merge main into amd-staging

clang-tools-extra/clang-tidy/ExpandModularHeadersPPCallbacks.h

@@ -137,7 +137,7 @@ class ExpandModularHeadersPPCallbacks : public PPCallbacks {
   std::unique_ptr<Preprocessor> PP;
   bool EnteredMainFile = false;
   bool StartedLexing = false;
-  Token CurrentToken;
+  Token CurrentToken = Token();
 };
 
 } // namespace tooling

clang-tools-extra/clang-tidy/bugprone/FloatLoopCounterCheck.cpp

@@ -31,6 +31,7 @@ void FloatLoopCounterCheck::registerMatchers(MatchFinder *Finder) {
 
 void FloatLoopCounterCheck::check(const MatchFinder::MatchResult &Result) {
   const auto *FS = Result.Nodes.getNodeAs<ForStmt>("for");
+  assert(FS && "FS should not be null");
 
   diag(FS->getInc()->getBeginLoc(), "loop induction expression should not have "
                                     "floating-point type")

clang-tools-extra/clang-tidy/cppcoreguidelines/ProBoundsAvoidUncheckedContainerAccessCheck.cpp

@@ -176,7 +176,7 @@ void ProBoundsAvoidUncheckedContainerAccessCheck::check(
     }
   } else if (const auto *MCE = dyn_cast<CXXMemberCallExpr>(MatchedExpr)) {
     // Case: a.operator[](i) or a->operator[](i)
-    const auto *Callee = dyn_cast<MemberExpr>(MCE->getCallee());
+    const auto *Callee = cast<MemberExpr>(MCE->getCallee());
 
     if (FixMode == At) {
       // Cases: a.operator[](i) => a.at(i) and a->operator[](i) => a->at(i)

clang/include/clang/AST/ASTConsumer.h

@@ -27,6 +27,7 @@ namespace clang {
   class VarDecl;
   class FunctionDecl;
   class ImportDecl;
+  class OpenACCRoutineDecl;
 
 /// ASTConsumer - This is an abstract interface that should be implemented by
 /// clients that read ASTs.  This abstraction layer allows the client to be
@@ -116,6 +117,11 @@ class ASTConsumer {
   // variable has been instantiated.
   virtual void HandleCXXStaticMemberVarInstantiation(VarDecl *D) {}
 
+  /// Callback to handle the end-of-translation unit attachment of OpenACC
+  /// routine declaration information.
+  virtual void HandleOpenACCRoutineReference(const FunctionDecl *FD,
+                                             const OpenACCRoutineDecl *RD) {}
+
   /// Callback involved at the end of a translation unit to
   /// notify the consumer that a vtable for the given C++ class is
   /// required.

clang/include/clang/Basic/arm_mve.td

@@ -783,15 +783,15 @@ let params = T.Unsigned in {
 }
 let params = T.Float in {
   def vminnmq: Intrinsic<Vector, (args Vector:$a, Vector:$b),
-                                 (IRIntBase<"minnum", [Vector]> $a, $b)>;
+                                 (fminnm $a, $b)>;
   def vmaxnmq: Intrinsic<Vector, (args Vector:$a, Vector:$b),
-                                 (IRIntBase<"maxnum", [Vector]> $a, $b)>;
+                                 (fmaxnm $a, $b)>;
   def vminnmaq: Intrinsic<Vector, (args Vector:$a, Vector:$b),
-                                  (IRIntBase<"minnum", [Vector]>
+                                  (fminnm
                                    (IRIntBase<"fabs", [Vector]> $a),
                                    (IRIntBase<"fabs", [Vector]> $b))>;
   def vmaxnmaq: Intrinsic<Vector, (args Vector:$a, Vector:$b),
-                                  (IRIntBase<"maxnum", [Vector]>
+                                  (fmaxnm
                                    (IRIntBase<"fabs", [Vector]> $a),
                                    (IRIntBase<"fabs", [Vector]> $b))>;
 }

clang/include/clang/Basic/arm_mve_defs.td

@@ -589,6 +589,10 @@ def fsub: strictFPAlt<fsub_node,
                       IRInt<"vsub", [Vector]>>;
 def fmul: strictFPAlt<fmul_node,
                       IRInt<"vmul", [Vector]>>;
+def fminnm : strictFPAlt<IRIntBase<"minnum", [Vector]>,
+                         IRInt<"vminnm", [Vector]>>;
+def fmaxnm : strictFPAlt<IRIntBase<"maxnum", [Vector]>,
+                         IRInt<"vmaxnm", [Vector]>>;
 
