[X86][Clang] VectorExprEvaluator::VisitCallExpr / InterpretBuiltin - allow element extraction/insertion intrinsics to be used in constexpr #159753 (#161302)

FIXES: #159753

Enable constexpr evaluation for X86 vector element extract/insert builtins. and adds corresponding tests
Index is masked with `(Idx & (NumElts - 1))`, matching existing CodeGen.

Author: AdityaC4

clang/include/clang/Basic/BuiltinsX86.td

@@ -52,7 +52,7 @@ def emms : X86Builtin<"void()"> {
   let Features = "mmx";
 }
 
-let Attributes = [NoThrow, Const, RequiredVectorWidth<64>], Features = "sse" in {
+let Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<64>], Features = "sse" in {
   def vec_ext_v4hi : X86Builtin<"short(_Vector<4, short>, _Constant int)">;
   def vec_set_v4hi : X86Builtin<"_Vector<4, short>(_Vector<4, short>, short, _Constant int)">;
 }
@@ -92,13 +92,6 @@ let Attributes = [Const, NoThrow, RequiredVectorWidth<128>] in {
     def cmpsd : X86Builtin<"_Vector<2, double>(_Vector<2, double>, _Vector<2, double>, _Constant char)">;
   }
 
-  let Features = "sse2" in {
-    def vec_ext_v2di : X86Builtin<"long long int(_Vector<2, long long int>, _Constant int)">;
-    def vec_ext_v4si : X86Builtin<"int(_Vector<4, int>, _Constant int)">;
-    def vec_ext_v4sf : X86Builtin<"float(_Vector<4, float>, _Constant int)">;
-    def vec_ext_v8hi : X86Builtin<"short(_Vector<8, short>, _Constant int)">;
-    def vec_set_v8hi : X86Builtin<"_Vector<8, short>(_Vector<8, short>, short, _Constant int)">;
-  }
 
   let Features = "sse2", Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<128>] in {
     def pavgb128 : X86Builtin<"_Vector<16, unsigned char>(_Vector<16, unsigned char>, _Vector<16, unsigned char>)">;
@@ -108,6 +101,12 @@ let Attributes = [Const, NoThrow, RequiredVectorWidth<128>] in {
     def packsswb128 : X86Builtin<"_Vector<16, char>(_Vector<8, short>, _Vector<8, short>)">;
     def packssdw128 : X86Builtin<"_Vector<8, short>(_Vector<4, int>, _Vector<4, int>)">;
     def packuswb128 : X86Builtin<"_Vector<16, char>(_Vector<8, short>, _Vector<8, short>)">;
+
+    def vec_ext_v2di : X86Builtin<"long long int(_Vector<2, long long int>, _Constant int)">;
+    def vec_ext_v4si : X86Builtin<"int(_Vector<4, int>, _Constant int)">;
+    def vec_ext_v4sf : X86Builtin<"float(_Vector<4, float>, _Constant int)">;
+    def vec_ext_v8hi : X86Builtin<"short(_Vector<8, short>, _Constant int)">;
+    def vec_set_v8hi : X86Builtin<"_Vector<8, short>(_Vector<8, short>, short, _Constant int)">;
   }
 
   let Features = "sse3" in {
@@ -323,9 +322,6 @@ let Features = "sse4.1", Attributes = [NoThrow, Const, RequiredVectorWidth<128>]
   def ptestnzc128 : X86Builtin<"int(_Vector<2, long long int>, _Vector<2, long long int>)">;
   def mpsadbw128 : X86Builtin<"_Vector<16, char>(_Vector<16, char>, _Vector<16, char>, _Constant char)">;
   def phminposuw128 : X86Builtin<"_Vector<8, short>(_Vector<8, short>)">;
-  def vec_ext_v16qi : X86Builtin<"char(_Vector<16, char>, _Constant int)">;
-  def vec_set_v16qi : X86Builtin<"_Vector<16, char>(_Vector<16, char>, char, _Constant int)">;
-  def vec_set_v4si : X86Builtin<"_Vector<4, int>(_Vector<4, int>, int, _Constant int)">;
 }
 
 let Features = "sse4.1", Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<128>] in {
@@ -338,6 +334,10 @@ let Features = "sse4.1", Attributes = [NoThrow, Const, Constexpr, RequiredVector
 
   def pmuldq128 : X86Builtin<"_Vector<2, long long int>(_Vector<4, int>, _Vector<4, int>)">;
   def packusdw128 : X86Builtin<"_Vector<8, short>(_Vector<4, int>, _Vector<4, int>)">;
+
+  def vec_ext_v16qi : X86Builtin<"char(_Vector<16, char>, _Constant int)">;
+  def vec_set_v16qi : X86Builtin<"_Vector<16, char>(_Vector<16, char>, char, _Constant int)">;
+  def vec_set_v4si : X86Builtin<"_Vector<4, int>(_Vector<4, int>, int, _Constant int)">;
 }
 
 let Features = "sse4.2", Attributes = [NoThrow, Const, RequiredVectorWidth<128>] in {
@@ -560,7 +560,7 @@ let Features = "avx", Attributes = [NoThrow, RequiredVectorWidth<128>] in {
   def maskstoreps : X86Builtin<"void(_Vector<4, float *>, _Vector<4, int>, _Vector<4, float>)">;
 }
 
-let Features = "avx", Attributes = [NoThrow, Const, RequiredVectorWidth<256>] in {
+let Features = "avx", Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<256>] in {
   def vec_ext_v32qi : X86Builtin<"char(_Vector<32, char>, _Constant int)">;
   def vec_ext_v16hi : X86Builtin<"short(_Vector<16, short>, _Constant int)">;
   def vec_ext_v8si : X86Builtin<"int(_Vector<8, int>, _Constant int)">;

clang/include/clang/Basic/BuiltinsX86_64.td

@@ -56,15 +56,15 @@ let Features = "sse2", Attributes = [NoThrow] in {
   def movnti64 : X86Builtin<"void(long long int *, long long int)">;
 }
 
-let Features = "sse4.1", Attributes = [NoThrow, Const, RequiredVectorWidth<128>] in {
+let Features = "sse4.1", Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<128>] in {
   def vec_set_v2di : X86Builtin<"_Vector<2, long long int>(_Vector<2, long long int>, long long int, _Constant int)">;
 }
 
 let Features = "crc32", Attributes = [NoThrow, Const] in {
   def crc32di : X86Builtin<"unsigned long long int(unsigned long long int, unsigned long long int)">;
 }
 
-let Features = "avx", Attributes = [NoThrow, Const, RequiredVectorWidth<256>] in {
+let Features = "avx", Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<256>] in {
   def vec_ext_v4di : X86Builtin<"long long int(_Vector<4, long long int>, _Constant int)">;
   def vec_set_v4di : X86Builtin<"_Vector<4, long long int>(_Vector<4, long long int>, long long int, _Constant int)">;
 }

clang/lib/AST/ByteCode/InterpBuiltin.cpp

@@ -2878,6 +2878,61 @@ static bool interp__builtin_x86_insert_subvector(InterpState &S, CodePtr OpPC,
   return true;
 }
 
+static bool interp__builtin_vec_ext(InterpState &S, CodePtr OpPC,
+                                    const CallExpr *Call, unsigned ID) {
+  assert(Call->getNumArgs() == 2);
+
+  APSInt ImmAPS = popToAPSInt(S, Call->getArg(1));
+  const Pointer &Vec = S.Stk.pop<Pointer>();
+  if (!Vec.getFieldDesc()->isPrimitiveArray())
+    return false;
+
+  unsigned NumElems = Vec.getNumElems();
+  unsigned Index =
+      static_cast<unsigned>(ImmAPS.getZExtValue() & (NumElems - 1));
+
+  PrimType ElemPT = Vec.getFieldDesc()->getPrimType();
+  // FIXME(#161685): Replace float+int split with a numeric-only type switch
+  if (ElemPT == PT_Float) {
+    S.Stk.push<Floating>(Vec.elem<Floating>(Index));
+    return true;
+  }
+  INT_TYPE_SWITCH_NO_BOOL(ElemPT, {
+    APSInt V = Vec.elem<T>(Index).toAPSInt();
+    pushInteger(S, V, Call->getType());
+  });
+
+  return true;
+}
+
+static bool interp__builtin_vec_set(InterpState &S, CodePtr OpPC,
+                                    const CallExpr *Call, unsigned ID) {
+  assert(Call->getNumArgs() == 3);
+
+  APSInt ImmAPS = popToAPSInt(S, Call->getArg(2));
+  APSInt ValAPS = popToAPSInt(S, Call->getArg(1));
+
+  const Pointer &Base = S.Stk.pop<Pointer>();
+  if (!Base.getFieldDesc()->isPrimitiveArray())
+    return false;
+
+  const Pointer &Dst = S.Stk.peek<Pointer>();
+
+  unsigned NumElems = Base.getNumElems();
+  unsigned Index =
+      static_cast<unsigned>(ImmAPS.getZExtValue() & (NumElems - 1));
+
+  PrimType ElemPT = Base.getFieldDesc()->getPrimType();
+  INT_TYPE_SWITCH_NO_BOOL(ElemPT, {
+    for (unsigned I = 0; I != NumElems; ++I)
+      Dst.elem<T>(I) = Base.elem<T>(I);
+    Dst.elem<T>(Index) = static_cast<T>(ValAPS);
+  });
+
+  Dst.initializeAllElements();
+  return true;
+}
+
 bool InterpretBuiltin(InterpState &S, CodePtr OpPC, const CallExpr *Call,
                       uint32_t BuiltinID) {
   if (!S.getASTContext().BuiltinInfo.isConstantEvaluated(BuiltinID))
@@ -3686,6 +3741,29 @@ bool InterpretBuiltin(InterpState &S, CodePtr OpPC, const CallExpr *Call,
   case X86::BI__builtin_ia32_insert128i256:
     return interp__builtin_x86_insert_subvector(S, OpPC, Call, BuiltinID);
 
+  case X86::BI__builtin_ia32_vec_ext_v4hi:
+  case X86::BI__builtin_ia32_vec_ext_v16qi:
+  case X86::BI__builtin_ia32_vec_ext_v8hi:
+  case X86::BI__builtin_ia32_vec_ext_v4si:
+  case X86::BI__builtin_ia32_vec_ext_v2di:
+  case X86::BI__builtin_ia32_vec_ext_v32qi:
+  case X86::BI__builtin_ia32_vec_ext_v16hi:
+  case X86::BI__builtin_ia32_vec_ext_v8si:
+  case X86::BI__builtin_ia32_vec_ext_v4di:
+  case X86::BI__builtin_ia32_vec_ext_v4sf:
+    return interp__builtin_vec_ext(S, OpPC, Call, BuiltinID);
+
+  case X86::BI__builtin_ia32_vec_set_v4hi:
+  case X86::BI__builtin_ia32_vec_set_v16qi:
+  case X86::BI__builtin_ia32_vec_set_v8hi:
+  case X86::BI__builtin_ia32_vec_set_v4si:
+  case X86::BI__builtin_ia32_vec_set_v2di:
+  case X86::BI__builtin_ia32_vec_set_v32qi:
+  case X86::BI__builtin_ia32_vec_set_v16hi:
+  case X86::BI__builtin_ia32_vec_set_v8si:
+  case X86::BI__builtin_ia32_vec_set_v4di:
+    return interp__builtin_vec_set(S, OpPC, Call, BuiltinID);
+
   default:
     S.FFDiag(S.Current->getLocation(OpPC),
              diag::note_invalid_subexpr_in_const_expr)

clang/lib/AST/ExprConstant.cpp

@@ -12235,6 +12235,41 @@ bool VectorExprEvaluator::VisitCallExpr(const CallExpr *E) {
 
     return Success(APValue(ResultElements.data(), ResultElements.size()), E);
   }
+
+  case clang::X86::BI__builtin_ia32_vec_set_v4hi:
+  case clang::X86::BI__builtin_ia32_vec_set_v16qi:
+  case clang::X86::BI__builtin_ia32_vec_set_v8hi:
+  case clang::X86::BI__builtin_ia32_vec_set_v4si:
+  case clang::X86::BI__builtin_ia32_vec_set_v2di:
+  case clang::X86::BI__builtin_ia32_vec_set_v32qi:
+  case clang::X86::BI__builtin_ia32_vec_set_v16hi:
+  case clang::X86::BI__builtin_ia32_vec_set_v8si:
+  case clang::X86::BI__builtin_ia32_vec_set_v4di: {
+    APValue VecVal;
+    APSInt Scalar, IndexAPS;
+    if (!EvaluateVector(E->getArg(0), VecVal, Info) ||
+        !EvaluateInteger(E->getArg(1), Scalar, Info) ||
+        !EvaluateInteger(E->getArg(2), IndexAPS, Info))
+      return false;
+
+    QualType ElemTy = E->getType()->castAs<VectorType>()->getElementType();
+    unsigned ElemWidth = Info.Ctx.getIntWidth(ElemTy);
+    bool ElemUnsigned = ElemTy->isUnsignedIntegerOrEnumerationType();
+    Scalar.setIsUnsigned(ElemUnsigned);
+    APSInt ElemAPS = Scalar.extOrTrunc(ElemWidth);
+    APValue ElemAV(ElemAPS);
+
+    unsigned NumElems = VecVal.getVectorLength();
+    unsigned Index =
+        static_cast<unsigned>(IndexAPS.getZExtValue() & (NumElems - 1));
+
+    SmallVector<APValue, 4> Elems;
+    Elems.reserve(NumElems);
+    for (unsigned ElemNum = 0; ElemNum != NumElems; ++ElemNum)
+      Elems.push_back(ElemNum == Index ? ElemAV : VecVal.getVectorElt(ElemNum));
+
+    return Success(APValue(Elems.data(), NumElems), E);
+  }
   }
 }
 
@@ -14822,6 +14857,25 @@ bool IntExprEvaluator::VisitBuiltinCallExpr(const CallExpr *E,
     return HandleMaskBinOp(
         [](const APSInt &LHS, const APSInt &RHS) { return LHS + RHS; });
   }
+
+  case clang::X86::BI__builtin_ia32_vec_ext_v4hi:
+  case clang::X86::BI__builtin_ia32_vec_ext_v16qi:
+  case clang::X86::BI__builtin_ia32_vec_ext_v8hi:
+  case clang::X86::BI__builtin_ia32_vec_ext_v4si:
+  case clang::X86::BI__builtin_ia32_vec_ext_v2di:
+  case clang::X86::BI__builtin_ia32_vec_ext_v32qi:
+  case clang::X86::BI__builtin_ia32_vec_ext_v16hi:
+  case clang::X86::BI__builtin_ia32_vec_ext_v8si:
+  case clang::X86::BI__builtin_ia32_vec_ext_v4di: {
+    APValue Vec;
+    APSInt IdxAPS;
+    if (!EvaluateVector(E->getArg(0), Vec, Info) ||
+        !EvaluateInteger(E->getArg(1), IdxAPS, Info))
+      return false;
+    unsigned N = Vec.getVectorLength();
+    unsigned Idx = static_cast<unsigned>(IdxAPS.getZExtValue() & (N - 1));
+    return Success(Vec.getVectorElt(Idx).getInt(), E);
+  }
   }
 }
 
@@ -16638,6 +16692,17 @@ bool FloatExprEvaluator::VisitCallExpr(const CallExpr *E) {
     (void)Result.fusedMultiplyAdd(SourceY, SourceZ, RM);
     return true;
   }
+
+  case clang::X86::BI__builtin_ia32_vec_ext_v4sf: {
+    APValue Vec;
+    APSInt IdxAPS;
+    if (!EvaluateVector(E->getArg(0), Vec, Info) ||
+        !EvaluateInteger(E->getArg(1), IdxAPS, Info))
+      return false;
+    unsigned N = Vec.getVectorLength();
+    unsigned Idx = static_cast<unsigned>(IdxAPS.getZExtValue() & (N - 1));
+    return Success(Vec.getVectorElt(Idx), E);
+  }
   }
 }
 

clang/test/CodeGen/X86/avx-builtins.c

@@ -1039,26 +1039,30 @@ int test_mm256_extract_epi8(__m256i A) {
   // CHECK: zext i8 %{{.*}} to i32
   return _mm256_extract_epi8(A, 31);
 }
+TEST_CONSTEXPR(_mm256_extract_epi8(((__m256i)(__v32qs){0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31}), 45) == 13);
 
 int test_mm256_extract_epi16(__m256i A) {
   // CHECK-LABEL: test_mm256_extract_epi16
   // CHECK: extractelement <16 x i16> %{{.*}}, {{i32|i64}} 15
   // CHECK: zext i16 %{{.*}} to i32
   return _mm256_extract_epi16(A, 15);
 }
+TEST_CONSTEXPR(_mm256_extract_epi16(((__m256i)(__v16hi){0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30}), 50) == 4);
 
 int test_mm256_extract_epi32(__m256i A) {
   // CHECK-LABEL: test_mm256_extract_epi32
   // CHECK: extractelement <8 x i32> %{{.*}}, {{i32|i64}} 7
   return _mm256_extract_epi32(A, 7);
 }
+TEST_CONSTEXPR(_mm256_extract_epi32(((__m256i)(__v8si){0, 5, 10, 15, 20, 25, 30, 35}), 18) == 10);
 
 #if __x86_64__
 long long test_mm256_extract_epi64(__m256i A) {
   // X64-LABEL: test_mm256_extract_epi64
   // X64: extractelement <4 x i64> %{{.*}}, {{i32|i64}} 3
   return _mm256_extract_epi64(A, 3);
 }
+TEST_CONSTEXPR(_mm256_extract_epi64(((__m256i)(__v4di){5, 15, 25, 35}), 14) == 25);
 #endif
 
 __m128d test_mm256_extractf128_pd(__m256d A) {
@@ -1120,25 +1124,29 @@ __m256i test_mm256_insert_epi8(__m256i x, char b) {
   // CHECK: insertelement <32 x i8> %{{.*}}, i8 %{{.*}}, {{i32|i64}} 14
   return _mm256_insert_epi8(x, b, 14);
 }
+TEST_CONSTEXPR(match_v32qi(_mm256_insert_epi8(((__m256i)(__v32qs){0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31}), 77, 47), 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 77, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31));
 
 __m256i test_mm256_insert_epi16(__m256i x, int b) {
   // CHECK-LABEL: test_mm256_insert_epi16
   // CHECK: insertelement <16 x i16> %{{.*}}, i16 %{{.*}}, {{i32|i64}} 4
   return _mm256_insert_epi16(x, b, 4);
 }
+TEST_CONSTEXPR(match_v16hi(_mm256_insert_epi16(((__m256i)(__v16hi){0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30}), 909, 62), 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 909, 30));
 
 __m256i test_mm256_insert_epi32(__m256i x, int b) {
   // CHECK-LABEL: test_mm256_insert_epi32
   // CHECK: insertelement <8 x i32> %{{.*}}, i32 %{{.*}}, {{i32|i64}} 5
   return _mm256_insert_epi32(x, b, 5);
 }
+TEST_CONSTEXPR(match_v8si(_mm256_insert_epi32(((__m256i)(__v8si){ 0, 5, 10, 15, 20, 25, 30, 35}), 4321, 18), 0, 5, 4321, 15, 20, 25, 30, 35));
 
 #if __x86_64__
 __m256i test_mm256_insert_epi64(__m256i x, long long b) {
   // X64-LABEL: test_mm256_insert_epi64
   // X64: insertelement <4 x i64> %{{.*}}, i64 %{{.*}}, {{i32|i64}} 2
   return _mm256_insert_epi64(x, b, 2);
 }
+TEST_CONSTEXPR(match_v4di(_mm256_insert_epi64(((__m256i)(__v4di){5, 15, 25, 35}), -123456789LL, 10), 5, 15, -123456789LL, 35));
 #endif
 
 __m256d test_mm256_insertf128_pd(__m256d A, __m128d B) {

clang/test/CodeGen/X86/mmx-builtins.c

@@ -292,6 +292,7 @@ int test_mm_extract_pi16(__m64 a) {
   // CHECK: extractelement <4 x i16> {{%.*}}, i64 2
   return _mm_extract_pi16(a, 2);
 }
+TEST_CONSTEXPR(_mm_extract_pi16(((__m64)(__v4hi){10, 20, 30, 40}), 7) == 40);
 
 __m64 test_m_from_int(int a) {
   // CHECK-LABEL: test_m_from_int
@@ -347,6 +348,7 @@ __m64 test_mm_insert_pi16(__m64 a, int d) {
   // CHECK: insertelement <4 x i16>
   return _mm_insert_pi16(a, d, 2);
 }
+TEST_CONSTEXPR(match_v4hi(_mm_insert_pi16(((__m64)(__v4hi){0, 1, 2, 3}), 77, 10), 0, 1, 77, 3));
 
 __m64 test_mm_madd_pi16(__m64 a, __m64 b) {
   // CHECK-LABEL: test_mm_madd_pi16

clang/test/CodeGen/X86/sse2-builtins.c

@@ -723,12 +723,14 @@ int test_mm_extract_epi16(__m128i A) {
   // CHECK: zext i16 %{{.*}} to i32
   return _mm_extract_epi16(A, 1);
 }
+TEST_CONSTEXPR(_mm_extract_epi16(((__m128i)(__v8hi){0, 10, 20, 30, 40, 50, 60, 70}), 25) == 10);
 
 __m128i test_mm_insert_epi16(__m128i A, int B) {
   // CHECK-LABEL: test_mm_insert_epi16
   // CHECK: insertelement <8 x i16> %{{.*}}, {{i32|i64}} 0
   return _mm_insert_epi16(A, B, 0);
 }
+TEST_CONSTEXPR(match_v8hi(_mm_insert_epi16(((__m128i)(__v8hi){0, 10, 20, 30, 40, 50, 60, 70}), 555, 17), 0, 555, 20, 30, 40, 50, 60, 70));
 
 void test_mm_lfence(void) {
   // CHECK-LABEL: test_mm_lfence