Merge branch 'main' into x86-demandedelts-vpermv-vpermv3

Author: RKSimon

clang-tools-extra/clang-tidy/ClangTidyDiagnosticConsumer.cpp

@@ -311,18 +311,7 @@ ClangTidyDiagnosticConsumer::ClangTidyDiagnosticConsumer(
     : Context(Ctx), ExternalDiagEngine(ExternalDiagEngine),
       RemoveIncompatibleErrors(RemoveIncompatibleErrors),
       GetFixesFromNotes(GetFixesFromNotes),
-      EnableNolintBlocks(EnableNolintBlocks) {
-
-  if (Context.getOptions().HeaderFilterRegex &&
-      !Context.getOptions().HeaderFilterRegex->empty())
-    HeaderFilter =
-        std::make_unique<llvm::Regex>(*Context.getOptions().HeaderFilterRegex);
-
-  if (Context.getOptions().ExcludeHeaderFilterRegex &&
-      !Context.getOptions().ExcludeHeaderFilterRegex->empty())
-    ExcludeHeaderFilter = std::make_unique<llvm::Regex>(
-        *Context.getOptions().ExcludeHeaderFilterRegex);
-}
+      EnableNolintBlocks(EnableNolintBlocks) {}
 
 void ClangTidyDiagnosticConsumer::finalizeLastError() {
   if (!Errors.empty()) {
@@ -571,17 +560,30 @@ void ClangTidyDiagnosticConsumer::checkFilters(SourceLocation Location,
   }
 
   StringRef FileName(File->getName());
-  LastErrorRelatesToUserCode =
-      LastErrorRelatesToUserCode || Sources.isInMainFile(Location) ||
-      (HeaderFilter &&
-       (HeaderFilter->match(FileName) &&
-        !(ExcludeHeaderFilter && ExcludeHeaderFilter->match(FileName))));
+  LastErrorRelatesToUserCode = LastErrorRelatesToUserCode ||
+                               Sources.isInMainFile(Location) ||
+                               (getHeaderFilter()->match(FileName) &&
+                                !getExcludeHeaderFilter()->match(FileName));
 
   unsigned LineNumber = Sources.getExpansionLineNumber(Location);
   LastErrorPassesLineFilter =
       LastErrorPassesLineFilter || passesLineFilter(FileName, LineNumber);
 }
 
+llvm::Regex *ClangTidyDiagnosticConsumer::getHeaderFilter() {
+  if (!HeaderFilter)
+    HeaderFilter =
+        std::make_unique<llvm::Regex>(*Context.getOptions().HeaderFilterRegex);
+  return HeaderFilter.get();
+}
+
+llvm::Regex *ClangTidyDiagnosticConsumer::getExcludeHeaderFilter() {
+  if (!ExcludeHeaderFilter)
+    ExcludeHeaderFilter = std::make_unique<llvm::Regex>(
+        *Context.getOptions().ExcludeHeaderFilterRegex);
+  return ExcludeHeaderFilter.get();
+}
+
 void ClangTidyDiagnosticConsumer::removeIncompatibleErrors() {
   // Each error is modelled as the set of intervals in which it applies
   // replacements. To detect overlapping replacements, we use a sweep line

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

@@ -302,6 +302,10 @@ class ClangTidyDiagnosticConsumer : public DiagnosticConsumer {
   /// context.
   llvm::Regex *getHeaderFilter();
 
+  /// Returns the \c ExcludeHeaderFilter constructed for the options set in the
+  /// context.
+  llvm::Regex *getExcludeHeaderFilter();
+
   /// Updates \c LastErrorRelatesToUserCode and LastErrorPassesLineFilter
   /// according to the diagnostic \p Location.
   void checkFilters(SourceLocation Location, const SourceManager &Sources);

clang-tools-extra/clang-tidy/ClangTidyOptions.cpp

@@ -194,8 +194,8 @@ ClangTidyOptions ClangTidyOptions::getDefaults() {
   Options.WarningsAsErrors = "";
   Options.HeaderFileExtensions = {"", "h", "hh", "hpp", "hxx"};
   Options.ImplementationFileExtensions = {"c", "cc", "cpp", "cxx"};
-  Options.HeaderFilterRegex = std::nullopt;
-  Options.ExcludeHeaderFilterRegex = std::nullopt;
+  Options.HeaderFilterRegex = "";
+  Options.ExcludeHeaderFilterRegex = "";
   Options.SystemHeaders = false;
   Options.FormatStyle = "none";
   Options.User = std::nullopt;

clang-tools-extra/docs/ReleaseNotes.rst

@@ -94,6 +94,9 @@ Improvements to clang-tidy
 - Improved :program:`clang-tidy-diff.py` script. Add the `-warnings-as-errors`
   argument to treat warnings as errors.
 
+- Fixed bug in :program:`clang-tidy` by which `HeaderFilterRegex` did not take
+  effect when passed via the `.clang-tidy` file.
+
 New checks
 ^^^^^^^^^^
 

clang-tools-extra/test/clang-tidy/infrastructure/header-filter-from-config-file/inheritance/.clang-tidy

@@ -0,0 +1 @@
+HeaderFilterRegex: '.*'

clang-tools-extra/test/clang-tidy/infrastructure/header-filter-from-config-file/inheritance/foo.cpp

@@ -0,0 +1,3 @@
+// RUN: clang-tidy -checks=-*,google-explicit-constructor %s 2>&1 | FileCheck %s
+#include "foo.h"
+// CHECK: foo.h:1:12: warning: single-argument constructors must be marked explicit

clang-tools-extra/test/clang-tidy/infrastructure/header-filter-from-config-file/inheritance/foo.h

@@ -0,0 +1 @@
+struct X { X(int); };

clang-tools-extra/test/clang-tidy/infrastructure/header-filter-from-config-file/inheritance/subfolder/.clang-tidy

@@ -0,0 +1,2 @@
+InheritParentConfig: true
+HeaderFilterRegex: 'subfolder/.*'

clang-tools-extra/test/clang-tidy/infrastructure/header-filter-from-config-file/inheritance/subfolder/bar.cpp

@@ -0,0 +1,8 @@
+// shell is required for the "dirname" command
+// REQUIRES: shell
+// RUN: clang-tidy -checks=-*,google-explicit-constructor %s -- -I "$(dirname %S)" 2>&1 | FileCheck %s
+#include "foo.h"
+// CHECK-NOT: foo.h:1:12: warning: single-argument constructors must be marked explicit
+
+#include "bar.h"
+// CHECK: bar.h:1:13: warning: single-argument constructors must be marked explicit

clang-tools-extra/test/clang-tidy/infrastructure/header-filter-from-config-file/inheritance/subfolder/bar.h

@@ -0,0 +1 @@
+struct XX { XX(int); };

clang-tools-extra/test/clang-tidy/infrastructure/header-filter-from-config-file/simple/.clang-tidy

@@ -0,0 +1 @@
+HeaderFilterRegex: '.*'

clang-tools-extra/test/clang-tidy/infrastructure/header-filter-from-config-file/simple/foo.cpp

@@ -0,0 +1,3 @@
+// RUN: clang-tidy -checks=-*,google-explicit-constructor %s 2>&1 | FileCheck %s
+#include "foo.h"
+// CHECK: foo.h:1:12: warning: single-argument constructors must be marked explicit

clang-tools-extra/test/clang-tidy/infrastructure/header-filter-from-config-file/simple/foo.h

@@ -0,0 +1 @@
+struct X { X(int); };

clang/include/clang/AST/ExternalASTSource.h

@@ -191,6 +191,10 @@ class ExternalASTSource : public RefCountedBase<ExternalASTSource> {
 
   virtual ExtKind hasExternalDefinitions(const Decl *D);
 