 // -----------------------------------------------------------------------------
 // Convenience lists of parameter types. 'T' is just a container record, so you

clang/include/clang/CIR/CIRGenerator.h

@@ -81,6 +81,9 @@ class CIRGenerator : public clang::ASTConsumer {
   void HandleTagDeclDefinition(clang::TagDecl *d) override;
   void HandleTagDeclRequiredDefinition(const clang::TagDecl *D) override;
   void HandleCXXStaticMemberVarInstantiation(clang::VarDecl *D) override;
+  void
+  HandleOpenACCRoutineReference(const clang::FunctionDecl *FD,
+                                const clang::OpenACCRoutineDecl *RD) override;
   void CompleteTentativeDefinition(clang::VarDecl *d) override;
   void HandleVTable(clang::CXXRecordDecl *rd) override;
 

clang/include/clang/CIR/Dialect/IR/CIROps.td

@@ -1173,6 +1173,35 @@ def CIR_SwitchOp : CIR_Op<"switch", [
   let hasLLVMLowering = false;
 }
 
+//===----------------------------------------------------------------------===//
+// IsConstantOp
+//===----------------------------------------------------------------------===//
+
+def CIR_IsConstantOp : CIR_Op<"is_constant", [Pure]> {
+  let summary = "Test for manifest compile-time constant";
+  let description = [{
+    Returns `true` if the argument is known to be a manifest compile-time
+    constant otherwise returns `false`. If the argument is a constant expression
+    which refers to a global (the address of which _is_ a constant, but not
+    manifest during the compile), then the intrinsic evaluates to `false`.
+
+    This is used to represent `__builtin_constant_p` in cases where the argument
+    isn't known to be constant during initial translation of the source code but
+    might be proven to be constant after later optimizations.
+
+    Example:
+    ```
+    %1 = cir.is_constant %2 : !s32i -> !cir.bool
+    ```
+  }];
+  let arguments = (ins CIR_AnyType:$val);
+  let results = (outs CIR_BoolType:$result);
+
+  let assemblyFormat = [{
+    $val `:` qualified(type($val)) `->` qualified(type($result)) attr-dict
+  }];
+}
+
 //===----------------------------------------------------------------------===//
 // SwitchFlatOp
 //===----------------------------------------------------------------------===//

clang/include/clang/Options/Options.td

@@ -4870,25 +4870,25 @@ def ggdb3 : Flag<["-"], "ggdb3">, Group<ggdbN_Group>;
 def glldb : Flag<["-"], "glldb">, Group<gTune_Group>;
 def gsce : Flag<["-"], "gsce">, Group<gTune_Group>;
 def gdbx : Flag<["-"], "gdbx">, Group<gTune_Group>;
-// Equivalent to our default dwarf version. Forces usual dwarf emission when
+// Equivalent to our default DWARF version. Forces usual DWARF emission when
 // CodeView is enabled.
 def gdwarf : Flag<["-"], "gdwarf">, Group<g_Group>,
   Visibility<[ClangOption, CLOption, DXCOption, FlangOption]>,
-  HelpText<"Generate source-level debug information with the default dwarf version">;
+  HelpText<"Generate source-level debug information with the default DWARF version">;
 
 let Visibility = [ClangOption, FlangOption] in {
 def gdwarf_2 : Flag<["-"], "gdwarf-2">, Group<g_Group>,
-  HelpText<"Generate source-level debug information with dwarf version 2">;
+  HelpText<"Generate source-level debug information with DWARF version 2">;
 def gdwarf_3 : Flag<["-"], "gdwarf-3">, Group<g_Group>,
-  HelpText<"Generate source-level debug information with dwarf version 3">;
+  HelpText<"Generate source-level debug information with DWARF version 3">;
 def gdwarf_4 : Flag<["-"], "gdwarf-4">, Group<g_Group>,
-  HelpText<"Generate source-level debug information with dwarf version 4">;
+  HelpText<"Generate source-level debug information with DWARF version 4">;
 def gdwarf_5 : Flag<["-"], "gdwarf-5">, Group<g_Group>,
-  HelpText<"Generate source-level debug information with dwarf version 5">;
+  HelpText<"Generate source-level debug information with DWARF version 5">;
 def gdwarf_6
     : Flag<["-"], "gdwarf-6">,
       Group<g_Group>,
-      HelpText<"Generate source-level debug information with dwarf version 6">;
+      HelpText<"Generate source-level debug information with DWARF version 6">;
 }
 def gdwarf64 : Flag<["-"], "gdwarf64">, Group<g_Group>,
   Visibility<[ClangOption, CC1Option, CC1AsOption]>,
@@ -4915,25 +4915,28 @@ def gno_heterogeneous_dwarf : Flag<["-"], "gno-heterogeneous-dwarf">,
   HelpText<"Disable DWARF extensions for heterogeneous debugging">,
   Alias<gheterogeneous_dwarf_EQ>, AliasArgs<["disabled"]>;
 
-def gcodeview : Flag<["-"], "gcodeview">,
+def gcodeview : Flag<["-"], "gcodeview">, Group<g_Group>,
   HelpText<"Generate CodeView debug information">,
   Visibility<[ClangOption, CC1Option, CC1AsOption, CLOption, DXCOption]>,
   MarshallingInfoFlag<CodeGenOpts<"EmitCodeView">>;
 defm codeview_ghash : BoolOption<"g", "codeview-ghash",
   CodeGenOpts<"CodeViewGHash">, DefaultFalse,
   PosFlag<SetTrue, [], [ClangOption, CC1Option],
           "Emit type record hashes in a .debug$H section">,
-  NegFlag<SetFalse>, BothFlags<[], [ClangOption, CLOption, DXCOption]>>;
+  NegFlag<SetFalse>, BothFlags<[], [ClangOption, CLOption, DXCOption]>>,
+  Group<g_flags_Group>;
 defm codeview_command_line : BoolOption<"g", "codeview-command-line",
   CodeGenOpts<"CodeViewCommandLine">, DefaultTrue,
   PosFlag<SetTrue, [], [ClangOption], "Emit compiler path and command line into CodeView debug information">,
   NegFlag<SetFalse, [], [ClangOption], "Don't emit compiler path and command line into CodeView debug information">,
-  BothFlags<[], [ClangOption, CLOption, DXCOption, CC1Option]>>;
+  BothFlags<[], [ClangOption, CLOption, DXCOption, CC1Option]>>,
+  Group<g_flags_Group>;
 defm inline_line_tables : BoolGOption<"inline-line-tables",
   CodeGenOpts<"NoInlineLineTables">, DefaultFalse,
   NegFlag<SetTrue, [], [ClangOption, CC1Option],
           "Don't emit inline line tables.">,
-  PosFlag<SetFalse>, BothFlags<[], [ClangOption, CLOption, DXCOption]>>;
+  PosFlag<SetFalse>, BothFlags<[], [ClangOption, CLOption, DXCOption]>>,
+  Group<g_flags_Group>;
 
 def gfull : Flag<["-"], "gfull">, Group<g_Group>;
 def gused : Flag<["-"], "gused">, Group<g_Group>;
@@ -4958,7 +4961,8 @@ defm strict_dwarf : BoolOption<"g", "strict-dwarf",
 defm omit_unreferenced_methods : BoolGOption<"omit-unreferenced-methods",
   CodeGenOpts<"DebugOmitUnreferencedMethods">, DefaultFalse,
   NegFlag<SetFalse>,
-  PosFlag<SetTrue, [], [CC1Option]>, BothFlags<[], [ClangOption, CLOption, DXCOption]>>;
+  PosFlag<SetTrue, [], [CC1Option]>, BothFlags<[], [ClangOption, CLOption, DXCOption]>>,
+  Group<g_flags_Group>;
 defm column_info : BoolOption<"g", "column-info",
   CodeGenOpts<"DebugColumnInfo">, DefaultTrue,
   NegFlag<SetFalse, [], [ClangOption, CC1Option]>,
@@ -5027,6 +5031,7 @@ defm structor_decl_linkage_names
                           "Attach linkage names to C++ constructor/destructor "
                           "declarations in DWARF.">,
                   BothFlags<[], [ClangOption, CLOption, CC1Option]>>,
+                  Group<g_flags_Group>,
                   DocBrief<[{On some ABIs (e.g., Itanium), constructors and destructors may have multiple variants. Historically, when generating DWARF, Clang did not attach ``DW_AT_linkage_name`` to structor DIEs because there were multiple possible manglings (depending on the structor variant) that could be used. With ``-gstructor-decl-linkage-names``, for ABIs with structor variants, we attach a "unified" mangled name to structor declarations DIEs which debuggers can use to look up all the definitions for a structor declaration. E.g., a "unified" mangled name ``_ZN3FooC4Ev`` may have multiple definitions associated with it such as ``_ZN3FooC1Ev`` and ``_ZN3FooC2Ev``.
 
 Enabling this flag results in a better interactive debugging experience (both GDB and LLDB have support for understanding these "unified" linkage names). However, it comes with a significant increase in debug-info size (particularly the `.debug_str` section). As an escape hatch, users can disable this feature using ``-gno-structor-decl-linkage-names``.}]>;
@@ -5035,7 +5040,8 @@ defm key_instructions : BoolGOption<"key-instructions",
     NegFlag<SetFalse>, PosFlag<SetTrue, [], [],
         "Enable Key Instructions, which reduces the jumpiness of debug stepping in optimized C/C++ code"
         " in some debuggers. DWARF only.">,
-    BothFlags<[], [ClangOption, CLOption, CC1Option]>>;
+    BothFlags<[], [ClangOption, CLOption, CC1Option]>>,
+  Group<g_flags_Group>;
 def headerpad__max__install__names : Joined<["-"], "headerpad_max_install_names">;
 def help : Flag<["-", "--"], "help">,
     Visibility<[ClangOption, CC1Option, CC1AsOption,
@@ -8690,7 +8696,7 @@ def main_file_name : Separate<["-"], "main-file-name">,
   Visibility<[CC1Option, CC1AsOption]>,
   MarshallingInfoString<CodeGenOpts<"MainFileName">>;
 def split_dwarf_output : Separate<["-"], "split-dwarf-output">,
-  HelpText<"File name to use for split dwarf debug info output">,
+  HelpText<"File name to use for split DWARF debug info output">,
   Visibility<[CC1Option, CC1AsOption, FC1Option]>,
   MarshallingInfoString<CodeGenOpts<"SplitDwarfOutput">>;
 
@@ -8724,7 +8730,7 @@ def dependent_lib : Joined<["--"], "dependent-lib=">,
   MarshallingInfoStringVector<CodeGenOpts<"DependentLibraries">>;
 
 def split_dwarf_file : Separate<["-"], "split-dwarf-file">,
-  HelpText<"Name of the split dwarf debug info file to encode in the object file">,
+  HelpText<"Name of the split DWARF debug info file to encode in the object file">,
   MarshallingInfoString<CodeGenOpts<"SplitDwarfFile">>;
 
 } // let Visibility = [CC1Option, FC1Option]

clang/include/clang/Sema/SemaOpenACC.h

@@ -37,8 +37,16 @@ class Scope;
 class SemaOpenACC : public SemaBase {
 public:
   using DeclGroupPtrTy = OpaquePtr<DeclGroupRef>;
+  using RoutineRefListTy = std::pair<FunctionDecl *, OpenACCRoutineDecl *>;
 
 private:
+  // We save a list of routine clauses that refer to a different function(that
+  // is, routine-with-a-name) so that we can do the emission at the 'end'.  We
+  // have to do this, since functions can be emitted before they are referenced,
+  // and the OpenACCRoutineDecl isn't necessarily emitted, as it might be in a
+  // function/etc. So we do these emits at the end of the TU.
+  llvm::SmallVector<RoutineRefListTy> RoutineRefList;
+
   struct ComputeConstructInfo {
     /// Which type of compute construct we are inside of, which we can use to
     /// determine whether we should add loops to the above collection.  We can
@@ -752,6 +760,7 @@ class SemaOpenACC : public SemaBase {
   };
 
   SemaOpenACC(Sema &S);
+  void ActOnEndOfTranslationUnit(TranslationUnitDecl *TU);
 
   // Called when we encounter a 'while' statement, before looking at its 'body'.
   void ActOnWhileStmt(SourceLocation WhileLoc);

clang/lib/Basic/Targets/AMDGPU.h

@@ -316,8 +316,10 @@ class LLVM_LIBRARY_VISIBILITY AMDGPUTargetInfo final : public TargetInfo {
       Opts["cl_amd_media_ops"] = true;
       Opts["cl_amd_media_ops2"] = true;
 
+      // FIXME: Check subtarget for image support.
       Opts["__opencl_c_images"] = true;
       Opts["__opencl_c_3d_image_writes"] = true;
+      Opts["__opencl_c_read_write_images"] = true;
       Opts["cl_khr_3d_image_writes"] = true;
       Opts["__opencl_c_program_scope_global_variables"] = true;
       Opts["__opencl_c_atomic_order_seq_cst"] = true;

clang/lib/CIR/CodeGen/CIRGenBuiltin.cpp

@@ -542,6 +542,45 @@ RValue CIRGenFunction::emitBuiltinExpr(const GlobalDecl &gd, unsigned builtinID,
     return emitCall(e->getCallee()->getType(), CIRGenCallee::forDirect(fnOp), e,
                     returnValue);
   }
+
+  case Builtin::BI__builtin_constant_p: {
+    mlir::Type resultType = convertType(e->getType());
+
+    const Expr *arg = e->getArg(0);
+    QualType argType = arg->getType();
+    // FIXME: The allowance for Obj-C pointers and block pointers is historical
+    // and likely a mistake.
+    if (!argType->isIntegralOrEnumerationType() && !argType->isFloatingType() &&
+        !argType->isObjCObjectPointerType() && !argType->isBlockPointerType()) {
+      // Per the GCC documentation, only numeric constants are recognized after
+      // inlining.
+      return RValue::get(
+          builder.getConstInt(getLoc(e->getSourceRange()),
+                              mlir::cast<cir::IntType>(resultType), 0));
+    }
+
+    if (arg->HasSideEffects(getContext())) {
+      // The argument is unevaluated, so be conservative if it might have
+      // side-effects.
+      return RValue::get(
+          builder.getConstInt(getLoc(e->getSourceRange()),
+                              mlir::cast<cir::IntType>(resultType), 0));
+    }
+
+    mlir::Value argValue = emitScalarExpr(arg);
+    if (argType->isObjCObjectPointerType()) {
+      cgm.errorNYI(e->getSourceRange(),
+                   "__builtin_constant_p: Obj-C object pointer");
+      return {};
+    }
+    argValue = builder.createBitcast(argValue, convertType(argType));
+
+    mlir::Value result = cir::IsConstantOp::create(
+        builder, getLoc(e->getSourceRange()), argValue);
+    // IsConstantOp returns a bool, but __builtin_constant_p returns an int.
+    result = builder.createBoolToInt(result, resultType);
+    return RValue::get(result);
+  }
   case Builtin::BI__builtin_dynamic_object_size:
   case Builtin::BI__builtin_object_size: {
     unsigned type =

clang/lib/CIR/CodeGen/CIRGenBuiltinX86.cpp

@@ -100,6 +100,44 @@ static mlir::Value emitX86MaskAddLogic(CIRGenBuilderTy &builder,
   return builder.createBitcast(resVec, ops[0].getType());
 }
 
