[llvm:ir] Add support for constant data exceeding 4GiB (#126481)

The test file is over 4GiB, which is too big, so I didn’t submit it.

Author: pzzp

clang/lib/CodeGen/CGExprConstant.cpp

@@ -364,14 +364,14 @@ bool ConstantAggregateBuilder::split(size_t Index, CharUnits Hint) {
     // FIXME: If possible, split into two ConstantDataSequentials at Hint.
     CharUnits ElemSize = getSize(CDS->getElementType());
     replace(Elems, Index, Index + 1,
-            llvm::map_range(llvm::seq(0u, CDS->getNumElements()),
-                            [&](unsigned Elem) {
+            llvm::map_range(llvm::seq(uint64_t(0u), CDS->getNumElements()),
+                            [&](uint64_t Elem) {
                               return CDS->getElementAsConstant(Elem);
                             }));
     replace(Offsets, Index, Index + 1,
             llvm::map_range(
-                llvm::seq(0u, CDS->getNumElements()),
-                [&](unsigned Elem) { return Offset + Elem * ElemSize; }));
+                llvm::seq(uint64_t(0u), CDS->getNumElements()),
+                [&](uint64_t Elem) { return Offset + Elem * ElemSize; }));
     return true;
   }
 

llvm/include/llvm/IR/Constants.h

@@ -617,34 +617,34 @@ class ConstantDataSequential : public ConstantData {
 
   /// If this is a sequential container of integers (of any size), return the
   /// specified element in the low bits of a uint64_t.
-  uint64_t getElementAsInteger(unsigned i) const;
+  uint64_t getElementAsInteger(uint64_t i) const;
 
   /// If this is a sequential container of integers (of any size), return the
   /// specified element as an APInt.
-  APInt getElementAsAPInt(unsigned i) const;
+  APInt getElementAsAPInt(uint64_t i) const;
 
   /// If this is a sequential container of floating point type, return the
   /// specified element as an APFloat.
-  APFloat getElementAsAPFloat(unsigned i) const;
+  APFloat getElementAsAPFloat(uint64_t i) const;
 
   /// If this is an sequential container of floats, return the specified element
   /// as a float.
-  float getElementAsFloat(unsigned i) const;
+  float getElementAsFloat(uint64_t i) const;
 
   /// If this is an sequential container of doubles, return the specified
   /// element as a double.
-  double getElementAsDouble(unsigned i) const;
+  double getElementAsDouble(uint64_t i) const;
 
   /// Return a Constant for a specified index's element.
   /// Note that this has to compute a new constant to return, so it isn't as
   /// efficient as getElementAsInteger/Float/Double.
-  Constant *getElementAsConstant(unsigned i) const;
+  Constant *getElementAsConstant(uint64_t i) const;
 
   /// Return the element type of the array/vector.
   Type *getElementType() const;
 
   /// Return the number of elements in the array or vector.
-  unsigned getNumElements() const;
+  uint64_t getNumElements() const;
 
   /// Return the size (in bytes) of each element in the array/vector.
   /// The size of the elements is known to be a multiple of one byte.
@@ -684,7 +684,7 @@ class ConstantDataSequential : public ConstantData {
   }
 
 private:
-  const char *getElementPointer(unsigned Elt) const;
+  const char *getElementPointer(uint64_t Elt) const;
 };
 
 //===----------------------------------------------------------------------===//

llvm/lib/Analysis/TargetTransformInfo.cpp

@@ -919,7 +919,7 @@ TargetTransformInfo::getOperandInfo(const Value *V) {
   } else if (const auto *CDS = dyn_cast<ConstantDataSequential>(V)) {
     OpInfo = OK_NonUniformConstantValue;
     bool AllPow2 = true, AllNegPow2 = true;
-    for (unsigned I = 0, E = CDS->getNumElements(); I != E; ++I) {
+    for (uint64_t I = 0, E = CDS->getNumElements(); I != E; ++I) {
       if (auto *CI = dyn_cast<ConstantInt>(CDS->getElementAsConstant(I))) {
         AllPow2 &= CI->getValue().isPowerOf2();
         AllNegPow2 &= CI->getValue().isNegatedPowerOf2();

llvm/lib/Analysis/ValueTracking.cpp

@@ -6355,7 +6355,7 @@ Value *llvm::isBytewiseValue(Value *V, const DataLayout &DL) {
 
   if (ConstantDataSequential *CA = dyn_cast<ConstantDataSequential>(C)) {
     Value *Val = UndefInt8;
-    for (unsigned I = 0, E = CA->getNumElements(); I != E; ++I)
+    for (uint64_t I = 0, E = CA->getNumElements(); I != E; ++I)
       if (!(Val = Merge(Val, isBytewiseValue(CA->getElementAsConstant(I), DL))))
         return nullptr;
     return Val;

llvm/lib/Bitcode/Writer/BitcodeWriter.cpp

@@ -2803,7 +2803,7 @@ void ModuleBitcodeWriter::writeConstants(unsigned FirstVal, unsigned LastVal,
                cast<ConstantDataSequential>(C)->isString()) {
       const ConstantDataSequential *Str = cast<ConstantDataSequential>(C);
       // Emit constant strings specially.
-      unsigned NumElts = Str->getNumElements();
+      uint64_t NumElts = Str->getNumElements();
       // If this is a null-terminated string, use the denser CSTRING encoding.
       if (Str->isCString()) {
         Code = bitc::CST_CODE_CSTRING;
@@ -2814,7 +2814,7 @@ void ModuleBitcodeWriter::writeConstants(unsigned FirstVal, unsigned LastVal,
       }
       bool isCStr7 = Code == bitc::CST_CODE_CSTRING;
       bool isCStrChar6 = Code == bitc::CST_CODE_CSTRING;
-      for (unsigned i = 0; i != NumElts; ++i) {
+      for (uint64_t i = 0; i != NumElts; ++i) {
         unsigned char V = Str->getElementAsInteger(i);
         Record.push_back(V);
         isCStr7 &= (V & 128) == 0;
@@ -2831,10 +2831,10 @@ void ModuleBitcodeWriter::writeConstants(unsigned FirstVal, unsigned LastVal,
       Code = bitc::CST_CODE_DATA;
       Type *EltTy = CDS->getElementType();
       if (isa<IntegerType>(EltTy)) {
-        for (unsigned i = 0, e = CDS->getNumElements(); i != e; ++i)
+        for (uint64_t i = 0, e = CDS->getNumElements(); i != e; ++i)
           Record.push_back(CDS->getElementAsInteger(i));
       } else {
-        for (unsigned i = 0, e = CDS->getNumElements(); i != e; ++i)
+        for (uint64_t i = 0, e = CDS->getNumElements(); i != e; ++i)
           Record.push_back(
               CDS->getElementAsAPFloat(i).bitcastToAPInt().getLimitedValue());
       }

llvm/lib/CodeGen/AsmPrinter/AsmPrinter.cpp

@@ -3612,9 +3612,9 @@ static void emitGlobalConstantDataSequential(
     return AP.OutStreamer->emitBytes(CDS->getAsString());
 
   // Otherwise, emit the values in successive locations.
-  unsigned ElementByteSize = CDS->getElementByteSize();
+  uint64_t ElementByteSize = CDS->getElementByteSize();
   if (isa<IntegerType>(CDS->getElementType())) {
-    for (unsigned I = 0, E = CDS->getNumElements(); I != E; ++I) {
+    for (uint64_t I = 0, E = CDS->getNumElements(); I != E; ++I) {
       emitGlobalAliasInline(AP, ElementByteSize * I, AliasList);
       if (AP.isVerbose())
         AP.OutStreamer->getCommentOS()
@@ -3624,7 +3624,7 @@ static void emitGlobalConstantDataSequential(
     }
   } else {
     Type *ET = CDS->getElementType();
-    for (unsigned I = 0, E = CDS->getNumElements(); I != E; ++I) {
+    for (uint64_t I = 0, E = CDS->getNumElements(); I != E; ++I) {
       emitGlobalAliasInline(AP, ElementByteSize * I, AliasList);
       emitGlobalConstantFP(CDS->getElementAsAPFloat(I), ET, AP);
     }

llvm/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp

@@ -1844,7 +1844,7 @@ SDValue SelectionDAGBuilder::getValueImpl(const Value *V) {
     if (const ConstantDataSequential *CDS =
           dyn_cast<ConstantDataSequential>(C)) {
       SmallVector<SDValue, 4> Ops;
-      for (unsigned i = 0, e = CDS->getNumElements(); i != e; ++i) {
+      for (uint64_t i = 0, e = CDS->getNumElements(); i != e; ++i) {
         SDNode *Val = getValue(CDS->getElementAsConstant(i)).getNode();
         // Add each leaf value from the operand to the Constants list
         // to form a flattened list of all the values.

llvm/lib/IR/AsmWriter.cpp

@@ -1679,7 +1679,7 @@ static void WriteConstantInternal(raw_ostream &Out, const Constant *CV,
     WriterCtx.TypePrinter->print(ETy, Out);
     Out << ' ';
     WriteAsOperandInternal(Out, CA->getElementAsConstant(0), WriterCtx);
-    for (unsigned i = 1, e = CA->getNumElements(); i != e; ++i) {
+    for (uint64_t i = 1, e = CA->getNumElements(); i != e; ++i) {
       Out << ", ";
       WriterCtx.TypePrinter->print(ETy, Out);
       Out << ' ';

llvm/lib/IR/Constants.cpp

@@ -2850,24 +2850,22 @@ bool ConstantDataSequential::isElementTypeCompatible(Type *Ty) {
   return false;
 }
 
-unsigned ConstantDataSequential::getNumElements() const {
+uint64_t ConstantDataSequential::getNumElements() const {
   if (ArrayType *AT = dyn_cast<ArrayType>(getType()))
     return AT->getNumElements();
   return cast<FixedVectorType>(getType())->getNumElements();
 }
 
-
 uint64_t ConstantDataSequential::getElementByteSize() const {
-  return getElementType()->getPrimitiveSizeInBits()/8;
+  return getElementType()->getPrimitiveSizeInBits() / 8;
 }
 
 /// Return the start of the specified element.
-const char *ConstantDataSequential::getElementPointer(unsigned Elt) const {
+const char *ConstantDataSequential::getElementPointer(uint64_t Elt) const {
   assert(Elt < getNumElements() && "Invalid Elt");
-  return DataElements+Elt*getElementByteSize();
+  return DataElements + Elt * getElementByteSize();
 }
 
-
 /// Return true if the array is empty or all zeros.
 static bool isAllZeros(StringRef Arr) {
   for (char I : Arr)
@@ -3106,8 +3104,7 @@ Constant *ConstantDataVector::getSplat(unsigned NumElts, Constant *V) {
   return ConstantVector::getSplat(ElementCount::getFixed(NumElts), V);
 }
 
-
-uint64_t ConstantDataSequential::getElementAsInteger(unsigned Elt) const {
+uint64_t ConstantDataSequential::getElementAsInteger(uint64_t Elt) const {
   assert(isa<IntegerType>(getElementType()) &&
          "Accessor can only be used when element is an integer");
   const char *EltPtr = getElementPointer(Elt);
@@ -3127,7 +3124,7 @@ uint64_t ConstantDataSequential::getElementAsInteger(unsigned Elt) const {
   }
 }
 
-APInt ConstantDataSequential::getElementAsAPInt(unsigned Elt) const {
+APInt ConstantDataSequential::getElementAsAPInt(uint64_t Elt) const {
   assert(isa<IntegerType>(getElementType()) &&
          "Accessor can only be used when element is an integer");
   const char *EltPtr = getElementPointer(Elt);
@@ -3155,7 +3152,7 @@ APInt ConstantDataSequential::getElementAsAPInt(unsigned Elt) const {
   }
 }
 
-APFloat ConstantDataSequential::getElementAsAPFloat(unsigned Elt) const {
+APFloat ConstantDataSequential::getElementAsAPFloat(uint64_t Elt) const {
   const char *EltPtr = getElementPointer(Elt);
 
   switch (getElementType()->getTypeID()) {
@@ -3180,19 +3177,19 @@ APFloat ConstantDataSequential::getElementAsAPFloat(unsigned Elt) const {
   }
 }
 
-float ConstantDataSequential::getElementAsFloat(unsigned Elt) const {
+float ConstantDataSequential::getElementAsFloat(uint64_t Elt) const {
   assert(getElementType()->isFloatTy() &&
          "Accessor can only be used when element is a 'float'");
   return *reinterpret_cast<const float *>(getElementPointer(Elt));
 }
 
-double ConstantDataSequential::getElementAsDouble(unsigned Elt) const {
+double ConstantDataSequential::getElementAsDouble(uint64_t Elt) const {
   assert(getElementType()->isDoubleTy() &&
          "Accessor can only be used when element is a 'float'");
   return *reinterpret_cast<const double *>(getElementPointer(Elt));
 }
 
-Constant *ConstantDataSequential::getElementAsConstant(unsigned Elt) const {
+Constant *ConstantDataSequential::getElementAsConstant(uint64_t Elt) const {
   if (getElementType()->isHalfTy() || getElementType()->isBFloatTy() ||
       getElementType()->isFloatTy() || getElementType()->isDoubleTy())
     return ConstantFP::get(getContext(), getElementAsAPFloat(Elt));

llvm/lib/Target/TargetLoweringObjectFile.cpp

@@ -104,14 +104,14 @@ static bool isSuitableForBSS(const GlobalVariable *GV) {
 static bool IsNullTerminatedString(const Constant *C) {
   // First check: is we have constant array terminated with zero
   if (const ConstantDataSequential *CDS = dyn_cast<ConstantDataSequential>(C)) {
-    unsigned NumElts = CDS->getNumElements();
+    uint64_t NumElts = CDS->getNumElements();
     assert(NumElts != 0 && "Can't have an empty CDS");
 
     if (CDS->getElementAsInteger(NumElts-1) != 0)
       return false; // Not null terminated.
 
     // Verify that the null doesn't occur anywhere else in the string.
-    for (unsigned i = 0; i != NumElts-1; ++i)
+    for (uint64_t i = 0; i != NumElts - 1; ++i)
       if (CDS->getElementAsInteger(i) == 0)
         return false;
     return true;