[OpenMPIRBuilder] Introduce OMPRegionInfo managing the stack of OpenMP region constructs.

clang/lib/CodeGen/CGOpenMPRuntime.cpp

@@ -1184,51 +1184,6 @@ CGOpenMPRuntime::getUserDefinedReduction(const OMPDeclareReductionDecl *D) {
   return UDRMap.lookup(D);
 }
 
-namespace {
-// Temporary RAII solution to perform a push/pop stack event on the OpenMP IR
-// Builder if one is present.
-struct PushAndPopStackRAII {
-  PushAndPopStackRAII(llvm::OpenMPIRBuilder *OMPBuilder, CodeGenFunction &CGF,
-                      bool HasCancel, llvm::omp::Directive Kind)
-      : OMPBuilder(OMPBuilder) {
-    if (!OMPBuilder)
-      return;
-
-    // The following callback is the crucial part of clangs cleanup process.
-    //
-    // NOTE:
-    // Once the OpenMPIRBuilder is used to create parallel regions (and
-    // similar), the cancellation destination (Dest below) is determined via
-    // IP. That means if we have variables to finalize we split the block at IP,
-    // use the new block (=BB) as destination to build a JumpDest (via
-    // getJumpDestInCurrentScope(BB)) which then is fed to
-    // EmitBranchThroughCleanup. Furthermore, there will not be the need
-    // to push & pop an FinalizationInfo object.
-    // The FiniCB will still be needed but at the point where the
-    // OpenMPIRBuilder is asked to construct a parallel (or similar) construct.
-    auto FiniCB = [&CGF](llvm::OpenMPIRBuilder::InsertPointTy IP) {
-      assert(IP.getBlock()->end() == IP.getPoint() &&
-             "Clang CG should cause non-terminated block!");
-      CGBuilderTy::InsertPointGuard IPG(CGF.Builder);
-      CGF.Builder.restoreIP(IP);
-      CodeGenFunction::JumpDest Dest =
-          CGF.getOMPCancelDestination(OMPD_parallel);
-      CGF.EmitBranchThroughCleanup(Dest);
-    };
-
-    // TODO: Remove this once we emit parallel regions through the
-    //       OpenMPIRBuilder as it can do this setup internally.
-    llvm::OpenMPIRBuilder::FinalizationInfo FI({FiniCB, Kind, HasCancel});
-    OMPBuilder->pushFinalizationCB(std::move(FI));
-  }
-  ~PushAndPopStackRAII() {
-    if (OMPBuilder)
-      OMPBuilder->popFinalizationCB();
-  }
-  llvm::OpenMPIRBuilder *OMPBuilder;
-};
-} // namespace
-
 static llvm::Function *emitParallelOrTeamsOutlinedFunction(
     CodeGenModule &CGM, const OMPExecutableDirective &D, const CapturedStmt *CS,
     const VarDecl *ThreadIDVar, OpenMPDirectiveKind InnermostKind,
@@ -1256,10 +1211,6 @@ static llvm::Function *emitParallelOrTeamsOutlinedFunction(
                dyn_cast<OMPTargetTeamsDistributeParallelForDirective>(&D))
     HasCancel = OPFD->hasCancel();
 
-  // TODO: Temporarily inform the OpenMPIRBuilder, if any, about the new
-  //       parallel region to make cancellation barriers work properly.
-  llvm::OpenMPIRBuilder &OMPBuilder = CGM.getOpenMPRuntime().getOMPBuilder();
-  PushAndPopStackRAII PSR(&OMPBuilder, CGF, HasCancel, InnermostKind);
   CGOpenMPOutlinedRegionInfo CGInfo(*CS, ThreadIDVar, CodeGen, InnermostKind,
                                     HasCancel, OutlinedHelperName);
   CodeGenFunction::CGCapturedStmtRAII CapInfoRAII(CGF, &CGInfo);

clang/lib/CodeGen/CGStmtOpenMP.cpp

@@ -1540,6 +1540,8 @@ static void emitCommonOMPParallelDirective(
     CodeGenFunction &CGF, const OMPExecutableDirective &S,
     OpenMPDirectiveKind InnermostKind, const RegionCodeGenTy &CodeGen,
     const CodeGenBoundParametersTy &CodeGenBoundParameters) {
+  CodeGenFunction::NonOpenMPIRBuilderRegion NonBuilderScope(CGF);
+
   const CapturedStmt *CS = S.getCapturedStmt(OMPD_parallel);
   llvm::Value *NumThreads = nullptr;
   llvm::Function *OutlinedFn =
@@ -1713,7 +1715,8 @@ void CodeGenFunction::OMPBuilderCBHelpers::EmitOMPOutlinedRegionBody(
 }
 
 void CodeGenFunction::EmitOMPParallelDirective(const OMPParallelDirective &S) {
-  if (CGM.getLangOpts().OpenMPIRBuilder) {
+  if (CGM.getLangOpts().OpenMPIRBuilder &&
+      !IsInsideNonOpenMPIRBuilderHandledRegion) {
     llvm::OpenMPIRBuilder &OMPBuilder = CGM.getOpenMPRuntime().getOMPBuilder();
     // Check if we have any if clause associated with the directive.
     llvm::Value *IfCond = nullptr;
@@ -3729,6 +3732,8 @@ static void emitScanBasedDirective(
 static bool emitWorksharingDirective(CodeGenFunction &CGF,
                                      const OMPLoopDirective &S,
                                      bool HasCancel) {
+  CodeGenFunction::NonOpenMPIRBuilderRegion NonOmpBuilderScope(CGF);
+
   bool HasLastprivates;
   if (llvm::any_of(S.getClausesOfKind<OMPReductionClause>(),
                    [](const OMPReductionClause *C) {
@@ -3906,6 +3911,8 @@ static LValue createSectionLVal(CodeGenFunction &CGF, QualType Ty,
 }
 
 void CodeGenFunction::EmitSections(const OMPExecutableDirective &S) {
+  NonOpenMPIRBuilderRegion NonOmpBuilderScope(*this);
+
   const Stmt *CapturedStmt = S.getInnermostCapturedStmt()->getCapturedStmt();
   const auto *CS = dyn_cast<CompoundStmt>(CapturedStmt);
   bool HasLastprivates = false;
@@ -4121,7 +4128,8 @@ void CodeGenFunction::EmitOMPSectionsDirective(const OMPSectionsDirective &S) {
 }
 
 void CodeGenFunction::EmitOMPSectionDirective(const OMPSectionDirective &S) {
-  if (CGM.getLangOpts().OpenMPIRBuilder) {
+  if (CGM.getLangOpts().OpenMPIRBuilder &&
+      !IsInsideNonOpenMPIRBuilderHandledRegion) {
     llvm::OpenMPIRBuilder &OMPBuilder = CGM.getOpenMPRuntime().getOMPBuilder();
     using InsertPointTy = llvm::OpenMPIRBuilder::InsertPointTy;
 
@@ -4132,8 +4140,14 @@ void CodeGenFunction::EmitOMPSectionDirective(const OMPSectionDirective &S) {
 
     auto BodyGenCB = [SectionRegionBodyStmt, this](InsertPointTy AllocaIP,
                                                    InsertPointTy CodeGenIP) {
-      OMPBuilderCBHelpers::EmitOMPInlinedRegionBody(
-          *this, SectionRegionBodyStmt, AllocaIP, CodeGenIP, "section");
+      Builder.restoreIP(CodeGenIP);
+      llvm::BasicBlock *FiniBB =
+          splitBBWithSuffix(Builder, false, ".sectionfini");
+
+      OMPBuilderCBHelpers::InlinedRegionBodyRAII IRB(*this, AllocaIP, *FiniBB);
+      EmitStmt(SectionRegionBodyStmt);
+
+      Builder.CreateBr(FiniBB);
     };
 
     LexicalScope Scope(*this, S.getSourceRange());
@@ -4485,6 +4499,8 @@ void CodeGenFunction::EmitOMPTaskBasedDirective(
     const OMPExecutableDirective &S, const OpenMPDirectiveKind CapturedRegion,
     const RegionCodeGenTy &BodyGen, const TaskGenTy &TaskGen,
     OMPTaskDataTy &Data) {
+  NonOpenMPIRBuilderRegion NonOmpBuilderScope(*this);
+
   // Emit outlined function for task construct.
   const CapturedStmt *CS = S.getCapturedStmt(CapturedRegion);
   auto I = CS->getCapturedDecl()->param_begin();
@@ -5027,6 +5043,8 @@ void CodeGenFunction::EmitOMPTargetTaskBasedDirective(
 }
 
 void CodeGenFunction::EmitOMPTaskDirective(const OMPTaskDirective &S) {
+  NonOpenMPIRBuilderRegion NonOmpBuilderScope(*this);
+
   // Emit outlined function for task construct.
   const CapturedStmt *CS = S.getCapturedStmt(OMPD_task);
   Address CapturedStruct = GenerateCapturedStmtArgument(*CS);
@@ -5044,6 +5062,7 @@ void CodeGenFunction::EmitOMPTaskDirective(const OMPTaskDirective &S) {
   // Check if we should emit tied or untied task.
   Data.Tied = !S.getSingleClause<OMPUntiedClause>();
   auto &&BodyGen = [CS](CodeGenFunction &CGF, PrePostActionTy &) {
+    NonOpenMPIRBuilderRegion NonOmpBuilderScope(CGF);
     CGF.EmitStmt(CS->getCapturedStmt());
   };
   auto &&TaskGen = [&S, SharedsTy, CapturedStruct,
@@ -6928,20 +6947,14 @@ void CodeGenFunction::EmitOMPCancelDirective(const OMPCancelDirective &S) {
       break;
     }
   }
-  if (CGM.getLangOpts().OpenMPIRBuilder) {
+  if (CGM.getLangOpts().OpenMPIRBuilder &&
+      !IsInsideNonOpenMPIRBuilderHandledRegion) {
     llvm::OpenMPIRBuilder &OMPBuilder = CGM.getOpenMPRuntime().getOMPBuilder();
-    // TODO: This check is necessary as we only generate `omp parallel` through
-    // the OpenMPIRBuilder for now.
-    if (S.getCancelRegion() == OMPD_parallel ||
-        S.getCancelRegion() == OMPD_sections ||
-        S.getCancelRegion() == OMPD_section) {
-      llvm::Value *IfCondition = nullptr;
-      if (IfCond)
-        IfCondition = EmitScalarExpr(IfCond,
-                                     /*IgnoreResultAssign=*/true);
-      return Builder.restoreIP(
-          OMPBuilder.createCancel(Builder, IfCondition, S.getCancelRegion()));
-    }
+    llvm::Value *IfCondition = nullptr;
+    if (IfCond)
+      IfCondition = EvaluateExprAsBool(IfCond);
+    return Builder.restoreIP(
+        OMPBuilder.createCancel(Builder, IfCondition, S.getCancelRegion()));
   }
 
   CGM.getOpenMPRuntime().emitCancelCall(*this, S.getBeginLoc(), IfCond,

clang/lib/CodeGen/CodeGenFunction.h

@@ -489,6 +489,32 @@ class CodeGenFunction : public CodeGenTypeCache {
     ~CGCapturedStmtRAII() { CGF.CapturedStmtInfo = PrevCapturedStmtInfo; }
   };
 
+  /// While in a region handled by Clang's CGOpenMPRuntime, do not use the
+  /// OpenMPIRBuilder which requires all surrounding regions to be handled by
+  /// OpenMPIRBuilder as well.
+  ///
+  /// Required until everything can be handled by OpenMPIRBuilder.
+  /// Isn't the ultimate solution to mixing OpenMPIRBuilder and
+  /// non-OpenMPIRBuilder codegen either, but works with the current regression
+  /// tests so far.
+  bool IsInsideNonOpenMPIRBuilderHandledRegion = false;
+  class NonOpenMPIRBuilderRegion {
+  private:
+    CodeGenFunction &CGF;
+    bool PreviousIsInsideNonOpenMPIRBuilderHandledRegion;
+
+  public:
+    NonOpenMPIRBuilderRegion(CodeGenFunction &CGF)
+        : CGF(CGF), PreviousIsInsideNonOpenMPIRBuilderHandledRegion(
+                        CGF.IsInsideNonOpenMPIRBuilderHandledRegion) {
+      CGF.IsInsideNonOpenMPIRBuilderHandledRegion = true;
+    }
+    ~NonOpenMPIRBuilderRegion() {
+      CGF.IsInsideNonOpenMPIRBuilderHandledRegion =
+          PreviousIsInsideNonOpenMPIRBuilderHandledRegion;
+    }
+  };
+
   /// An abstract representation of regular/ObjC call/message targets.
   class AbstractCallee {
     /// The function declaration of the callee.
@@ -1776,16 +1802,11 @@ class CodeGenFunction : public CodeGenTypeCache {
     /// \param IP	Insertion point for generating the finalization code.
     static void FinalizeOMPRegion(CodeGenFunction &CGF, InsertPointTy IP) {
       CGBuilderTy::InsertPointGuard IPG(CGF.Builder);
-      assert(IP.getBlock()->end() != IP.getPoint() &&
-             "OpenMP IR Builder should cause terminated block!");
 
-      llvm::BasicBlock *IPBB = IP.getBlock();
-      llvm::BasicBlock *DestBB = IPBB->getUniqueSuccessor();
-      assert(DestBB && "Finalization block should have one successor!");
+      CGF.Builder.restoreIP(IP);
+      llvm::BasicBlock *DestBB =
+          llvm::splitBB(CGF.Builder, /*CreateBranch*/ false, ".ompfinalize");
 
-      // erase and replace with cleanup branch.
-      IPBB->getTerminator()->eraseFromParent();
-      CGF.Builder.SetInsertPoint(IPBB);
       CodeGenFunction::JumpDest Dest = CGF.getJumpDestInCurrentScope(DestBB);
       CGF.EmitBranchThroughCleanup(Dest);
     }