+  /// True if this function declaration was a definition before in its own
+  /// module.
+  virtual bool wasThisDeclarationADefinition(const FunctionDecl *FD);
+
   /// Finds all declarations lexically contained within the given
   /// DeclContext, after applying an optional filter predicate.
   ///

clang/include/clang/Basic/BuiltinsAMDGPU.def

@@ -163,6 +163,8 @@ BUILTIN(__builtin_amdgcn_raw_buffer_load_b64, "V2UiQbiiIi", "n")
 BUILTIN(__builtin_amdgcn_raw_buffer_load_b96, "V3UiQbiiIi", "n")
 BUILTIN(__builtin_amdgcn_raw_buffer_load_b128, "V4UiQbiiIi", "n")
 
+TARGET_BUILTIN(__builtin_amdgcn_raw_ptr_buffer_load_lds, "vQbv*3IUiiiIiIi", "t", "vmem-to-lds-load-insts")
+
 //===----------------------------------------------------------------------===//
 // Ballot builtins.
 //===----------------------------------------------------------------------===//

clang/include/clang/Basic/DiagnosticSemaKinds.td

@@ -13056,6 +13056,6 @@ def err_acc_decl_for_routine
     : Error<"expected function or lambda declaration for 'routine' construct">;
 
 // AMDGCN builtins diagnostics
-def err_amdgcn_global_load_lds_size_invalid_value : Error<"invalid size value">;
-def note_amdgcn_global_load_lds_size_valid_value : Note<"size must be %select{1, 2, or 4|1, 2, 4, 12 or 16}0">;
+def err_amdgcn_load_lds_size_invalid_value : Error<"invalid size value">;
+def note_amdgcn_load_lds_size_valid_value : Note<"size must be %select{1, 2, or 4|1, 2, 4, 12 or 16}0">;
 } // end of sema component.

clang/include/clang/Sema/MultiplexExternalSemaSource.h

@@ -92,6 +92,8 @@ class MultiplexExternalSemaSource : public ExternalSemaSource {
 
   ExtKind hasExternalDefinitions(const Decl *D) override;
 
+  bool wasThisDeclarationADefinition(const FunctionDecl *FD) override;
+
   /// Find all declarations with the given name in the
   /// given context.
   bool FindExternalVisibleDeclsByName(const DeclContext *DC,

clang/include/clang/Serialization/ASTReader.h

@@ -1392,6 +1392,10 @@ class ASTReader
 
   llvm::DenseMap<const Decl *, bool> DefinitionSource;
 
+  /// Friend functions that were defined but might have had their bodies
+  /// removed.
+  llvm::DenseSet<const FunctionDecl *> ThisDeclarationWasADefinitionSet;
+
   bool shouldDisableValidationForFile(const serialization::ModuleFile &M) const;
 
   /// Reads a statement from the specified cursor.
@@ -2374,6 +2378,8 @@ class ASTReader
 
   ExtKind hasExternalDefinitions(const Decl *D) override;
 
+  bool wasThisDeclarationADefinition(const FunctionDecl *FD) override;
+
   /// Retrieve a selector from the given module with its local ID
   /// number.
   Selector getLocalSelector(ModuleFile &M, unsigned LocalID);

clang/lib/AST/ExternalASTSource.cpp

@@ -38,6 +38,10 @@ ExternalASTSource::hasExternalDefinitions(const Decl *D) {
   return EK_ReplyHazy;
 }
 
+bool ExternalASTSource::wasThisDeclarationADefinition(const FunctionDecl *FD) {
+  return false;
+}
+
 void ExternalASTSource::FindFileRegionDecls(FileID File, unsigned Offset,
                                             unsigned Length,
                                             SmallVectorImpl<Decl *> &Decls) {}

clang/lib/Sema/MultiplexExternalSemaSource.cpp

@@ -107,6 +107,14 @@ MultiplexExternalSemaSource::hasExternalDefinitions(const Decl *D) {
   return EK_ReplyHazy;
 }
 
+bool MultiplexExternalSemaSource::wasThisDeclarationADefinition(
+    const FunctionDecl *FD) {
+  for (const auto &S : Sources)
+    if (S->wasThisDeclarationADefinition(FD))
+      return true;
+  return false;
+}
+
 bool MultiplexExternalSemaSource::FindExternalVisibleDeclsByName(
     const DeclContext *DC, DeclarationName Name,
     const DeclContext *OriginalDC) {

clang/lib/Sema/SemaAMDGPU.cpp

@@ -35,6 +35,7 @@ bool SemaAMDGPU::CheckAMDGCNBuiltinFunctionCall(unsigned BuiltinID,
       Builtin::evaluateRequiredTargetFeatures("gfx950-insts", CallerFeatureMap);
 
   switch (BuiltinID) {
+  case AMDGPU::BI__builtin_amdgcn_raw_ptr_buffer_load_lds:
   case AMDGPU::BI__builtin_amdgcn_global_load_lds: {
     constexpr const int SizeIdx = 2;
     llvm::APSInt Size;
@@ -54,11 +55,9 @@ bool SemaAMDGPU::CheckAMDGCNBuiltinFunctionCall(unsigned BuiltinID,
       [[fallthrough]];
     }
     default:
-      Diag(ArgExpr->getExprLoc(),
-           diag::err_amdgcn_global_load_lds_size_invalid_value)
+      Diag(ArgExpr->getExprLoc(), diag::err_amdgcn_load_lds_size_invalid_value)
           << ArgExpr->getSourceRange();
-      Diag(ArgExpr->getExprLoc(),
-           diag::note_amdgcn_global_load_lds_size_valid_value)
+      Diag(ArgExpr->getExprLoc(), diag::note_amdgcn_load_lds_size_valid_value)
           << HasGFX950Insts << ArgExpr->getSourceRange();
       return true;
     }

clang/lib/Sema/SemaTemplateInstantiateDecl.cpp

@@ -2604,11 +2604,13 @@ Decl *TemplateDeclInstantiator::VisitFunctionDecl(
   // Friend function defined withing class template may stop being function
   // definition during AST merges from different modules, in this case decl
   // with function body should be used for instantiation.
-  if (isFriend) {
-    const FunctionDecl *Defn = nullptr;
-    if (D->hasBody(Defn)) {
-      D = const_cast<FunctionDecl *>(Defn);
-      FunctionTemplate = Defn->getDescribedFunctionTemplate();
+  if (ExternalASTSource *Source = SemaRef.Context.getExternalSource()) {
+    if (isFriend && Source->wasThisDeclarationADefinition(D)) {
+      const FunctionDecl *Defn = nullptr;
+      if (D->hasBody(Defn)) {
+        D = const_cast<FunctionDecl *>(Defn);
+        FunctionTemplate = Defn->getDescribedFunctionTemplate();
+      }
     }
   }
 

clang/lib/Serialization/ASTReader.cpp

@@ -9661,6 +9661,10 @@ ExternalASTSource::ExtKind ASTReader::hasExternalDefinitions(const Decl *FD) {
   return I->second ? EK_Never : EK_Always;
 }
 
+bool ASTReader::wasThisDeclarationADefinition(const FunctionDecl *FD) {
+  return ThisDeclarationWasADefinitionSet.contains(FD);
+}
+
 Selector ASTReader::getLocalSelector(ModuleFile &M, unsigned LocalID) {
   return DecodeSelector(getGlobalSelectorID(M, LocalID));
 }

clang/lib/Serialization/ASTReaderDecl.cpp

@@ -523,6 +523,9 @@ void ASTDeclReader::ReadFunctionDefinition(FunctionDecl *FD) {
   }
   // Store the offset of the body so we can lazily load it later.
   Reader.PendingBodies[FD] = GetCurrentCursorOffset();
+  // For now remember ThisDeclarationWasADefinition only for friend functions.
+  if (FD->getFriendObjectKind())
+    Reader.ThisDeclarationWasADefinitionSet.insert(FD);
 }
 
 void ASTDeclReader::Visit(Decl *D) {

clang/test/CodeGenOpenCL/builtins-amdgcn-raw-buffer-load.cl

@@ -170,3 +170,12 @@ v3u32 test_amdgcn_raw_ptr_buffer_load_b96_non_const_soffset(__amdgpu_buffer_rsrc
 v4u32 test_amdgcn_raw_ptr_buffer_load_b128_non_const_soffset(__amdgpu_buffer_rsrc_t rsrc, int offset, int soffset) {
   return __builtin_amdgcn_raw_buffer_load_b128(rsrc, /*offset=*/0, soffset, /*aux=*/0);
 }
+
+// CHECK-LABEL: @test_amdgcn_raw_ptr_buffer_load_lds(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    tail call void @llvm.amdgcn.raw.ptr.buffer.load.lds(ptr addrspace(8) [[RSRC:%.*]], ptr addrspace(3) [[LDS:%.*]], i32 1, i32 [[OFFSET:%.*]], i32 [[SOFFSET:%.*]], i32 2, i32 3)
+// CHECK-NEXT:    ret void
+//
+void test_amdgcn_raw_ptr_buffer_load_lds(__amdgpu_buffer_rsrc_t rsrc, __local void * lds, int offset, int soffset) {
+    __builtin_amdgcn_raw_ptr_buffer_load_lds(rsrc, lds, 1, offset, soffset, 2, 3);
+}

clang/test/SemaCXX/friend-default-parameters-modules.cpp

@@ -0,0 +1,39 @@
+// RUN: rm -fR %t
+// RUN: split-file %s %t
+// RUN: cd %t
+// RUN: %clang_cc1 -std=c++20 -fmodule-map-file=modules.map -xc++ -emit-module -fmodule-name=foo modules.map -o foo.pcm
+// RUN: %clang_cc1 -std=c++20 -fmodule-map-file=modules.map -O1 -emit-obj main.cc -verify -fmodule-file=foo.pcm
+
+//--- modules.map
+module "foo" {
+  export *
+  module "foo.h" {
+    export *
+    header "foo.h"
+  }
+}
+
+//--- foo.h
+#pragma once
+
+template <int>
+void Create(const void* = nullptr);
+
+template <int>
+struct ObjImpl {
+  template <int>
+  friend void ::Create(const void*);
+};
+
+template <int I>
+void Create(const void*) {
+  (void) ObjImpl<I>{};
+}
+
+//--- main.cc
+// expected-no-diagnostics
+#include "foo.h"
+
+int main() {
+  Create<42>();
+}

clang/test/SemaCXX/friend-default-parameters.cpp

@@ -0,0 +1,21 @@
+// RUN: %clang_cc1 -std=c++20 -verify -emit-llvm-only %s
+
+template <int>
+void Create(const void* = nullptr);
+
+template <int>
+struct ObjImpl {
+  template <int>
+  friend void ::Create(const void*);
+};
+
+template <int I>
+void Create(const void*) {
+  (void) ObjImpl<I>{};
+}
+
+int main() {
+  Create<42>();
+}
+
+// expected-no-diagnostics

clang/test/SemaOpenCL/builtins-amdgcn-raw-ptr-buffer-load-lds-error.cl

@@ -0,0 +1,10 @@
+// RUN: %clang_cc1 -triple amdgcn-unknown-unknown -target-cpu gfx90a -S -verify=gfx90a,expected -o - %s
+// RUN: %clang_cc1 -triple amdgcn-unknown-unknown -target-cpu gfx950 -S -verify=gfx950,expected  -o - %s
+// REQUIRES: amdgpu-registered-target
+
+void test_amdgcn_raw_ptr_buffer_load_lds(__amdgpu_buffer_rsrc_t rsrc, __local void* lds, int offset, int soffset, int x) {
+  __builtin_amdgcn_raw_ptr_buffer_load_lds(rsrc, lds, x, offset, soffset, 0, 0); //expected-error{{argument to '__builtin_amdgcn_raw_ptr_buffer_load_lds' must be a constant integer}}
+  __builtin_amdgcn_raw_ptr_buffer_load_lds(rsrc, lds, 4, offset, soffset, x, 0); //expected-error{{argument to '__builtin_amdgcn_raw_ptr_buffer_load_lds' must be a constant integer}}
+  __builtin_amdgcn_raw_ptr_buffer_load_lds(rsrc, lds, 4, offset, soffset, 0, x); //expected-error{{argument to '__builtin_amdgcn_raw_ptr_buffer_load_lds' must be a constant integer}}
+  __builtin_amdgcn_raw_ptr_buffer_load_lds(rsrc, lds, 3, offset, soffset, 0, 0); //expected-error{{invalid size value}} gfx950-note{{size must be 1, 2, 4, 12 or 16}} gfx90a-note{{size must be 1, 2, or 4}}
+}

clang/test/SemaOpenCL/builtins-amdgcn-raw-ptr-buffer-load-lds-target-error.cl

@@ -0,0 +1,6 @@
+// RUN: %clang_cc1 -triple amdgcn-unknown-unknown -target-cpu gfx1100 -S -verify -o - %s
+// REQUIRES: amdgpu-registered-target
+
+void test_amdgcn_raw_ptr_buffer_load_lds(__amdgpu_buffer_rsrc_t rsrc, __local void* lds, int offset, int soffset, int x) {
+  __builtin_amdgcn_raw_ptr_buffer_load_lds(rsrc, lds, 4, offset, soffset, 0, 0); //expected-error{{needs target feature vmem-to-lds-load-insts}}
+}

llvm/include/llvm/IR/IntrinsicsAMDGPU.td

@@ -1863,7 +1863,9 @@ class AMDGPURawBufferLoadLDS : Intrinsic <
    ImmArg<ArgIndex<6>>, IntrNoCallback, IntrNoFree], "", [SDNPMemOperand]>, AMDGPURsrcIntrinsic<0>;
 def int_amdgcn_raw_buffer_load_lds : AMDGPURawBufferLoadLDS;
 
-class AMDGPURawPtrBufferLoadLDS : Intrinsic <
+class AMDGPURawPtrBufferLoadLDS :
+  ClangBuiltin<"__builtin_amdgcn_raw_ptr_buffer_load_lds">,
+  Intrinsic <
   [],
   [AMDGPUBufferRsrcTy,        // rsrc(SGPR)
    LLVMQualPointerType<3>,    // LDS base offset

llvm/lib/Target/X86/X86ISelLowering.cpp

@@ -6413,9 +6413,7 @@ static bool getFauxShuffleMask(SDValue N, const APInt &DemandedElts,
   case X86ISD::VTRUNC: {
     SDValue Src = N.getOperand(0);
     EVT SrcVT = Src.getValueType();
-    // Truncated source must be a simple vector.
-    if (!SrcVT.isSimple() || (SrcVT.getSizeInBits() % 128) != 0 ||
-        (SrcVT.getScalarSizeInBits() % 8) != 0)
+    if (SrcVT.getSizeInBits() != NumSizeInBits)
       return false;
     unsigned NumSrcElts = SrcVT.getVectorNumElements();
     unsigned NumBitsPerSrcElt = SrcVT.getScalarSizeInBits();

llvm/lib/Transforms/Utils/SimplifyCFG.cpp

@@ -6228,7 +6228,6 @@ static bool initializeUniqueCases(SwitchInst *SI, PHINode *&PHI,
   }
   // Find the default result value.
   SmallVector<std::pair<PHINode *, Constant *>, 1> DefaultResults;
-  BasicBlock *DefaultDest = SI->getDefaultDest();
   getCaseResults(SI, nullptr, SI->getDefaultDest(), &CommonDest, DefaultResults,
                  DL, TTI);
   // If the default value is not found abort unless the default destination