+static mlir::Value emitX86MaskUnpack(CIRGenBuilderTy &builder,
+                                     mlir::Location loc,
+                                     const std::string &intrinsicName,
+                                     SmallVectorImpl<mlir::Value> &ops) {
+  unsigned numElems = cast<cir::IntType>(ops[0].getType()).getWidth();
+
+  // Convert both operands to mask vectors.
+  mlir::Value lhs = getMaskVecValue(builder, loc, ops[0], numElems);
+  mlir::Value rhs = getMaskVecValue(builder, loc, ops[1], numElems);
+
+  mlir::Type i32Ty = builder.getSInt32Ty();
+
+  // Create indices for extracting the first half of each vector.
+  SmallVector<mlir::Attribute, 32> halfIndices;
+  for (auto i : llvm::seq<unsigned>(0, numElems / 2))
+    halfIndices.push_back(cir::IntAttr::get(i32Ty, i));
+
+  // Extract first half of each vector. This gives better codegen than
+  // doing it in a single shuffle.
+  mlir::Value lhsHalf = builder.createVecShuffle(loc, lhs, lhs, halfIndices);
+  mlir::Value rhsHalf = builder.createVecShuffle(loc, rhs, rhs, halfIndices);
+
+  // Create indices for concatenating the vectors.
+  // NOTE: Operands are swapped to match the intrinsic definition.
+  // After the half extraction, both vectors have numElems/2 elements.
+  // In createVecShuffle(rhsHalf, lhsHalf, indices), indices [0..numElems/2-1]
+  // select from rhsHalf, and indices [numElems/2..numElems-1] select from
+  // lhsHalf.
+  SmallVector<mlir::Attribute, 64> concatIndices;
+  for (auto i : llvm::seq<unsigned>(0, numElems))
+    concatIndices.push_back(cir::IntAttr::get(i32Ty, i));
+
+  // Concat the vectors (RHS first, then LHS).
+  mlir::Value res =
+      builder.createVecShuffle(loc, rhsHalf, lhsHalf, concatIndices);
+  return builder.createBitcast(res, ops[0].getType());
+}
+
 static mlir::Value emitX86MaskLogic(CIRGenBuilderTy &builder,
                                     mlir::Location loc,
                                     cir::BinOpKind binOpKind,
@@ -257,7 +295,15 @@ mlir::Value CIRGenFunction::emitX86BuiltinExpr(unsigned builtinID,
     return emitVecInsert(builder, getLoc(expr->getExprLoc()), ops[0], ops[1],
                          ops[2]);
   }
-
+  case X86::BI__builtin_ia32_kunpckhi:
+    return emitX86MaskUnpack(builder, getLoc(expr->getExprLoc()),
+                             "x86.avx512.kunpackb", ops);
+  case X86::BI__builtin_ia32_kunpcksi:
+    return emitX86MaskUnpack(builder, getLoc(expr->getExprLoc()),
+                             "x86.avx512.kunpackw", ops);
+  case X86::BI__builtin_ia32_kunpckdi:
+    return emitX86MaskUnpack(builder, getLoc(expr->getExprLoc()),
+                             "x86.avx512.kunpackd", ops);
   case X86::BI_mm_setcsr:
   case X86::BI__builtin_ia32_ldmxcsr: {
     mlir::Location loc = getLoc(expr->getExprLoc());
@@ -947,9 +993,6 @@ mlir::Value CIRGenFunction::emitX86BuiltinExpr(unsigned builtinID,
         getMaskVecValue(builder, getLoc(expr->getExprLoc()), ops[0], numElts);
     return builder.createBitcast(resVec, ops[0].getType());
   }
-  case X86::BI__builtin_ia32_kunpckdi:
-  case X86::BI__builtin_ia32_kunpcksi:
-  case X86::BI__builtin_ia32_kunpckhi:
   case X86::BI__builtin_ia32_sqrtsh_round_mask:
   case X86::BI__builtin_ia32_sqrtsd_round_mask:
   case X86::BI__builtin_ia32_sqrtss_round_mask: