[OPENMP]Call __kmpc_push_tripcount in task context.

Runtime function __kmpc_push_tripcount better to call inside of the task
context for target regions. Otherwise, the libomptarget is unable to
link the provided tripcount value for nowait target regions and
completely looses this information.

llvm-svn: 372609

3 files changed, 71 insertions, 58 deletions

diff --git a/clang/lib/CodeGen/CGOpenMPRuntime.cpp b/clang/lib/CodeGen/CGOpenMPRuntime.cpp
index 45833e11390..46a9ec3ac2d 100644
--- a/clang/lib/CodeGen/CGOpenMPRuntime.cpp
+++ b/clang/lib/CodeGen/CGOpenMPRuntime.cpp
@@ -9176,9 +9176,11 @@ void CGOpenMPRuntime::emitUDMapperArrayInitOrDel(
 }
 
 void CGOpenMPRuntime::emitTargetNumIterationsCall(
-    CodeGenFunction &CGF, const OMPExecutableDirective &D, const Expr *Device,
-    const llvm::function_ref<llvm::Value *(
-        CodeGenFunction &CGF, const OMPLoopDirective &D)> &SizeEmitter) {
+    CodeGenFunction &CGF, const OMPExecutableDirective &D,
+    llvm::Value *DeviceID,
+    llvm::function_ref<llvm::Value *(CodeGenFunction &CGF,
+                                     const OMPLoopDirective &D)>
+        SizeEmitter) {
   OpenMPDirectiveKind Kind = D.getDirectiveKind();
   const OMPExecutableDirective *TD = &D;
   // Get nested teams distribute kind directive, if any.
@@ -9187,30 +9189,24 @@ void CGOpenMPRuntime::emitTargetNumIterationsCall(
   if (!TD)
     return;
   const auto *LD = cast<OMPLoopDirective>(TD);
-  auto &&CodeGen = [LD, &Device, &SizeEmitter, this](CodeGenFunction &CGF,
+  auto &&CodeGen = [LD, DeviceID, SizeEmitter, this](CodeGenFunction &CGF,
                                                      PrePostActionTy &) {
-    llvm::Value *NumIterations = SizeEmitter(CGF, *LD);
-
-    // Emit device ID if any.
-    llvm::Value *DeviceID;
-    if (Device)
-      DeviceID = CGF.Builder.CreateIntCast(CGF.EmitScalarExpr(Device),
-                                           CGF.Int64Ty, /*isSigned=*/true);
-    else
-      DeviceID = CGF.Builder.getInt64(OMP_DEVICEID_UNDEF);
-
-    llvm::Value *Args[] = {DeviceID, NumIterations};
-    CGF.EmitRuntimeCall(
-        createRuntimeFunction(OMPRTL__kmpc_push_target_tripcount), Args);
+    if (llvm::Value *NumIterations = SizeEmitter(CGF, *LD)) {
+      llvm::Value *Args[] = {DeviceID, NumIterations};
+      CGF.EmitRuntimeCall(
+          createRuntimeFunction(OMPRTL__kmpc_push_target_tripcount), Args);
+    }
   };
   emitInlinedDirective(CGF, OMPD_unknown, CodeGen);
 }
 
-void CGOpenMPRuntime::emitTargetCall(CodeGenFunction &CGF,
-                                     const OMPExecutableDirective &D,
-                                     llvm::Function *OutlinedFn,
-                                     llvm::Value *OutlinedFnID,
-                                     const Expr *IfCond, const Expr *Device) {
+void CGOpenMPRuntime::emitTargetCall(
+    CodeGenFunction &CGF, const OMPExecutableDirective &D,
+    llvm::Function *OutlinedFn, llvm::Value *OutlinedFnID, const Expr *IfCond,
+    const Expr *Device,
+    llvm::function_ref<llvm::Value *(CodeGenFunction &CGF,
+                                     const OMPLoopDirective &D)>
+        SizeEmitter) {
   if (!CGF.HaveInsertPoint())
     return;
 
@@ -9229,8 +9225,8 @@ void CGOpenMPRuntime::emitTargetCall(CodeGenFunction &CGF,
   llvm::Value *MapTypesArray = nullptr;
   // Fill up the pointer arrays and transfer execution to the device.
   auto &&ThenGen = [this, Device, OutlinedFn, OutlinedFnID, &D, &InputInfo,
-                    &MapTypesArray, &CS, RequiresOuterTask,
-                    &CapturedVars](CodeGenFunction &CGF, PrePostActionTy &) {
+                    &MapTypesArray, &CS, RequiresOuterTask, &CapturedVars,
+                    SizeEmitter](CodeGenFunction &CGF, PrePostActionTy &) {
     // On top of the arrays that were filled up, the target offloading call
     // takes as arguments the device id as well as the host pointer. The host
     // pointer is used by the runtime library to identify the current target
@@ -9262,6 +9258,9 @@ void CGOpenMPRuntime::emitTargetCall(CodeGenFunction &CGF,
     llvm::Value *NumTeams = emitNumTeamsForTargetDirective(CGF, D);
     llvm::Value *NumThreads = emitNumThreadsForTargetDirective(CGF, D);
 
+    // Emit tripcount for the target loop-based directive.
+    emitTargetNumIterationsCall(CGF, D, DeviceID, SizeEmitter);
+
     bool HasNowait = D.hasClausesOfKind<OMPNowaitClause>();
     // The target region is an outlined function launched by the runtime
     // via calls __tgt_target() or __tgt_target_teams().
@@ -11285,12 +11284,13 @@ void CGOpenMPSIMDRuntime::emitTargetOutlinedFunction(
   llvm_unreachable("Not supported in SIMD-only mode");
 }
 
-void CGOpenMPSIMDRuntime::emitTargetCall(CodeGenFunction &CGF,
-                                         const OMPExecutableDirective &D,
-                                         llvm::Function *OutlinedFn,
-                                         llvm::Value *OutlinedFnID,
-                                         const Expr *IfCond,
-                                         const Expr *Device) {
+void CGOpenMPSIMDRuntime::emitTargetCall(
+    CodeGenFunction &CGF, const OMPExecutableDirective &D,
+    llvm::Function *OutlinedFn, llvm::Value *OutlinedFnID, const Expr *IfCond,
+    const Expr *Device,
+    llvm::function_ref<llvm::Value *(CodeGenFunction &CGF,
+                                     const OMPLoopDirective &D)>
+        SizeEmitter) {
   llvm_unreachable("Not supported in SIMD-only mode");
 }
 
diff --git a/clang/lib/CodeGen/CGOpenMPRuntime.h b/clang/lib/CodeGen/CGOpenMPRuntime.h
index 92e946c670c..afec628feb2 100644
--- a/clang/lib/CodeGen/CGOpenMPRuntime.h
+++ b/clang/lib/CodeGen/CGOpenMPRuntime.h
@@ -793,6 +793,17 @@ private:
   /// default.
   virtual unsigned getDefaultFirstprivateAddressSpace() const { return 0; }
 
+  /// Emit code that pushes the trip count of loops associated with constructs
+  /// 'target teams distribute' and 'teams distribute parallel for'.
+  /// \param SizeEmitter Emits the int64 value for the number of iterations of
+  /// the associated loop.
+  void emitTargetNumIterationsCall(
+      CodeGenFunction &CGF, const OMPExecutableDirective &D,
+      llvm::Value *DeviceID,
+      llvm::function_ref<llvm::Value *(CodeGenFunction &CGF,
+                                       const OMPLoopDirective &D)>
+          SizeEmitter);
+
 public:
   explicit CGOpenMPRuntime(CodeGenModule &CGM)
       : CGOpenMPRuntime(CGM, ".", ".") {}
@@ -1414,15 +1425,6 @@ public:
                                           bool IsOffloadEntry,
                                           const RegionCodeGenTy &CodeGen);
 
-  /// Emit code that pushes the trip count of loops associated with constructs
-  /// 'target teams distribute' and 'teams distribute parallel for'.
-  /// \param SizeEmitter Emits the int64 value for the number of iterations of
-  /// the associated loop.
-  virtual void emitTargetNumIterationsCall(
-      CodeGenFunction &CGF, const OMPExecutableDirective &D, const Expr *Device,
-      const llvm::function_ref<llvm::Value *(
-          CodeGenFunction &CGF, const OMPLoopDirective &D)> &SizeEmitter);
-
   /// Emit the target offloading code associated with \a D. The emitted
   /// code attempts offloading the execution to the device, an the event of
   /// a failure it executes the host version outlined in \a OutlinedFn.
@@ -1433,11 +1435,15 @@ public:
   /// directive, or null if no if clause is used.
   /// \param Device Expression evaluated in device clause associated with the
   /// target directive, or null if no device clause is used.
-  virtual void emitTargetCall(CodeGenFunction &CGF,
-                              const OMPExecutableDirective &D,
-                              llvm::Function *OutlinedFn,
-                              llvm::Value *OutlinedFnID, const Expr *IfCond,
-                              const Expr *Device);
+  /// \param SizeEmitter Callback to emit number of iterations for loop-based
+  /// directives.
+  virtual void
+  emitTargetCall(CodeGenFunction &CGF, const OMPExecutableDirective &D,
+                 llvm::Function *OutlinedFn, llvm::Value *OutlinedFnID,
+                 const Expr *IfCond, const Expr *Device,
+                 llvm::function_ref<llvm::Value *(CodeGenFunction &CGF,
+                                                  const OMPLoopDirective &D)>
+                     SizeEmitter);
 
   /// Emit the target regions enclosed in \a GD function definition or
   /// the function itself in case it is a valid device function. Returns true if
@@ -2117,9 +2123,13 @@ public:
   /// directive, or null if no if clause is used.
   /// \param Device Expression evaluated in device clause associated with the
   /// target directive, or null if no device clause is used.
-  void emitTargetCall(CodeGenFunction &CGF, const OMPExecutableDirective &D,
-                      llvm::Function *OutlinedFn, llvm::Value *OutlinedFnID,
-                      const Expr *IfCond, const Expr *Device) override;
+  void
+  emitTargetCall(CodeGenFunction &CGF, const OMPExecutableDirective &D,
+                 llvm::Function *OutlinedFn, llvm::Value *OutlinedFnID,
+                 const Expr *IfCond, const Expr *Device,
+                 llvm::function_ref<llvm::Value *(CodeGenFunction &CGF,
+                                                  const OMPLoopDirective &D)>
+                     SizeEmitter) override;
 
   /// Emit the target regions enclosed in \a GD function definition or
   /// the function itself in case it is a valid device function. Returns true if
diff --git a/clang/lib/CodeGen/CGStmtOpenMP.cpp b/clang/lib/CodeGen/CGStmtOpenMP.cpp
index e34b820625c..b41ad553226 100644
--- a/clang/lib/CodeGen/CGStmtOpenMP.cpp
+++ b/clang/lib/CodeGen/CGStmtOpenMP.cpp
@@ -4137,18 +4137,21 @@ static void emitCommonOMPTargetDirective(CodeGenFunction &CGF,
   CGM.getOpenMPRuntime().emitTargetOutlinedFunction(S, ParentName, Fn, FnID,
                                                     IsOffloadEntry, CodeGen);
   OMPLexicalScope Scope(CGF, S, OMPD_task);
-  auto &&SizeEmitter = [](CodeGenFunction &CGF, const OMPLoopDirective &D) {
-    OMPLoopScope(CGF, D);
-    // Emit calculation of the iterations count.
-    llvm::Value *NumIterations = CGF.EmitScalarExpr(D.getNumIterations());
-    NumIterations = CGF.Builder.CreateIntCast(NumIterations, CGF.Int64Ty,
-                                              /*isSigned=*/false);
-    return NumIterations;
+  auto &&SizeEmitter =
+      [IsOffloadEntry](CodeGenFunction &CGF,
+                       const OMPLoopDirective &D) -> llvm::Value * {
+    if (IsOffloadEntry) {
+      OMPLoopScope(CGF, D);
+      // Emit calculation of the iterations count.
+      llvm::Value *NumIterations = CGF.EmitScalarExpr(D.getNumIterations());
+      NumIterations = CGF.Builder.CreateIntCast(NumIterations, CGF.Int64Ty,
+                                                /*isSigned=*/false);
+      return NumIterations;
+    }
+    return nullptr;
   };
-  if (IsOffloadEntry)
-    CGM.getOpenMPRuntime().emitTargetNumIterationsCall(CGF, S, Device,
-                                                       SizeEmitter);
-  CGM.getOpenMPRuntime().emitTargetCall(CGF, S, Fn, FnID, IfCond, Device);
+  CGM.getOpenMPRuntime().emitTargetCall(CGF, S, Fn, FnID, IfCond, Device,
+                                        SizeEmitter);
 }
 
 static void emitTargetRegion(CodeGenFunction &CGF, const OMPTargetDirective &S,