diff --git a/clang/lib/CodeGen/CGOpenMPRuntime.cpp b/clang/lib/CodeGen/CGOpenMPRuntime.cpp
--- a/clang/lib/CodeGen/CGOpenMPRuntime.cpp
+++ b/clang/lib/CodeGen/CGOpenMPRuntime.cpp
@@ -2098,6 +2098,14 @@
     OutlinedFnArgs.push_back(ThreadIDAddr.getPointer());
     OutlinedFnArgs.push_back(ZeroAddrBound.getPointer());
     OutlinedFnArgs.append(CapturedVars.begin(), CapturedVars.end());
+
+    // Ensure we do not inline the function. This is trivially true for the ones
+    // passed to __kmpc_fork_call but the ones calles in serialized regions
+    // could be inlined. This is not a perfect but it is closer to the invariant
+    // we want, namely, every data environment starts with a new function.
+    // TODO: We should pass the if condition to the runtime function and do the
+    //       handling there. Much cleaner code.
+    OutlinedFn->addFnAttr(llvm::Attribute::NoInline);
     RT.emitOutlinedFunctionCall(CGF, Loc, OutlinedFn, OutlinedFnArgs);
 
     // __kmpc_end_serialized_parallel(&Loc, GTid);
diff --git a/clang/lib/CodeGen/CGOpenMPRuntimeGPU.cpp b/clang/lib/CodeGen/CGOpenMPRuntimeGPU.cpp
--- a/clang/lib/CodeGen/CGOpenMPRuntimeGPU.cpp
+++ b/clang/lib/CodeGen/CGOpenMPRuntimeGPU.cpp
@@ -1576,11 +1576,6 @@
   auto *OutlinedFun =
       cast<llvm::Function>(CGOpenMPRuntime::emitParallelOutlinedFunction(
           D, ThreadIDVar, InnermostKind, CodeGen));
-  if (CGM.getLangOpts().Optimize) {
-    OutlinedFun->removeFnAttr(llvm::Attribute::NoInline);
-    OutlinedFun->removeFnAttr(llvm::Attribute::OptimizeNone);
-    OutlinedFun->addFnAttr(llvm::Attribute::AlwaysInline);
-  }
   IsInTargetMasterThreadRegion = PrevIsInTargetMasterThreadRegion;
   IsInTTDRegion = PrevIsInTTDRegion;
   if (getExecutionMode() != CGOpenMPRuntimeGPU::EM_SPMD &&
@@ -1698,11 +1693,6 @@
   CodeGen.setAction(Action);
   llvm::Function *OutlinedFun = CGOpenMPRuntime::emitTeamsOutlinedFunction(
       D, ThreadIDVar, InnermostKind, CodeGen);
-  if (CGM.getLangOpts().Optimize) {
-    OutlinedFun->removeFnAttr(llvm::Attribute::NoInline);
-    OutlinedFun->removeFnAttr(llvm::Attribute::OptimizeNone);
-    OutlinedFun->addFnAttr(llvm::Attribute::AlwaysInline);
-  }
 
   return OutlinedFun;
 }
@@ -2102,6 +2092,14 @@
   // Force inline this outlined function at its call site.
   Fn->setLinkage(llvm::GlobalValue::InternalLinkage);
 
+  // Ensure we do not inline the function. This is trivially true for the ones
+  // passed to __kmpc_fork_call but the ones calles in serialized regions
+  // could be inlined. This is not a perfect but it is closer to the invariant
+  // we want, namely, every data environment starts with a new function.
+  // TODO: We should pass the if condition to the runtime function and do the
+  //       handling there. Much cleaner code.
+  cast<llvm::Function>(OutlinedFn)->addFnAttr(llvm::Attribute::NoInline);
+
   Address ZeroAddr = CGF.CreateDefaultAlignTempAlloca(CGF.Int32Ty,
                                                       /*Name=*/".zero.addr");
   CGF.InitTempAlloca(ZeroAddr, CGF.Builder.getInt32(/*C*/ 0));
@@ -3134,11 +3132,6 @@
       "_omp_reduction_shuffle_and_reduce_func", &CGM.getModule());
   CGM.SetInternalFunctionAttributes(GlobalDecl(), Fn, CGFI);
   Fn->setDoesNotRecurse();
-  if (CGM.getLangOpts().Optimize) {
-    Fn->removeFnAttr(llvm::Attribute::NoInline);
-    Fn->removeFnAttr(llvm::Attribute::OptimizeNone);
-    Fn->addFnAttr(llvm::Attribute::AlwaysInline);
-  }
 
   CodeGenFunction CGF(CGM);
   CGF.StartFunction(GlobalDecl(), C.VoidTy, Fn, CGFI, Args, Loc, Loc);
diff --git a/clang/test/OpenMP/parallel_if_codegen.cpp b/clang/test/OpenMP/parallel_if_codegen.cpp
--- a/clang/test/OpenMP/parallel_if_codegen.cpp
+++ b/clang/test/OpenMP/parallel_if_codegen.cpp
@@ -7,6 +7,7 @@
 // RUN: %clang_cc1 -fopenmp-simd -fopenmp-version=45 -x c++ -triple %itanium_abi_triple -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck --check-prefix SIMD-ONLY0 %s
 // SIMD-ONLY0-NOT: {{__kmpc|__tgt}}
 
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple %itanium_abi_triple -emit-llvm %s -disable-O0-optnone -o - | FileCheck %s --check-prefix=WOPT
 // RUN: %clang_cc1 -verify -fopenmp -x c++ -triple %itanium_abi_triple -emit-llvm %s -o - | FileCheck %s
 // RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple %itanium_abi_triple -emit-pch -o %t %s
 // RUN: %clang_cc1 -fopenmp -x c++ -triple %itanium_abi_triple -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
@@ -96,14 +97,20 @@
   return tmain(Arg);
 }
 
+// WOPT:     noinline
+// WOPT-NOT: optnone
 // CHECK: define internal {{.*}}void [[CAP_FN4]]
 // CHECK: call {{.*}}void @{{.+}}fn4
 // CHECK: ret void
 
+// WOPT:     noinline
+// WOPT-NOT: optnone
 // CHECK: define internal {{.*}}void [[CAP_FN5]]
 // CHECK: call {{.*}}void @{{.+}}fn5
 // CHECK: ret void
 
+// WOPT:     noinline
+// WOPT-NOT: optnone
 // CHECK: define internal {{.*}}void [[CAP_FN6]]
 // CHECK: call {{.*}}void @{{.+}}fn6
 // CHECK: ret void
@@ -129,14 +136,20 @@
 // CHECK: br label %[[OMP_END]]
 // CHECK: [[OMP_END]]
 
+// WOPT:     noinline
+// WOPT-NOT: optnone
 // CHECK: define internal {{.*}}void [[CAP_FN1]]
 // CHECK: call {{.*}}void @{{.+}}fn1
 // CHECK: ret void
 
+// WOPT:     noinline
+// WOPT-NOT: optnone
 // CHECK: define internal {{.*}}void [[CAP_FN2]]
 // CHECK: call {{.*}}void @{{.+}}fn2
 // CHECK: ret void
 
+// WOPT:     noinline
+// WOPT-NOT: optnone
 // CHECK: define internal {{.*}}void [[CAP_FN3]]
 // CHECK: call {{.*}}void @{{.+}}fn3
 // CHECK: ret void