From: Anastasia Stulova <anastasia.stulova@arm.com>
Date: Thu, 16 Feb 2017 12:27:47 +0000 (+0000)
Subject: [OpenCL] Correct ndrange_t implementation
X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=f5802a7c28a2be35565dc66a16679ba383f4dfd2;p=clang

[OpenCL] Correct ndrange_t implementation

Removed ndrange_t as Clang builtin type and added
as a struct type in the OpenCL header.

Use type name to do the Sema checking in enqueue_kernel
and modify IR generation accordingly.

Review: D28058

Patch by Dmitry Borisenkov!



git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@295311 91177308-0d34-0410-b5e6-96231b3b80d8
---

diff --git a/include/clang/AST/ASTContext.h b/include/clang/AST/ASTContext.h
index 4366b79c1c..5fa515d02d 100644
--- a/include/clang/AST/ASTContext.h
+++ b/include/clang/AST/ASTContext.h
@@ -976,7 +976,7 @@ public:
   CanQualType SingletonId;
 #include "clang/Basic/OpenCLImageTypes.def"
   CanQualType OCLSamplerTy, OCLEventTy, OCLClkEventTy;
-  CanQualType OCLQueueTy, OCLNDRangeTy, OCLReserveIDTy;
+  CanQualType OCLQueueTy, OCLReserveIDTy;
   CanQualType OMPArraySectionTy;
 
   // Types for deductions in C++0x [stmt.ranged]'s desugaring. Built on demand.
diff --git a/include/clang/AST/BuiltinTypes.def b/include/clang/AST/BuiltinTypes.def
index c0c6819280..181131aba0 100644
--- a/include/clang/AST/BuiltinTypes.def
+++ b/include/clang/AST/BuiltinTypes.def
@@ -169,9 +169,6 @@ BUILTIN_TYPE(OCLClkEvent, OCLClkEventTy)
 // OpenCL queue_t.
 BUILTIN_TYPE(OCLQueue, OCLQueueTy)
 
-// OpenCL ndrange_t.
-BUILTIN_TYPE(OCLNDRange, OCLNDRangeTy)
-
 // OpenCL reserve_id_t.
 BUILTIN_TYPE(OCLReserveID, OCLReserveIDTy)
 
diff --git a/include/clang/AST/Type.h b/include/clang/AST/Type.h
index 0db0905cba..f88d0435b2 100644
--- a/include/clang/AST/Type.h
+++ b/include/clang/AST/Type.h
@@ -1744,7 +1744,6 @@ public:
   bool isEventT() const;                        // OpenCL event_t
   bool isClkEventT() const;                     // OpenCL clk_event_t
   bool isQueueT() const;                        // OpenCL queue_t
-  bool isNDRangeT() const;                      // OpenCL ndrange_t
   bool isReserveIDT() const;                    // OpenCL reserve_id_t
 
   bool isPipeType() const;                      // OpenCL pipe type
@@ -5801,10 +5800,6 @@ inline bool Type::isQueueT() const {
   return isSpecificBuiltinType(BuiltinType::OCLQueue);
 }
 
-inline bool Type::isNDRangeT() const {
-  return isSpecificBuiltinType(BuiltinType::OCLNDRange);
-}
-
 inline bool Type::isReserveIDT() const {
   return isSpecificBuiltinType(BuiltinType::OCLReserveID);
 }
@@ -5822,7 +5817,7 @@ inline bool Type::isPipeType() const {
 
 inline bool Type::isOpenCLSpecificType() const {
   return isSamplerT() || isEventT() || isImageType() || isClkEventT() ||
-         isQueueT() || isNDRangeT() || isReserveIDT() || isPipeType();
+         isQueueT() || isReserveIDT() || isPipeType();
 }
 
 inline bool Type::isTemplateTypeParmType() const {
diff --git a/include/clang/Serialization/ASTBitCodes.h b/include/clang/Serialization/ASTBitCodes.h
index 24fefc74ed..c361e80fcd 100644
--- a/include/clang/Serialization/ASTBitCodes.h
+++ b/include/clang/Serialization/ASTBitCodes.h
@@ -803,14 +803,12 @@ namespace clang {
       PREDEF_TYPE_SAMPLER_ID    = 39,
       /// \brief OpenCL queue type.
       PREDEF_TYPE_QUEUE_ID      = 40,
-      /// \brief OpenCL ndrange type.
-      PREDEF_TYPE_NDRANGE_ID    = 41,
       /// \brief OpenCL reserve_id type.
-      PREDEF_TYPE_RESERVE_ID_ID = 42,
+      PREDEF_TYPE_RESERVE_ID_ID = 41,
       /// \brief The placeholder type for OpenMP array section.
-      PREDEF_TYPE_OMP_ARRAY_SECTION = 43,
+      PREDEF_TYPE_OMP_ARRAY_SECTION = 42,
       /// \brief The '__float128' type
-      PREDEF_TYPE_FLOAT128_ID = 44,
+      PREDEF_TYPE_FLOAT128_ID = 43,
       /// \brief OpenCL image types with auto numeration
 #define IMAGE_TYPE(ImgType, Id, SingletonId, Access, Suffix) \
       PREDEF_TYPE_##Id##_ID,
diff --git a/lib/AST/ASTContext.cpp b/lib/AST/ASTContext.cpp
index 84161409ef..7416894b81 100644
--- a/lib/AST/ASTContext.cpp
+++ b/lib/AST/ASTContext.cpp
@@ -1167,7 +1167,6 @@ void ASTContext::InitBuiltinTypes(const TargetInfo &Target,
     InitBuiltinType(OCLEventTy, BuiltinType::OCLEvent);
     InitBuiltinType(OCLClkEventTy, BuiltinType::OCLClkEvent);
     InitBuiltinType(OCLQueueTy, BuiltinType::OCLQueue);
-    InitBuiltinType(OCLNDRangeTy, BuiltinType::OCLNDRange);
     InitBuiltinType(OCLReserveIDTy, BuiltinType::OCLReserveID);
   }
   
@@ -1775,7 +1774,6 @@ TypeInfo ASTContext::getTypeInfoImpl(const Type *T) const {
     case BuiltinType::OCLEvent:
     case BuiltinType::OCLClkEvent:
     case BuiltinType::OCLQueue:
-    case BuiltinType::OCLNDRange:
     case BuiltinType::OCLReserveID:
       // Currently these types are pointers to opaque types.
       Width = Target->getPointerWidth(0);
@@ -5945,7 +5943,6 @@ static char getObjCEncodingForPrimitiveKind(const ASTContext *C,
     case BuiltinType::OCLEvent:
     case BuiltinType::OCLClkEvent:
     case BuiltinType::OCLQueue:
-    case BuiltinType::OCLNDRange:
     case BuiltinType::OCLReserveID:
     case BuiltinType::OCLSampler:
     case BuiltinType::Dependent:
diff --git a/lib/AST/ExprConstant.cpp b/lib/AST/ExprConstant.cpp
index 337eaa40b9..f364f6f310 100644
--- a/lib/AST/ExprConstant.cpp
+++ b/lib/AST/ExprConstant.cpp
@@ -7089,7 +7089,6 @@ static int EvaluateBuiltinClassifyType(const CallExpr *E,
     case BuiltinType::OCLEvent:
     case BuiltinType::OCLClkEvent:
     case BuiltinType::OCLQueue:
-    case BuiltinType::OCLNDRange:
     case BuiltinType::OCLReserveID:
     case BuiltinType::Dependent:
       llvm_unreachable("CallExpr::isBuiltinClassifyType(): unimplemented type");
diff --git a/lib/AST/ItaniumMangle.cpp b/lib/AST/ItaniumMangle.cpp
index 1aa9ad26e9..084b6c3d73 100644
--- a/lib/AST/ItaniumMangle.cpp
+++ b/lib/AST/ItaniumMangle.cpp
@@ -2500,9 +2500,6 @@ void CXXNameMangler::mangleType(const BuiltinType *T) {
   case BuiltinType::OCLQueue:
     Out << "9ocl_queue";
     break;
-  case BuiltinType::OCLNDRange:
-    Out << "11ocl_ndrange";
-    break;
   case BuiltinType::OCLReserveID:
     Out << "13ocl_reserveid";
     break;
diff --git a/lib/AST/MicrosoftMangle.cpp b/lib/AST/MicrosoftMangle.cpp
index 007a358957..6e14dd055c 100644
--- a/lib/AST/MicrosoftMangle.cpp
+++ b/lib/AST/MicrosoftMangle.cpp
@@ -1800,10 +1800,6 @@ void MicrosoftCXXNameMangler::mangleType(const BuiltinType *T, Qualifiers,
     Out << "PA";
     mangleArtificalTagType(TTK_Struct, "ocl_queue");
     break;
-  case BuiltinType::OCLNDRange:
-    Out << "PA";
-    mangleArtificalTagType(TTK_Struct, "ocl_ndrange");
-    break;
   case BuiltinType::OCLReserveID:
     Out << "PA";
     mangleArtificalTagType(TTK_Struct, "ocl_reserveid");
diff --git a/lib/AST/NSAPI.cpp b/lib/AST/NSAPI.cpp
index ac2a8d3542..e7c8c16b01 100644
--- a/lib/AST/NSAPI.cpp
+++ b/lib/AST/NSAPI.cpp
@@ -453,7 +453,6 @@ NSAPI::getNSNumberFactoryMethodKind(QualType T) const {
   case BuiltinType::OCLEvent:
   case BuiltinType::OCLClkEvent:
   case BuiltinType::OCLQueue:
-  case BuiltinType::OCLNDRange:
   case BuiltinType::OCLReserveID:
   case BuiltinType::BoundMember:
   case BuiltinType::Dependent:
diff --git a/lib/AST/Type.cpp b/lib/AST/Type.cpp
index 8f22a1b197..50bebd5f91 100644
--- a/lib/AST/Type.cpp
+++ b/lib/AST/Type.cpp
@@ -2648,8 +2648,6 @@ StringRef BuiltinType::getName(const PrintingPolicy &Policy) const {
     return "clk_event_t";
   case OCLQueue:
     return "queue_t";
-  case OCLNDRange:
-    return "ndrange_t";
   case OCLReserveID:
     return "reserve_id_t";
   case OMPArraySection:
@@ -3661,7 +3659,6 @@ bool Type::canHaveNullability() const {
     case BuiltinType::OCLEvent:
     case BuiltinType::OCLClkEvent:
     case BuiltinType::OCLQueue:
-    case BuiltinType::OCLNDRange:
     case BuiltinType::OCLReserveID:
     case BuiltinType::BuiltinFn:
     case BuiltinType::NullPtr:
diff --git a/lib/AST/TypeLoc.cpp b/lib/AST/TypeLoc.cpp
index 7242858f21..c9a2686557 100644
--- a/lib/AST/TypeLoc.cpp
+++ b/lib/AST/TypeLoc.cpp
@@ -340,7 +340,6 @@ TypeSpecifierType BuiltinTypeLoc::getWrittenTypeSpec() const {
   case BuiltinType::OCLEvent:
   case BuiltinType::OCLClkEvent:
   case BuiltinType::OCLQueue:
-  case BuiltinType::OCLNDRange:
   case BuiltinType::OCLReserveID:
   case BuiltinType::BuiltinFn:
   case BuiltinType::OMPArraySection:
diff --git a/lib/CodeGen/CGBuiltin.cpp b/lib/CodeGen/CGBuiltin.cpp
index 5bdb5ee8d1..f8c6f2bc7d 100644
--- a/lib/CodeGen/CGBuiltin.cpp
+++ b/lib/CodeGen/CGBuiltin.cpp
@@ -2529,13 +2529,14 @@ RValue CodeGenFunction::EmitBuiltinExpr(const FunctionDecl *FD,
     unsigned NumArgs = E->getNumArgs();
 
     llvm::Type *QueueTy = ConvertType(getContext().OCLQueueTy);
-    llvm::Type *RangeTy = ConvertType(getContext().OCLNDRangeTy);
     llvm::Type *GenericVoidPtrTy = Builder.getInt8PtrTy(
         getContext().getTargetAddressSpace(LangAS::opencl_generic));
 
     llvm::Value *Queue = EmitScalarExpr(E->getArg(0));
     llvm::Value *Flags = EmitScalarExpr(E->getArg(1));
-    llvm::Value *Range = EmitScalarExpr(E->getArg(2));
+    LValue NDRangeL = EmitAggExprToLValue(E->getArg(2));
+    llvm::Value *Range = NDRangeL.getAddress().getPointer();
+    llvm::Type *RangeTy = NDRangeL.getAddress().getType();
 
     if (NumArgs == 4) {
       // The most basic form of the call with parameters:
@@ -2548,8 +2549,16 @@ RValue CodeGenFunction::EmitBuiltinExpr(const FunctionDecl *FD,
       llvm::Value *Block = Builder.CreatePointerCast(
           EmitScalarExpr(E->getArg(3)), GenericVoidPtrTy);
 
