSum == 0) {
     errs() << "  Alias Analysis Evaluator Pointer Alias Summary: "
            << NoAliasCount * 100 / AliasSum << "%/"
            << MayAliasCount * 100 / AliasSum << "%/"
            << PartialAliasCount * 100 / AliasSum << "%/"
            << MustAliasCount * 100 / AliasSum << "%\n";
   }
 
   // Display the summary for mod/ref analysis
-  int64_t ModRefSum = NoModRefCount + ModCount + RefCount + ModRefCount;
+  int64_t ModRefSum = NoModRefCount + RefCount + ModCount + ModRefCount;
   if (ModRefSum == 0) {
     errs() << "  Alias Analysis Mod/Ref Evaluator Summary: no "
               "mod/ref!\n";
   } else {
     errs() << "  " << ModRefSum << " Total ModRef Queries Performed\n";
     errs() << "  " << NoModRefCount << " no mod/ref responses ";
     PrintPercent(NoModRefCount, ModRefSum);
     errs() << "  " << ModCount << " mod responses ";

lib/Analysis/AliasSetTracker.cpp

for (iterator I = begin(), E = end(); I != E; ++I)
     if (AA.alias(MemoryLocation(Ptr, Size, AAInfo),
                  MemoryLocation(I.getPointer(), I.getSize(), I.getAAInfo())))
       return true;
 
   // Check the unknown instructions...
   if (!UnknownInsts.empty()) {
     for (unsigned i = 0, e = UnknownInsts.size(); i != e; ++i)
       if (auto *Inst = getUnknownInst(i))
-        if (AA.getModRefInfo(Inst, MemoryLocation(Ptr, Size, AAInfo)) !=
-            MRI_NoModRef)
+        if (modOrRefSet(
+                AA.getModRefInfo(Inst, MemoryLocation(Ptr, Size, AAInfo))))
           return true;
   }
 
   return false;
 }
 
 bool AliasSet::aliasesUnknownInst(const Instruction *Inst,
                                   AliasAnalysis &AA) const {
 
   if (AliasAny)
     return true;
 
   if (!Inst->mayReadOrWriteMemory())
     return false;
 
   for (unsigned i = 0, e = UnknownInsts.size(); i != e; ++i) {
     if (auto *UnknownInst = getUnknownInst(i)) {
       ImmutableCallSite C1(UnknownInst), C2(Inst);
-      if (!C1 || !C2 || AA.getModRefInfo(C1, C2) != MRI_NoModRef ||
-          AA.getModRefInfo(C2, C1) != MRI_NoModRef)
+      if (!C1 || !C2 || modOrRefSet(AA.getModRefInfo(C1, C2)) ||
+          modOrRefSet(AA.getModRefInfo(C2, C1)))
         return true;
     }
   }
 
   for (iterator I = begin(), E = end(); I != E; ++I)
-    if (AA.getModRefInfo(Inst, MemoryLocation(I.getPointer(), I.getSize(),
-                                              I.getAAInfo())) != MRI_NoModRef)
+    if (modOrRefSet(AA.getModRefInfo(
+            Inst, MemoryLocation(I.getPointer(), I.getSize(), I.getAAInfo()))))
       return true;
 
   return false;
 }
 
 void AliasSetTracker::clear() {
   // Delete all the PointerRec entries.
   for (PointerMapType::iterator I = PointerMap.begin(), E = PointerMap.end();
AliasSet &AliasSetTracker::mergeAllAliasSets() {
   // sets to it.
   AliasSets.push_back(new AliasSet());
   AliasAnyAS = &AliasSets.back();
   AliasAnyAS->Alias = AliasSet::SetMayAlias;
   AliasAnyAS->Access = AliasSet::ModRefAccess;
   AliasAnyAS->AliasAny = true;
 
   for (auto Cur : ASVector) {
-    
     // If Cur was already forwarding, just forward to the new AS instead.
     AliasSet *FwdTo = Cur->Forward;
     if (FwdTo) {
       Cur->Forward = AliasAnyAS;
-      AliasAnyAS->addRef();      
+      AliasAnyAS->addRef();
       FwdTo->dropRef(*this);
       continue;
     }
 
     // Otherwise, perform the actual merge.
     AliasAnyAS->mergeSetIn(*Cur, *this);
   }
 

lib/Analysis/BasicAliasAnalysis.cpp

for (auto CI = CS.data_operands_begin(), CE = CS.data_operands_end();
           getBestAAResults().alias(MemoryLocation(*CI), MemoryLocation(Object));
 
       // Operand doesnt alias 'Object', continue looking for other aliases
       if (AR == NoAlias)
         continue;
       // Operand aliases 'Object', but call doesn't modify it. Strengthen
       // initial assumption and keep looking in case if there are more aliases.
       if (CS.onlyReadsMemory(OperandNo)) {
-        Result = static_cast<ModRefInfo>(Result | MRI_Ref);
+        Result = setRef(Result);
         continue;
       }
       // Operand aliases 'Object' but call only writes into it.
       if (CS.doesNotReadMemory(OperandNo)) {
-        Result = static_cast<ModRefInfo>(Result | MRI_Mod);
+        Result = setMod(Result);
         continue;
       }
       // This operand aliases 'Object' and call reads and writes into it.
       Result = MRI_ModRef;
       break;
     }
 
