Make `@llvm.objectsize` more conservative with null pointers.

[george.burgess.iv]

Summary

In pursuit of fixing https://llvm.org/bugs/show_bug.cgi?id=23277, we'd like @llvm.objectsize to treat null pointers as opaque/unknown pointers, rather than treating them as pointers to 0 bytes of memory. This patch adds an argument to @llvm.objectsize that lets us control how @llvm.objectsize treats null.

As Sanjoy noted in http://lists.llvm.org/pipermail/llvm-dev/2017-January/108623.html, we can also make a backwards-incompatible change that just makes @llvm.objectsize adopt this null-is-unknown-size behavior in all cases. I dunno what the better option is here, but we could easily do that by throwing basically half of this patch out (all except for the changes to MemoryBuiltins).

If we want the absolutely-minimal patch, we could cut the changes to MemoryBuiltins.cpp down to a few lines (just returning unknown() in ObjectSizeOffsetVisitor::visitConstantPointerNull). Doing so would cause us to delay lowering the calls to @llvm.objectsize as much as possible, which may inhibit some optimizations.

There's a handful of tests with @llvm.objectsize in them that this patch doesn't update, since they don't CHECK for objectsize with a specific set of arguments. If we want, I'm happy to update them to use the three-arg version, as well.

I wasn't entirely sure who to poke with this, since I can't find a code owner for intrinsics. :)

[ahatanak]

lib/Analysis/MemoryBuiltins.cpp

439 ↗

(On Diff #83710)

Is there a possibility that, when NullIsUnknownSize=true and MustSucceed=false, returning a ConstantInt instead of a nullptr can cause InstCombine to prematurely optimize away objectsize?

I'm thinking about a case where the first operand of objectsize is "select %cmp, null, %someptr". If %cmp turns out to be false later and the size of the object "%someptr" points to is known, objectsize can be replaced with the exact size of the object.

[george.burgess.iv]

Addressed feedback

[george.burgess.iv]

lib/Analysis/MemoryBuiltins.cpp

439 ↗

(On Diff #83710)

Good point. I rewrote this to just return unknown, then. :)

[george.burgess.iv]

Ping :)

[george.burgess.iv]

Ping :)

[echristo]

One inline request, otherwise I think this is ok. Not sure if we want to use an i2/i8 rather than a pair of i1, but either way.

-eric

include/llvm/Analysis/MemoryBuiltins.h

146 ↗

(On Diff #84868)

As an enhancement I'd love to see this encapsulated into a preferences struct or something rather than two bools and an enum.

[arsenm]

lib/Analysis/MemoryBuiltins.cpp

602–603 ↗

(On Diff #84868)

I think this will do the wrong thing with non-0 address spaces

test/Transforms/InstCombine/objsize.ll

268 ↗

(On Diff #84868)

Should add some tests with another address space to make sure it isn't treating 0 as an invalid pointer

[george.burgess.iv]

• Swapped to using an options struct for getObjectSize configuration
• Only use relaxed checks in addrspace 0

(also removed tstellarAMD as a reviewer, since their account was deactivated)

[george.burgess.iv]

Thanks for the comments!

include/llvm/Analysis/MemoryBuiltins.h

146 ↗

(On Diff #84868)

Done.

FWIW, most users only care about RoundToAlign, so I played with having a ObjectSizeOpts::withRoundToAlign(bool) method, but I think getObjectSize(..., ObjectSizeOpts{}.withRoundToAlign(RoundToAlign)) ends up being a bit uglier than just making an ObjectSizeOpts and setting RoundToAlign directly. Happy to do withRoundToAlign instead if you think that's better.

lib/Analysis/MemoryBuiltins.cpp

602–603 ↗

(On Diff #84868)

Good point; fixed.

[echristo]

This is OK with me.

[george.burgess.iv]

Thanks for the review and comments, everyone! :)

llvm/trunk/docs/LangRef.rst

 '``llvm.objectsize``' Intrinsic
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
 Syntax:
 """""""
 
 ::
 
-      declare i32 @llvm.objectsize.i32(i8* <object>, i1 <min>)
-      declare i64 @llvm.objectsize.i64(i8* <object>, i1 <min>)
+      declare i32 @llvm.objectsize.i32(i8* <object>, i1 <min>, i1 <nullunknown>)
+      declare i64 @llvm.objectsize.i64(i8* <object>, i1 <min>, i1 <nullunknown>)
 
 Overview:
 """""""""
 
 The ``llvm.objectsize`` intrinsic is designed to provide information to
 the optimizers to determine at compile time whether a) an operation
 (like memcpy) will overflow a buffer that corresponds to an object, or
 b) that a runtime check for overflow isn't necessary. An object in this
 context means an allocation of a specific class, structure, array, or
 other object.
 
 Arguments:
 """"""""""
 
-The ``llvm.objectsize`` intrinsic takes two arguments. The first
-argument is a pointer to or into the ``object``. The second argument is
-a boolean and determines whether ``llvm.objectsize`` returns 0 (if true)
-or -1 (if false) when the object size is unknown. The second argument
-only accepts constants.
+The ``llvm.objectsize`` intrinsic takes three arguments. The first argument is
+a pointer to or into the ``object``. The second argument determines whether
+``llvm.objectsize`` returns 0 (if true) or -1 (if false) when the object size
+is unknown. The third argument controls how ``llvm.objectsize`` acts when
+``null`` is used as its pointer argument. If it's true and the pointer is in
+address space 0, ``null`` is treated as an opaque value with an unknown number
+of bytes. Otherwise, ``llvm.objectsize`` reports 0 bytes available when given
+``null``.
+
+The second and third arguments only accept constants.
 
 Semantics:
 """"""""""
 
 The ``llvm.objectsize`` intrinsic is lowered to a constant representing
 the size of the object concerned. If the size cannot be determined at
 compile time, ``llvm.objectsize`` returns ``i32/i64 -1 or 0`` (depending
 on the ``min`` argument).

llvm/trunk/include/llvm/Analysis/MemoryBuiltins.h

 namespace llvm {
 class CallInst;
 class PointerType;
 class DataLayout;
 class TargetLibraryInfo;
 class Type;
 class Value;
 
-enum class ObjSizeMode {
-  Exact = 0,
-  Min = 1,
-  Max = 2
-};
-
 /// \brief Tests if a value is a call or invoke to a library function that
 /// allocates or reallocates memory (either malloc, calloc, realloc, or strdup
 /// like).
 bool isAllocationFn(const Value *V, const TargetLibraryInfo *TLI,
                     bool LookThroughBitCast = false);
 
 /// \brief Tests if a value is a call or invoke to a function that returns a
 /// NoAlias pointer (including malloc/calloc/realloc/strdup-like functions).
static inline CallInst *isFreeCall(Value *I, const TargetLibraryInfo *TLI) {
   return const_cast<CallInst*>(isFreeCall((const Value*)I, TLI));
 }
 
 
 //===----------------------------------------------------------------------===//
 //  Utility functions to compute size of objects.
 //
 
+/// Various options to control the behavior of getObjectSize.
+struct ObjectSizeOpts {
+  /// Controls how we handle conditional statements with unknown conditions.
+  enum class Mode : uint8_t {
+    /// Fail to evaluate an unknown condition.
+    Exact,
+    /// Evaluate all branches of an unknown condition. If all evaluations
+    /// succeed, pick the minimum size.
+    Min,
+    /// Same as Min, except we pick the maximum size of all of the branches.
+    Max
+  };
+
+  /// How we want to evaluate this object's size.
+  Mode EvalMode = Mode::Exact;
+  /// Whether to round the result up to the alignment of allocas, byval
+  /// arguments, and global variables.
+  bool RoundToAlign = false;
+  /// If this is true, null pointers in address space 0 will be treated as
+  /// though they can't be evaluated. Otherwise, null is always considered to
+  /// point to a 0 byte region of memory.
+  bool NullIsUnknownSize = false;
+};
+
 /// \brief Compute the size of the object pointed by Ptr. Returns true and the
 /// object size in Size if successful, and false otherwise. In this context, by
 /// object we mean the region of memory starting at Ptr to the end of the
 /// underlying object pointed to by Ptr.
-/// If RoundToAlign is true, then Size is rounded up to the aligment of allocas,
-/// byval arguments, and global variables.
-/// If Mode is Min or Max the size will be evaluated even if it depends on
-/// a condition and corresponding value will be returned (min or max).
 bool getObjectSize(const Value *Ptr, uint64_t &Size, const DataLayout &DL,
-                   const TargetLibraryInfo *TLI, bool RoundToAlign = false,
-                   ObjSizeMode Mode = ObjSizeMode::Exact);
+                   const TargetLibraryInfo *TLI, ObjectSizeOpts Opts = {});
 
 /// Try to turn a call to @llvm.objectsize into an integer value of the given
 /// Type. Returns null on failure.
 /// If MustSucceed is true, this function will not return null, and may return
 /// conservative values governed by the second argument of the call to
 /// objectsize.
 ConstantInt *lowerObjectSizeCall(IntrinsicInst *ObjectSize,
                                  const DataLayout &DL,
                                  const TargetLibraryInfo *TLI,
                                  bool MustSucceed);
 
 typedef std::pair<APInt, APInt> SizeOffsetType;
 
 /// \brief Evaluate the size and offset of an object pointed to by a Value*
 /// statically. Fails if size or offset are not known at compile time.
 class ObjectSizeOffsetVisitor
   : public InstVisitor<ObjectSizeOffsetVisitor, SizeOffsetType> {
 
   const DataLayout &DL;
   const TargetLibraryInfo *TLI;
-  bool RoundToAlign;
-  ObjSizeMode Mode;
+  ObjectSizeOpts Options;
   unsigned IntTyBits;
   APInt Zero;
   SmallPtrSet<Instruction *, 8> SeenInsts;
 
   APInt align(APInt Size, uint64_t Align);
 
   SizeOffsetType unknown() {
     return std::make_pair(APInt(), APInt());
   }
 
 public:
   ObjectSizeOffsetVisitor(const DataLayout &DL, const TargetLibraryInfo *TLI,
-                          LLVMContext &Context, bool RoundToAlign = false,
-                          ObjSizeMode Mode = ObjSizeMode::Exact);
+                          LLVMContext &Context, ObjectSizeOpts Options = {});
 
   SizeOffsetType compute(Value *V);
 
   static bool knownSize(const SizeOffsetType &SizeOffset) {
     return SizeOffset.first.getBitWidth() > 1;
   }
 
   static bool knownOffset(const SizeOffsetType &SizeOffset) {

llvm/trunk/include/llvm/IR/Intrinsics.td

 
 // NOTE: these are internal interfaces.
 def int_setjmp     : Intrinsic<[llvm_i32_ty],  [llvm_ptr_ty]>;
 def int_longjmp    : Intrinsic<[], [llvm_ptr_ty, llvm_i32_ty], [IntrNoReturn]>;
 def int_sigsetjmp  : Intrinsic<[llvm_i32_ty] , [llvm_ptr_ty, llvm_i32_ty]>;
 def int_siglongjmp : Intrinsic<[], [llvm_ptr_ty, llvm_i32_ty], [IntrNoReturn]>;
 
 // Internal interface for object size checking
-def int_objectsize : Intrinsic<[llvm_anyint_ty], [llvm_anyptr_ty, llvm_i1_ty],
+def int_objectsize : Intrinsic<[llvm_anyint_ty],
+                               [llvm_anyptr_ty, llvm_i1_ty, llvm_i1_ty],
                                [IntrNoMem]>,
                                GCCBuiltin<"__builtin_object_size">;
 
 //===--------------- Constrained Floating Point Intrinsics ----------------===//
 //
 
 let IntrProperties = [IntrInaccessibleMemOnly] in {
   def int_experimental_constrained_fadd : Intrinsic<[ llvm_anyfloat_ty ],

llvm/trunk/lib/Analysis/BasicAliasAnalysis.cpp

static bool isEscapeSource(const Value *V) {
   return false;
 }
 
 /// Returns the size of the object specified by V or UnknownSize if unknown.
 static uint64_t getObjectSize(const Value *V, const DataLayout &DL,
                               const TargetLibraryInfo &TLI,
                               bool RoundToAlign = false) {
   uint64_t Size;
-  if (getObjectSize(V, Size, DL, &TLI, RoundToAlign))
+  ObjectSizeOpts Opts;
+  Opts.RoundToAlign = RoundToAlign;
+  if (getObjectSize(V, Size, DL, &TLI, Opts))
     return Size;
   return MemoryLocation::UnknownSize;
 }
 
 /// Returns true if we can prove that the object specified by V is smaller than
 /// Size.
 static bool isObjectSmallerThan(const Value *V, uint64_t Size,
                                 const DataLayout &DL,

llvm/trunk/lib/Analysis/MemoryBuiltins.cpp

static APInt getSizeWithOverflow(const SizeOffsetType &Data) {
   return Data.first - Data.second;
 }
 
 /// \brief Compute the size of the object pointed by Ptr. Returns true and the
 /// object size in Size if successful, and false otherwise.
 /// If RoundToAlign is true, then Size is rounded up to the aligment of allocas,
 /// byval arguments, and global variables.
 bool llvm::getObjectSize(const Value *Ptr, uint64_t &Size, const DataLayout &DL,
-                         const TargetLibraryInfo *TLI, bool RoundToAlign,
-                         llvm::ObjSizeMode Mode) {
-  ObjectSizeOffsetVisitor Visitor(DL, TLI, Ptr->getContext(),
-                                  RoundToAlign, Mode);
+                         const TargetLibraryInfo *TLI, ObjectSizeOpts Opts) {
+  ObjectSizeOffsetVisitor Visitor(DL, TLI, Ptr->getContext(), Opts);
   SizeOffsetType Data = Visitor.compute(const_cast<Value*>(Ptr));
   if (!Visitor.bothKnown(Data))
     return false;
 
   Size = getSizeWithOverflow(Data).getZExtValue();
   return true;
 }
 
 ConstantInt *llvm::lowerObjectSizeCall(IntrinsicInst *ObjectSize,
                                        const DataLayout &DL,
                                        const TargetLibraryInfo *TLI,
                                        bool MustSucceed) {
   assert(ObjectSize->getIntrinsicID() == Intrinsic::objectsize &&
          "ObjectSize must be a call to llvm.objectsize!");
 
   bool MaxVal = cast<ConstantInt>(ObjectSize->getArgOperand(1))->isZero();
-  ObjSizeMode Mode;
+  ObjectSizeOpts EvalOptions;
   // Unless we have to fold this to something, try to be as accurate as
   // possible.
   if (MustSucceed)
-    Mode = MaxVal ? ObjSizeMode::Max : ObjSizeMode::Min;
+    EvalOptions.EvalMode =
+        MaxVal ? ObjectSizeOpts::Mode::Max : ObjectSizeOpts::Mode::Min;
   else
-    Mode = ObjSizeMode::Exact;
+    EvalOptions.EvalMode = ObjectSizeOpts::Mode::Exact;
+
+  EvalOptions.NullIsUnknownSize =
+      cast<ConstantInt>(ObjectSize->getArgOperand(2))->isOne();
 
   // FIXME: Does it make sense to just return a failure value if the size won't
   // fit in the output and `!MustSucceed`?
   uint64_t Size;
   auto *ResultType = cast<IntegerType>(ObjectSize->getType());
-  if (getObjectSize(ObjectSize->getArgOperand(0), Size, DL, TLI, false, Mode) &&
+  if (getObjectSize(ObjectSize->getArgOperand(0), Size, DL, TLI, EvalOptions) &&
       isUIntN(ResultType->getBitWidth(), Size))
     return ConstantInt::get(ResultType, Size);
 
   if (!MustSucceed)
     return nullptr;
 
   return ConstantInt::get(ResultType, MaxVal ? -1ULL : 0);
 }
 
 STATISTIC(ObjectVisitorArgument,
           "Number of arguments with unsolved size and offset");
 STATISTIC(ObjectVisitorLoad,
           "Number of load instructions with unsolved size and offset");
 
 
 APInt ObjectSizeOffsetVisitor::align(APInt Size, uint64_t Align) {
-  if (RoundToAlign && Align)
+  if (Options.RoundToAlign && Align)
     return APInt(IntTyBits, alignTo(Size.getZExtValue(), Align));
   return Size;
 }
 
 ObjectSizeOffsetVisitor::ObjectSizeOffsetVisitor(const DataLayout &DL,
                                                  const TargetLibraryInfo *TLI,
                                                  LLVMContext &Context,
-                                                 bool RoundToAlign,
-                                                 ObjSizeMode Mode)
-    : DL(DL), TLI(TLI), RoundToAlign(RoundToAlign), Mode(Mode) {
+                                                 ObjectSizeOpts Options)
+    : DL(DL), TLI(TLI), Options(Options) {
   // Pointer size must be rechecked for each object visited since it could have
   // a different address space.
 }
 
 SizeOffsetType ObjectSizeOffsetVisitor::compute(Value *V) {
   IntTyBits = DL.getPointerTypeSizeInBits(V->getType());
   Zero = APInt::getNullValue(IntTyBits);
 
SizeOffsetType ObjectSizeOffsetVisitor::visitCallSite(CallSite CS) {
   // - strcpy / strncpy
   // - strcat / strncat
   // - memcpy / memmove
   // - strcat / strncat
   // - memset
 }
 
 SizeOffsetType
-ObjectSizeOffsetVisitor::visitConstantPointerNull(ConstantPointerNull&) {
+ObjectSizeOffsetVisitor::visitConstantPointerNull(ConstantPointerNull& CPN) {
+  if (Options.NullIsUnknownSize && CPN.getType()->getAddressSpace() == 0)
+    return unknown();
   return std::make_pair(Zero, Zero);
 }
 
 SizeOffsetType
 ObjectSizeOffsetVisitor::visitExtractElementInst(ExtractElementInst&) {
   return unknown();
 }
 
if (bothKnown(TrueSide) && bothKnown(FalseSide)) {
     }
 
     APInt TrueResult = getSizeWithOverflow(TrueSide);
     APInt FalseResult = getSizeWithOverflow(FalseSide);
 
     if (TrueResult == FalseResult) {
       return TrueSide;
     }
-    if (Mode == ObjSizeMode::Min) {
+    if (Options.EvalMode == ObjectSizeOpts::Mode::Min) {
       if (TrueResult.slt(FalseResult))
         return TrueSide;
       return FalseSide;
     }
-    if (Mode == ObjSizeMode::Max) {
+    if (Options.EvalMode == ObjectSizeOpts::Mode::Max) {
       if (TrueResult.sgt(FalseResult))
         return TrueSide;
       return FalseSide;
     }
   }
   return unknown();
 }
 
if (!bothKnown(Result)) {
     }
   }
 
   SeenVals.clear();
   return Result;
 }
 
 SizeOffsetEvalType ObjectSizeOffsetEvaluator::compute_(Value *V) {
-  ObjectSizeOffsetVisitor Visitor(DL, TLI, Context, RoundToAlign);
+  ObjectSizeOpts ObjSizeOptions;
+  ObjSizeOptions.RoundToAlign = RoundToAlign;
+
+  ObjectSizeOffsetVisitor Visitor(DL, TLI, Context, ObjSizeOptions);
   SizeOffsetType Const = Visitor.compute(V);
   if (Visitor.bothKnown(Const))
     return std::make_pair(ConstantInt::get(Context, Const.first),
                           ConstantInt::get(Context, Const.second));
 
   V = V->stripPointerCasts();
 
   // Check cache.

llvm/trunk/lib/IR/AutoUpgrade.cpp

if (Name.startswith("nvvm.")) {
         NewFn = nullptr;
         return true;
       }
     }
   }
   case 'o':
     // We only need to change the name to match the mangling including the
     // address space.
-    if (F->arg_size() == 2 && Name.startswith("objectsize.")) {
+    if (Name.startswith("objectsize.")) {
       Type *Tys[2] = { F->getReturnType(), F->arg_begin()->getType() };
-      if (F->getName() != Intrinsic::getName(Intrinsic::objectsize, Tys)) {
+      if (F->arg_size() == 2 ||
+          F->getName() != Intrinsic::getName(Intrinsic::objectsize, Tys)) {
         rename(F);
-        NewFn = Intrinsic::getDeclaration(F->getParent(),
-                                          Intrinsic::objectsize, Tys);
+        NewFn = Intrinsic::getDeclaration(F->getParent(), Intrinsic::objectsize,
+                                          Tys);
         return true;
       }
     }
     break;
 
   case 's':
     if (Name == "stackprotectorcheck") {
       NewFn = nullptr;

   case Intrinsic::ctlz:
   case Intrinsic::cttz:
     assert(CI->getNumArgOperands() == 1 &&
            "Mismatch between function args and call args");
     NewCall =
         Builder.CreateCall(NewFn, {CI->getArgOperand(0), Builder.getFalse()});
     break;
 
-  case Intrinsic::objectsize:
-    NewCall =
-        Builder.CreateCall(NewFn, {CI->getArgOperand(0), CI->getArgOperand(1)});
+  case Intrinsic::objectsize: {
+    Value *NullIsUnknownSize = CI->getNumArgOperands() == 2
+                                   ? Builder.getFalse()
+                                   : CI->getArgOperand(2);
+    NewCall = Builder.CreateCall(
+        NewFn, {CI->getArgOperand(0), CI->getArgOperand(1), NullIsUnknownSize});
     break;
+  }
 
   case Intrinsic::ctpop:
     NewCall = Builder.CreateCall(NewFn, {CI->getArgOperand(0)});
     break;
 
   case Intrinsic::convert_from_fp16:
     NewCall = Builder.CreateCall(NewFn, {CI->getArgOperand(0)});
     break;

llvm/trunk/lib/Target/AMDGPU/AMDGPUPromoteAlloca.cpp

for (Value *V : WorkList) {
     case Intrinsic::objectsize: {
       Value *Src = Intr->getOperand(0);
       Type *SrcTy = Src->getType()->getPointerElementType();
       Function *ObjectSize = Intrinsic::getDeclaration(Mod,
         Intrinsic::objectsize,
         { Intr->getType(), PointerType::get(SrcTy, AMDGPUAS::LOCAL_ADDRESS) }
       );
 
-      CallInst *NewCall
-        = Builder.CreateCall(ObjectSize, { Src, Intr->getOperand(1) });
+      CallInst *NewCall = Builder.CreateCall(
+          ObjectSize, {Src, Intr->getOperand(1), Intr->getOperand(2)});
       Intr->replaceAllUsesWith(NewCall);
       Intr->eraseFromParent();
       continue;
     }
     default:
       Intr->print(errs());
       llvm_unreachable("Don't know how to promote alloca intrinsic use.");
     }
   }
 }
 
 FunctionPass *llvm::createAMDGPUPromoteAlloca(const TargetMachine *TM) {
   return new AMDGPUPromoteAlloca(TM);
 }

llvm/trunk/lib/Transforms/Instrumentation/AddressSanitizer.cpp

for (auto &BB : F) {
     }
   }
 
   bool UseCalls =
       CompileKernel ||
       (ClInstrumentationWithCallsThreshold >= 0 &&
        ToInstrument.size() > (unsigned)ClInstrumentationWithCallsThreshold);
   const DataLayout &DL = F.getParent()->getDataLayout();
-  ObjectSizeOffsetVisitor ObjSizeVis(DL, TLI, F.getContext(),
-                                     /*RoundToAlign=*/true);
+  ObjectSizeOpts ObjSizeOpts;
+  ObjSizeOpts.RoundToAlign = true;
+  ObjectSizeOffsetVisitor ObjSizeVis(DL, TLI, F.getContext(), ObjSizeOpts);
 
   // Instrument.
   int NumInstrumented = 0;
   for (auto Inst : ToInstrument) {
     if (ClDebugMin < 0 || ClDebugMax < 0 ||
         (NumInstrumented >= ClDebugMin && NumInstrumented <= ClDebugMax)) {
       if (isInterestingMemoryAccess(Inst, &IsWrite, &TypeSize, &Alignment))
         instrumentMop(ObjSizeVis, Inst, UseCalls,

llvm/trunk/test/Assembler/auto_upgrade_intrinsics.ll

entry:
   ret void
 }
 
 
 @a = private global [60 x i8] zeroinitializer, align 1
 
 define i32 @test.objectsize() {
 ; CHECK-LABEL: @test.objectsize(
-; CHECK: @llvm.objectsize.i32.p0i8
+; CHECK: @llvm.objectsize.i32.p0i8(i8* getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i1 false, i1 false)
   %s = call i32 @llvm.objectsize.i32(i8* getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i1 false)
   ret i32 %s
 }
 
+declare i64 @llvm.objectsize.i64.p0i8(i8*, i1) nounwind readonly
+define i64 @test.objectsize.2() {
+; CHECK-LABEL: @test.objectsize.2(
+; CHECK: @llvm.objectsize.i64.p0i8(i8* getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i1 false, i1 false)
+  %s = call i64 @llvm.objectsize.i64.p0i8(i8* getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i1 false)
+  ret i64 %s
+}
+
+
 declare <2 x double> @llvm.masked.load.v2f64(<2 x double>* %ptrs, i32, <2 x i1> %mask, <2 x double> %src0)
 
 define <2 x double> @tests.masked.load(<2 x double>* %ptr, <2 x i1> %mask, <2 x double> %passthru)  {
 ; CHECK-LABEL: @tests.masked.load(
 ; CHECK: @llvm.masked.load.v2f64.p0v2f64
   %res = call <2 x double> @llvm.masked.load.v2f64(<2 x double>* %ptr, i32 1, <2 x i1> %mask, <2 x double> %passthru)
   ret <2 x double> %res
 }

llvm/trunk/test/CodeGen/AMDGPU/promote-alloca-mem-intrinsics.ll

 ; RUN: opt -S -mtriple=amdgcn-unknown-amdhsa -amdgpu-promote-alloca < %s | FileCheck %s
 
 declare void @llvm.memcpy.p0i8.p1i8.i32(i8* nocapture, i8 addrspace(1)* nocapture, i32, i32, i1) #0
 declare void @llvm.memcpy.p1i8.p0i8.i32(i8 addrspace(1)* nocapture, i8* nocapture, i32, i32, i1) #0
 
 declare void @llvm.memmove.p0i8.p1i8.i32(i8* nocapture, i8 addrspace(1)* nocapture, i32, i32, i1) #0
 declare void @llvm.memmove.p1i8.p0i8.i32(i8 addrspace(1)* nocapture, i8* nocapture, i32, i32, i1) #0
 
 declare void @llvm.memset.p0i8.i32(i8* nocapture, i8, i32, i32, i1) #0
 
-declare i32 @llvm.objectsize.i32.p0i8(i8*, i1) #1
+declare i32 @llvm.objectsize.i32.p0i8(i8*, i1, i1) #1
 
 ; CHECK-LABEL: @promote_with_memcpy(
 ; CHECK: getelementptr inbounds [64 x [17 x i32]], [64 x [17 x i32]] addrspace(3)* @promote_with_memcpy.alloca, i32 0, i32 %{{[0-9]+}}
 ; CHECK: call void @llvm.memcpy.p3i8.p1i8.i32(i8 addrspace(3)* %alloca.bc, i8 addrspace(1)* %in.bc, i32 68, i32 4, i1 false)
 ; CHECK: call void @llvm.memcpy.p1i8.p3i8.i32(i8 addrspace(1)* %out.bc, i8 addrspace(3)* %alloca.bc, i32 68, i32 4, i1 false)
 define void @promote_with_memcpy(i32 addrspace(1)* %out, i32 addrspace(1)* %in) #0 {
   %alloca = alloca [17 x i32], align 4
   %alloca.bc = bitcast [17 x i32]* %alloca to i8*
define void @promote_with_memset(i32 addrspace(1)* %out, i32 addrspace(1)* %in) #0 {
   %in.bc = bitcast i32 addrspace(1)* %in to i8 addrspace(1)*
   %out.bc = bitcast i32 addrspace(1)* %out to i8 addrspace(1)*
   call void @llvm.memset.p0i8.i32(i8* %alloca.bc, i8 7, i32 68, i32 4, i1 false)
   ret void
 }
 
 ; CHECK-LABEL: @promote_with_objectsize(
 ; CHECK: [[PTR:%[0-9]+]] = getelementptr inbounds [64 x [17 x i32]], [64 x [17 x i32]] addrspace(3)* @promote_with_objectsize.alloca, i32 0, i32 %{{[0-9]+}}
-; CHECK: call i32 @llvm.objectsize.i32.p3i8(i8 addrspace(3)* %alloca.bc, i1 false)
+; CHECK: call i32 @llvm.objectsize.i32.p3i8(i8 addrspace(3)* %alloca.bc, i1 false, i1 false)
 define void @promote_with_objectsize(i32 addrspace(1)* %out) #0 {
   %alloca = alloca [17 x i32], align 4
   %alloca.bc = bitcast [17 x i32]* %alloca to i8*
-  %size = call i32 @llvm.objectsize.i32.p0i8(i8* %alloca.bc, i1 false)
+  %size = call i32 @llvm.objectsize.i32.p0i8(i8* %alloca.bc, i1 false, i1 false)
   store i32 %size, i32 addrspace(1)* %out
   ret void
 }
 
 attributes #0 = { nounwind "amdgpu-flat-work-group-size"="64,64" "amdgpu-waves-per-eu"="1,3" }
 attributes #1 = { nounwind readnone }

llvm/trunk/test/Transforms/CodeGenPrepare/basic.ll

 ; RUN: opt -codegenprepare -S < %s | FileCheck %s
 
 target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64"
 target triple = "x86_64-apple-darwin10.0.0"
 
 ; CHECK-LABEL: @test1(
 ; objectsize should fold to a constant, which causes the branch to fold to an
 ; uncond branch. Next, we fold the control flow alltogether.
 ; rdar://8785296
 define i32 @test1(i8* %ptr) nounwind ssp noredzone align 2 {
 entry:
-  %0 = tail call i64 @llvm.objectsize.i64(i8* %ptr, i1 false)
+  %0 = tail call i64 @llvm.objectsize.i64(i8* %ptr, i1 false, i1 false)
   %1 = icmp ugt i64 %0, 3
   br i1 %1, label %T, label %trap
 
 ; CHECK: T:
 ; CHECK-NOT: br label %
 
 trap:                                             ; preds = %0, %entry
   tail call void @llvm.trap() noreturn nounwind
   unreachable
 
 T:
 ; CHECK: ret i32 4
   ret i32 4
 }
 
-declare i64 @llvm.objectsize.i64(i8*, i1) nounwind readonly
+; CHECK-LABEL: @test_objectsize_null_flag(
+define i64 @test_objectsize_null_flag(i8* %ptr) {
+entry:
+  ; CHECK: ret i64 -1
+  %0 = tail call i64 @llvm.objectsize.i64(i8* null, i1 false, i1 true)
+  ret i64 %0
+}
+
+; CHECK-LABEL: @test_objectsize_null_flag_min(
+define i64 @test_objectsize_null_flag_min(i8* %ptr) {
+entry:
+  ; CHECK: ret i64 0
+  %0 = tail call i64 @llvm.objectsize.i64(i8* null, i1 true, i1 true)
+  ret i64 %0
+}
+
+; Test foldable null pointers because we evaluate them with non-exact modes in
+; CodeGenPrepare.
+; CHECK-LABEL: @test_objectsize_null_flag_noas0(
+define i64 @test_objectsize_null_flag_noas0() {
+entry:
+  ; CHECK: ret i64 0
+  %0 = tail call i64 @llvm.objectsize.i64.p1i8(i8 addrspace(1)* null, i1 false,
+                                               i1 true)
+  ret i64 %0
+}
+
+; CHECK-LABEL: @test_objectsize_null_flag_min_noas0(
+define i64 @test_objectsize_null_flag_min_noas0() {
+entry:
+  ; CHECK: ret i64 0
+  %0 = tail call i64 @llvm.objectsize.i64.p1i8(i8 addrspace(1)* null, i1 true,
+                                               i1 true)
+  ret i64 %0
+}
+
+
+declare i64 @llvm.objectsize.i64(i8*, i1, i1) nounwind readonly
+declare i64 @llvm.objectsize.i64.p1i8(i8 addrspace(1)*, i1, i1) nounwind readonly
 
 declare void @llvm.trap() nounwind

llvm/trunk/test/Transforms/InferAddressSpaces/AMDGPU/intrinsics.ll

 ; RUN: opt -S -mtriple=amdgcn-amd-amdhsa -infer-address-spaces %s | FileCheck %s
 
 ; CHECK-LABEL: @objectsize_group_to_flat_i32(
-; CHECK: %val = call i32 @llvm.objectsize.i32.p3i8(i8 addrspace(3)* %group.ptr, i1 true)
+; CHECK: %val = call i32 @llvm.objectsize.i32.p3i8(i8 addrspace(3)* %group.ptr, i1 true, i1 false)
 define i32 @objectsize_group_to_flat_i32(i8 addrspace(3)* %group.ptr) #0 {
   %cast = addrspacecast i8 addrspace(3)* %group.ptr to i8 addrspace(4)*
-  %val = call i32 @llvm.objectsize.i32.p4i8(i8 addrspace(4)* %cast, i1 true)
+  %val = call i32 @llvm.objectsize.i32.p4i8(i8 addrspace(4)* %cast, i1 true, i1 false)
   ret i32 %val
 }
 
 ; CHECK-LABEL: @objectsize_global_to_flat_i64(
-; CHECK: %val = call i64 @llvm.objectsize.i64.p3i8(i8 addrspace(3)* %global.ptr, i1 true)
+; CHECK: %val = call i64 @llvm.objectsize.i64.p3i8(i8 addrspace(3)* %global.ptr, i1 true, i1 false)
 define i64 @objectsize_global_to_flat_i64(i8 addrspace(3)* %global.ptr) #0 {
   %cast = addrspacecast i8 addrspace(3)* %global.ptr to i8 addrspace(4)*
-  %val = call i64 @llvm.objectsize.i64.p4i8(i8 addrspace(4)* %cast, i1 true)
+  %val = call i64 @llvm.objectsize.i64.p4i8(i8 addrspace(4)* %cast, i1 true, i1 false)
   ret i64 %val
 }
 
 ; CHECK-LABEL: @atomicinc_global_to_flat_i32(
 ; CHECK: call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %global.ptr, i32 %y)
 define i32 @atomicinc_global_to_flat_i32(i32 addrspace(1)* %global.ptr, i32 %y) #0 {
   %cast = addrspacecast i32 addrspace(1)* %global.ptr to i32 addrspace(4)*
   %ret = call i32 @llvm.amdgcn.atomic.inc.i32.p4i32(i32 addrspace(4)* %cast, i32 %y)

 ; CHECK-LABEL: @atomicdec_group_to_flat_i64(
 ; CHECK: call i64 @llvm.amdgcn.atomic.dec.i64.p3i64(i64 addrspace(3)* %group.ptr, i64 %y)
 define i64 @atomicdec_group_to_flat_i64(i64 addrspace(3)* %group.ptr, i64 %y) #0 {
   %cast = addrspacecast i64 addrspace(3)* %group.ptr to i64 addrspace(4)*
   %ret = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %cast, i64 %y)
   ret i64 %ret
 }
 
-declare i32 @llvm.objectsize.i32.p4i8(i8 addrspace(4)*, i1) #1
-declare i64 @llvm.objectsize.i64.p4i8(i8 addrspace(4)*, i1) #1
+declare i32 @llvm.objectsize.i32.p4i8(i8 addrspace(4)*, i1, i1) #1
+declare i64 @llvm.objectsize.i64.p4i8(i8 addrspace(4)*, i1, i1) #1
 declare i32 @llvm.amdgcn.atomic.inc.i32.p4i32(i32 addrspace(4)* nocapture, i32) #2
 declare i64 @llvm.amdgcn.atomic.inc.i64.p4i64(i64 addrspace(4)* nocapture, i64) #2
 declare i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* nocapture, i32) #2
 declare i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* nocapture, i64) #2
 
 attributes #0 = { nounwind }
 attributes #1 = { nounwind readnone }
 attributes #2 = { nounwind argmemonly }

llvm/trunk/test/Transforms/InstCombine/memset_chk-1.ll

 }
 
 ; Test that RAUW in SimplifyLibCalls for __memset_chk generates valid IR
 define i32 @test_rauw(i8* %a, i8* %b, i8** %c) {
 ; CHECK-LABEL: test_rauw
 entry:
   %call49 = call i64 @strlen(i8* %a)
   %add180 = add i64 %call49, 1
-  %yo107 = call i64 @llvm.objectsize.i64.p0i8(i8* %b, i1 false)
+  %yo107 = call i64 @llvm.objectsize.i64.p0i8(i8* %b, i1 false, i1 false)
   %call50 = call i8* @__memmove_chk(i8* %b, i8* %a, i64 %add180, i64 %yo107)
 ; CHECK: %strlen = call i64 @strlen(i8* %b)
 ; CHECK-NEXT: %strchr2 = getelementptr i8, i8* %b, i64 %strlen
   %call51i = call i8* @strrchr(i8* %b, i32 0)
   %d = load i8*, i8** %c, align 8
   %sub182 = ptrtoint i8* %d to i64
   %sub183 = ptrtoint i8* %b to i64
   %sub184 = sub i64 %sub182, %sub183
   %add52.i.i = add nsw i64 %sub184, 1
 ; CHECK: call void @llvm.memset.p0i8.i64(i8* %strchr2
   %call185 = call i8* @__memset_chk(i8* %call51i, i32 0, i64 %add52.i.i, i64 -1)
   ret i32 4
 }
 
 declare i8* @__memmove_chk(i8*, i8*, i64, i64)
 declare i8* @strrchr(i8*, i32)
 declare i64 @strlen(i8* nocapture)
-declare i64 @llvm.objectsize.i64.p0i8(i8*, i1)
+declare i64 @llvm.objectsize.i64.p0i8(i8*, i1, i1)
 
 declare i8* @__memset_chk(i8*, i32, i64, i64)
 
 ; FIXME: memset(malloc(x), 0, x) -> calloc(1, x)
 
 define float* @pr25892(i64 %size) #0 {
 entry:
   %call = tail call i8* @malloc(i64 %size) #1
   %cmp = icmp eq i8* %call, null
   br i1 %cmp, label %cleanup, label %if.end
 if.end:
   %bc = bitcast i8* %call to float*
-  %call2 = tail call i64 @llvm.objectsize.i64.p0i8(i8* nonnull %call, i1 false)
+  %call2 = tail call i64 @llvm.objectsize.i64.p0i8(i8* nonnull %call, i1 false, i1 false)
   %call3 = tail call i8* @__memset_chk(i8* nonnull %call, i32 0, i64 %size, i64 %call2) #1
   br label %cleanup
 cleanup:
   %retval.0 = phi float* [ %bc, %if.end ], [ null, %entry ]
   ret float* %retval.0
 