-      return RValue::get(Builder.CreateCall(
-          CGM.CreateRuntimeFunction(FTy, Name), {Queue, Flags, Range, Block}));
+      AttrBuilder B;
+      B.addAttribute(Attribute::ByVal);
+      AttributeSet ByValAttrSet =
+          AttributeSet::get(CGM.getModule().getContext(), 3U, B);
+
+      auto RTCall =
+          Builder.CreateCall(CGM.CreateRuntimeFunction(FTy, Name, ByValAttrSet),
+                             {Queue, Flags, Range, Block});
+      RTCall->setAttributes(ByValAttrSet);
+      return RValue::get(RTCall);
     }
     assert(NumArgs >= 5 && "Invalid enqueue_kernel signature");
 
diff --git a/lib/CodeGen/CGDebugInfo.cpp b/lib/CodeGen/CGDebugInfo.cpp
index b48714f69b..252ee8d71d 100644
--- a/lib/CodeGen/CGDebugInfo.cpp
+++ b/lib/CodeGen/CGDebugInfo.cpp
@@ -582,8 +582,6 @@ llvm::DIType *CGDebugInfo::CreateType(const BuiltinType *BT) {
     return getOrCreateStructPtrType("opencl_clk_event_t", OCLClkEventDITy);
   case BuiltinType::OCLQueue:
     return getOrCreateStructPtrType("opencl_queue_t", OCLQueueDITy);
-  case BuiltinType::OCLNDRange:
-    return getOrCreateStructPtrType("opencl_ndrange_t", OCLNDRangeDITy);
   case BuiltinType::OCLReserveID:
     return getOrCreateStructPtrType("opencl_reserve_id_t", OCLReserveIDDITy);
 
diff --git a/lib/CodeGen/CGOpenCLRuntime.cpp b/lib/CodeGen/CGOpenCLRuntime.cpp
index 9062936fdd..db02c631c9 100644
--- a/lib/CodeGen/CGOpenCLRuntime.cpp
+++ b/lib/CodeGen/CGOpenCLRuntime.cpp
@@ -58,9 +58,6 @@ llvm::Type *CGOpenCLRuntime::convertOpenCLSpecificType(const Type *T) {
   case BuiltinType::OCLQueue:
     return llvm::PointerType::get(
         llvm::StructType::create(Ctx, "opencl.queue_t"), 0);
-  case BuiltinType::OCLNDRange:
-    return llvm::PointerType::get(
-        llvm::StructType::create(Ctx, "opencl.ndrange_t"), 0);
   case BuiltinType::OCLReserveID:
     return llvm::PointerType::get(
         llvm::StructType::create(Ctx, "opencl.reserve_id_t"), 0);
diff --git a/lib/CodeGen/CodeGenTypes.cpp b/lib/CodeGen/CodeGenTypes.cpp
index c9e49a6af6..dc24b2040f 100644
--- a/lib/CodeGen/CodeGenTypes.cpp
+++ b/lib/CodeGen/CodeGenTypes.cpp
@@ -472,7 +472,6 @@ llvm::Type *CodeGenTypes::ConvertType(QualType T) {
     case BuiltinType::OCLEvent:
     case BuiltinType::OCLClkEvent:
     case BuiltinType::OCLQueue:
-    case BuiltinType::OCLNDRange:
     case BuiltinType::OCLReserveID:
       ResultType = CGM.getOpenCLRuntime().convertOpenCLSpecificType(Ty);
       break;
diff --git a/lib/CodeGen/ItaniumCXXABI.cpp b/lib/CodeGen/ItaniumCXXABI.cpp
index 6a5fd1916b..38918c32bc 100644
--- a/lib/CodeGen/ItaniumCXXABI.cpp
+++ b/lib/CodeGen/ItaniumCXXABI.cpp
@@ -2637,7 +2637,6 @@ static bool TypeInfoIsInStandardLibrary(const BuiltinType *Ty) {
     case BuiltinType::OCLEvent:
     case BuiltinType::OCLClkEvent:
     case BuiltinType::OCLQueue:
-    case BuiltinType::OCLNDRange:
     case BuiltinType::OCLReserveID:
       return false;
 
diff --git a/lib/Headers/opencl-c.h b/lib/Headers/opencl-c.h
index d68db8605b..3bcb9a3af1 100644
--- a/lib/Headers/opencl-c.h
+++ b/lib/Headers/opencl-c.h
@@ -16748,16 +16748,12 @@ typedef int clk_profiling_info;
 
 #define MAX_WORK_DIM        3
 
-// ToDo: Remove definition of ndrange_t in Clang as an opaque type and add back
-// the following ndrange_t definition.
-#if 0
 typedef struct {
     unsigned int workDimension;
     size_t globalWorkOffset[MAX_WORK_DIM];
     size_t globalWorkSize[MAX_WORK_DIM];
     size_t localWorkSize[MAX_WORK_DIM];
 } ndrange_t;
-#endif
 
 ndrange_t __ovld ndrange_1D(size_t);
 ndrange_t __ovld ndrange_1D(size_t, size_t);
diff --git a/lib/Index/USRGeneration.cpp b/lib/Index/USRGeneration.cpp
index 3960074b5a..73dddd9a8b 100644
--- a/lib/Index/USRGeneration.cpp
+++ b/lib/Index/USRGeneration.cpp
@@ -654,7 +654,6 @@ void USRGenerator::VisitType(QualType T) {
         case BuiltinType::OCLEvent:
         case BuiltinType::OCLClkEvent:
         case BuiltinType::OCLQueue:
-        case BuiltinType::OCLNDRange:
         case BuiltinType::OCLReserveID:
         case BuiltinType::OCLSampler:
           IgnoreResults = true;
diff --git a/lib/Sema/Sema.cpp b/lib/Sema/Sema.cpp
index 412f944f89..11d1bac5d3 100644
--- a/lib/Sema/Sema.cpp
+++ b/lib/Sema/Sema.cpp
@@ -217,7 +217,6 @@ void Sema::Initialize() {
     if (getLangOpts().OpenCLVersion >= 200) {
       addImplicitTypedef("clk_event_t", Context.OCLClkEventTy);
       addImplicitTypedef("queue_t", Context.OCLQueueTy);
-      addImplicitTypedef("ndrange_t", Context.OCLNDRangeTy);
       addImplicitTypedef("reserve_id_t", Context.OCLReserveIDTy);
       addImplicitTypedef("atomic_int", Context.getAtomicType(Context.IntTy));
       addImplicitTypedef("atomic_uint",
diff --git a/lib/Sema/SemaChecking.cpp b/lib/Sema/SemaChecking.cpp
index 3aedb2a8c9..463b73d098 100644
--- a/lib/Sema/SemaChecking.cpp
+++ b/lib/Sema/SemaChecking.cpp
@@ -408,10 +408,10 @@ static bool SemaOpenCLBuiltinEnqueueKernel(Sema &S, CallExpr *TheCall) {
   }
 
   // Third argument is always an ndrange_t type.
-  if (!Arg2->getType()->isNDRangeT()) {
+  if (Arg2->getType().getAsString() != "ndrange_t") {
     S.Diag(TheCall->getArg(2)->getLocStart(),
            diag::err_opencl_enqueue_kernel_expected_type)
-        << S.Context.OCLNDRangeTy;
+        << "'ndrange_t'";
     return true;
   }
 
diff --git a/lib/Serialization/ASTCommon.cpp b/lib/Serialization/ASTCommon.cpp
index ecd249cc50..65792c7e4d 100644
--- a/lib/Serialization/ASTCommon.cpp
+++ b/lib/Serialization/ASTCommon.cpp
@@ -147,9 +147,6 @@ serialization::TypeIdxFromBuiltin(const BuiltinType *BT) {
   case BuiltinType::OCLQueue:
     ID = PREDEF_TYPE_QUEUE_ID;
     break;
-  case BuiltinType::OCLNDRange:
-    ID = PREDEF_TYPE_NDRANGE_ID;
-    break;
   case BuiltinType::OCLReserveID:
     ID = PREDEF_TYPE_RESERVE_ID_ID;
     break;
diff --git a/lib/Serialization/ASTReader.cpp b/lib/Serialization/ASTReader.cpp
index 1b65ecbe98..452f5410a8 100644
--- a/lib/Serialization/ASTReader.cpp
+++ b/lib/Serialization/ASTReader.cpp
@@ -6378,9 +6378,6 @@ QualType ASTReader::GetType(TypeID ID) {
     case PREDEF_TYPE_QUEUE_ID:
       T = Context.OCLQueueTy;
       break;
-    case PREDEF_TYPE_NDRANGE_ID:
-      T = Context.OCLNDRangeTy;
-      break;
     case PREDEF_TYPE_RESERVE_ID_ID:
       T = Context.OCLReserveIDTy;
       break;
diff --git a/test/CodeGenOpenCL/cl20-device-side-enqueue.cl b/test/CodeGenOpenCL/cl20-device-side-enqueue.cl
index fdfb199944..def2906615 100644
--- a/test/CodeGenOpenCL/cl20-device-side-enqueue.cl
+++ b/test/CodeGenOpenCL/cl20-device-side-enqueue.cl
@@ -2,6 +2,8 @@
 // RUN: %clang_cc1 %s -cl-std=CL2.0 -ffake-address-space-map -O0 -emit-llvm -o - -triple "spir64-unknown-unknown" | FileCheck %s --check-prefix=COMMON --check-prefix=B64
 
 typedef void (^bl_t)(local void *);
+typedef struct {int a;} ndrange_t;
+
 // N.B. The check here only exists to set BL_GLOBAL
 // COMMON: @block_G =  addrspace(1) constant void (i8 addrspace(3)*) addrspace(4)* addrspacecast (void (i8 addrspace(3)*) addrspace(1)* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(2)* } addrspace(1)* [[BL_GLOBAL:@__block_literal_global(\.[0-9]+)?]] to void (i8 addrspace(3)*) addrspace(1)*) to void (i8 addrspace(3)*) addrspace(4)*)
 const bl_t block_G = (bl_t) ^ (local void *a) {};
@@ -11,7 +13,7 @@ kernel void device_side_enqueue(global int *a, global int *b, int i) {
   queue_t default_queue;
   // COMMON: %flags = alloca i32
   unsigned flags = 0;
-  // COMMON: %ndrange = alloca %opencl.ndrange_t*
+  // COMMON: %ndrange = alloca %struct.ndrange_t
   ndrange_t ndrange;
   // COMMON: %clk_event = alloca %opencl.clk_event_t*
   clk_event_t clk_event;
@@ -20,12 +22,12 @@ kernel void device_side_enqueue(global int *a, global int *b, int i) {
   // COMMON: %event_wait_list2 = alloca [1 x %opencl.clk_event_t*]
   clk_event_t event_wait_list2[] = {clk_event};
 
+  // COMMON: [[NDR:%[a-z0-9]+]] = alloca %struct.ndrange_t, align 4
   // COMMON: [[DEF_Q:%[0-9]+]] = load %opencl.queue_t{{.*}}*, %opencl.queue_t{{.*}}** %default_queue
   // COMMON: [[FLAGS:%[0-9]+]] = load i32, i32* %flags
-  // COMMON: [[NDR:%[0-9]+]] = load %opencl.ndrange_t*, %opencl.ndrange_t** %ndrange
   // COMMON: [[BL:%[0-9]+]] = bitcast <{ i8*, i32, i32, i8*, %struct.__block_descriptor addrspace(2)*, i32{{.*}}, i32{{.*}}, i32{{.*}} }>* %block to void ()*
   // COMMON: [[BL_I8:%[0-9]+]] = addrspacecast void ()* [[BL]] to i8 addrspace(4)*
-  // COMMON: call i32 @__enqueue_kernel_basic(%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %opencl.ndrange_t* [[NDR]], i8 addrspace(4)* [[BL_I8]])
+  // COMMON: call i32 @__enqueue_kernel_basic(%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %struct.ndrange_t* byval [[NDR]]{{(.[0-9]+)?}}, i8 addrspace(4)* [[BL_I8]])
   enqueue_kernel(default_queue, flags, ndrange,
                  ^(void) {
                    a[i] = b[i];
@@ -33,12 +35,11 @@ kernel void device_side_enqueue(global int *a, global int *b, int i) {
 
   // COMMON: [[DEF_Q:%[0-9]+]] = load %opencl.queue_t{{.*}}*, %opencl.queue_t{{.*}}** %default_queue
   // COMMON: [[FLAGS:%[0-9]+]] = load i32, i32* %flags
-  // COMMON: [[NDR:%[0-9]+]] = load %opencl.ndrange_t*, %opencl.ndrange_t** %ndrange
   // COMMON: [[WAIT_EVNT:%[0-9]+]] = addrspacecast %opencl.clk_event_t{{.*}}** %event_wait_list to %opencl.clk_event_t{{.*}}* addrspace(4)*
   // COMMON: [[EVNT:%[0-9]+]] = addrspacecast %opencl.clk_event_t{{.*}}** %clk_event to %opencl.clk_event_t{{.*}}* addrspace(4)*
   // COMMON: [[BL:%[0-9]+]] = bitcast <{ i8*, i32, i32, i8*, %struct.__block_descriptor addrspace(2)*, i32{{.*}}, i32{{.*}}, i32{{.*}} }>* %block3 to void ()*
   // COMMON: [[BL_I8:%[0-9]+]] = addrspacecast void ()* [[BL]] to i8 addrspace(4)*
-  // COMMON: call i32 @__enqueue_kernel_basic_events(%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]],  %opencl.ndrange_t* {{.*}}, i32 2, %opencl.clk_event_t{{.*}}* addrspace(4)* [[WAIT_EVNT]], %opencl.clk_event_t{{.*}}* addrspace(4)* [[EVNT]], i8 addrspace(4)* [[BL_I8]])
+  // COMMON: call i32 @__enqueue_kernel_basic_events(%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]],  %struct.ndrange_t* {{.*}}, i32 2, %opencl.clk_event_t{{.*}}* addrspace(4)* [[WAIT_EVNT]], %opencl.clk_event_t{{.*}}* addrspace(4)* [[EVNT]], i8 addrspace(4)* [[BL_I8]])
   enqueue_kernel(default_queue, flags, ndrange, 2, &event_wait_list, &clk_event,
                  ^(void) {
                    a[i] = b[i];
@@ -46,9 +47,8 @@ kernel void device_side_enqueue(global int *a, global int *b, int i) {
 
   // COMMON: [[DEF_Q:%[0-9]+]] = load %opencl.queue_t{{.*}}*, %opencl.queue_t{{.*}}** %default_queue
   // COMMON: [[FLAGS:%[0-9]+]] = load i32, i32* %flags
-  // COMMON: [[NDR:%[0-9]+]] = load %opencl.ndrange_t*, %opencl.ndrange_t** %ndrange
-  // B32: call i32 (%opencl.queue_t{{.*}}*, i32, %opencl.ndrange_t*, i8 addrspace(4)*, i32, ...) @__enqueue_kernel_vaargs(%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %opencl.ndrange_t* [[NDR]], i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(2)* } addrspace(1)* @__block_literal_global{{(.[0-9]+)?}} to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1, i32 256)
-  // B64: call i32 (%opencl.queue_t{{.*}}*, i32, %opencl.ndrange_t*, i8 addrspace(4)*, i32, ...) @__enqueue_kernel_vaargs(%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %opencl.ndrange_t* [[NDR]], i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(2)* } addrspace(1)* @__block_literal_global{{(.[0-9]+)?}} to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1, i64 256)
+  // B32: call i32 (%opencl.queue_t{{.*}}*, i32, %struct.ndrange_t*, i8 addrspace(4)*, i32, ...) @__enqueue_kernel_vaargs(%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %struct.ndrange_t* [[NDR]]{{(.[0-9]+)?}}, i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(2)* } addrspace(1)* @__block_literal_global{{(.[0-9]+)?}} to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1, i32 256)
+  // B64: call i32 (%opencl.queue_t{{.*}}*, i32, %struct.ndrange_t*, i8 addrspace(4)*, i32, ...) @__enqueue_kernel_vaargs(%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %struct.ndrange_t* [[NDR]]{{(.[0-9]+)?}}, i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(2)* } addrspace(1)* @__block_literal_global{{(.[0-9]+)?}} to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1, i64 256)
   enqueue_kernel(default_queue, flags, ndrange,
                  ^(local void *p) {
                    return;
@@ -57,11 +57,10 @@ kernel void device_side_enqueue(global int *a, global int *b, int i) {
   char c;
   // COMMON: [[DEF_Q:%[0-9]+]] = load %opencl.queue_t{{.*}}*, %opencl.queue_t{{.*}}** %default_queue
   // COMMON: [[FLAGS:%[0-9]+]] = load i32, i32* %flags
-  // COMMON: [[NDR:%[0-9]+]] = load %opencl.ndrange_t*, %opencl.ndrange_t** %ndrange
   // B32: [[SIZE:%[0-9]+]] = zext i8 {{%[0-9]+}} to i32
-  // B32: call i32 (%opencl.queue_t{{.*}}*, i32, %opencl.ndrange_t*, i8 addrspace(4)*, i32, ...) @__enqueue_kernel_vaargs(%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %opencl.ndrange_t* [[NDR]], i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(2)* } addrspace(1)* @__block_literal_global{{(.[0-9]+)?}} to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1, i32 [[SIZE]])
+  // B32: call i32 (%opencl.queue_t{{.*}}*, i32, %struct.ndrange_t*, i8 addrspace(4)*, i32, ...) @__enqueue_kernel_vaargs(%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %struct.ndrange_t* [[NDR]]{{(.[0-9]+)?}}, i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(2)* } addrspace(1)* @__block_literal_global{{(.[0-9]+)?}} to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1, i32 [[SIZE]])
   // B64: [[SIZE:%[0-9]+]] = zext i8 {{%[0-9]+}} to i64
-  // B64: call i32 (%opencl.queue_t{{.*}}*, i32, %opencl.ndrange_t*, i8 addrspace(4)*, i32, ...) @__enqueue_kernel_vaargs(%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %opencl.ndrange_t* [[NDR]], i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(2)* } addrspace(1)* @__block_literal_global{{(.[0-9]+)?}} to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1, i64 [[SIZE]])
+  // B64: call i32 (%opencl.queue_t{{.*}}*, i32, %struct.ndrange_t*, i8 addrspace(4)*, i32, ...) @__enqueue_kernel_vaargs(%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %struct.ndrange_t* [[NDR]]{{(.[0-9]+)?}}, i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(2)* } addrspace(1)* @__block_literal_global{{(.[0-9]+)?}} to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1, i64 [[SIZE]])
   enqueue_kernel(default_queue, flags, ndrange,
                  ^(local void *p) {
                    return;
@@ -70,12 +69,11 @@ kernel void device_side_enqueue(global int *a, global int *b, int i) {
 
   // COMMON: [[DEF_Q:%[0-9]+]] = load %opencl.queue_t{{.*}}*, %opencl.queue_t{{.*}}** %default_queue
   // COMMON: [[FLAGS:%[0-9]+]] = load i32, i32* %flags
-  // COMMON: [[NDR:%[0-9]+]] = load %opencl.ndrange_t*, %opencl.ndrange_t** %ndrange
   // COMMON: [[AD:%arraydecay[0-9]*]] = getelementptr inbounds [1 x %opencl.clk_event_t*], [1 x %opencl.clk_event_t*]* %event_wait_list2, i32 0, i32 0
   // COMMON: [[WAIT_EVNT:%[0-9]+]] = addrspacecast %opencl.clk_event_t{{.*}}** [[AD]] to %opencl.clk_event_t{{.*}}* addrspace(4)*
   // COMMON: [[EVNT:%[0-9]+]]  = addrspacecast %opencl.clk_event_t{{.*}}** %clk_event to %opencl.clk_event_t{{.*}}* addrspace(4)*
-  // B32: call i32 (%opencl.queue_t{{.*}}*, i32,  %opencl.ndrange_t*, i32, %opencl.clk_event_t{{.*}}* addrspace(4)*, %opencl.clk_event_t{{.*}}* addrspace(4)*, i8 addrspace(4)*, i32, ...) @__enqueue_kernel_events_vaargs(%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]],  %opencl.ndrange_t* {{.*}}, i32 2, %opencl.clk_event_t{{.*}} [[WAIT_EVNT]], %opencl.clk_event_t{{.*}} [[EVNT]], i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(2)* } addrspace(1)* @__block_literal_global{{(.[0-9]+)?}} to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1, i32 256)
-  // B64: call i32 (%opencl.queue_t{{.*}}*, i32,  %opencl.ndrange_t*, i32, %opencl.clk_event_t{{.*}}* addrspace(4)*, %opencl.clk_event_t{{.*}}* addrspace(4)*, i8 addrspace(4)*, i32, ...) @__enqueue_kernel_events_vaargs(%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]],  %opencl.ndrange_t* {{.*}}, i32 2, %opencl.clk_event_t{{.*}} [[WAIT_EVNT]], %opencl.clk_event_t{{.*}} [[EVNT]], i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(2)* } addrspace(1)* @__block_literal_global{{(.[0-9]+)?}} to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1, i64 256)
+  // B32: call i32 (%opencl.queue_t{{.*}}*, i32,  %struct.ndrange_t*, i32, %opencl.clk_event_t{{.*}}* addrspace(4)*, %opencl.clk_event_t{{.*}}* addrspace(4)*, i8 addrspace(4)*, i32, ...) @__enqueue_kernel_events_vaargs(%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]],  %struct.ndrange_t* {{.*}}, i32 2, %opencl.clk_event_t{{.*}} [[WAIT_EVNT]], %opencl.clk_event_t{{.*}} [[EVNT]], i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(2)* } addrspace(1)* @__block_literal_global{{(.[0-9]+)?}} to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1, i32 256)
+  // B64: call i32 (%opencl.queue_t{{.*}}*, i32,  %struct.ndrange_t*, i32, %opencl.clk_event_t{{.*}}* addrspace(4)*, %opencl.clk_event_t{{.*}}* addrspace(4)*, i8 addrspace(4)*, i32, ...) @__enqueue_kernel_events_vaargs(%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]],  %struct.ndrange_t* {{.*}}, i32 2, %opencl.clk_event_t{{.*}} [[WAIT_EVNT]], %opencl.clk_event_t{{.*}} [[EVNT]], i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(2)* } addrspace(1)* @__block_literal_global{{(.[0-9]+)?}} to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1, i64 256)
   enqueue_kernel(default_queue, flags, ndrange, 2, event_wait_list2, &clk_event,
                  ^(local void *p) {
                    return;
@@ -84,14 +82,13 @@ kernel void device_side_enqueue(global int *a, global int *b, int i) {
 
   // COMMON: [[DEF_Q:%[0-9]+]] = load %opencl.queue_t{{.*}}*, %opencl.queue_t{{.*}}** %default_queue
   // COMMON: [[FLAGS:%[0-9]+]] = load i32, i32* %flags
-  // COMMON: [[NDR:%[0-9]+]] = load %opencl.ndrange_t*, %opencl.ndrange_t** %ndrange
   // COMMON: [[AD:%arraydecay[0-9]*]] = getelementptr inbounds [1 x %opencl.clk_event_t*], [1 x %opencl.clk_event_t*]* %event_wait_list2, i32 0, i32 0
   // COMMON: [[WAIT_EVNT:%[0-9]+]] = addrspacecast %opencl.clk_event_t{{.*}}** [[AD]] to %opencl.clk_event_t{{.*}}* addrspace(4)*
   // COMMON: [[EVNT:%[0-9]+]]  = addrspacecast %opencl.clk_event_t{{.*}}** %clk_event to %opencl.clk_event_t{{.*}}* addrspace(4)*
   // B32: [[SIZE:%[0-9]+]] = zext i8 {{%[0-9]+}} to i32
-  // B32: call i32 (%opencl.queue_t{{.*}}*, i32,  %opencl.ndrange_t*, i32, %opencl.clk_event_t{{.*}}* addrspace(4)*, %opencl.clk_event_t{{.*}}* addrspace(4)*, i8 addrspace(4)*, i32, ...) @__enqueue_kernel_events_vaargs(%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]],  %opencl.ndrange_t* {{.*}}, i32 2, %opencl.clk_event_t{{.*}}* addrspace(4)* [[WAIT_EVNT]], %opencl.clk_event_t{{.*}}* addrspace(4)* [[EVNT]], i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(2)* } addrspace(1)* @__block_literal_global{{(.[0-9]+)?}} to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1, i32 [[SIZE]])
+  // B32: call i32 (%opencl.queue_t{{.*}}*, i32,  %struct.ndrange_t*, i32, %opencl.clk_event_t{{.*}}* addrspace(4)*, %opencl.clk_event_t{{.*}}* addrspace(4)*, i8 addrspace(4)*, i32, ...) @__enqueue_kernel_events_vaargs(%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]],  %struct.ndrange_t* {{.*}}, i32 2, %opencl.clk_event_t{{.*}}* addrspace(4)* [[WAIT_EVNT]], %opencl.clk_event_t{{.*}}* addrspace(4)* [[EVNT]], i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(2)* } addrspace(1)* @__block_literal_global{{(.[0-9]+)?}} to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1, i32 [[SIZE]])
   // B64: [[SIZE:%[0-9]+]] = zext i8 {{%[0-9]+}} to i64
-  // B64: call i32 (%opencl.queue_t{{.*}}*, i32,  %opencl.ndrange_t*, i32, %opencl.clk_event_t{{.*}}* addrspace(4)*, %opencl.clk_event_t{{.*}}* addrspace(4)*, i8 addrspace(4)*, i32, ...) @__enqueue_kernel_events_vaargs(%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]],  %opencl.ndrange_t* {{.*}}, i32 2, %opencl.clk_event_t{{.*}}* addrspace(4)* [[WAIT_EVNT]], %opencl.clk_event_t{{.*}}* addrspace(4)* [[EVNT]], i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(2)* } addrspace(1)* @__block_literal_global{{(.[0-9]+)?}} to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1, i64 [[SIZE]])
+  // B64: call i32 (%opencl.queue_t{{.*}}*, i32,  %struct.ndrange_t*, i32, %opencl.clk_event_t{{.*}}* addrspace(4)*, %opencl.clk_event_t{{.*}}* addrspace(4)*, i8 addrspace(4)*, i32, ...) @__enqueue_kernel_events_vaargs(%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]],  %struct.ndrange_t* {{.*}}, i32 2, %opencl.clk_event_t{{.*}}* addrspace(4)* [[WAIT_EVNT]], %opencl.clk_event_t{{.*}}* addrspace(4)* [[EVNT]], i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(2)* } addrspace(1)* @__block_literal_global{{(.[0-9]+)?}} to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1, i64 [[SIZE]])
   enqueue_kernel(default_queue, flags, ndrange, 2, event_wait_list2, &clk_event,
                  ^(local void *p) {
                    return;
@@ -101,11 +98,10 @@ kernel void device_side_enqueue(global int *a, global int *b, int i) {
   long l;
   // COMMON: [[DEF_Q:%[0-9]+]] = load %opencl.queue_t{{.*}}*, %opencl.queue_t{{.*}}** %default_queue
   // COMMON: [[FLAGS:%[0-9]+]] = load i32, i32* %flags
-  // COMMON: [[NDR:%[0-9]+]] = load %opencl.ndrange_t*, %opencl.ndrange_t** %ndrange
   // B32: [[SIZE:%[0-9]+]] = trunc i64 {{%[0-9]+}} to i32
-  // B32: call i32 (%opencl.queue_t{{.*}}*, i32, %opencl.ndrange_t*, i8 addrspace(4)*, i32, ...) @__enqueue_kernel_vaargs(%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %opencl.ndrange_t* [[NDR]], i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(2)* } addrspace(1)* @__block_literal_global{{(.[0-9]+)?}} to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1, i32 [[SIZE]])
+  // B32: call i32 (%opencl.queue_t{{.*}}*, i32, %struct.ndrange_t*, i8 addrspace(4)*, i32, ...) @__enqueue_kernel_vaargs(%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %struct.ndrange_t* [[NDR]]{{(.[0-9]+)?}}, i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(2)* } addrspace(1)* @__block_literal_global{{(.[0-9]+)?}} to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1, i32 [[SIZE]])
   // B64: [[SIZE:%[0-9]+]] = load i64, i64* %l
-  // B64: call i32 (%opencl.queue_t{{.*}}*, i32, %opencl.ndrange_t*, i8 addrspace(4)*, i32, ...) @__enqueue_kernel_vaargs(%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %opencl.ndrange_t* [[NDR]], i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(2)* } addrspace(1)* @__block_literal_global{{(.[0-9]+)?}} to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1, i64 [[SIZE]])
+  // B64: call i32 (%opencl.queue_t{{.*}}*, i32, %struct.ndrange_t*, i8 addrspace(4)*, i32, ...) @__enqueue_kernel_vaargs(%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %struct.ndrange_t* [[NDR]]{{(.[0-9]+)?}}, i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(2)* } addrspace(1)* @__block_literal_global{{(.[0-9]+)?}} to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1, i64 [[SIZE]])
   enqueue_kernel(default_queue, flags, ndrange,
                  ^(local void *p) {
                    return;
@@ -114,9 +110,8 @@ kernel void device_side_enqueue(global int *a, global int *b, int i) {
 
   // COMMON: [[DEF_Q:%[0-9]+]] = load %opencl.queue_t*, %opencl.queue_t** %default_queue
   // COMMON: [[FLAGS:%[0-9]+]] = load i32, i32* %flags
-  // COMMON: [[NDR:%[0-9]+]] = load %opencl.ndrange_t*, %opencl.ndrange_t** %ndrange
-  // B32: call i32 (%opencl.queue_t{{.*}}*, i32, %opencl.ndrange_t*, i8 addrspace(4)*, i32, ...) @__enqueue_kernel_vaargs(%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %opencl.ndrange_t* [[NDR]], i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(2)* } addrspace(1)* @__block_literal_global{{(.[0-9]+)?}} to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1, i32 0)
-  // B64: call i32 (%opencl.queue_t{{.*}}*, i32, %opencl.ndrange_t*, i8 addrspace(4)*, i32, ...) @__enqueue_kernel_vaargs(%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %opencl.ndrange_t* [[NDR]], i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(2)* } addrspace(1)* @__block_literal_global{{(.[0-9]+)?}} to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1, i64 4294967296)
+  // B32: call i32 (%opencl.queue_t{{.*}}*, i32, %struct.ndrange_t*, i8 addrspace(4)*, i32, ...) @__enqueue_kernel_vaargs(%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %struct.ndrange_t* [[NDR]]{{(.[0-9]+)?}}, i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(2)* } addrspace(1)* @__block_literal_global{{(.[0-9]+)?}} to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1, i32 0)
+  // B64: call i32 (%opencl.queue_t{{.*}}*, i32, %struct.ndrange_t*, i8 addrspace(4)*, i32, ...) @__enqueue_kernel_vaargs(%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %struct.ndrange_t* [[NDR]]{{(.[0-9]+)?}}, i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(2)* } addrspace(1)* @__block_literal_global{{(.[0-9]+)?}} to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1, i64 4294967296)
   enqueue_kernel(default_queue, flags, ndrange,
                  ^(local void *p) {
                    return;
diff --git a/test/Headers/opencl-c-header.cl b/test/Headers/opencl-c-header.cl
index fffbd7a8ed..2c28be19fa 100644
--- a/test/Headers/opencl-c-header.cl
+++ b/test/Headers/opencl-c-header.cl
@@ -1,5 +1,5 @@
-// RUN: %clang_cc1 -triple spir-unknown-unknown -internal-isystem ../../lib/Headers -include opencl-c.h -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple spir-unknown-unknown -internal-isystem ../../lib/Headers -include opencl-c.h -emit-llvm -o - %s -cl-std=CL1.1| FileCheck %s
+// RUN: %clang_cc1 -O0 -triple spir-unknown-unknown -internal-isystem ../../lib/Headers -include opencl-c.h -emit-llvm -o - %s| FileCheck %s
+// RUN: %clang_cc1 -O0 -triple spir-unknown-unknown -internal-isystem ../../lib/Headers -include opencl-c.h -emit-llvm -o - %s -cl-std=CL1.1| FileCheck %s
 
 // Test including the default header as a module.
 // The module should be compiled only once and loaded from cache afterwards.
@@ -25,26 +25,28 @@
 
 // ===
 // Compile for OpenCL 2.0 for the first time. The module should change.
-// RUN: %clang_cc1 -triple spir-unknown-unknown -emit-llvm -o - -cl-std=CL2.0 -finclude-default-header -fmodules -fimplicit-module-maps -fmodules-cache-path=%t -fdisable-module-hash -ftime-report %s 2>&1 | FileCheck --check-prefix=CHECK20 --check-prefix=CHECK-MOD %s
+// RUN: %clang_cc1 -triple spir-unknown-unknown -O0 -emit-llvm -o - -cl-std=CL2.0 -finclude-default-header -fmodules -fimplicit-module-maps -fmodules-cache-path=%t -fdisable-module-hash -ftime-report %s 2>&1 | FileCheck --check-prefix=CHECK20 --check-prefix=CHECK-MOD %s
 // RUN: not diff %t/1_0.pcm %t/opencl_c.pcm
 // RUN: chmod u-w %t/opencl_c.pcm
 
 // ===
 // Compile for OpenCL 2.0 for the second time. The module should not change.
-// RUN: %clang_cc1 -triple spir-unknown-unknown -emit-llvm -o - -cl-std=CL2.0 -finclude-default-header -fmodules -fimplicit-module-maps -fmodules-cache-path=%t -fdisable-module-hash -ftime-report %s 2>&1 | FileCheck --check-prefix=CHECK20 --check-prefix=CHECK-MOD %s
+// RUN: %clang_cc1 -triple spir-unknown-unknown -O0 -emit-llvm -o - -cl-std=CL2.0 -finclude-default-header -fmodules -fimplicit-module-maps -fmodules-cache-path=%t -fdisable-module-hash -ftime-report %s 2>&1 | FileCheck --check-prefix=CHECK20 --check-prefix=CHECK-MOD %s
 
 // Check cached module works for different OpenCL versions.
 // RUN: rm -rf %t
 // RUN: mkdir -p %t
 // RUN: %clang_cc1 -triple spir64-unknown-unknown -emit-llvm -o - -cl-std=CL1.2 -finclude-default-header -fmodules -fimplicit-module-maps -fmodules-cache-path=%t -ftime-report %s 2>&1 | FileCheck --check-prefix=CHECK --check-prefix=CHECK-MOD %s
-// RUN: %clang_cc1 -triple amdgcn--amdhsa -emit-llvm -o - -cl-std=CL2.0 -finclude-default-header -fmodules -fimplicit-module-maps -fmodules-cache-path=%t -ftime-report %s 2>&1 | FileCheck --check-prefix=CHECK20 --check-prefix=CHECK-MOD %s
+// RUN: %clang_cc1 -triple amdgcn--amdhsa -O0 -emit-llvm -o - -cl-std=CL2.0 -finclude-default-header -fmodules -fimplicit-module-maps -fmodules-cache-path=%t -ftime-report %s 2>&1 | FileCheck --check-prefix=CHECK20 --check-prefix=CHECK-MOD %s
 // RUN: chmod u-w %t 
 // RUN: %clang_cc1 -triple spir64-unknown-unknown -emit-llvm -o - -cl-std=CL1.2 -finclude-default-header -fmodules -fimplicit-module-maps -fmodules-cache-path=%t -ftime-report %s 2>&1 | FileCheck --check-prefix=CHECK --check-prefix=CHECK-MOD %s
-// RUN: %clang_cc1 -triple amdgcn--amdhsa -emit-llvm -o - -cl-std=CL2.0 -finclude-default-header -fmodules -fimplicit-module-maps -fmodules-cache-path=%t -ftime-report %s 2>&1 | FileCheck --check-prefix=CHECK20 --check-prefix=CHECK-MOD %s
+// RUN: %clang_cc1 -triple amdgcn--amdhsa -O0 -emit-llvm -o - -cl-std=CL2.0 -finclude-default-header -fmodules -fimplicit-module-maps -fmodules-cache-path=%t -ftime-report %s 2>&1 | FileCheck --check-prefix=CHECK20 --check-prefix=CHECK-MOD %s
 // RUN: chmod u+w %t
 
 // Verify that called builtins occur in the generated IR.
 
+// CHECK-NOT: ndrange_t
+// CHECK20: ndrange_t
 // CHECK: _Z16convert_char_rtec
 // CHECK-NOT: _Z3ctzc
 // CHECK20: _Z3ctzc
@@ -52,11 +54,9 @@
 char f(char x) {
 #if __OPENCL_C_VERSION__ != CL_VERSION_2_0
   return convert_char_rte(x);
-#ifdef NO_HEADER
-  //expected-warning@-2{{implicit declaration of function 'convert_char_rte' is invalid in C99}}
-#endif //NO_HEADER
 
 #else //__OPENCL_C_VERSION__
+  ndrange_t t;
   return ctz(x);
 #endif //__OPENCL_C_VERSION__
 }
diff --git a/test/PCH/ocl_types.h b/test/PCH/ocl_types.h
index bdc4bb11c1..96391289f7 100644
--- a/test/PCH/ocl_types.h
+++ b/test/PCH/ocl_types.h
@@ -32,9 +32,6 @@ typedef clk_event_t clkevt_t;
 // queue_t
 typedef queue_t q_t;
 
-// ndrange_t
-typedef ndrange_t range_t;
-
 // reserve_id_t
 typedef reserve_id_t reserveid_t;
 
diff --git a/test/SemaOpenCL/cl20-device-side-enqueue.cl b/test/SemaOpenCL/cl20-device-side-enqueue.cl
index 6d8d375c99..3e87cfcddf 100644
--- a/test/SemaOpenCL/cl20-device-side-enqueue.cl
+++ b/test/SemaOpenCL/cl20-device-side-enqueue.cl
@@ -1,6 +1,7 @@
 // RUN: %clang_cc1 %s -cl-std=CL2.0 -triple "spir-unknown-unknown" -verify -pedantic -fsyntax-only -DB32
 // RUN: %clang_cc1 %s -cl-std=CL2.0 -triple "spir64-unknown-unknown" -verify -pedantic -fsyntax-only -Wconversion -DWCONV
 
+typedef struct {int a;} ndrange_t;
 // Diagnostic tests for different overloads of enqueue_kernel from Table 6.13.17.1 of OpenCL 2.0 Spec.
 kernel void enqueue_kernel_tests() {
   queue_t default_queue;
diff --git a/tools/libclang/CIndex.cpp b/tools/libclang/CIndex.cpp
index f4661ca206..673d96374b 100644
--- a/tools/libclang/CIndex.cpp
+++ b/tools/libclang/CIndex.cpp
@@ -1493,7 +1493,6 @@ bool CursorVisitor::VisitBuiltinTypeLoc(BuiltinTypeLoc TL) {
   case BuiltinType::OCLEvent:
   case BuiltinType::OCLClkEvent:
   case BuiltinType::OCLQueue:
-  case BuiltinType::OCLNDRange:
   case BuiltinType::OCLReserveID:
 #define BUILTIN_TYPE(Id, SingletonId)
 #define SIGNED_TYPE(Id, SingletonId) case BuiltinType::Id: