Merge branch 'main' into an/avx-masked-shuffles

Author: RKSimon

clang/cmake/modules/CMakeLists.txt

@@ -8,15 +8,19 @@ include(FindPrefixFromConfig)
 # the usual CMake convention seems to be ${Project}Targets.cmake.
 set(CLANG_INSTALL_PACKAGE_DIR "${CMAKE_INSTALL_PACKAGEDIR}/clang" CACHE STRING
   "Path for CMake subdirectory for Clang (defaults to '${CMAKE_INSTALL_PACKAGEDIR}/clang')")
-# CMAKE_INSTALL_PACKAGEDIR might be absolute, so don't reuse below.
-set(clang_cmake_builddir "${CMAKE_BINARY_DIR}/lib${LLVM_LIBDIR_SUFFIX}/cmake/clang")
 
 # Keep this in sync with llvm/cmake/CMakeLists.txt!
 set(LLVM_INSTALL_PACKAGE_DIR "${CMAKE_INSTALL_PACKAGEDIR}/llvm" CACHE STRING
   "Path for CMake subdirectory for LLVM (defaults to '${CMAKE_INSTALL_PACKAGEDIR}/llvm')")
 # CMAKE_INSTALL_PACKAGEDIR might be absolute, so don't reuse below.
-string(REPLACE "${CMAKE_CFG_INTDIR}" "." llvm_cmake_builddir "${LLVM_LIBRARY_DIR}")
-set(llvm_cmake_builddir "${llvm_cmake_builddir}/cmake/llvm")
+string(REPLACE "${CMAKE_CFG_INTDIR}" "." llvm_builddir "${LLVM_LIBRARY_DIR}")
+set(llvm_cmake_builddir "${llvm_builddir}/cmake/llvm")
+if(CLANG_BUILT_STANDALONE)
+  # CMAKE_INSTALL_PACKAGEDIR might be absolute, so don't reuse below.
+  set(clang_cmake_builddir "${CMAKE_BINARY_DIR}/lib${LLVM_LIBDIR_SUFFIX}/cmake/clang")
+else()
+  set(clang_cmake_builddir "${llvm_builddir}/cmake/clang")
+endif()
 
 get_property(CLANG_EXPORTS GLOBAL PROPERTY CLANG_EXPORTS)
 export(TARGETS ${CLANG_EXPORTS} FILE ${clang_cmake_builddir}/ClangTargets.cmake)

clang/docs/ReleaseNotes.rst

@@ -518,6 +518,7 @@ X86 Support
 - Remove `[no-]evex512` feature request from intrinsics and builtins.
 - Change features `avx10.x-[256,512]` to `avx10.x`.
 - `-march=wildcatlake` is now supported.
+- `-march=novalake` is now supported.
 
 Arm and AArch64 Support
 ^^^^^^^^^^^^^^^^^^^^^^^

clang/docs/analyzer/checkers.rst

@@ -3465,12 +3465,6 @@ Check for an out-of-bound pointer being returned to callers.
    return x; // warn: undefined or garbage returned
  }
 
-
-alpha.security.cert
-^^^^^^^^^^^^^^^^^^^
-
-SEI CERT checkers which tries to find errors based on their `C coding rules <https://wiki.sei.cmu.edu/confluence/display/c/2+Rules>`_.
-
 alpha.unix
 ^^^^^^^^^^
 

clang/include/clang/AST/TypeBase.h

@@ -4378,8 +4378,6 @@ class ConstantMatrixType final : public MatrixType {
   unsigned NumRows;
   unsigned NumColumns;
 
-  static constexpr unsigned MaxElementsPerDimension = (1 << 20) - 1;
-
   ConstantMatrixType(QualType MatrixElementType, unsigned NRows,
                      unsigned NColumns, QualType CanonElementType);
 
@@ -4398,16 +4396,6 @@ class ConstantMatrixType final : public MatrixType {
     return getNumRows() * getNumColumns();
   }
 
-  /// Returns true if \p NumElements is a valid matrix dimension.
-  static constexpr bool isDimensionValid(size_t NumElements) {
-    return NumElements > 0 && NumElements <= MaxElementsPerDimension;
-  }
-
-  /// Returns the maximum number of elements per dimension.
-  static constexpr unsigned getMaxElementsPerDimension() {
-    return MaxElementsPerDimension;
-  }
-
   void Profile(llvm::FoldingSetNodeID &ID) {
     Profile(ID, getElementType(), getNumRows(), getNumColumns(),
             getTypeClass());

clang/include/clang/Basic/BuiltinsX86.td

@@ -1359,23 +1359,17 @@ let Features = "avx512bw", Attributes = [NoThrow, Const, Constexpr, RequiredVect
   def pshufb512 : X86Builtin<"_Vector<64, char>(_Vector<64, char>, _Vector<64, char>)">;
 }
 
-let Features = "avx512cd,avx512vl", Attributes = [NoThrow, Const, RequiredVectorWidth<128>] in {
+let Features = "avx512cd,avx512vl", Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<128>] in {
   def vpconflictdi_128 : X86Builtin<"_Vector<2, long long int>(_Vector<2, long long int>)">;
-}
-
-let Features = "avx512cd,avx512vl", Attributes = [NoThrow, Const, RequiredVectorWidth<256>] in {
-  def vpconflictdi_256 : X86Builtin<"_Vector<4, long long int>(_Vector<4, long long int>)">;
-}
-
-let Features = "avx512cd,avx512vl", Attributes = [NoThrow, Const, RequiredVectorWidth<128>] in {
   def vpconflictsi_128 : X86Builtin<"_Vector<4, int>(_Vector<4, int>)">;
 }
 
-let Features = "avx512cd,avx512vl", Attributes = [NoThrow, Const, RequiredVectorWidth<256>] in {
+let Features = "avx512cd,avx512vl", Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<256>] in {
+  def vpconflictdi_256 : X86Builtin<"_Vector<4, long long int>(_Vector<4, long long int>)">;
   def vpconflictsi_256 : X86Builtin<"_Vector<8, int>(_Vector<8, int>)">;
 }
 
-let Features = "avx512cd", Attributes = [NoThrow, Const, RequiredVectorWidth<512>] in {
+let Features = "avx512cd", Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<512>] in {
   def vpconflictdi_512 : X86Builtin<"_Vector<8, long long int>(_Vector<8, long long int>)">;
   def vpconflictsi_512 : X86Builtin<"_Vector<16, int>(_Vector<16, int>)">;
 }

clang/include/clang/Basic/LangOptions.def

@@ -433,6 +433,7 @@ ENUM_LANGOPT(RegisterStaticDestructors, RegisterStaticDestructorsKind, 2,
 LANGOPT(RegCall4, 1, 0, NotCompatible, "Set __regcall4 as a default calling convention to respect __regcall ABI v.4")
 
 LANGOPT(MatrixTypes, 1, 0, NotCompatible, "Enable or disable the builtin matrix type")
+VALUE_LANGOPT(MaxMatrixDimension, 32, (1 << 20) - 1, NotCompatible, "maximum allowed matrix dimension")
 
 LANGOPT(CXXAssumptions, 1, 1, NotCompatible, "Enable or disable codegen and compile-time checks for C++23's [[assume]] attribute")
 

clang/include/clang/StaticAnalyzer/Checkers/Checkers.td

@@ -76,9 +76,6 @@ def SecurityAlpha : Package<"security">, ParentPackage<Alpha>;
 def CERT : Package<"cert">, ParentPackage<Security>;
 def ENV : Package<"env">, ParentPackage<CERT>;
 
-def CERTAlpha : Package<"cert">, ParentPackage<SecurityAlpha>;
-def POSAlpha : Package<"pos">, ParentPackage<CERTAlpha>;
-
 def Unix : Package<"unix">;
 def UnixAlpha : Package<"unix">, ParentPackage<Alpha>;
 def CString : Package<"cstring">, ParentPackage<Unix>;

clang/lib/AST/ASTContext.cpp

@@ -4712,8 +4712,8 @@ QualType ASTContext::getConstantMatrixType(QualType ElementTy, unsigned NumRows,
 
   assert(MatrixType::isValidElementType(ElementTy) &&
          "need a valid element type");
-  assert(ConstantMatrixType::isDimensionValid(NumRows) &&
-         ConstantMatrixType::isDimensionValid(NumColumns) &&
+  assert(NumRows > 0 && NumRows <= LangOpts.MaxMatrixDimension &&
+         NumColumns > 0 && NumColumns <= LangOpts.MaxMatrixDimension &&
          "need valid matrix dimensions");
   void *InsertPos = nullptr;
   if (ConstantMatrixType *MTP = MatrixTypes.FindNodeOrInsertPos(ID, InsertPos))

clang/lib/AST/ByteCode/InterpBuiltin.cpp

@@ -3101,6 +3101,33 @@ static bool interp__builtin_vec_set(InterpState &S, CodePtr OpPC,
   return true;
 }
 
+static bool interp__builtin_ia32_vpconflict(InterpState &S, CodePtr OpPC,
+                                            const CallExpr *Call) {
+  assert(Call->getNumArgs() == 1);
+
+  QualType Arg0Type = Call->getArg(0)->getType();
+  const auto *VecT = Arg0Type->castAs<VectorType>();
+  PrimType ElemT = *S.getContext().classify(VecT->getElementType());
+  unsigned NumElems = VecT->getNumElements();
+  bool DestUnsigned = Call->getType()->isUnsignedIntegerOrEnumerationType();
+  const Pointer &Src = S.Stk.pop<Pointer>();
+  const Pointer &Dst = S.Stk.peek<Pointer>();
+
+  for (unsigned I = 0; I != NumElems; ++I) {
+    INT_TYPE_SWITCH_NO_BOOL(ElemT, {
+      APSInt ElemI = Src.elem<T>(I).toAPSInt();
+      APInt ConflictMask(ElemI.getBitWidth(), 0);
+      for (unsigned J = 0; J != I; ++J) {
+        APSInt ElemJ = Src.elem<T>(J).toAPSInt();
+        ConflictMask.setBitVal(J, ElemI == ElemJ);
+      }
+      Dst.elem<T>(I) = static_cast<T>(APSInt(ConflictMask, DestUnsigned));
+    });
+  }
+  Dst.initializeAllElements();
+  return true;
+}
+
 bool InterpretBuiltin(InterpState &S, CodePtr OpPC, const CallExpr *Call,
                       uint32_t BuiltinID) {
   if (!S.getASTContext().BuiltinInfo.isConstantEvaluated(BuiltinID))
@@ -3891,7 +3918,13 @@ bool InterpretBuiltin(InterpState &S, CodePtr OpPC, const CallExpr *Call,
         [](const APSInt &Lo, const APSInt &Hi, const APSInt &Amt) {
           return llvm::APIntOps::fshr(Hi, Lo, Amt);
         });
-
+  case X86::BI__builtin_ia32_vpconflictsi_128:
+  case X86::BI__builtin_ia32_vpconflictsi_256:
+  case X86::BI__builtin_ia32_vpconflictsi_512:
+  case X86::BI__builtin_ia32_vpconflictdi_128:
+  case X86::BI__builtin_ia32_vpconflictdi_256:
+  case X86::BI__builtin_ia32_vpconflictdi_512:
+    return interp__builtin_ia32_vpconflict(S, OpPC, Call);
   case clang::X86::BI__builtin_ia32_blendpd:
   case clang::X86::BI__builtin_ia32_blendpd256:
   case clang::X86::BI__builtin_ia32_blendps:

clang/lib/AST/ExprConstant.cpp

@@ -12179,6 +12179,37 @@ bool VectorExprEvaluator::VisitCallExpr(const CallExpr *E) {
 
     return Success(APValue(ResultElements.data(), ResultElements.size()), E);
   }
+  case X86::BI__builtin_ia32_vpconflictsi_128:
+  case X86::BI__builtin_ia32_vpconflictsi_256:
+  case X86::BI__builtin_ia32_vpconflictsi_512:
+  case X86::BI__builtin_ia32_vpconflictdi_128:
+  case X86::BI__builtin_ia32_vpconflictdi_256:
+  case X86::BI__builtin_ia32_vpconflictdi_512: {
+    APValue Source;
+
+    if (!EvaluateAsRValue(Info, E->getArg(0), Source))
+      return false;
+
+    unsigned SourceLen = Source.getVectorLength();
+    SmallVector<APValue, 32> ResultElements;
+    ResultElements.reserve(SourceLen);
+
+    const auto *VecT = E->getType()->castAs<VectorType>();
+    bool DestUnsigned =
+        VecT->getElementType()->isUnsignedIntegerOrEnumerationType();
+
+    for (unsigned I = 0; I != SourceLen; ++I) {
+      const APValue &EltI = Source.getVectorElt(I);
+
+      APInt ConflictMask(EltI.getInt().getBitWidth(), 0);
+      for (unsigned J = 0; J != I; ++J) {
+        const APValue &EltJ = Source.getVectorElt(J);
+        ConflictMask.setBitVal(J, EltI.getInt() == EltJ.getInt());
+      }
+      ResultElements.push_back(APValue(APSInt(ConflictMask, DestUnsigned)));
+    }
+    return Success(APValue(ResultElements.data(), ResultElements.size()), E);
+  }
   case X86::BI__builtin_ia32_blendpd:
   case X86::BI__builtin_ia32_blendpd256:
   case X86::BI__builtin_ia32_blendps: