[clang][Interp] Add IntegralAP for arbitrary-precision integers (#65844)

This adds `IntegralAP` backing the two new primtypes `IntAP` (unsigned
arbitrary-precision int) and `IntAPS` (same but signed).

We use this for `int128` support (which isn't available on all host
systems we support AFAIK) and I think we can also use this for `_BitInt`
later.

Author: tbaederr

clang/lib/AST/Interp/ByteCodeExprGen.cpp

@@ -171,14 +171,17 @@ bool ByteCodeExprGen<Emitter>::VisitCastExpr(const CastExpr *CE) {
       return this->discard(SubExpr);
     std::optional<PrimType> FromT = classify(SubExpr->getType());
     std::optional<PrimType> ToT = classify(CE->getType());
+
     if (!FromT || !ToT)
       return false;
 
     if (!this->visit(SubExpr))
       return false;
 
-    if (FromT == ToT)
+    if (FromT == ToT) {
+      assert(ToT != PT_IntAP && ToT != PT_IntAPS);
       return true;
+    }
 
     return this->emitCast(*FromT, *ToT, CE);
   }
@@ -1638,6 +1641,9 @@ bool ByteCodeExprGen<Emitter>::visitZeroInitializer(QualType QT,
     return this->emitZeroSint64(E);
   case PT_Uint64:
     return this->emitZeroUint64(E);
+  case PT_IntAP:
+  case PT_IntAPS:
+    assert(false);
   case PT_Ptr:
     return this->emitNullPtr(E);
   case PT_FnPtr:
@@ -1877,6 +1883,9 @@ bool ByteCodeExprGen<Emitter>::emitConst(T Value, PrimType Ty, const Expr *E) {
     return this->emitConstSint64(Value, E);
   case PT_Uint64:
     return this->emitConstUint64(Value, E);
+  case PT_IntAP:
+  case PT_IntAPS:
+    assert(false);
   case PT_Bool:
     return this->emitConstBool(Value, E);
   case PT_Ptr:

clang/lib/AST/Interp/Context.cpp

@@ -103,7 +103,7 @@ std::optional<PrimType> Context::classify(QualType T) const {
     case 8:
       return PT_Sint8;
     default:
-      return std::nullopt;
+      return PT_IntAPS;
     }
   }
 
@@ -118,7 +118,7 @@ std::optional<PrimType> Context::classify(QualType T) const {
     case 8:
       return PT_Uint8;
     default:
-      return std::nullopt;
+      return PT_IntAP;
     }
   }
 

clang/lib/AST/Interp/Descriptor.cpp

@@ -10,6 +10,7 @@
 #include "Boolean.h"
 #include "Floating.h"
 #include "FunctionPointer.h"
+#include "IntegralAP.h"
 #include "Pointer.h"
 #include "PrimType.h"
 #include "Record.h"
@@ -182,6 +183,10 @@ static BlockCtorFn getCtorPrim(PrimType Type) {
   // constructor called.
   if (Type == PT_Float)
     return ctorTy<PrimConv<PT_Float>::T>;
+  if (Type == PT_IntAP)
+    return ctorTy<PrimConv<PT_IntAP>::T>;
+  if (Type == PT_IntAPS)
+    return ctorTy<PrimConv<PT_IntAPS>::T>;
 
   COMPOSITE_TYPE_SWITCH(Type, return ctorTy<T>, return nullptr);
 }
@@ -191,6 +196,10 @@ static BlockDtorFn getDtorPrim(PrimType Type) {
   // destructor called, since they might allocate memory.
   if (Type == PT_Float)
     return dtorTy<PrimConv<PT_Float>::T>;
+  if (Type == PT_IntAP)
+    return dtorTy<PrimConv<PT_IntAP>::T>;
+  if (Type == PT_IntAPS)
+    return dtorTy<PrimConv<PT_IntAPS>::T>;
 
   COMPOSITE_TYPE_SWITCH(Type, return dtorTy<T>, return nullptr);
 }

clang/lib/AST/Interp/EvalEmitter.cpp

@@ -9,6 +9,7 @@
 #include "EvalEmitter.h"
 #include "ByteCodeGenError.h"
 #include "Context.h"
+#include "IntegralAP.h"
 #include "Interp.h"
 #include "Opcode.h"
 #include "clang/AST/DeclCXX.h"

clang/lib/AST/Interp/Integral.h

@@ -29,6 +29,8 @@ namespace interp {
 using APInt = llvm::APInt;
 using APSInt = llvm::APSInt;
 
+template <bool Signed> class IntegralAP;
+
 // Helper structure to select the representation.
 template <unsigned Bits, bool Signed> struct Repr;
 template <> struct Repr<8, false> { using Type = uint8_t; };
@@ -61,6 +63,8 @@ template <unsigned Bits, bool Signed> class Integral final {
   template <typename T> explicit Integral(T V) : V(V) {}
 
 public:
+  using AsUnsigned = Integral<Bits, false>;
+
   /// Zero-initializes an integral.
   Integral() : V(0) {}
 

clang/lib/AST/Interp/IntegralAP.h

@@ -0,0 +1,256 @@
+//===--- Integral.h - Wrapper for numeric types for the VM ------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// Defines the VM types and helpers operating on types.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_CLANG_AST_INTERP_INTEGRAL_AP_H
+#define LLVM_CLANG_AST_INTERP_INTEGRAL_AP_H
+
+#include "clang/AST/APValue.h"
+#include "clang/AST/ComparisonCategories.h"
+#include "llvm/ADT/APSInt.h"
+#include "llvm/Support/MathExtras.h"
+#include "llvm/Support/raw_ostream.h"
+#include <cstddef>
+#include <cstdint>
+
+#include "Primitives.h"
+
+namespace clang {
+namespace interp {
+
+using APInt = llvm::APInt;
+using APSInt = llvm::APSInt;
+template <unsigned Bits, bool Signed> class Integral;
+class Boolean;
+
+template <bool Signed> class IntegralAP final {
+public:
+  APSInt V;
+
+public:
+  using AsUnsigned = IntegralAP<false>;
+
+  template <typename T>
+  IntegralAP(T Value) : V(APInt(sizeof(T) * 8, Value, std::is_signed_v<T>)) {}
+
+  IntegralAP(APInt V) : V(V) {}
+  IntegralAP(APSInt V) : V(V) {}
+  /// Arbitrary value for initialized variables.
+  IntegralAP() : V(APSInt::getMaxValue(1024, Signed)) {}
+
+  IntegralAP operator-() const { return IntegralAP(-V); }
+  bool operator>(IntegralAP RHS) const { return V > RHS.V; }
+  bool operator>=(IntegralAP RHS) const { return V >= RHS.V; }
+  bool operator<(IntegralAP RHS) const { return V < RHS.V; }
+  bool operator<=(IntegralAP RHS) const { return V <= RHS.V; }
+
+  explicit operator bool() const { return !V.isZero(); }
+  explicit operator int8_t() const { return V.getSExtValue(); }
+  explicit operator uint8_t() const { return V.getZExtValue(); }
+  explicit operator int16_t() const { return V.getSExtValue(); }
+  explicit operator uint16_t() const { return V.getZExtValue(); }
+  explicit operator int32_t() const { return V.getSExtValue(); }
+  explicit operator uint32_t() const { return V.getZExtValue(); }
+  explicit operator int64_t() const { return V.getSExtValue(); }
+  explicit operator uint64_t() const { return V.getZExtValue(); }
+
+  template <typename T> static IntegralAP from(T Value, unsigned NumBits = 0) {
+    assert(NumBits > 0);
+    APSInt Copy = APSInt(APInt(NumBits, Value, Signed), !Signed);
+
+    return IntegralAP<Signed>(Copy);
+  }
+
+  template <bool InputSigned>
+  static IntegralAP from(IntegralAP<InputSigned> V, unsigned NumBits = 0) {
+    if constexpr (Signed == InputSigned)
+      return V;
+
+    APSInt Copy = V.V;
+    Copy.setIsSigned(Signed);
+
+    return IntegralAP<Signed>(Copy);
+  }
+
+  template <unsigned Bits, bool InputSigned>
+  static IntegralAP from(Integral<Bits, InputSigned> I) {
+    // FIXME: Take bits parameter.
+    APSInt Copy =
+        APSInt(APInt(128, static_cast<int64_t>(I), InputSigned), !Signed);
+    Copy.setIsSigned(Signed);
+
+    assert(Copy.isSigned() == Signed);
+    return IntegralAP<Signed>(Copy);
+  }
+  static IntegralAP from(const Boolean &B) {
+    assert(false);
+    return IntegralAP::zero();
+  }
+
+  static IntegralAP zero() {
+    assert(false);
+    return IntegralAP(0);
+  }
+
+  // FIXME: This can't be static if the bitwidth depends on V.
+  static constexpr unsigned bitWidth() { return 128; }
+
+  APSInt toAPSInt(unsigned Bits = 0) const { return V; }
+  APValue toAPValue() const { return APValue(V); }
+
+  bool isZero() const { return V.isZero(); }
+  bool isPositive() const { return V.isNonNegative(); }
+  bool isNegative() const { return !V.isNonNegative(); }
+  bool isMin() const { return V.isMinValue(); }
+  bool isMax() const { return V.isMaxValue(); }
+  static bool isSigned() { return Signed; }
+  bool isMinusOne() const { return Signed && V == -1; }
+
+  unsigned countLeadingZeros() const { return V.countl_zero(); }
+
+  void print(llvm::raw_ostream &OS) const { OS << V; }
+
+  IntegralAP truncate(unsigned bitWidth) const {
+    assert(false);
+    return V;
+  }
+
+  IntegralAP<false> toUnsigned() const {
+    APSInt Copy = V;
+    Copy.setIsSigned(false);
+    return IntegralAP<false>(Copy);
+  }
+
+  ComparisonCategoryResult compare(const IntegralAP &RHS) const {
+    return Compare(V, RHS.V);
+  }
+
+  static bool increment(IntegralAP A, IntegralAP *R) {
+    assert(false);
+    *R = IntegralAP(A.V + 1);
+    return false;
+  }
+
+  static bool decrement(IntegralAP A, IntegralAP *R) {
+    assert(false);
+    *R = IntegralAP(A.V - 1);
+    return false;
+  }
+
+  static bool add(IntegralAP A, IntegralAP B, unsigned OpBits, IntegralAP *R) {
+    return CheckAddUB(A, B, OpBits, R);
+  }
+
+  static bool sub(IntegralAP A, IntegralAP B, unsigned OpBits, IntegralAP *R) {
+    /// FIXME: Gotta check if the result fits into OpBits bits.
+    return CheckSubUB(A, B, R);
+  }
+
+  static bool mul(IntegralAP A, IntegralAP B, unsigned OpBits, IntegralAP *R) {
+    assert(false);
+    // return CheckMulUB(A.V, B.V, R->V);
+    return false;
+  }
+
+  static bool rem(IntegralAP A, IntegralAP B, unsigned OpBits, IntegralAP *R) {
+    assert(false);
+    *R = IntegralAP(A.V % B.V);
+    return false;
+  }
+
+  static bool div(IntegralAP A, IntegralAP B, unsigned OpBits, IntegralAP *R) {
+    assert(false);
+    *R = IntegralAP(A.V / B.V);
+    return false;
+  }
+
+  static bool bitAnd(IntegralAP A, IntegralAP B, unsigned OpBits,
+                     IntegralAP *R) {
+    assert(false);
+    *R = IntegralAP(A.V & B.V);
+    return false;
+  }
+
+  static bool bitOr(IntegralAP A, IntegralAP B, unsigned OpBits,
+                    IntegralAP *R) {
+    assert(false);
+    *R = IntegralAP(A.V | B.V);
+    return false;
+  }
+
+  static bool bitXor(IntegralAP A, IntegralAP B, unsigned OpBits,
+                     IntegralAP *R) {
+    assert(false);
+    *R = IntegralAP(A.V ^ B.V);
+    return false;
+  }
+
+  static bool neg(const IntegralAP &A, IntegralAP *R) {
+    APSInt AI = A.V;
+
+    AI.setIsSigned(Signed);
+    *R = IntegralAP(AI);
+    return false;
+  }
+
+  static bool comp(IntegralAP A, IntegralAP *R) {
+    assert(false);
+    *R = IntegralAP(~A.V);
+    return false;
+  }
+
+  static void shiftLeft(const IntegralAP A, const IntegralAP B, unsigned OpBits,
+                        IntegralAP *R) {
+    *R = IntegralAP(A.V << B.V.getZExtValue());
+  }
+
+  static void shiftRight(const IntegralAP A, const IntegralAP B,
+                         unsigned OpBits, IntegralAP *R) {
+    *R = IntegralAP(A.V >> B.V.getZExtValue());
+  }
+
+private:
+  static bool CheckAddUB(const IntegralAP &A, const IntegralAP &B,
+                         unsigned BitWidth, IntegralAP *R) {
+    if (!A.isSigned()) {
+      R->V = A.V + B.V;
+      return false;
+    }
+
+    const APSInt &LHS = A.V;
+    const APSInt &RHS = B.V;
+
+    APSInt Value(LHS.extend(BitWidth) + RHS.extend(BitWidth), false);
+    APSInt Result = Value.trunc(LHS.getBitWidth());
+    if (Result.extend(BitWidth) != Value)
+      return true;
+
+    R->V = Result;
+    return false;
+  }
+  static bool CheckSubUB(const IntegralAP &A, const IntegralAP &B,
+                         IntegralAP *R) {
+    R->V = A.V - B.V;
+    return false; // Success!
+  }
+};
+
+template <bool Signed>
+inline llvm::raw_ostream &operator<<(llvm::raw_ostream &OS,
+                                     IntegralAP<Signed> I) {
+  I.print(OS);
+  return OS;
+}
+
+} // namespace interp
+} // namespace clang
+
+#endif