 ; CHECK-LABEL: @pr25892(
 ; CHECK:       entry:
 ; CHECK-NEXT:    %call = tail call i8* @malloc(i64 %size)
 ; CHECK-NEXT:    %cmp = icmp eq i8* %call, null
 ; CHECK-NEXT:    br i1 %cmp, label %cleanup, label %if.end
 ; CHECK:       if.end:
 ; CHECK-NEXT:    %bc = bitcast i8* %call to float*
-; CHECK-NEXT:    %call2 = tail call i64 @llvm.objectsize.i64.p0i8(i8* nonnull %call, i1 false)
+; CHECK-NEXT:    %call2 = tail call i64 @llvm.objectsize.i64.p0i8(i8* nonnull %call, i1 false, i1 false)
 ; CHECK-NEXT:    %call3 = tail call i8* @__memset_chk(i8* nonnull %call, i32 0, i64 %size, i64 %call2)
 ; CHECK-NEXT:    br label %cleanup
 ; CHECK:       cleanup:
 ; CHECK-NEXT:    %retval.0 = phi float* [ %bc, %if.end ], [ null, %entry ]
 ; CHECK-NEXT:    ret float* %retval.0
 }
 
 declare noalias i8* @malloc(i64) #1
 
 attributes #0 = { nounwind ssp uwtable }
 attributes #1 = { nounwind }
 attributes #2 = { nounwind readnone }
 

llvm/trunk/test/Transforms/InstCombine/objsize.ll

 ; Test a pile of objectsize bounds checking.
 ; RUN: opt < %s -instcombine -S | FileCheck %s
 ; We need target data to get the sizes of the arrays and structures.
 target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128"
 
 @a = private global [60 x i8] zeroinitializer, align 1 ; <[60 x i8]*>
 @.str = private constant [8 x i8] c"abcdefg\00"   ; <[8 x i8]*>
 define i32 @foo() nounwind {
 ; CHECK-LABEL: @foo(
 ; CHECK-NEXT: ret i32 60
-  %1 = call i32 @llvm.objectsize.i32.p0i8(i8* getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i1 false)
+  %1 = call i32 @llvm.objectsize.i32.p0i8(i8* getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i1 false, i1 false)
   ret i32 %1
 }
 
 define i8* @bar() nounwind {
 ; CHECK-LABEL: @bar(
 entry:
   %retval = alloca i8*
-  %0 = call i32 @llvm.objectsize.i32.p0i8(i8* getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i1 false)
+  %0 = call i32 @llvm.objectsize.i32.p0i8(i8* getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i1 false, i1 false)
   %cmp = icmp ne i32 %0, -1
 ; CHECK: br i1 true
   br i1 %cmp, label %cond.true, label %cond.false
 
 cond.true:
   %1 = load i8*, i8** %retval
   ret i8* %1
 
 cond.false:
   %2 = load i8*, i8** %retval
   ret i8* %2
 }
 
 define i32 @f() nounwind {
 ; CHECK-LABEL: @f(
 ; CHECK-NEXT: ret i32 0
-  %1 = call i32 @llvm.objectsize.i32.p0i8(i8* getelementptr ([60 x i8], [60 x i8]* @a, i32 1, i32 0), i1 false)
+  %1 = call i32 @llvm.objectsize.i32.p0i8(i8* getelementptr ([60 x i8], [60 x i8]* @a, i32 1, i32 0), i1 false, i1 false)
   ret i32 %1
 }
 
 @window = external global [0 x i8]
 
 define i1 @baz() nounwind {
 ; CHECK-LABEL: @baz(
 ; CHECK-NEXT: objectsize
-  %1 = tail call i32 @llvm.objectsize.i32.p0i8(i8* getelementptr inbounds ([0 x i8], [0 x i8]* @window, i32 0, i32 0), i1 false)
+  %1 = tail call i32 @llvm.objectsize.i32.p0i8(i8* getelementptr inbounds ([0 x i8], [0 x i8]* @window, i32 0, i32 0), i1 false, i1 false)
   %2 = icmp eq i32 %1, -1
   ret i1 %2
 }
 
 define void @test1(i8* %q, i32 %x) nounwind noinline {
 ; CHECK-LABEL: @test1(
 ; CHECK: objectsize.i32.p0i8
 entry:
-  %0 = call i32 @llvm.objectsize.i32.p0i8(i8* getelementptr inbounds ([0 x i8], [0 x i8]* @window, i32 0, i32 10), i1 false) ; <i64> [#uses=1]
+  %0 = call i32 @llvm.objectsize.i32.p0i8(i8* getelementptr inbounds ([0 x i8], [0 x i8]* @window, i32 0, i32 10), i1 false, i1 false) ; <i64> [#uses=1]
   %1 = icmp eq i32 %0, -1                         ; <i1> [#uses=1]
   br i1 %1, label %"47", label %"46"
 
 "46":                                             ; preds = %entry
   unreachable
 
 "47":                                             ; preds = %entry
   unreachable
 }
 
 @.str5 = private constant [9 x i32] [i32 97, i32 98, i32 99, i32 100, i32 0, i32
  101, i32 102, i32 103, i32 0], align 4
 define i32 @test2() nounwind {
 ; CHECK-LABEL: @test2(
 ; CHECK-NEXT: ret i32 34
-  %1 = call i32 @llvm.objectsize.i32.p0i8(i8* getelementptr (i8, i8* bitcast ([9 x i32]* @.str5 to i8*), i32 2), i1 false)
+  %1 = call i32 @llvm.objectsize.i32.p0i8(i8* getelementptr (i8, i8* bitcast ([9 x i32]* @.str5 to i8*), i32 2), i1 false, i1 false)
   ret i32 %1
 }
 
 ; rdar://7674946
 @array = internal global [480 x float] zeroinitializer ; <[480 x float]*> [#uses=1]
 
 declare i8* @__memcpy_chk(i8*, i8*, i32, i32) nounwind
 
-declare i32 @llvm.objectsize.i32.p0i8(i8*, i1) nounwind readonly
+declare i32 @llvm.objectsize.i32.p0i8(i8*, i1, i1) nounwind readonly
+
+declare i32 @llvm.objectsize.i32.p1i8(i8 addrspace(1)*, i1, i1) nounwind readonly
 
 declare i8* @__inline_memcpy_chk(i8*, i8*, i32) nounwind inlinehint
 
 define void @test3() nounwind {
 ; CHECK-LABEL: @test3(
 entry:
   br i1 undef, label %bb11, label %bb12
 
 bb11:
   %0 = getelementptr inbounds float, float* getelementptr inbounds ([480 x float], [480 x float]* @array, i32 0, i32 128), i32 -127 ; <float*> [#uses=1]
   %1 = bitcast float* %0 to i8*                   ; <i8*> [#uses=1]
-  %2 = call i32 @llvm.objectsize.i32.p0i8(i8* %1, i1 false) ; <i32> [#uses=1]
+  %2 = call i32 @llvm.objectsize.i32.p0i8(i8* %1, i1 false, i1 false) ; <i32> [#uses=1]
   %3 = call i8* @__memcpy_chk(i8* undef, i8* undef, i32 512, i32 %2) nounwind ; <i8*> [#uses=0]
 ; CHECK: unreachable
   unreachable
 
 bb12:
   %4 = getelementptr inbounds float, float* getelementptr inbounds ([480 x float], [480 x float]* @array, i32 0, i32 128), i32 -127 ; <float*> [#uses=1]
   %5 = bitcast float* %4 to i8*                   ; <i8*> [#uses=1]
   %6 = call i8* @__inline_memcpy_chk(i8* %5, i8* undef, i32 512) nounwind inlinehint ; <i8*> [#uses=0]
 ; CHECK: @__inline_memcpy_chk
   unreachable
 }
 
 ; rdar://7718857
 
 %struct.data = type { [100 x i32], [100 x i32], [1024 x i8] }
 
 define i32 @test4(i8** %esc) nounwind ssp {
 ; CHECK-LABEL: @test4(
 entry:
   %0 = alloca %struct.data, align 8
   %1 = bitcast %struct.data* %0 to i8*
-  %2 = call i32 @llvm.objectsize.i32.p0i8(i8* %1, i1 false) nounwind
+  %2 = call i32 @llvm.objectsize.i32.p0i8(i8* %1, i1 false, i1 false) nounwind
 ; CHECK-NOT: @llvm.objectsize
 ; CHECK: @llvm.memset.p0i8.i32(i8* %1, i8 0, i32 1824, i32 8, i1 false)
   %3 = call i8* @__memset_chk(i8* %1, i32 0, i32 1824, i32 %2) nounwind
   store i8* %1, i8** %esc
   ret i32 0
 }
 
 ; rdar://7782496
 @s = external global i8*
 
 define i8* @test5(i32 %n) nounwind ssp {
 ; CHECK-LABEL: @test5(
 entry:
   %0 = tail call noalias i8* @malloc(i32 20) nounwind
-  %1 = tail call i32 @llvm.objectsize.i32.p0i8(i8* %0, i1 false)
+  %1 = tail call i32 @llvm.objectsize.i32.p0i8(i8* %0, i1 false, i1 false)
   %2 = load i8*, i8** @s, align 8
 ; CHECK-NOT: @llvm.objectsize
 ; CHECK: @llvm.memcpy.p0i8.p0i8.i32(i8* %0, i8* %1, i32 10, i32 1, i1 false)
   %3 = tail call i8* @__memcpy_chk(i8* %0, i8* %2, i32 10, i32 %1) nounwind
   ret i8* %0
 }
 
 define void @test6(i32 %n) nounwind ssp {
 ; CHECK-LABEL: @test6(
 entry:
   %0 = tail call noalias i8* @malloc(i32 20) nounwind
-  %1 = tail call i32 @llvm.objectsize.i32.p0i8(i8* %0, i1 false)
+  %1 = tail call i32 @llvm.objectsize.i32.p0i8(i8* %0, i1 false, i1 false)
   %2 = load i8*, i8** @s, align 8
 ; CHECK-NOT: @llvm.objectsize
 ; CHECK: @__memcpy_chk(i8* %0, i8* %1, i32 30, i32 20)
   %3 = tail call i8* @__memcpy_chk(i8* %0, i8* %2, i32 30, i32 %1) nounwind
   ret void
 }
 
 declare i8* @__memset_chk(i8*, i32, i32, i32) nounwind
 
 declare noalias i8* @malloc(i32) nounwind
 
 define i32 @test7(i8** %esc) {
 ; CHECK-LABEL: @test7(
   %alloc = call noalias i8* @malloc(i32 48) nounwind
   store i8* %alloc, i8** %esc
   %gep = getelementptr inbounds i8, i8* %alloc, i32 16
-  %objsize = call i32 @llvm.objectsize.i32.p0i8(i8* %gep, i1 false) nounwind readonly
+  %objsize = call i32 @llvm.objectsize.i32.p0i8(i8* %gep, i1 false, i1 false) nounwind readonly
 ; CHECK: ret i32 32
   ret i32 %objsize
 }
 
 declare noalias i8* @calloc(i32, i32) nounwind
 
 define i32 @test8(i8** %esc) {
 ; CHECK-LABEL: @test8(
   %alloc = call noalias i8* @calloc(i32 5, i32 7) nounwind
   store i8* %alloc, i8** %esc
   %gep = getelementptr inbounds i8, i8* %alloc, i32 5
-  %objsize = call i32 @llvm.objectsize.i32.p0i8(i8* %gep, i1 false) nounwind readonly
+  %objsize = call i32 @llvm.objectsize.i32.p0i8(i8* %gep, i1 false, i1 false) nounwind readonly
 ; CHECK: ret i32 30
   ret i32 %objsize
 }
 
 declare noalias i8* @strdup(i8* nocapture) nounwind
 declare noalias i8* @strndup(i8* nocapture, i32) nounwind
 
 ; CHECK-LABEL: @test9(
 define i32 @test9(i8** %esc) {
   %call = tail call i8* @strdup(i8* getelementptr inbounds ([8 x i8], [8 x i8]* @.str, i64 0, i64 0)) nounwind
   store i8* %call, i8** %esc, align 8
-  %1 = tail call i32 @llvm.objectsize.i32.p0i8(i8* %call, i1 true)
+  %1 = tail call i32 @llvm.objectsize.i32.p0i8(i8* %call, i1 true, i1 false)
 ; CHECK: ret i32 8
   ret i32 %1
 }
 
 ; CHECK-LABEL: @test10(
 define i32 @test10(i8** %esc) {
   %call = tail call i8* @strndup(i8* getelementptr inbounds ([8 x i8], [8 x i8]* @.str, i64 0, i64 0), i32 3) nounwind
   store i8* %call, i8** %esc, align 8
-  %1 = tail call i32 @llvm.objectsize.i32.p0i8(i8* %call, i1 true)
+  %1 = tail call i32 @llvm.objectsize.i32.p0i8(i8* %call, i1 true, i1 false)
 ; CHECK: ret i32 4
   ret i32 %1
 }
 
 ; CHECK-LABEL: @test11(
 define i32 @test11(i8** %esc) {
   %call = tail call i8* @strndup(i8* getelementptr inbounds ([8 x i8], [8 x i8]* @.str, i64 0, i64 0), i32 7) nounwind
   store i8* %call, i8** %esc, align 8
-  %1 = tail call i32 @llvm.objectsize.i32.p0i8(i8* %call, i1 true)
+  %1 = tail call i32 @llvm.objectsize.i32.p0i8(i8* %call, i1 true, i1 false)
 ; CHECK: ret i32 8
   ret i32 %1
 }
 
 ; CHECK-LABEL: @test12(
 define i32 @test12(i8** %esc) {
   %call = tail call i8* @strndup(i8* getelementptr inbounds ([8 x i8], [8 x i8]* @.str, i64 0, i64 0), i32 8) nounwind
   store i8* %call, i8** %esc, align 8
-  %1 = tail call i32 @llvm.objectsize.i32.p0i8(i8* %call, i1 true)
+  %1 = tail call i32 @llvm.objectsize.i32.p0i8(i8* %call, i1 true, i1 false)
 ; CHECK: ret i32 8
   ret i32 %1
 }
 
 ; CHECK-LABEL: @test13(
 define i32 @test13(i8** %esc) {
   %call = tail call i8* @strndup(i8* getelementptr inbounds ([8 x i8], [8 x i8]* @.str, i64 0, i64 0), i32 57) nounwind
   store i8* %call, i8** %esc, align 8
-  %1 = tail call i32 @llvm.objectsize.i32.p0i8(i8* %call, i1 true)
+  %1 = tail call i32 @llvm.objectsize.i32.p0i8(i8* %call, i1 true, i1 false)
 ; CHECK: ret i32 8
   ret i32 %1
 }
 
 @globalalias = internal alias [60 x i8], [60 x i8]* @a
 
 ; CHECK-LABEL: @test18(
 ; CHECK-NEXT: ret i32 60
 define i32 @test18() {
   %bc = bitcast [60 x i8]* @globalalias to i8*
-  %1 = call i32 @llvm.objectsize.i32.p0i8(i8* %bc, i1 false)
+  %1 = call i32 @llvm.objectsize.i32.p0i8(i8* %bc, i1 false, i1 false)
   ret i32 %1
 }
 
 @globalalias2 = weak alias [60 x i8], [60 x i8]* @a
 
 ; CHECK-LABEL: @test19(
 ; CHECK: llvm.objectsize
 define i32 @test19() {
   %bc = bitcast [60 x i8]* @globalalias2 to i8*
-  %1 = call i32 @llvm.objectsize.i32.p0i8(i8* %bc, i1 false)
+  %1 = call i32 @llvm.objectsize.i32.p0i8(i8* %bc, i1 false, i1 false)
+  ret i32 %1
+}
+
+; CHECK-LABEL: @test20(
+; CHECK: ret i32 0
+define i32 @test20() {
+  %1 = call i32 @llvm.objectsize.i32.p0i8(i8* null, i1 false, i1 false)
+  ret i32 %1
+}
+
+; CHECK-LABEL: @test21(
+; CHECK: ret i32 0
+define i32 @test21() {
+  %1 = call i32 @llvm.objectsize.i32.p0i8(i8* null, i1 true, i1 false)
+  ret i32 %1
+}
+
+; CHECK-LABEL: @test22(
+; CHECK: llvm.objectsize
+define i32 @test22() {
+  %1 = call i32 @llvm.objectsize.i32.p0i8(i8* null, i1 false, i1 true)
+  ret i32 %1
+}
+
+; CHECK-LABEL: @test23(
+; CHECK: llvm.objectsize
+define i32 @test23() {
+  %1 = call i32 @llvm.objectsize.i32.p0i8(i8* null, i1 true, i1 true)
   ret i32 %1
 }
 
+; 1 is an arbitrary non-zero address space.
+; CHECK-LABEL: @test24(
+; CHECK: ret i32 0
+define i32 @test24() {
+  %1 = call i32 @llvm.objectsize.i32.p1i8(i8 addrspace(1)* null, i1 false,
+                                          i1 false)
+  ret i32 %1
+}
+
+; CHECK-LABEL: @test25(
+; CHECK: ret i32 0
+define i32 @test25() {
+  %1 = call i32 @llvm.objectsize.i32.p1i8(i8 addrspace(1)* null, i1 true,
+                                          i1 false)
+  ret i32 %1
+}
+
+; CHECK-LABEL: @test26(
+; CHECK: ret i32 0
+define i32 @test26() {
+  %1 = call i32 @llvm.objectsize.i32.p1i8(i8 addrspace(1)* null, i1 false,
+                                          i1 true)
+  ret i32 %1
+}
+
+; CHECK-LABEL: @test27(
+; CHECK: ret i32 0
+define i32 @test27() {
+  %1 = call i32 @llvm.objectsize.i32.p1i8(i8 addrspace(1)* null, i1 true,
+                                          i1 true)
+  ret i32 %1
+}

llvm/trunk/test/Transforms/InstCombine/stpcpy_chk-1.ll

 
 ; Check case where the string length is not constant.
 
 define i8* @test_simplify5() {
 ; CHECK-LABEL: @test_simplify5(
   %dst = getelementptr inbounds [60 x i8], [60 x i8]* @a, i32 0, i32 0
   %src = getelementptr inbounds [12 x i8], [12 x i8]* @.str, i32 0, i32 0
 
-; CHECK-NEXT: %len = call i32 @llvm.objectsize.i32.p0i8(i8* getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i1 false)
+; CHECK-NEXT: %len = call i32 @llvm.objectsize.i32.p0i8(i8* getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i1 false, i1 false)
 ; CHECK-NEXT: %1 = call i8* @__memcpy_chk(i8* getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i8* getelementptr inbounds ([12 x i8], [12 x i8]* @.str, i32 0, i32 0), i32 12, i32 %len)
 ; CHECK-NEXT: ret i8* getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 11)
-  %len = call i32 @llvm.objectsize.i32.p0i8(i8* %dst, i1 false)
+  %len = call i32 @llvm.objectsize.i32.p0i8(i8* %dst, i1 false, i1 false)
   %ret = call i8* @__stpcpy_chk(i8* %dst, i8* %src, i32 %len)
   ret i8* %ret
 }
 
 ; Check case where the source and destination are the same.
 
 define i8* @test_simplify6() {
 ; CHECK-LABEL: @test_simplify6(
   %dst = getelementptr inbounds [60 x i8], [60 x i8]* @a, i32 0, i32 0
 
 ; CHECK-NEXT: %strlen = call i32 @strlen(i8* getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0))
 ; CHECK-NEXT: %1 = getelementptr inbounds [60 x i8], [60 x i8]* @a, i32 0, i32 %strlen
 ; CHECK-NEXT: ret i8* %1
-  %len = call i32 @llvm.objectsize.i32.p0i8(i8* %dst, i1 false)
+  %len = call i32 @llvm.objectsize.i32.p0i8(i8* %dst, i1 false, i1 false)
   %ret = call i8* @__stpcpy_chk(i8* %dst, i8* %dst, i32 %len)
   ret i8* %ret
 }
 
 ; Check case where slen < strlen (src).
 
 define i8* @test_no_simplify1() {
 ; CHECK-LABEL: @test_no_simplify1(
   %dst = getelementptr inbounds [60 x i8], [60 x i8]* @a, i32 0, i32 0
   %src = getelementptr inbounds [60 x i8], [60 x i8]* @b, i32 0, i32 0
 
 ; CHECK-NEXT: %ret = call i8* @__stpcpy_chk(i8* getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i8* getelementptr inbounds ([60 x i8], [60 x i8]* @b, i32 0, i32 0), i32 8)
 ; CHECK-NEXT: ret i8* %ret
   %ret = call i8* @__stpcpy_chk(i8* %dst, i8* %src, i32 8)
   ret i8* %ret
 }
 
 declare i8* @__stpcpy_chk(i8*, i8*, i32) nounwind
-declare i32 @llvm.objectsize.i32.p0i8(i8*, i1) nounwind readonly
+declare i32 @llvm.objectsize.i32.p0i8(i8*, i1, i1) nounwind readonly

llvm/trunk/test/Transforms/InstCombine/strcpy_chk-1.ll

 
 ; Check case where the string length is not constant.
 
 define i8* @test_simplify5() {
 ; CHECK-LABEL: @test_simplify5(
   %dst = getelementptr inbounds [60 x i8], [60 x i8]* @a, i32 0, i32 0
   %src = getelementptr inbounds [12 x i8], [12 x i8]* @.str, i32 0, i32 0
 
-; CHECK-NEXT: %len = call i32 @llvm.objectsize.i32.p0i8(i8* getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i1 false)
+; CHECK-NEXT: %len = call i32 @llvm.objectsize.i32.p0i8(i8* getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i1 false, i1 false)
 ; CHECK-NEXT: %1 = call i8* @__memcpy_chk(i8* getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i8* getelementptr inbounds ([12 x i8], [12 x i8]* @.str, i32 0, i32 0), i32 12, i32 %len)
 ; CHECK-NEXT: ret i8* %1
-  %len = call i32 @llvm.objectsize.i32.p0i8(i8* %dst, i1 false)
+  %len = call i32 @llvm.objectsize.i32.p0i8(i8* %dst, i1 false, i1 false)
   %ret = call i8* @__strcpy_chk(i8* %dst, i8* %src, i32 %len)
   ret i8* %ret
 }
 
 ; Check case where the source and destination are the same.
 
 define i8* @test_simplify6() {
 ; CHECK-LABEL: @test_simplify6(
   %dst = getelementptr inbounds [60 x i8], [60 x i8]* @a, i32 0, i32 0
 
-; CHECK-NEXT: %len = call i32 @llvm.objectsize.i32.p0i8(i8* getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i1 false)
+; CHECK-NEXT: %len = call i32 @llvm.objectsize.i32.p0i8(i8* getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i1 false, i1 false)
 ; CHECK-NEXT: %ret = call i8* @__strcpy_chk(i8* getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i8* getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i32 %len)
 ; CHECK-NEXT: ret i8* %ret
-  %len = call i32 @llvm.objectsize.i32.p0i8(i8* %dst, i1 false)
+  %len = call i32 @llvm.objectsize.i32.p0i8(i8* %dst, i1 false, i1 false)
   %ret = call i8* @__strcpy_chk(i8* %dst, i8* %dst, i32 %len)
   ret i8* %ret
 }
 
 ; Check case where slen < strlen (src).
 
 define i8* @test_no_simplify1() {
 ; CHECK-LABEL: @test_no_simplify1(
   %dst = getelementptr inbounds [60 x i8], [60 x i8]* @a, i32 0, i32 0
   %src = getelementptr inbounds [60 x i8], [60 x i8]* @b, i32 0, i32 0
 
 ; CHECK-NEXT: %ret = call i8* @__strcpy_chk(i8* getelementptr inbounds ([60 x i8], [60 x i8]* @a, i32 0, i32 0), i8* getelementptr inbounds ([60 x i8], [60 x i8]* @b, i32 0, i32 0), i32 8)
 ; CHECK-NEXT: ret i8* %ret
   %ret = call i8* @__strcpy_chk(i8* %dst, i8* %src, i32 8)
   ret i8* %ret
 }
 
 declare i8* @__strcpy_chk(i8*, i8*, i32) nounwind
-declare i32 @llvm.objectsize.i32.p0i8(i8*, i1) nounwind readonly
+declare i32 @llvm.objectsize.i32.p0i8(i8*, i1, i1) nounwind readonly