     // Early return if we improved mod ref information
     if (Result != MRI_ModRef)
       return Result;
   }
 
   // If the CallSite is to malloc or calloc, we can assume that it doesn't
   // modify any IR visible value.  This is only valid because we assume these
   // routines do not read values visible in the IR.  TODO: Consider special
   // casing realloc and strdup routines which access only their arguments as
   // well.  Or alternatively, replace all of this with inaccessiblememonly once
-  // that's implemented fully. 
+  // that's implemented fully.
   auto *Inst = CS.getInstruction();
   if (isMallocOrCallocLikeFn(Inst, &TLI)) {
     // Be conservative if the accessed pointer may alias the allocation -
     // fallback to the generic handling below.
     if (getBestAAResults().alias(MemoryLocation(Inst), Loc) == NoAlias)
       return MRI_NoModRef;
   }
 
if (auto *Inst = dyn_cast<MemCpyInst>(CS.getInstruction())) {
     if ((DestAA = getBestAAResults().alias(MemoryLocation::getForDest(Inst),
                                            Loc)) == MustAlias)
       // The converse case.
       return MRI_Mod;
 
     // It's also possible for Loc to alias both src and dest, or neither.
     ModRefInfo rv = MRI_NoModRef;
     if (SrcAA != NoAlias)
-      rv = static_cast<ModRefInfo>(rv | MRI_Ref);
+      rv = setRef(rv);
     if (DestAA != NoAlias)
-      rv = static_cast<ModRefInfo>(rv | MRI_Mod);
+      rv = setMod(rv);
     return rv;
   }
 
   // While the assume intrinsic is marked as arbitrarily writing so that
   // proper control dependencies will be maintained, it never aliases any
   // particular memory location.
   if (isIntrinsicCall(CS, Intrinsic::assume))
     return MRI_NoModRef;
ModRefInfo BasicAAResult::getModRefInfo(ImmutableCallSite CS1,
   // *Unlike* assumes, guard intrinsics are modeled as reading memory since the
   // heap state at the point the guard is issued needs to be consistent in case
   // the guard invokes the "deopt" continuation.
 
   // NB! This function is *not* commutative, so we specical case two
   // possibilities for guard intrinsics.
 
   if (isIntrinsicCall(CS1, Intrinsic::experimental_guard))
-    return getModRefBehavior(CS2) & MRI_Mod ? MRI_Ref : MRI_NoModRef;
+    return modSet(ModRefInfo(getModRefBehavior(CS2))) ? MRI_Ref : MRI_NoModRef;
 
   if (isIntrinsicCall(CS2, Intrinsic::experimental_guard))
-    return getModRefBehavior(CS1) & MRI_Mod ? MRI_Mod : MRI_NoModRef;
+    return modSet(ModRefInfo(getModRefBehavior(CS1))) ? MRI_Mod : MRI_NoModRef;
 
   // The AAResultBase base class has some smarts, lets use them.
   return AAResultBase::getModRefInfo(CS1, CS2);
 }
 
 /// Provide ad-hoc rules to disambiguate accesses through two GEP operators,
 /// both having the exact same pointer operand.
 static AliasResult aliasSameBasePointerGEPs(const GEPOperator *GEP1,

lib/Analysis/GlobalsModRef.cpp

struct AlignedMapPointerTraits {
     }
     enum { NumLowBitsAvailable = 3 };
     static_assert(alignof(AlignedMap) >= (1 << NumLowBitsAvailable),
                   "AlignedMap insufficiently aligned to have enough low bits.");
   };
 
   /// The bit that flags that this function may read any global. This is
   /// chosen to mix together with ModRefInfo bits.
+  /// FIXME: This assumes ModRefInfo lattice will remain 4 bits!
   enum { MayReadAnyGlobal = 4 };
 
   /// Checks to document the invariants of the bit packing here.
   static_assert((MayReadAnyGlobal & MRI_ModRef) == 0,
                 "ModRef and the MayReadAnyGlobal flag bits overlap.");
   static_assert(((MayReadAnyGlobal | MRI_ModRef) >>
                  AlignedMapPointerTraits::NumLowBitsAvailable) == 0,
                 "Insufficient low bits to store our flag and ModRef info.");
void GlobalsAAResult::DeletionCallbackHandle::deleted() {
   GAR->Handles.erase(I);
   // This object is now destroyed!
 }
 
 FunctionModRefBehavior GlobalsAAResult::getModRefBehavior(const Function *F) {
   FunctionModRefBehavior Min = FMRB_UnknownModRefBehavior;
 
   if (FunctionInfo *FI = getFunctionInfo(F)) {
-    if (FI->getModRefInfo() == MRI_NoModRef)
+    if (!modOrRefSet(FI->getModRefInfo()))
       Min = FMRB_DoesNotAccessMemory;
-    else if ((FI->getModRefInfo() & MRI_Mod) == 0)
+    else if (!modSet(FI->getModRefInfo()))
       Min = FMRB_OnlyReadsMemory;
   }
 
   return FunctionModRefBehavior(AAResultBase::getModRefBehavior(F) & Min);
 }
 
 FunctionModRefBehavior
 GlobalsAAResult::getModRefBehavior(ImmutableCallSite CS) {
   FunctionModRefBehavior Min = FMRB_UnknownModRefBehavior;
 
   if (!CS.hasOperandBundles())
     if (const Function *F = CS.getCalledFunction())
       if (FunctionInfo *FI = getFunctionInfo(F)) {
-        if (FI->getModRefInfo() == MRI_NoModRef)
+        if (!modOrRefSet(FI->getModRefInfo()))
           Min = FMRB_DoesNotAccessMemory;
-        else if ((FI->getModRefInfo() & MRI_Mod) == 0)
+        else if (!modSet(FI->getModRefInfo()))
           Min = FMRB_OnlyReadsMemory;
       }
 
   return FunctionModRefBehavior(AAResultBase::getModRefBehavior(CS) & Min);
 }
 
 /// Returns the function info for the function, or null if we don't have
 /// anything useful to say about it.
for (scc_iterator<CallGraph *> I = scc_begin(&CG); !I.isAtEnd(); ++I) {
     if (KnowNothing) {
       for (auto *Node : SCC)
         FunctionInfos.erase(Node->getFunction());
       continue;
     }
 
     // Scan the function bodies for explicit loads or stores.
     for (auto *Node : SCC) {
-      if (FI.getModRefInfo() == MRI_ModRef)
+      if (modAndRefSet(FI.getModRefInfo()))
         break; // The mod/ref lattice saturates here.
 
       // Don't prove any properties based on the implementation of an optnone
       // function. Function attributes were already used as a best approximation
       // above.
       if (Node->getFunction()->hasFnAttribute(Attribute::OptimizeNone))
         continue;
 
       for (Instruction &I : instructions(Node->getFunction())) {
-        if (FI.getModRefInfo() == MRI_ModRef)
+        if (modAndRefSet(FI.getModRefInfo()))
           break; // The mod/ref lattice saturates here.
 
         // We handle calls specially because the graph-relevant aspects are
         // handled above.
         if (auto CS = CallSite(&I)) {
           if (isAllocationFn(&I, &TLI) || isFreeCall(&I, &TLI)) {
             // FIXME: It is completely unclear why this is necessary and not
             // handled by the above graph code.
for (auto *Node : SCC) {
         // thay read or write memory.
         if (I.mayReadFromMemory())
           FI.addModRefInfo(MRI_Ref);
         if (I.mayWriteToMemory())
           FI.addModRefInfo(MRI_Mod);
       }
     }
 
-    if ((FI.getModRefInfo() & MRI_Mod) == 0)
+    if (!modSet(FI.getModRefInfo()))
       ++NumReadMemFunctions;
-    if (FI.getModRefInfo() == MRI_NoModRef)
+    if (!modOrRefSet(FI.getModRefInfo()))
       ++NumNoMemFunctions;
 
     // Finally, now that we know the full effect on this SCC, clone the
     // information to each function in the SCC.
     // FI is a reference into FunctionInfos, so copy it now so that it doesn't
     // get invalidated if DenseMap decides to re-hash.
     FunctionInfo CachedFI = FI;
     for (unsigned i = 1, e = SCC.size(); i != e; ++i)
ModRefInfo GlobalsAAResult::getModRefInfoForArgument(ImmutableCallSite CS,
   }
 
   // We identified all objects in the argument list, and none of them were GV.
   return MRI_NoModRef;
 }
 
 ModRefInfo GlobalsAAResult::getModRefInfo(ImmutableCallSite CS,
                                           const MemoryLocation &Loc) {
-  unsigned Known = MRI_ModRef;
+  ModRefInfo Known = MRI_ModRef;
 
   // If we are asking for mod/ref info of a direct call with a pointer to a
   // global we are tracking, return information if we have it.
   if (const GlobalValue *GV =
           dyn_cast<GlobalValue>(GetUnderlyingObject(Loc.Ptr, DL)))
     if (GV->hasLocalLinkage())
       if (const Function *F = CS.getCalledFunction())
         if (NonAddressTakenGlobals.count(GV))
           if (const FunctionInfo *FI = getFunctionInfo(F))
-            Known = FI->getModRefInfoForGlobal(*GV) |
-              getModRefInfoForArgument(CS, GV);
+            Known = modRefUnion(FI->getModRefInfoForGlobal(*GV),
+                                getModRefInfoForArgument(CS, GV));
 
-  if (Known == MRI_NoModRef)
+  if (!modOrRefSet(Known))
     return MRI_NoModRef; // No need to query other mod/ref analyses
-  return ModRefInfo(Known & AAResultBase::getModRefInfo(CS, Loc));
+  return modRefIntersection(Known, AAResultBase::getModRefInfo(CS, Loc));
 }
 
 GlobalsAAResult::GlobalsAAResult(const DataLayout &DL,
                                  const TargetLibraryInfo &TLI)
     : AAResultBase(), DL(DL), TLI(TLI) {}
 
 GlobalsAAResult::GlobalsAAResult(GlobalsAAResult &&Arg)
     : AAResultBase(std::move(Arg)), DL(Arg.DL), TLI(Arg.TLI),

lib/Analysis/Loads.cpp

if (StoreInst *SI = dyn_cast<StoreInst>(Inst)) {
       // of alias analysis that is important for reg2mem'd code.
       if ((isa<AllocaInst>(StrippedPtr) || isa<GlobalVariable>(StrippedPtr)) &&
           (isa<AllocaInst>(StorePtr) || isa<GlobalVariable>(StorePtr)) &&
           StrippedPtr != StorePtr)
         continue;
 
       // If we have alias analysis and it says the store won't modify the loaded
       // value, ignore the store.
-      if (AA && (AA->getModRefInfo(SI, StrippedPtr, AccessSize) & MRI_Mod) == 0)
+      if (AA && !modSet(AA->getModRefInfo(SI, StrippedPtr, AccessSize)))
         continue;
 
       // Otherwise the store that may or may not alias the pointer, bail out.
       ++ScanFrom;
       return nullptr;
     }
 
     // If this is some other instruction that may clobber Ptr, bail out.
     if (Inst->mayWriteToMemory()) {
       // If alias analysis claims that it really won't modify the load,
       // ignore it.
-      if (AA &&
-          (AA->getModRefInfo(Inst, StrippedPtr, AccessSize) & MRI_Mod) == 0)
+      if (AA && !modSet(AA->getModRefInfo(Inst, StrippedPtr, AccessSize)))
         continue;
 
       // May modify the pointer, bail out.
       ++ScanFrom;
       return nullptr;
     }
   }
 
   // Got to the start of the block, we didn't find it, but are done for this
   // block.
   return nullptr;
 }

lib/Analysis/MemoryDependenceAnalysis.cpp

while (ScanIt != BB->begin()) {
     if (!Limit)
       return MemDepResult::getUnknown();
 
     // If this inst is a memory op, get the pointer it accessed
     MemoryLocation Loc;
     ModRefInfo MR = GetLocation(Inst, Loc, TLI);
     if (Loc.Ptr) {
       // A simple instruction.
-      if (AA.getModRefInfo(CS, Loc) != MRI_NoModRef)
+      if (modOrRefSet(AA.getModRefInfo(CS, Loc)))
         return MemDepResult::getClobber(Inst);
       continue;
     }
 
     if (auto InstCS = CallSite(Inst)) {
       // If these two calls do not interfere, look past it.
-      switch (AA.getModRefInfo(CS, InstCS)) {
-      case MRI_NoModRef:
+      if (isNoModRef(AA.getModRefInfo(CS, InstCS))) {
         // If the two calls are the same, return InstCS as a Def, so that
         // CS can be found redundant and eliminated.
-        if (isReadOnlyCall && !(MR & MRI_Mod) &&
+        if (isReadOnlyCall && !modSet(MR) &&
             CS.getInstruction()->isIdenticalToWhenDefined(Inst))
           return MemDepResult::getDef(Inst);
 
         // Otherwise if the two calls don't interact (e.g. InstCS is readnone)
         // keep scanning.
         continue;
-      default:
+      } else
         return MemDepResult::getClobber(Inst);
-      }
-    }
 
     // If we could not obtain a pointer for the instruction and the instruction
     // touches memory then assume that this is a dependency.
-    if (MR != MRI_NoModRef)
+    if (modOrRefSet(MR))
       return MemDepResult::getClobber(Inst);
   }
 
   // No dependence found.  If this is the entry block of the function, it is
   // unknown, otherwise it is non-local.
   if (BB != &BB->getParent()->getEntryBlock())
     return MemDepResult::getNonLocal();
   return MemDepResult::getNonFuncLocal();
if (StoreInst *SI = dyn_cast<StoreInst>(Inst)) {
       if (SI->isVolatile())
         if (!QueryInst || isNonSimpleLoadOrStore(QueryInst) ||
             isOtherMemAccess(QueryInst))
           return MemDepResult::getClobber(SI);
 
       // If alias analysis can tell that this store is guaranteed to not modify
       // the query pointer, ignore it.  Use getModRefInfo to handle cases where
       // the query pointer points to constant memory etc.
-      if (AA.getModRefInfo(SI, MemLoc) == MRI_NoModRef)
+      if (!modOrRefSet(AA.getModRefInfo(SI, MemLoc)))
         continue;
 
       // Ok, this store might clobber the query pointer.  Check to see if it is
       // a must alias: in this case, we want to return this as a def.
       MemoryLocation StoreLoc = MemoryLocation::get(SI);
 
       // If we found a pointer, check if it could be the same as our pointer.
       AliasResult R = AA.alias(StoreLoc, MemLoc);
#endif
     // can't bypass the fence if the query instruction is a store.
     if (FenceInst *FI = dyn_cast<FenceInst>(Inst))
       if (isLoad && FI->getOrdering() == AtomicOrdering::Release)
         continue;
 
     // See if this instruction (e.g. a call or vaarg) mod/ref's the pointer.
     ModRefInfo MR = AA.getModRefInfo(Inst, MemLoc);
     // If necessary, perform additional analysis.
-    if (MR == MRI_ModRef)
+    if (modAndRefSet(MR))
       MR = AA.callCapturesBefore(Inst, MemLoc, &DT, &OBB);
     switch (MR) {
     case MRI_NoModRef:
       // If the call has no effect on the queried pointer, just ignore it.
       continue;
     case MRI_Mod:
       return MemDepResult::getClobber(Inst);
     case MRI_Ref:

lib/Analysis/MemorySSA.cpp

case Intrinsic::assume:
       return false;
     default:
       break;
     }
   }
 
   if (UseCS) {
     ModRefInfo I = AA.getModRefInfo(DefInst, UseCS);
-    return I != MRI_NoModRef;
+    return modOrRefSet(I);
   }
 
   if (auto *DefLoad = dyn_cast<LoadInst>(DefInst)) {
     if (auto *UseLoad = dyn_cast<LoadInst>(UseInst)) {
       switch (getLoadReorderability(UseLoad, DefLoad)) {
       case Reorderability::Always:
         return false;
       case Reorderability::Never:
         return true;
       case Reorderability::IfNoAlias:
         return !AA.isNoAlias(UseLoc, MemoryLocation::get(DefLoad));
       }
     }
   }
 
-  return AA.getModRefInfo(DefInst, UseLoc) & MRI_Mod;
+  return modSet(AA.getModRefInfo(DefInst, UseLoc));
 }
 
 static bool instructionClobbersQuery(MemoryDef *MD, const MemoryUseOrDef *MU,
                                      const MemoryLocOrCall &UseMLOC,
                                      AliasAnalysis &AA) {
   // FIXME: This is a temporary hack to allow a single instructionClobbersQuery
   // to exist while MemoryLocOrCall is pushed through places.
   if (UseMLOC.IsCall)
MemoryUseOrDef *MemorySSA::createNewAccess(Instruction *I) {
   // The isOrdered check is used to ensure that volatiles end up as defs
   // (atomics end up as ModRef right now anyway).  Until we separate the
   // ordering chain from the memory chain, this enables people to see at least
   // some relative ordering to volatiles.  Note that getClobberingMemoryAccess
   // will still give an answer that bypasses other volatile loads.  TODO:
   // Separate memory aliasing and ordering into two different chains so that we
   // can precisely represent both "what memory will this read/write/is clobbered
   // by" and "what instructions can I move this past".
-  bool Def = bool(ModRef & MRI_Mod) || isOrdered(I);
-  bool Use = bool(ModRef & MRI_Ref);
+  bool Def = modSet(ModRef) || isOrdered(I);
+  bool Use = refSet(ModRef);
 
   // It's possible for an instruction to not modify memory at all. During
   // construction, we ignore them.
   if (!Def && !Use)
     return nullptr;
 
   assert((Def || Use) &&
          "Trying to create a memory access with a non-memory instruction");

lib/Target/Hexagon/HexagonLoopIdiomRecognition.cpp

mayLoopAccessLocation(Value *Ptr, ModRefInfo Access, Loop *L,
   // TODO: For this to be really effective, we have to dive into the pointer
   // operand in the store.  Store to &A[i] of 100 will always return may alias
   // with store of &A[100], we need to StoreLoc to be "A" with size of 100,
   // which will then no-alias a store to &A[100].
   MemoryLocation StoreLoc(Ptr, AccessSize);
 
   for (auto *B : L->blocks())
     for (auto &I : *B)
-      if (Ignored.count(&I) == 0 && (AA.getModRefInfo(&I, StoreLoc) & Access))
+      if (Ignored.count(&I) == 0 &&
+          modRefIntersection(AA.getModRefInfo(&I, StoreLoc), Access))
         return true;
 
   return false;
 }
 
 void HexagonLoopIdiomRecognize::collectStores(Loop *CurLoop, BasicBlock *BB,
       SmallVectorImpl<StoreInst*> &Stores) {
   Stores.clear();

lib/Transforms/IPO/FunctionAttrs.cpp

if (CS) {
       // Ignore calls to functions in the same SCC, as long as the call sites
       // don't have operand bundles.  Calls with operand bundles are allowed to
       // have memory effects not described by the memory effects of the call
       // target.
       if (!CS.hasOperandBundles() && CS.getCalledFunction() &&
           SCCNodes.count(CS.getCalledFunction()))
         continue;
       FunctionModRefBehavior MRB = AAR.getModRefBehavior(CS);
+      ModRefInfo MRI = ModRefInfo(MRB);
 
       // If the call doesn't access memory, we're done.
-      if (!(MRB & MRI_ModRef))
+      if (isNoModRef(MRI))
         continue;
 
       if (!AliasAnalysis::onlyAccessesArgPointees(MRB)) {
         // The call could access any memory. If that includes writes, give up.
-        if (MRB & MRI_Mod)
+        if (modSet(MRI))
           return MAK_MayWrite;
         // If it reads, note it.
-        if (MRB & MRI_Ref)
+        if (refSet(MRI))
           ReadsMemory = true;
         continue;
       }
 
       // Check whether all pointer arguments point to local memory, and
       // ignore calls that only access local memory.
       for (CallSite::arg_iterator CI = CS.arg_begin(), CE = CS.arg_end();
            CI != CE; ++CI) {
         Value *Arg = *CI;
         if (!Arg->getType()->isPtrOrPtrVectorTy())
           continue;
 
         AAMDNodes AAInfo;
         I->getAAMetadata(AAInfo);
         MemoryLocation Loc(Arg, MemoryLocation::UnknownSize, AAInfo);
 
         // Skip accesses to local or constant memory as they don't impact the
         // externally visible mod/ref behavior.
         if (AAR.pointsToConstantMemory(Loc, /*OrLocal=*/true))
           continue;
 
-        if (MRB & MRI_Mod)
+        if (modSet(MRI))
           // Writes non-local memory.  Give up.
           return MAK_MayWrite;
-        if (MRB & MRI_Ref)
+        if (refSet(MRI))
           // Ok, it reads non-local memory.
           ReadsMemory = true;
       }
       continue;
     } else if (LoadInst *LI = dyn_cast<LoadInst>(I)) {
       // Ignore non-volatile loads from local memory. (Atomic is okay here.)
       if (!LI->isVolatile()) {
         MemoryLocation Loc = MemoryLocation::get(LI);

lib/Transforms/ObjCARC/ObjCARCContract.cpp

if (Store) {
       }
 
       // Otherwise, be conservative and return nullptr.
       return nullptr;
     }
 
     // Ok, now we know we have not seen a store yet. See if Inst can write to
     // our load location, if it can not, just ignore the instruction.
-    if (!(AA->getModRefInfo(Inst, Loc) & MRI_Mod))
+    if (!modSet(AA->getModRefInfo(Inst, Loc)))
       continue;
 
     Store = dyn_cast<StoreInst>(Inst);
 
     // If Inst can, then check if Inst is a simple store. If Inst is not a
     // store or a store that is not simple, then we have some we do not
     // understand writing to this memory implying we can not move the load
     // over the write to any subsequent store that we may find.

lib/Transforms/Scalar/DeadStoreElimination.cpp

if (isFirstBlock) {
       isFirstBlock = false;
     } else {
       // It's not SecondBB or (in case of a loop) the second visit of SecondBB.
       // In this case we also have to look at instructions after SI.
       EI = B->end();
     }
     for (; BI != EI; ++BI) {
       Instruction *I = &*BI;
-      if (I->mayWriteToMemory() && I != SecondI) {
-        auto Res = AA->getModRefInfo(I, MemLoc);
-        if (Res & MRI_Mod)
+      if (I->mayWriteToMemory() && I != SecondI)
+        if (modSet(AA->getModRefInfo(I, MemLoc)))
           return false;
-      }
-    }
     if (B != FirstBB) {
       assert(B != &FirstBB->getParent()->getEntryBlock() &&
           "Should not hit the entry block because SI must be dominated by LI");
       for (auto PredI = pred_begin(B), PE = pred_end(B); PredI != PE; ++PredI) {
         if (!Visited.insert(*PredI).second)
           continue;
         WorkList.push_back(*PredI);
       }
if (auto CS = CallSite(&*BBI)) {
       // pointers.
       if (AA->doesNotAccessMemory(CS))
         continue;
 
       // If the call might load from any of our allocas, then any store above
       // the call is live.
       DeadStackObjects.remove_if([&](Value *I) {
         // See if the call site touches the value.
-        ModRefInfo A = AA->getModRefInfo(CS, I, getPointerSize(I, DL, *TLI));
-
-        return A == MRI_ModRef || A == MRI_Ref;
+        return refSet(AA->getModRefInfo(CS, I, getPointerSize(I, DL, *TLI)));
       });
 
       // If all of the allocas were clobbered by the call then we're not going
       // to find anything else to process.
       if (DeadStackObjects.empty())
         break;
 
       continue;
while (InstDep.isDef() || InstDep.isClobber()) {
       //   store -> P
       //   store -> Q
       //   store -> P
       // we can remove the first store to P even though we don't know if P and Q
       // alias.
       if (DepWrite == &BB.front()) break;
 
       // Can't look past this instruction if it might read 'Loc'.
-      if (AA->getModRefInfo(DepWrite, Loc) & MRI_Ref)
+      if (refSet(AA->getModRefInfo(DepWrite, Loc)))
         break;
 
       InstDep = MD->getPointerDependencyFrom(Loc, /*isLoad=*/ false,
                                              DepWrite->getIterator(), &BB,
                                              /*QueryInst=*/ nullptr, &Limit);
     }
   }
 

lib/Transforms/Scalar/LoopIdiomRecognize.cpp

mayLoopAccessLocation(Value *Ptr, ModRefInfo Access, Loop *L,
   // with store of &A[100], we need to StoreLoc to be "A" with size of 100,
   // which will then no-alias a store to &A[100].
   MemoryLocation StoreLoc(Ptr, AccessSize);
 
   for (Loop::block_iterator BI = L->block_begin(), E = L->block_end(); BI != E;
        ++BI)
     for (Instruction &I : **BI)
       if (IgnoredStores.count(&I) == 0 &&
-          (AA.getModRefInfo(&I, StoreLoc) & Access))
+          modRefIntersection(AA.getModRefInfo(&I, StoreLoc), Access))
         return true;
 
   return false;
 }
 
 // If we have a negative stride, Start refers to the end of the memory location
 // we're trying to memset.  Therefore, we need to recompute the base pointer,
 // which is just Start - BECount*Size.

lib/Transforms/Scalar/MemCpyOptimizer.cpp

 // This method try to lift a store instruction before position P.
 // It will lift the store and its argument + that anything that
 // may alias with these.
 // The method returns true if it was successful.
 static bool moveUp(AliasAnalysis &AA, StoreInst *SI, Instruction *P,
                    const LoadInst *LI) {
   // If the store alias this position, early bail out.
   MemoryLocation StoreLoc = MemoryLocation::get(SI);
-  if (AA.getModRefInfo(P, StoreLoc) != MRI_NoModRef)
+  if (modOrRefSet(AA.getModRefInfo(P, StoreLoc)))
     return false;
 
   // Keep track of the arguments of all instruction we plan to lift
   // so we can make sure to lift them as well if apropriate.
   DenseSet<Instruction*> Args;
   if (auto *Ptr = dyn_cast<Instruction>(SI->getPointerOperand()))
     if (Ptr->getParent() == SI->getParent())
       Args.insert(Ptr);
 
   // Instruction to lift before P.
   SmallVector<Instruction*, 8> ToLift;
 
   // Memory locations of lifted instructions.
   SmallVector<MemoryLocation, 8> MemLocs{StoreLoc};
 
   // Lifted callsites.
   SmallVector<ImmutableCallSite, 8> CallSites;
 
   const MemoryLocation LoadLoc = MemoryLocation::get(LI);
 
   for (auto I = --SI->getIterator(), E = P->getIterator(); I != E; --I) {
     auto *C = &*I;
 
-    bool MayAlias = AA.getModRefInfo(C, None) != MRI_NoModRef;
+    bool MayAlias = modOrRefSet(AA.getModRefInfo(C, None));
 
     bool NeedLift = false;
     if (Args.erase(C))
       NeedLift = true;
     else if (MayAlias) {
       NeedLift = llvm::any_of(MemLocs, [C, &AA](const MemoryLocation &ML) {
-        return AA.getModRefInfo(C, ML);
+        return modOrRefSet(AA.getModRefInfo(C, ML));
       });
 
       if (!NeedLift)
         NeedLift =
             llvm::any_of(CallSites, [C, &AA](const ImmutableCallSite &CS) {
-              return AA.getModRefInfo(C, CS);
+              return modOrRefSet(AA.getModRefInfo(C, CS));
             });
     }
 
     if (!NeedLift)
       continue;
 
     if (MayAlias) {
       // Since LI is implicitly moved downwards past the lifted instructions,
       // none of them may modify its source.
-      if (AA.getModRefInfo(C, LoadLoc) & MRI_Mod)
+      if (modSet(AA.getModRefInfo(C, LoadLoc)))
         return false;
       else if (auto CS = ImmutableCallSite(C)) {
         // If we can't lift this before P, it's game over.
-        if (AA.getModRefInfo(P, CS) != MRI_NoModRef)
+        if (modOrRefSet(AA.getModRefInfo(P, CS)))
           return false;
 
         CallSites.push_back(CS);
       } else if (isa<LoadInst>(C) || isa<StoreInst>(C) || isa<VAArgInst>(C)) {
         // If we can't lift this before P, it's game over.
         auto ML = MemoryLocation::get(C);
-        if (AA.getModRefInfo(P, ML) != MRI_NoModRef)
+        if (modOrRefSet(AA.getModRefInfo(P, ML)))
           return false;
 
         MemLocs.push_back(ML);
       } else
         // We don't know how to lift this instruction.
         return false;
     }
 
if (LI->isSimple() && LI->hasOneUse() &&
         MemoryLocation LoadLoc = MemoryLocation::get(LI);
 
         // We use alias analysis to check if an instruction may store to
         // the memory we load from in between the load and the store. If
         // such an instruction is found, we try to promote there instead
         // of at the store position.
         Instruction *P = SI;
         for (auto &I : make_range(++LI->getIterator(), SI->getIterator())) {
-          if (AA.getModRefInfo(&I, LoadLoc) & MRI_Mod) {
+          if (modSet(AA.getModRefInfo(&I, LoadLoc))) {
             P = &I;
             break;
           }
         }
 
         // We found an instruction that may write to the loaded memory.
         // We can try to promote at this position instead of the store
         // position if nothing alias the store memory after this and the store
if (LI->isSimple() && LI->hasOneUse() &&
         // Check that nothing touches the dest of the "copy" between
         // the call and the store.
         Value *CpyDest = SI->getPointerOperand()->stripPointerCasts();
         bool CpyDestIsLocal = isa<AllocaInst>(CpyDest);
         AliasAnalysis &AA = LookupAliasAnalysis();
         MemoryLocation StoreLoc = MemoryLocation::get(SI);
         for (BasicBlock::iterator I = --SI->getIterator(), E = C->getIterator();
              I != E; --I) {
-          if (AA.getModRefInfo(&*I, StoreLoc) != MRI_NoModRef) {
+          if (modOrRefSet(AA.getModRefInfo(&*I, StoreLoc))) {
             C = nullptr;
             break;
           }
           // The store to dest may never happen if an exception can be thrown
           // between the load and the store.
           if (I->mayThrow() && !CpyDestIsLocal) {
             C = nullptr;
             break;
bool MemCpyOptPass::performCallSlotOptzn(Instruction *cpy, Value *cpyDest,
 
   // In addition to knowing that the call does not access src in some
   // unexpected manner, for example via a global, which we deduce from
   // the use analysis, we also need to know that it does not sneakily
   // access dest.  We rely on AA to figure this out for us.
   AliasAnalysis &AA = LookupAliasAnalysis();
   ModRefInfo MR = AA.getModRefInfo(C, cpyDest, srcSize);
   // If necessary, perform additional analysis.
-  if (MR != MRI_NoModRef)
+  if (modOrRefSet(MR))
     MR = AA.callCapturesBefore(C, cpyDest, srcSize, &DT);
-  if (MR != MRI_NoModRef)
+  if (modOrRefSet(MR))
     return false;
 
   // We can't create address space casts here because we don't know if they're
   // safe for the target.
   if (cpySrc->getType()->getPointerAddressSpace() !=
       cpyDest->getType()->getPointerAddressSpace())
     return false;
   for (unsigned i = 0; i < CS.arg_size(); ++i)

lib/Transforms/Scalar/Sink.cpp

static bool isSafeToMove(Instruction *Inst, AliasAnalysis &AA,
   if (Inst->mayWriteToMemory()) {
     Stores.insert(Inst);
     return false;
   }
 
   if (LoadInst *L = dyn_cast<LoadInst>(Inst)) {
     MemoryLocation Loc = MemoryLocation::get(L);
     for (Instruction *S : Stores)
-      if (AA.getModRefInfo(S, Loc) & MRI_Mod)
+      if (modSet(AA.getModRefInfo(S, Loc)))
         return false;
   }
 
   if (isa<TerminatorInst>(Inst) || isa<PHINode>(Inst) || Inst->isEHPad() ||
       Inst->mayThrow())
     return false;
 
   if (auto CS = CallSite(Inst)) {
     // Convergent operations cannot be made control-dependent on additional
     // values.
     if (CS.hasFnAttr(Attribute::Convergent))
       return false;
 
     for (Instruction *S : Stores)
-      if (AA.getModRefInfo(S, CS) & MRI_Mod)
+      if (modSet(AA.getModRefInfo(S, CS)))
         return false;
   }
 
   return true;
 }
 
 /// IsAcceptableTarget - Return true if it is possible to sink the instruction
 /// in the specified basic block.

lib/Transforms/Scalar/TailRecursionElimination.cpp

static bool canMoveAboveCall(Instruction *I, CallInst *CI, AliasAnalysis *AA) {
   if (LoadInst *L = dyn_cast<LoadInst>(I)) {
     // Loads may always be moved above calls without side effects.
     if (CI->mayHaveSideEffects()) {
       // Non-volatile loads may be moved above a call with side effects if it
       // does not write to memory and the load provably won't trap.
       // Writes to memory only matter if they may alias the pointer
       // being loaded from.
       const DataLayout &DL = L->getModule()->getDataLayout();
-      if ((AA->getModRefInfo(CI, MemoryLocation::get(L)) & MRI_Mod) ||
+      if (modSet(AA->getModRefInfo(CI, MemoryLocation::get(L))) ||
           !isSafeToLoadUnconditionally(L->getPointerOperand(),
                                        L->getAlignment(), DL, L))
         return false;
     }
   }
 
   // Otherwise, if this is a side-effect free instruction, check to make sure
   // that it does not use the return value of the call.  If it doesn't use the