Merged master:7b774b7cd48 into amd-gfx:9333d68d986

Local branch amd-gfx 9333d68 Merged master:170ee645f4d into amd-gfx:817733143d1 Remote branch master 7b774b7 [OPENMP50]Add if clause in teams distribute simd directive.
jaebaek · Dec 11, 2019 · 296f1a2 · 296f1a2
2 parents 9333d68 + 7b774b7
commit 296f1a2
Show file tree

Hide file tree

Showing 169 changed files with 4,710 additions and 4,119 deletions.
diff --git a/clang/include/clang/Basic/DiagnosticParseKinds.td b/clang/include/clang/Basic/DiagnosticParseKinds.td
@@ -63,7 +63,7 @@ def warn_null_statement : Warning<
 
 def warn_misleading_indentation : Warning<
   "misleading indentation; statement is not part of "
-  "the previous '%select{if|else|for|while|else if}0'">,
+  "the previous '%select{if|else|for|while}0'">,
   InGroup<MisleadingIndentation>, DefaultIgnore;
 def note_previous_statement : Note<
   "previous statement is here">;

diff --git a/clang/include/clang/Basic/OpenMPKinds.def b/clang/include/clang/Basic/OpenMPKinds.def
@@ -878,6 +878,7 @@ OPENMP_TEAMS_DISTRIBUTE_SIMD_CLAUSE(aligned)
 OPENMP_TEAMS_DISTRIBUTE_SIMD_CLAUSE(safelen)
 OPENMP_TEAMS_DISTRIBUTE_SIMD_CLAUSE(simdlen)
 OPENMP_TEAMS_DISTRIBUTE_SIMD_CLAUSE(allocate)
+OPENMP_TEAMS_DISTRIBUTE_SIMD_CLAUSE(if)
 
 // Clauses allowed for OpenMP directive 'teams distribute parallel for simd'
 OPENMP_TEAMS_DISTRIBUTE_PARALLEL_FOR_SIMD_CLAUSE(firstprivate)

diff --git a/clang/include/clang/Basic/arm_mve.td b/clang/include/clang/Basic/arm_mve.td
@@ -609,30 +609,21 @@ defm vstrhq: scatter_offset_both<!listconcat(T.All16, T.Int32), u16, 1>;
 defm vstrwq: scatter_offset_both<T.All32, u32, 2>;
 defm vstrdq: scatter_offset_both<T.Int64, u64, 3>;
 
-multiclass PredicatedImmediateVectorShift<
-    Immediate immtype, string predIntrName, list<dag> unsignedFlag = []> {
-  foreach predIntr = [IRInt<predIntrName, [Vector, Predicate]>] in {
-    def _m_n: Intrinsic<Vector, (args Vector:$inactive, Vector:$v,
-                                      immtype:$sh, Predicate:$pred),
-       !con((predIntr $v, $sh), !dag(predIntr, unsignedFlag, ?),
-            (predIntr $pred, $inactive))>;
-    def _x_n: Intrinsic<Vector, (args Vector:$v, immtype:$sh,
-                                      Predicate:$pred),
-       !con((predIntr $v, $sh), !dag(predIntr, unsignedFlag, ?),
-            (predIntr $pred, (undef Vector)))>;
-  }
-}
-
 let params = T.Int in {
   def vshlq_n: Intrinsic<Vector, (args Vector:$v, imm_0toNm1:$sh),
                          (shl $v, (splat (Scalar $sh)))>;
-  defm vshlq: PredicatedImmediateVectorShift<imm_0toNm1, "shl_imm_predicated">;
+  defm vshlq: IntrinsicMX<Vector, (args Vector:$v, imm_0toNm1:$sh,
+                                        Predicate:$pred),
+      (IRInt<"shl_imm_predicated", [Vector, Predicate]>
+           $v, $sh, $pred, $inactive), "_n">;
 
   let pnt = PNT_NType in {
     def vshrq_n: Intrinsic<Vector, (args Vector:$v, imm_1toN:$sh),
                             (immshr $v, $sh, (unsignedflag Scalar))>;
-    defm vshrq: PredicatedImmediateVectorShift<imm_1toN, "shr_imm_predicated",
-                                     [(unsignedflag Scalar)]>;
+    defm vshrq: IntrinsicMX<Vector, (args Vector:$v, imm_1toN:$sh,
+                                          Predicate:$pred),
+        (IRInt<"shr_imm_predicated", [Vector, Predicate]>
+             $v, $sh, (unsignedflag Scalar), $pred, $inactive), "_n">;
   }
 }
 
@@ -713,25 +704,17 @@ def vadciq_m: Intrinsic<Vector, (args Vector:$inactive, Vector:$a, Vector:$b,
 multiclass VectorComplexAddPred<dag not_halving, dag angle> {
   def "" : Intrinsic<Vector, (args Vector:$a, Vector:$b),
      (IRInt<"vcaddq", [Vector]> not_halving, angle, $a, $b)>;
-  def _m : Intrinsic<Vector, (args Vector:$inactive, Vector:$a, Vector:$b,
-                              Predicate:$pred),
+  defm "" : IntrinsicMX<Vector, (args Vector:$a, Vector:$b, Predicate:$pred),
      (IRInt<"vcaddq_predicated", [Vector, Predicate]>
        not_halving, angle, $inactive, $a, $b, $pred)>;
-  def _x : Intrinsic<Vector, (args Vector:$a, Vector:$b, Predicate:$pred),
-     (IRInt<"vcaddq_predicated", [Vector, Predicate]>
-       not_halving, angle, (undef Vector), $a, $b, $pred)>;
 }
 
 multiclass VectorComplexMulPred<dag angle> {
   def "" : Intrinsic<Vector, (args Vector:$a, Vector:$b),
     (IRInt<"vcmulq", [Vector]> angle, $a, $b)>;
-  def _m : Intrinsic<Vector, (args Vector:$inactive, Vector:$a, Vector:$b,
-                              Predicate:$pred),
+  defm "" : IntrinsicMX<Vector, (args Vector:$a, Vector:$b, Predicate:$pred),
     (IRInt<"vcmulq_predicated", [Vector, Predicate]> angle, $inactive, $a, $b,
       $pred)>;
-  def _x : Intrinsic<Vector, (args Vector:$a, Vector:$b, Predicate:$pred),
-    (IRInt<"vcmulq_predicated", [Vector, Predicate]> angle, (undef Vector), $a,
-      $b, $pred)>;
 }
 
 multiclass VectorComplexMLAPred<dag angle> {

diff --git a/clang/include/clang/Basic/arm_mve_defs.td b/clang/include/clang/Basic/arm_mve_defs.td
@@ -432,6 +432,30 @@ class NameOverride<string basename_> {
   string basename = basename_;
 }
 
+// A wrapper to define both _m and _x versions of a predicated
+// intrinsic.
+multiclass IntrinsicMX<Type rettype, dag arguments, dag cg,
+                       string nameSuffix = "",
+                       PolymorphicNameType pnt_x = PNT_Type> {
+  // The _m variant takes an initial parameter called $inactive, which
+  // provides the input value of the output register, i.e. all the
+  // inactive lanes in the predicated operation take their values from
+  // this.
+  def "_m" # nameSuffix:
+     Intrinsic<rettype, !con((args rettype:$inactive), arguments), cg>;
+
+  // The _x variant leaves off that parameter, and simply uses an
+  // undef value of the same type.
+  def "_x" # nameSuffix:
+     Intrinsic<rettype, arguments, (seq (undef rettype):$inactive, cg)> {
+    // Allow overriding of the polymorphic name type, because
+    // sometimes the _m and _x variants polymorph differently
+    // (typically because the type of the inactive parameter can be
+    // used as a disambiguator if it's present).
+    let pnt = pnt_x;
+  }
+}
+
 // -----------------------------------------------------------------------------
 // Convenience lists of parameter types. 'T' is just a container record, so you
 // can define a typical intrinsic with 'let Params = T.Usual', or similar,

diff --git a/clang/lib/Basic/OpenMPKinds.cpp b/clang/lib/Basic/OpenMPKinds.cpp
@@ -808,6 +808,8 @@ bool clang::isAllowedClauseForDirective(OpenMPDirectiveKind DKind,
     }
     break;
   case OMPD_teams_distribute_simd:
+    if (OpenMPVersion < 50 && CKind == OMPC_if)
+      return false;
     switch (CKind) {
 #define OPENMP_TEAMS_DISTRIBUTE_SIMD_CLAUSE(Name)                              \
   case OMPC_##Name:                                                            \

diff --git a/clang/lib/CodeGen/BackendUtil.cpp b/clang/lib/CodeGen/BackendUtil.cpp
@@ -896,7 +896,7 @@ void EmitAssemblyHelper::EmitAssembly(BackendAction Action,
 
   {
     PrettyStackTraceString CrashInfo("Per-function optimization");
-    llvm::TimeTraceScope TimeScope("PerFunctionPasses", StringRef(""));
+    llvm::TimeTraceScope TimeScope("PerFunctionPasses");
 
     PerFunctionPasses.doInitialization();
     for (Function &F : *TheModule)
@@ -907,13 +907,13 @@ void EmitAssemblyHelper::EmitAssembly(BackendAction Action,
 
   {
     PrettyStackTraceString CrashInfo("Per-module optimization passes");
-    llvm::TimeTraceScope TimeScope("PerModulePasses", StringRef(""));
+    llvm::TimeTraceScope TimeScope("PerModulePasses");
     PerModulePasses.run(*TheModule);
   }
 
   {
     PrettyStackTraceString CrashInfo("Code generation");
-    llvm::TimeTraceScope TimeScope("CodeGenPasses", StringRef(""));
+    llvm::TimeTraceScope TimeScope("CodeGenPasses");
     CodeGenPasses.run(*TheModule);
   }
 
@@ -1499,7 +1499,7 @@ void clang::EmitBackendOutput(DiagnosticsEngine &Diags,
                               BackendAction Action,
                               std::unique_ptr<raw_pwrite_stream> OS) {
 
-  llvm::TimeTraceScope TimeScope("Backend", StringRef(""));
+  llvm::TimeTraceScope TimeScope("Backend");
 
   std::unique_ptr<llvm::Module> EmptyModule;
   if (!CGOpts.ThinLTOIndexFile.empty()) {

diff --git a/clang/lib/CodeGen/CodeGenAction.cpp b/clang/lib/CodeGen/CodeGenAction.cpp
@@ -249,7 +249,7 @@ namespace clang {
 
     void HandleTranslationUnit(ASTContext &C) override {
       {
-        llvm::TimeTraceScope TimeScope("Frontend", StringRef(""));
+        llvm::TimeTraceScope TimeScope("Frontend");
         PrettyStackTraceString CrashInfo("Per-file LLVM IR generation");
         if (FrontendTimesIsEnabled) {
           LLVMIRGenerationRefCount += 1;

diff --git a/clang/lib/CodeGen/CodeGenModule.cpp b/clang/lib/CodeGen/CodeGenModule.cpp
@@ -2212,9 +2212,15 @@ llvm::Constant *CodeGenModule::EmitAnnotateAttr(llvm::GlobalValue *GV,
                  *UnitGV = EmitAnnotationUnit(L),
                  *LineNoCst = EmitAnnotationLineNo(L);
 
+  llvm::Constant *ASZeroGV = GV;
+  if (GV->getAddressSpace() != 0) {
+    ASZeroGV = llvm::ConstantExpr::getAddrSpaceCast(
+                   GV, GV->getValueType()->getPointerTo(0));
+  }
+
   // Create the ConstantStruct for the global annotation.
   llvm::Constant *Fields[4] = {
-    llvm::ConstantExpr::getBitCast(GV, Int8PtrTy),
+    llvm::ConstantExpr::getBitCast(ASZeroGV, Int8PtrTy),
     llvm::ConstantExpr::getBitCast(AnnoGV, Int8PtrTy),
     llvm::ConstantExpr::getBitCast(UnitGV, Int8PtrTy),
     LineNoCst

diff --git a/clang/lib/Frontend/CompilerInstance.cpp b/clang/lib/Frontend/CompilerInstance.cpp
@@ -917,8 +917,9 @@ bool CompilerInstance::ExecuteAction(FrontendAction &Act) {
   if (!hasTarget())
     return false;
 
-  // Create TargetInfo for the other side of CUDA and OpenMP compilation.
-  if ((getLangOpts().CUDA || getLangOpts().OpenMPIsDevice) &&
+  // Create TargetInfo for the other side of CUDA/OpenMP/SYCL compilation.
+  if ((getLangOpts().CUDA || getLangOpts().OpenMPIsDevice ||
+       getLangOpts().SYCLIsDevice) &&
       !getFrontendOpts().AuxTriple.empty()) {
     auto TO = std::make_shared<TargetOptions>();
     TO->Triple = llvm::Triple::normalize(getFrontendOpts().AuxTriple);

diff --git a/clang/lib/Frontend/InitPreprocessor.cpp b/clang/lib/Frontend/InitPreprocessor.cpp
@@ -1124,7 +1124,8 @@ void clang::InitializePreprocessor(
   if (InitOpts.UsePredefines) {
     // FIXME: This will create multiple definitions for most of the predefined
     // macros. This is not the right way to handle this.
-    if ((LangOpts.CUDA || LangOpts.OpenMPIsDevice) && PP.getAuxTargetInfo())
+    if ((LangOpts.CUDA || LangOpts.OpenMPIsDevice || LangOpts.SYCLIsDevice) &&
+        PP.getAuxTargetInfo())
       InitializePredefinedMacros(*PP.getAuxTargetInfo(), LangOpts, FEOpts,
                                  PP.getPreprocessorOpts(), Builder);
 

diff --git a/clang/lib/Parse/ParseAST.cpp b/clang/lib/Parse/ParseAST.cpp
@@ -151,7 +151,7 @@ void clang::ParseAST(Sema &S, bool PrintStats, bool SkipFunctionBodies) {
   bool HaveLexer = S.getPreprocessor().getCurrentLexer();
 
   if (HaveLexer) {
-    llvm::TimeTraceScope TimeScope("Frontend", StringRef(""));
+    llvm::TimeTraceScope TimeScope("Frontend");
     P.Initialize();
     Parser::DeclGroupPtrTy ADecl;
     for (bool AtEOF = P.ParseFirstTopLevelDecl(ADecl); !AtEOF;

diff --git a/clang/lib/Sema/Sema.cpp b/clang/lib/Sema/Sema.cpp
@@ -924,8 +924,7 @@ void Sema::ActOnEndOfTranslationUnitFragment(TUFragmentKind Kind) {
   }
 
   {
-    llvm::TimeTraceScope TimeScope("PerformPendingInstantiations",
-                                   StringRef(""));
+    llvm::TimeTraceScope TimeScope("PerformPendingInstantiations");
     PerformPendingInstantiations();
   }
 

diff --git a/clang/lib/Sema/SemaOpenMP.cpp b/clang/lib/Sema/SemaOpenMP.cpp
@@ -4741,6 +4741,8 @@ StmtResult Sema::ActOnOpenMPExecutableDirective(
   case OMPD_teams_distribute_simd:
     Res = ActOnOpenMPTeamsDistributeSimdDirective(
         ClausesWithImplicit, AStmt, StartLoc, EndLoc, VarsWithInheritedDSA);
+    if (LangOpts.OpenMP >= 50)
+      AllowedNameModifiers.push_back(OMPD_simd);
     break;
   case OMPD_teams_distribute_parallel_for_simd:
     Res = ActOnOpenMPTeamsDistributeParallelForSimdDirective(
@@ -10768,6 +10770,11 @@ static OpenMPDirectiveKind getOpenMPCaptureRegionForClause(
           (NameModifier == OMPD_unknown || NameModifier == OMPD_simd))
         CaptureRegion = OMPD_target;
       break;
+    case OMPD_teams_distribute_simd:
+      if (OpenMPVersion >= 50 &&
+          (NameModifier == OMPD_unknown || NameModifier == OMPD_simd))
+        CaptureRegion = OMPD_teams;
+      break;
     case OMPD_cancel:
     case OMPD_parallel:
     case OMPD_parallel_master:
@@ -10812,7 +10819,6 @@ static OpenMPDirectiveKind getOpenMPCaptureRegionForClause(
     case OMPD_ordered:
     case OMPD_atomic:
     case OMPD_teams_distribute:
-    case OMPD_teams_distribute_simd:
     case OMPD_requires:
       llvm_unreachable("Unexpected OpenMP directive with if-clause");
     case OMPD_unknown:

diff --git a/clang/lib/Serialization/GlobalModuleIndex.cpp b/clang/lib/Serialization/GlobalModuleIndex.cpp
@@ -134,7 +134,7 @@ GlobalModuleIndex::GlobalModuleIndex(std::unique_ptr<llvm::MemoryBuffer> Buffer,
                        "' failed: " + toString(std::move(Err)));
   };
 
-  llvm::TimeTraceScope TimeScope("Module LoadIndex", StringRef(""));
+  llvm::TimeTraceScope TimeScope("Module LoadIndex");
   // Read the global index.
   bool InGlobalIndexBlock = false;
   bool Done = false;
@@ -770,7 +770,7 @@ bool GlobalModuleIndexBuilder::writeIndex(llvm::BitstreamWriter &Stream) {
   }
 
   using namespace llvm;
-  llvm::TimeTraceScope TimeScope("Module WriteIndex", StringRef(""));
+  llvm::TimeTraceScope TimeScope("Module WriteIndex");
 
   // Emit the file header.
   Stream.Emit((unsigned)'B', 8);

diff --git a/clang/lib/StaticAnalyzer/Checkers/CMakeLists.txt b/clang/lib/StaticAnalyzer/Checkers/CMakeLists.txt
@@ -28,6 +28,7 @@ add_clang_library(clangStaticAnalyzerCheckers
   CXXSelfAssignmentChecker.cpp
   DeadStoresChecker.cpp
   DebugCheckers.cpp
+  DebugIteratorModeling.cpp
   DeleteWithNonVirtualDtorChecker.cpp
   DereferenceChecker.cpp
   DirectIvarAssignment.cpp
@@ -42,7 +43,10 @@ add_clang_library(clangStaticAnalyzerCheckers
   GTestChecker.cpp
   IdenticalExprChecker.cpp
   InnerPointerChecker.cpp
-  IteratorChecker.cpp
+  InvalidatedIteratorChecker.cpp
+  Iterator.cpp
+  IteratorModeling.cpp
+  IteratorRangeChecker.cpp
   IvarInvalidationChecker.cpp
   LLVMConventionsChecker.cpp
   LocalizationChecker.cpp
@@ -51,6 +55,7 @@ add_clang_library(clangStaticAnalyzerCheckers
   MallocChecker.cpp
   MallocOverflowSecurityChecker.cpp
   MallocSizeofChecker.cpp
+  MismatchedIteratorChecker.cpp
   MmapWriteExecChecker.cpp
   MIGChecker.cpp
   MoveChecker.cpp