[OpenCL] Remove block invoke function from emitted block literal struct

[yaxunl]

Summary

OpenCL runtime tracks the invoke function emitted for
any block expression. Due to restrictions on blocks in
OpenCL (v2.0 s6.12.5), it is always possible to know the
block invoke function when emitting call of block expression
or __enqueue_kernel builtin functions. Since __enqueu_kernel
already has an argument for the invoke function, it is redundant
to have invoke function member in the llvm block literal structure.

This patch removes invoke function from the llvm block literal
structure. It also removes the bitcast of block invoke function
to the "generic block literal type" which is useless for OpenCL.

This will save some space for the kernel argument, and also
eliminate some store instructions.

[yaxunl]

ping

[bader]

Hi Sam,

Sorry for the delay. LGTM, I have only minor refactoring suggestion.

Thanks,
Alexey

lib/CodeGen/CGBlocks.cpp

1070–1072

I think it would be more readable if we merge this if statement with the if statement at the line #1103.
It's used to initialize FuncPtr for non-OpenCL languages and the first use of this variable is in the else block of if statement at the line #1103.
If I didn't miss something it should reasonable to combine this if block with 'else' block at the line #1106.

[yaxunl]

lib/CodeGen/CGBlocks.cpp

1070–1072

BlockPtr is used on line 1093, so it cannot be moved to line 1106.

[svenvh]

Sorry for digging up an old commit...

Apparently this broke block arguments, e.g. the following test case:

int foo(int (^ bl)(void)) {
  return bl();
}

int get21() {
  return foo(^{return 21;});
}

int get42() {
  return foo(^{return 42;});
}

In particular, the VarDecl that getBlockExpr() sees doesn't have an initializer when the called block comes from an argument (causing clang to crash).

[yaxunl]

In D43783#1204353, @svenvh wrote:

Sorry for digging up an old commit...

Apparently this broke block arguments, e.g. the following test case:

int foo(int (^ bl)(void)) {
  return bl();
}

int get21() {
  return foo(^{return 21;});
}

int get42() {
  return foo(^{return 42;});
}

In particular, the VarDecl that getBlockExpr() sees doesn't have an initializer when the called block comes from an argument (causing clang to crash).

Sorry for the delay. I think block should not be allowed as function argument since this generally leads indirect function calls therefore requires support of function pointer. It will rely on optimizations to get rid of indirect function calls.

[Anastasia]

In D43783#1212485, @yaxunl wrote:

In D43783#1204353, @svenvh wrote:

Sorry for digging up an old commit...

Apparently this broke block arguments, e.g. the following test case:

int foo(int (^ bl)(void)) {
  return bl();
}

int get21() {
  return foo(^{return 21;});
}

int get42() {
  return foo(^{return 42;});
}

In particular, the VarDecl that getBlockExpr() sees doesn't have an initializer when the called block comes from an argument (causing clang to crash).

Sorry for the delay. I think block should not be allowed as function argument since this generally leads indirect function calls therefore requires support of function pointer. It will rely on optimizations to get rid of indirect function calls.

The idea was to allow blocks as function parameters because they are statically known at each function call. This can be resolved later at IR level instead of frontend. I am also not sure there can be other corner cases where it wouldn't work in Clang since we can't leverage analysis passes here. I feel this might be a risky thing to do in Clang. Currently it doesn't work for the examples provided and therefore breaking the compliance with the spec.

[yaxunl]

In D43783#1215573, @Anastasia wrote:

In D43783#1212485, @yaxunl wrote:

In D43783#1204353, @svenvh wrote:

Sorry for digging up an old commit...

Apparently this broke block arguments, e.g. the following test case:

int foo(int (^ bl)(void)) {
  return bl();
}

int get21() {
  return foo(^{return 21;});
}

int get42() {
  return foo(^{return 42;});
}

In particular, the VarDecl that getBlockExpr() sees doesn't have an initializer when the called block comes from an argument (causing clang to crash).

Sorry for the delay. I think block should not be allowed as function argument since this generally leads indirect function calls therefore requires support of function pointer. It will rely on optimizations to get rid of indirect function calls.

The idea was to allow blocks as function parameters because they are statically known at each function call. This can be resolved later at IR level instead of frontend. I am also not sure there can be other corner cases where it wouldn't work in Clang since we can't leverage analysis passes here. I feel this might be a risky thing to do in Clang. Currently it doesn't work for the examples provided and therefore breaking the compliance with the spec.

I agree this patch should be reverted for conformance to spec. I will do that. Thanks.

[Anastasia]

Ping! Do you still plan to do this? :)

[yaxunl]

In D43783#1235090, @Anastasia wrote:

Ping! Do you still plan to do this? :)

Sorry I caught up in something else. Since there are some subsequent commits, it may take some efforts to revert it.

[Anastasia]

In D43783#1235126, @yaxunl wrote:

In D43783#1235090, @Anastasia wrote:

Ping! Do you still plan to do this? :)

Sorry I caught up in something else. Since there are some subsequent commits, it may take some efforts to revert it.

No problem. Let me know if help is needed. :)

In D43783#1235126, @yaxunl wrote:

In D43783#1235090, @Anastasia wrote:

Ping! Do you still plan to do this? :)

Sorry I caught up in something else. Since there are some subsequent commits, it may take some efforts to revert it.

Let me know if you need help. I am happy to do this myself.

[svenvh]

Reverted in r343582, test added in r343583.

[AlexeySotkin]

In D43783#1215573, @Anastasia wrote:

In D43783#1212485, @yaxunl wrote:

In D43783#1204353, @svenvh wrote:

Sorry for digging up an old commit...

Apparently this broke block arguments, e.g. the following test case:

int foo(int (^ bl)(void)) {
  return bl();
}

int get21() {
  return foo(^{return 21;});
}

int get42() {
  return foo(^{return 42;});
}

In particular, the VarDecl that getBlockExpr() sees doesn't have an initializer when the called block comes from an argument (causing clang to crash).

Sorry for the delay. I think block should not be allowed as function argument since this generally leads indirect function calls therefore requires support of function pointer. It will rely on optimizations to get rid of indirect function calls.

The idea was to allow blocks as function parameters because they are statically known at each function call. This can be resolved later at IR level instead of frontend. I am also not sure there can be other corner cases where it wouldn't work in Clang since we can't leverage analysis passes here. I feel this might be a risky thing to do in Clang. Currently it doesn't work for the examples provided and therefore breaking the compliance with the spec.

The spec is not clear about what is "statically determinable". To me, in the example provided we can not resolve the bl() call without inlining foo, even at IR level. As Sam noted, that leads to indirect call and require support of functions pointers.
I see a contradiction in the spec in allowing blocks to be a function argument and disallowing function pointers at the same time. I think maybe the spec should be changed to clarify existing restrictions or add more restrictions for cases when blocks are passed as function argument.

[Anastasia]

In D43783#1272001, @AlexeySotkin wrote:

In D43783#1215573, @Anastasia wrote:

In D43783#1212485, @yaxunl wrote:

In D43783#1204353, @svenvh wrote:

Sorry for digging up an old commit...

Apparently this broke block arguments, e.g. the following test case:

int foo(int (^ bl)(void)) {
  return bl();
}

int get21() {
  return foo(^{return 21;});
}

int get42() {
  return foo(^{return 42;});
}

In particular, the VarDecl that getBlockExpr() sees doesn't have an initializer when the called block comes from an argument (causing clang to crash).

Sorry for the delay. I think block should not be allowed as function argument since this generally leads indirect function calls therefore requires support of function pointer. It will rely on optimizations to get rid of indirect function calls.

The idea was to allow blocks as function parameters because they are statically known at each function call. This can be resolved later at IR level instead of frontend. I am also not sure there can be other corner cases where it wouldn't work in Clang since we can't leverage analysis passes here. I feel this might be a risky thing to do in Clang. Currently it doesn't work for the examples provided and therefore breaking the compliance with the spec.

The spec is not clear about what is "statically determinable". To me, in the example provided we can not resolve the bl() call without inlining foo, even at IR level. As Sam noted, that leads to indirect call and require support of functions pointers.
I see a contradiction in the spec in allowing blocks to be a function argument and disallowing function pointers at the same time. I think maybe the spec should be changed to clarify existing restrictions or add more restrictions for cases when blocks are passed as function argument.

I am not sure that blocks in function call parameters contradict the rule of them being statically determinable though. You can statically determine what blocks are being passed into a function. There might be multiple different ones though. However, it should be possible to lower the function pointers into the direct calls.

lib/CodeGen/CGBlocks.cpp

static CharUnits getLowBit(CharUnits v) {
   return CharUnits::fromQuantity(v.getQuantity() & (~v.getQuantity() + 1));
 }
 
 static void initializeForBlockHeader(CodeGenModule &CGM, CGBlockInfo &info,
                              SmallVectorImpl<llvm::Type*> &elementTypes) {
 
   assert(elementTypes.empty());
   if (CGM.getLangOpts().OpenCL) {
-    // The header is basically 'struct { int; int; generic void *;
+    // The header is basically 'struct { int; int;
     // custom_fields; }'. Assert that struct is packed.
-    auto GenericAS =
-        CGM.getContext().getTargetAddressSpace(LangAS::opencl_generic);
-    auto GenPtrAlign =
-        CharUnits::fromQuantity(CGM.getTarget().getPointerAlign(GenericAS) / 8);
-    auto GenPtrSize =
-        CharUnits::fromQuantity(CGM.getTarget().getPointerWidth(GenericAS) / 8);
-    assert(CGM.getIntSize() <= GenPtrSize);
-    assert(CGM.getIntAlign() <= GenPtrAlign);
-    assert((2 * CGM.getIntSize()).isMultipleOf(GenPtrAlign));
     elementTypes.push_back(CGM.IntTy); /* total size */
     elementTypes.push_back(CGM.IntTy); /* align */
-    elementTypes.push_back(
-        CGM.getOpenCLRuntime()
-            .getGenericVoidPointerType()); /* invoke function */
-    unsigned Offset =
-        2 * CGM.getIntSize().getQuantity() + GenPtrSize.getQuantity();
-    unsigned BlockAlign = GenPtrAlign.getQuantity();
+    unsigned Offset = 2 * CGM.getIntSize().getQuantity();
+    unsigned BlockAlign = CGM.getIntAlign().getQuantity();
     if (auto *Helper =
             CGM.getTargetCodeGenInfo().getTargetOpenCLBlockHelper()) {
       for (auto I : Helper->getCustomFieldTypes()) /* custom fields */ {
         // TargetOpenCLBlockHelp needs to make sure the struct is packed.
         // If necessary, add padding fields to the custom fields.
         unsigned Align = CGM.getDataLayout().getABITypeAlignment(I);
         if (BlockAlign < Align)
           BlockAlign = Align;
blockInfo.reset(findAndRemoveBlockInfo(&FirstBlockInfo,
                                          blockExpr->getBlockDecl()));
 
   blockInfo->BlockExpression = blockExpr;
   return EmitBlockLiteral(*blockInfo);
 }
 
 llvm::Value *CodeGenFunction::EmitBlockLiteral(const CGBlockInfo &blockInfo) {
   bool IsOpenCL = CGM.getContext().getLangOpts().OpenCL;
-  auto GenVoidPtrTy =
-      IsOpenCL ? CGM.getOpenCLRuntime().getGenericVoidPointerType() : VoidPtrTy;
-  LangAS GenVoidPtrAddr = IsOpenCL ? LangAS::opencl_generic : LangAS::Default;
-  auto GenVoidPtrSize = CharUnits::fromQuantity(
-      CGM.getTarget().getPointerWidth(
-          CGM.getContext().getTargetAddressSpace(GenVoidPtrAddr)) /
-      8);
   // Using the computed layout, generate the actual block function.
   bool isLambdaConv = blockInfo.getBlockDecl()->isConversionFromLambda();
   CodeGenFunction BlockCGF{CGM, true};
   BlockCGF.SanOpts = SanOpts;
   auto *InvokeFn = BlockCGF.GenerateBlockFunction(
       CurGD, blockInfo, LocalDeclMap, isLambdaConv, blockInfo.CanBeGlobal);
-  auto *blockFn = llvm::ConstantExpr::getPointerCast(InvokeFn, GenVoidPtrTy);
 
   // If there is nothing to capture, we can emit this as a global block.
   if (blockInfo.CanBeGlobal)
     return CGM.getAddrOfGlobalBlockIfEmitted(blockInfo.BlockExpression);
 
   // Otherwise, we have to emit this as a local block.
 
   Address blockAddr = blockInfo.LocalAddress;
// Initialize the block header.
     } else {
       addHeaderField(
           llvm::ConstantInt::get(IntTy, blockInfo.BlockSize.getQuantity()),
           getIntSize(), "block.size");
       addHeaderField(
           llvm::ConstantInt::get(IntTy, blockInfo.BlockAlign.getQuantity()),
           getIntSize(), "block.align");
     }
-    addHeaderField(blockFn, GenVoidPtrSize, "block.invoke");
-    if (!IsOpenCL)
+    if (!IsOpenCL) {
+      addHeaderField(llvm::ConstantExpr::getBitCast(InvokeFn, VoidPtrTy),
+                     getPointerSize(), "block.invoke");
       addHeaderField(descriptor, getPointerSize(), "block.descriptor");
-    else if (auto *Helper =
+    } else if (auto *Helper =
                    CGM.getTargetCodeGenInfo().getTargetOpenCLBlockHelper()) {
       for (auto I : Helper->getCustomFieldValues(*this, blockInfo)) {
         addHeaderField(
             I.first,
             CharUnits::fromQuantity(
                 CGM.getDataLayout().getTypeAllocSize(I.first->getType())),
             I.second);
       }
     }
llvm::Type *CodeGenModule::getBlockDescriptorType() {
   unsigned AddrSpace = 0;
   if (getLangOpts().OpenCL)
     AddrSpace = getContext().getTargetAddressSpace(LangAS::opencl_constant);
   BlockDescriptorType = llvm::PointerType::get(BlockDescriptorType, AddrSpace);
   return BlockDescriptorType;
 }
 
 llvm::Type *CodeGenModule::getGenericBlockLiteralType() {
+  assert(!getLangOpts().OpenCL && "OpenCL does not need this");
+
   if (GenericBlockLiteralType)
     return GenericBlockLiteralType;
 
   llvm::Type *BlockDescPtrTy = getBlockDescriptorType();
 
-  if (getLangOpts().OpenCL) {
-    // struct __opencl_block_literal_generic {
-    //   int __size;
-    //   int __align;
-    //   __generic void *__invoke;
-    //   /* custom fields */
-    // };
-    SmallVector<llvm::Type *, 8> StructFields(
-        {IntTy, IntTy, getOpenCLRuntime().getGenericVoidPointerType()});
-    if (auto *Helper = getTargetCodeGenInfo().getTargetOpenCLBlockHelper()) {
-      for (auto I : Helper->getCustomFieldTypes())
-        StructFields.push_back(I);
-    }
-    GenericBlockLiteralType = llvm::StructType::create(
-        StructFields, "struct.__opencl_block_literal_generic");
-  } else {
-    // struct __block_literal_generic {
-    //   void *__isa;
-    //   int __flags;
-    //   int __reserved;
-    //   void (*__invoke)(void *);
-    //   struct __block_descriptor *__descriptor;
-    // };
-    GenericBlockLiteralType =
-        llvm::StructType::create("struct.__block_literal_generic", VoidPtrTy,
-                                 IntTy, IntTy, VoidPtrTy, BlockDescPtrTy);
-  }
 
   return GenericBlockLiteralType;
 }
 
 RValue CodeGenFunction::EmitBlockCallExpr(const CallExpr *E,
                                           ReturnValueSlot ReturnValue) {
   const BlockPointerType *BPT =
     E->getCallee()->getType()->getAs<BlockPointerType>();
 
   llvm::Value *BlockPtr = EmitScalarExpr(E->getCallee());
+  llvm::Value *FuncPtr;
 
+  if (!CGM.getLangOpts().OpenCL) {
     // Get a pointer to the generic block literal.
-  // For OpenCL we generate generic AS void ptr to be able to reuse the same
-  // block definition for blocks with captures generated as private AS local
-  // variables and without captures generated as global AS program scope
-  // variables.
-  unsigned AddrSpace = 0;
-  if (getLangOpts().OpenCL)
-    AddrSpace = getContext().getTargetAddressSpace(LangAS::opencl_generic);
-
     llvm::Type *BlockLiteralTy =
-      llvm::PointerType::get(CGM.getGenericBlockLiteralType(), AddrSpace);
+        llvm::PointerType::get(CGM.getGenericBlockLiteralType(), 0);
 
     // Bitcast the callee to a block literal.
     BlockPtr =
         Builder.CreatePointerCast(BlockPtr, BlockLiteralTy, "block.literal");
 
     // Get the function pointer from the literal.
-  llvm::Value *FuncPtr =
-      Builder.CreateStructGEP(CGM.getGenericBlockLiteralType(), BlockPtr,
-                              CGM.getLangOpts().OpenCL ? 2 : 3);
+    FuncPtr =
+        Builder.CreateStructGEP(CGM.getGenericBlockLiteralType(), BlockPtr, 3);
+  }
 
   // Add the block literal.
   CallArgList Args;
 
   QualType VoidPtrQualTy = getContext().VoidPtrTy;
   llvm::Type *GenericVoidPtrTy = VoidPtrTy;
   if (getLangOpts().OpenCL) {
     GenericVoidPtrTy = CGM.getOpenCLRuntime().getGenericVoidPointerType();
     VoidPtrQualTy =
         getContext().getPointerType(getContext().getAddrSpaceQualType(
             getContext().VoidTy, LangAS::opencl_generic));
   }
 
   BlockPtr = Builder.CreatePointerCast(BlockPtr, GenericVoidPtrTy);
   Args.add(RValue::get(BlockPtr), VoidPtrQualTy);
 
   QualType FnType = BPT->getPointeeType();
 
   // And the rest of the arguments.
   EmitCallArgs(Args, FnType->getAs<FunctionProtoType>(), E->arguments());
 
   // Load the function.
-  llvm::Value *Func = Builder.CreateAlignedLoad(FuncPtr, getPointerAlign());
+  llvm::Value *Func;
+  if (CGM.getLangOpts().OpenCL)
+    Func = CGM.getOpenCLRuntime().getInvokeFunction(E->getCallee());
+  else
+    Func = Builder.CreateAlignedLoad(FuncPtr, getPointerAlign());
 
   const FunctionType *FuncTy = FnType->castAs<FunctionType>();
   const CGFunctionInfo &FnInfo =
     CGM.getTypes().arrangeBlockFunctionCall(Args, FuncTy);
 
   // Cast the function pointer to the right type.
   llvm::Type *BlockFTy = CGM.getTypes().GetFunctionType(FnInfo);
 
if (!IsOpenCL) {
     BlockFlags flags = BLOCK_IS_GLOBAL | BLOCK_HAS_SIGNATURE;
     if (blockInfo.UsesStret)
       flags |= BLOCK_USE_STRET;
 
     fields.addInt(CGM.IntTy, flags.getBitMask());
 
     // Reserved
     fields.addInt(CGM.IntTy, 0);
+
+    // Function
+    fields.add(blockFn);
   } else {
     fields.addInt(CGM.IntTy, blockInfo.BlockSize.getQuantity());
     fields.addInt(CGM.IntTy, blockInfo.BlockAlign.getQuantity());
   }
 
-  // Function
-  fields.add(blockFn);
-
   if (!IsOpenCL) {
     // Descriptor
     fields.add(buildBlockDescriptor(CGM, blockInfo));
   } else if (auto *Helper =
                  CGM.getTargetCodeGenInfo().getTargetOpenCLBlockHelper()) {
     for (auto I : Helper->getCustomFieldValues(CGM, blockInfo)) {
       fields.add(I);
     }

lib/CodeGen/CGOpenCLRuntime.h

public:
   /// \brief Record invoke function and block literal emitted during normal
   /// codegen for a block expression. The information is used by
   /// emitOpenCLEnqueuedBlock to emit wrapper kernel.
   ///
   /// \param InvokeF invoke function emitted for the block expression.
   /// \param Block block literal emitted for the block expression.
   void recordBlockInfo(const BlockExpr *E, llvm::Function *InvokeF,
                        llvm::Value *Block);
+
+  /// \return LLVM block invoke function emitted for an expression derived from
+  /// the block expression.
+  llvm::Function *getInvokeFunction(const Expr *E);
 };
 
 }
 }
 
 #endif

lib/CodeGen/CGOpenCLRuntime.cpp

 
 llvm::PointerType *CGOpenCLRuntime::getGenericVoidPointerType() {
   assert(CGM.getLangOpts().OpenCL);
   return llvm::IntegerType::getInt8PtrTy(
       CGM.getLLVMContext(),
       CGM.getContext().getTargetAddressSpace(LangAS::opencl_generic));
 }
 
+// Get the block literal from an expression derived from the block expression.
+// OpenCL v2.0 s6.12.5:
+// Block variable declarations are implicitly qualified with const. Therefore
+// all block variables must be initialized at declaration time and may not be
+// reassigned.
+static const BlockExpr *getBlockExpr(const Expr *E) {
+  if (auto Cast = dyn_cast<CastExpr>(E)) {
+    E = Cast->getSubExpr();
+  }
+  if (auto DR = dyn_cast<DeclRefExpr>(E)) {
+    E = cast<VarDecl>(DR->getDecl())->getInit();
+  }
+  E = E->IgnoreImplicit();
+  if (auto Cast = dyn_cast<CastExpr>(E)) {
+    E = Cast->getSubExpr();
+  }
+  return cast<BlockExpr>(E);
+}
+
 /// Record emitted llvm invoke function and llvm block literal for the
 /// corresponding block expression.
 void CGOpenCLRuntime::recordBlockInfo(const BlockExpr *E,
                                       llvm::Function *InvokeF,
                                       llvm::Value *Block) {
   assert(EnqueuedBlockMap.find(E) == EnqueuedBlockMap.end() &&
          "Block expression emitted twice");
   assert(isa<llvm::Function>(InvokeF) && "Invalid invoke function");
   assert(Block->getType()->isPointerTy() && "Invalid block literal type");
   EnqueuedBlockMap[E].InvokeFunc = InvokeF;
   EnqueuedBlockMap[E].BlockArg = Block;
   EnqueuedBlockMap[E].Kernel = nullptr;
 }
 
+llvm::Function *CGOpenCLRuntime::getInvokeFunction(const Expr *E) {
+  return EnqueuedBlockMap[getBlockExpr(E)].InvokeFunc;
+}
+
 CGOpenCLRuntime::EnqueuedBlockInfo
 CGOpenCLRuntime::emitOpenCLEnqueuedBlock(CodeGenFunction &CGF, const Expr *E) {
   CGF.EmitScalarExpr(E);
 
-  // The block literal may be assigned to a const variable. Chasing down
-  // to get the block literal.
-  if (auto DR = dyn_cast<DeclRefExpr>(E)) {
-    E = cast<VarDecl>(DR->getDecl())->getInit();
-  }
-  E = E->IgnoreImplicit();
-  if (auto Cast = dyn_cast<CastExpr>(E)) {
-    E = Cast->getSubExpr();
-  }
-  auto *Block = cast<BlockExpr>(E);
-
+  const BlockExpr *Block = getBlockExpr(E);
   assert(EnqueuedBlockMap.find(Block) != EnqueuedBlockMap.end() &&
          "Block expression not emitted");
 
   // Do not emit the block wrapper again if it has been emitted.
   if (EnqueuedBlockMap[Block].Kernel) {
     return EnqueuedBlockMap[Block];
   }
 

test/CodeGenOpenCL/amdgpu-enqueue-kernel.cl

-// RUN: %clang_cc1 %s -cl-std=CL2.0 -O0 -emit-llvm -o - -triple amdgcn | FileCheck %s --check-prefix=CHECK
+// RUN: %clang_cc1 %s -cl-std=CL2.0 -O0 -emit-llvm -o - -triple amdgcn | FileCheck %s
 
 typedef struct {int a;} ndrange_t;
 
 void callee(long id, global long *out) {
   out[id] = id;
 }
 
 // CHECK-LABEL: define amdgpu_kernel void @test
kernel void test(global char *a, char b, global long *c, long d) {
 
   void (^block)(void) = ^{
     callee(d, c);
   };
 
   enqueue_kernel(default_queue, flags, ndrange, block);
 }
 
-// CHECK-LABEL: define internal amdgpu_kernel void @__test_block_invoke_kernel(<{ i32, i32, i8*, i8 addrspace(1)*, i8 }>)
+// CHECK-LABEL: define internal amdgpu_kernel void @__test_block_invoke_kernel(<{ i32, i32, i8 addrspace(1)*, i8 }>)
 // CHECK-SAME: #[[ATTR:[0-9]+]] !kernel_arg_addr_space !{{.*}} !kernel_arg_access_qual !{{.*}} !kernel_arg_type !{{.*}} !kernel_arg_base_type !{{.*}} !kernel_arg_type_qual !{{.*}}
 // CHECK: entry:
-// CHECK:  %1 = alloca <{ i32, i32, i8*, i8 addrspace(1)*, i8 }>, align 8, addrspace(5)
-// CHECK:  store <{ i32, i32, i8*, i8 addrspace(1)*, i8 }> %0, <{ i32, i32, i8*, i8 addrspace(1)*, i8 }> addrspace(5)* %1, align 8
-// CHECK:  %2 = addrspacecast <{ i32, i32, i8*, i8 addrspace(1)*, i8 }> addrspace(5)* %1 to i8*
+// CHECK:  %1 = alloca <{ i32, i32, i8 addrspace(1)*, i8 }>, align 8, addrspace(5)
+// CHECK:  store <{ i32, i32, i8 addrspace(1)*, i8 }> %0, <{ i32, i32, i8 addrspace(1)*, i8 }> addrspace(5)* %1, align 8
+// CHECK:  %2 = addrspacecast <{ i32, i32, i8 addrspace(1)*, i8 }> addrspace(5)* %1 to i8*
 // CHECK:  call void @__test_block_invoke(i8* %2)
 // CHECK:  ret void
 // CHECK:}
 
-// CHECK-LABEL: define internal amdgpu_kernel void @__test_block_invoke_2_kernel(<{ i32, i32, i8*, i8 addrspace(1)*, i64 addrspace(1)*, i64, i8 }>)
+// CHECK-LABEL: define internal amdgpu_kernel void @__test_block_invoke_2_kernel(<{ i32, i32, i8 addrspace(1)*, i64 addrspace(1)*, i64, i8 }>)
 // CHECK-SAME: #[[ATTR]] !kernel_arg_addr_space !{{.*}} !kernel_arg_access_qual !{{.*}} !kernel_arg_type !{{.*}} !kernel_arg_base_type !{{.*}} !kernel_arg_type_qual !{{.*}}
 
-// CHECK-LABEL: define internal amdgpu_kernel void @__test_block_invoke_3_kernel(<{ i32, i32, i8*, i8 addrspace(1)*, i64 addrspace(1)*, i64, i8 }>, i8 addrspace(3)*)
+// CHECK-LABEL: define internal amdgpu_kernel void @__test_block_invoke_3_kernel(<{ i32, i32, i8 addrspace(1)*, i64 addrspace(1)*, i64, i8 }>, i8 addrspace(3)*)
 // CHECK-SAME: #[[ATTR]] !kernel_arg_addr_space !{{.*}} !kernel_arg_access_qual !{{.*}} !kernel_arg_type !{{.*}} !kernel_arg_base_type !{{.*}} !kernel_arg_type_qual !{{.*}}
 
-// CHECK-LABEL: define internal amdgpu_kernel void @__test_block_invoke_4_kernel(<{ i32, i32, i8*, i64, i64 addrspace(1)* }>)
+// CHECK-LABEL: define internal amdgpu_kernel void @__test_block_invoke_4_kernel(<{ i32, i32, i64, i64 addrspace(1)* }>)
 // CHECK-SAME: #[[ATTR]] !kernel_arg_addr_space !{{.*}} !kernel_arg_access_qual !{{.*}} !kernel_arg_type !{{.*}} !kernel_arg_base_type !{{.*}} !kernel_arg_type_qual !{{.*}}
 
 // CHECK: attributes #[[ATTR]] = { nounwind "enqueued-block" }

test/CodeGenOpenCL/blocks.cl

 // RUN: %clang_cc1 %s -cl-std=CL2.0 -emit-llvm -o - -O0 -triple spir-unknown-unknown | FileCheck -check-prefixes=COMMON,SPIR %s
 // RUN: %clang_cc1 %s -cl-std=CL2.0 -emit-llvm -o - -O0 -triple amdgcn-amd-amdhsa | FileCheck -check-prefixes=COMMON,AMDGCN %s
 
-// SPIR: %struct.__opencl_block_literal_generic = type { i32, i32, i8 addrspace(4)* }
-// AMDGCN: %struct.__opencl_block_literal_generic = type { i32, i32, i8* }
-// SPIR: @__block_literal_global = internal addrspace(1) constant { i32, i32, i8 addrspace(4)* } { i32 12, i32 4, i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*, i8 addrspace(3)*)* @block_A_block_invoke to i8*) to i8 addrspace(4)*) }
-// AMDGCN: @__block_literal_global = internal addrspace(1) constant { i32, i32, i8* } { i32 16, i32 8, i8* bitcast (void (i8*, i8 addrspace(3)*)* @block_A_block_invoke to i8*) }
+// COMMON: @__block_literal_global = internal addrspace(1) constant { i32, i32 } { i32 8, i32 4 }
 // COMMON-NOT: .str
 
 // SPIR-LABEL: define internal {{.*}}void @block_A_block_invoke(i8 addrspace(4)* %.block_descriptor, i8 addrspace(3)* %a)
 // AMDGCN-LABEL: define internal {{.*}}void @block_A_block_invoke(i8* %.block_descriptor, i8 addrspace(3)* %a)
 void (^block_A)(local void *) = ^(local void *a) {
   return;
 };
 
 // COMMON-LABEL: define {{.*}}void @foo()
 void foo(){
   int i;
   // COMMON-NOT: %block.isa
   // COMMON-NOT: %block.flags
   // COMMON-NOT: %block.reserved
   // COMMON-NOT: %block.descriptor
-  // SPIR: %[[block_size:.*]] = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 }>, <{ i32, i32, i8 addrspace(4)*, i32 }>* %block, i32 0, i32 0
-  // AMDGCN: %[[block_size:.*]] = getelementptr inbounds <{ i32, i32, i8*, i32 }>, <{ i32, i32, i8*, i32 }> addrspace(5)* %block, i32 0, i32 0
-  // SPIR: store i32 16, i32* %[[block_size]]
-  // AMDGCN: store i32 20, i32 addrspace(5)* %[[block_size]]
-  // SPIR: %[[block_align:.*]] = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 }>, <{ i32, i32, i8 addrspace(4)*, i32 }>* %block, i32 0, i32 1
-  // AMDGCN: %[[block_align:.*]] = getelementptr inbounds <{ i32, i32, i8*, i32 }>, <{ i32, i32, i8*, i32 }> addrspace(5)* %block, i32 0, i32 1
+  // SPIR: %[[block_size:.*]] = getelementptr inbounds <{ i32, i32, i32 }>, <{ i32, i32, i32 }>* %[[block:.*]], i32 0, i32 0
+  // AMDGCN: %[[block_size:.*]] = getelementptr inbounds <{ i32, i32, i32 }>, <{ i32, i32, i32 }> addrspace(5)* %[[block:.*]], i32 0, i32 0
+  // SPIR: store i32 12, i32* %[[block_size]]
+  // AMDGCN: store i32 12, i32 addrspace(5)* %[[block_size]]
+  // SPIR: %[[block_align:.*]] = getelementptr inbounds <{ i32, i32, i32 }>, <{ i32, i32, i32 }>* %[[block]], i32 0, i32 1
+  // AMDGCN: %[[block_align:.*]] = getelementptr inbounds <{ i32, i32, i32 }>, <{ i32, i32, i32 }> addrspace(5)* %[[block]], i32 0, i32 1
   // SPIR: store i32 4, i32* %[[block_align]]
-  // AMDGCN: store i32 8, i32 addrspace(5)* %[[block_align]]
-  // SPIR: %[[block_invoke:.*]] = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 }>, <{ i32, i32, i8 addrspace(4)*, i32 }>* %[[block:.*]], i32 0, i32 2
-  // SPIR: store i8 addrspace(4)* addrspacecast (i8* bitcast (i32 (i8 addrspace(4)*)* @__foo_block_invoke to i8*) to i8 addrspace(4)*), i8 addrspace(4)** %[[block_invoke]]
-  // SPIR: %[[block_captured:.*]] = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 }>, <{ i32, i32, i8 addrspace(4)*, i32 }>* %[[block]], i32 0, i32 3
+  // AMDGCN: store i32 4, i32 addrspace(5)* %[[block_align]]
+  // SPIR: %[[block_captured:.*]] = getelementptr inbounds <{ i32, i32, i32 }>, <{ i32, i32, i32 }>* %[[block]], i32 0, i32 2
   // SPIR: %[[i_value:.*]] = load i32, i32* %i
   // SPIR: store i32 %[[i_value]], i32* %[[block_captured]],
-  // SPIR: %[[blk_ptr:.*]] = bitcast <{ i32, i32, i8 addrspace(4)*, i32 }>* %[[block]] to i32 ()*
+  // SPIR: %[[blk_ptr:.*]] = bitcast <{ i32, i32, i32 }>* %[[block]] to i32 ()*
   // SPIR: %[[blk_gen_ptr:.*]] = addrspacecast i32 ()* %[[blk_ptr]] to i32 () addrspace(4)*
   // SPIR: store i32 () addrspace(4)* %[[blk_gen_ptr]], i32 () addrspace(4)** %[[block_B:.*]],
-  // SPIR: %[[blk_gen_ptr:.*]] = load i32 () addrspace(4)*, i32 () addrspace(4)** %[[block_B]]
-  // SPIR: %[[block_literal:.*]] = bitcast i32 () addrspace(4)* %[[blk_gen_ptr]] to %struct.__opencl_block_literal_generic addrspace(4)*
-  // SPIR: %[[invoke_addr:.*]] = getelementptr inbounds %struct.__opencl_block_literal_generic, %struct.__opencl_block_literal_generic addrspace(4)* %[[block_literal]], i32 0, i32 2
-  // SPIR: %[[blk_gen_ptr:.*]] = bitcast %struct.__opencl_block_literal_generic addrspace(4)* %[[block_literal]] to i8 addrspace(4)*
-  // SPIR: %[[invoke_func_ptr:.*]] = load i8 addrspace(4)*, i8 addrspace(4)* addrspace(4)* %[[invoke_addr]]
-  // SPIR: %[[invoke_func:.*]] = addrspacecast i8 addrspace(4)* %[[invoke_func_ptr]] to i32 (i8 addrspace(4)*)*
-  // SPIR: call {{.*}}i32 %[[invoke_func]](i8 addrspace(4)* %[[blk_gen_ptr]])
-  // AMDGCN: %[[block_invoke:.*]] = getelementptr inbounds <{ i32, i32, i8*, i32 }>, <{ i32, i32, i8*, i32 }> addrspace(5)* %[[block:.*]], i32 0, i32 2
-  // AMDGCN: store i8* bitcast (i32 (i8*)* @__foo_block_invoke to i8*), i8* addrspace(5)* %[[block_invoke]]
-  // AMDGCN: %[[block_captured:.*]] = getelementptr inbounds <{ i32, i32, i8*, i32 }>, <{ i32, i32, i8*, i32 }> addrspace(5)* %[[block]], i32 0, i32 3
+  // SPIR: %[[block_literal:.*]] = load i32 () addrspace(4)*, i32 () addrspace(4)** %[[block_B]]
+  // SPIR: %[[blk_gen_ptr:.*]] = bitcast i32 () addrspace(4)* %[[block_literal]] to i8 addrspace(4)*
+  // SPIR: call {{.*}}i32 @__foo_block_invoke(i8 addrspace(4)* %[[blk_gen_ptr]])
+  // AMDGCN: %[[block_captured:.*]] = getelementptr inbounds <{ i32, i32, i32 }>, <{ i32, i32, i32 }> addrspace(5)* %[[block]], i32 0, i32 2
   // AMDGCN: %[[i_value:.*]] = load i32, i32 addrspace(5)* %i
   // AMDGCN: store i32 %[[i_value]], i32 addrspace(5)* %[[block_captured]],
-  // AMDGCN: %[[blk_ptr:.*]] = bitcast <{ i32, i32, i8*, i32 }> addrspace(5)* %[[block]] to i32 () addrspace(5)*
+  // AMDGCN: %[[blk_ptr:.*]] = bitcast <{ i32, i32, i32 }> addrspace(5)* %[[block]] to i32 () addrspace(5)*
   // AMDGCN: %[[blk_gen_ptr:.*]] = addrspacecast i32 () addrspace(5)* %[[blk_ptr]] to i32 ()*
   // AMDGCN: store i32 ()* %[[blk_gen_ptr]], i32 ()* addrspace(5)* %[[block_B:.*]],
-  // AMDGCN: %[[blk_gen_ptr:.*]] = load i32 ()*, i32 ()* addrspace(5)* %[[block_B]]
-  // AMDGCN: %[[block_literal:.*]] = bitcast i32 ()* %[[blk_gen_ptr]] to %struct.__opencl_block_literal_generic*
-  // AMDGCN: %[[invoke_addr:.*]] = getelementptr inbounds %struct.__opencl_block_literal_generic, %struct.__opencl_block_literal_generic* %[[block_literal]], i32 0, i32 2
-  // AMDGCN: %[[blk_gen_ptr:.*]] = bitcast %struct.__opencl_block_literal_generic* %[[block_literal]] to i8*
-  // AMDGCN: %[[invoke_func_ptr:.*]] = load i8*, i8** %[[invoke_addr]]
-  // AMDGCN: %[[invoke_func:.*]] = bitcast i8* %[[invoke_func_ptr]] to i32 (i8*)*
-  // AMDGCN: call {{.*}}i32 %[[invoke_func]](i8* %[[blk_gen_ptr]])
+  // AMDGCN: %[[block_literal:.*]] = load i32 ()*, i32 ()* addrspace(5)* %[[block_B]]
+  // AMDGCN: %[[blk_gen_ptr:.*]] = bitcast i32 ()* %[[block_literal]] to i8*
+  // AMDGCN: call {{.*}}i32 @__foo_block_invoke(i8* %[[blk_gen_ptr]])
 
   int (^ block_B)(void) = ^{
     return i;
   };
   block_B();
 }
 
 // SPIR-LABEL: define internal {{.*}}i32 @__foo_block_invoke(i8 addrspace(4)* %.block_descriptor)
-// SPIR:  %[[block:.*]] = bitcast i8 addrspace(4)* %.block_descriptor to <{ i32, i32, i8 addrspace(4)*, i32 }> addrspace(4)*
-// SPIR:  %[[block_capture_addr:.*]] = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 }>, <{ i32, i32, i8 addrspace(4)*, i32 }> addrspace(4)* %[[block]], i32 0, i32 3
+// SPIR:  %[[block:.*]] = bitcast i8 addrspace(4)* %.block_descriptor to <{ i32, i32, i32 }> addrspace(4)*
+// SPIR:  %[[block_capture_addr:.*]] = getelementptr inbounds <{ i32, i32, i32 }>, <{ i32, i32, i32 }> addrspace(4)* %[[block]], i32 0, i32 2
 // SPIR:  %[[block_capture:.*]] = load i32, i32 addrspace(4)* %[[block_capture_addr]]
 // AMDGCN-LABEL: define internal {{.*}}i32 @__foo_block_invoke(i8* %.block_descriptor)
-// AMDGCN:  %[[block:.*]] = bitcast i8* %.block_descriptor to <{ i32, i32, i8*, i32 }>*
-// AMDGCN:  %[[block_capture_addr:.*]] = getelementptr inbounds <{ i32, i32, i8*, i32 }>, <{ i32, i32, i8*, i32 }>* %[[block]], i32 0, i32 3
+// AMDGCN:  %[[block:.*]] = bitcast i8* %.block_descriptor to <{ i32, i32, i32 }>*
+// AMDGCN:  %[[block_capture_addr:.*]] = getelementptr inbounds <{ i32, i32, i32 }>, <{ i32, i32, i32 }>* %[[block]], i32 0, i32 2
 // AMDGCN:  %[[block_capture:.*]] = load i32, i32* %[[block_capture_addr]]
 
 // COMMON-NOT: define{{.*}}@__foo_block_invoke_kernel

test/CodeGenOpenCL/cl20-device-side-enqueue.cl

 // RUN: %clang_cc1 %s -cl-std=CL2.0 -ffake-address-space-map -O0 -emit-llvm -o - -triple "spir-unknown-unknown" | FileCheck %s --check-prefix=COMMON --check-prefix=B32
 // 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
 
 #pragma OPENCL EXTENSION cl_khr_subgroups : enable
 
 typedef void (^bl_t)(local void *);
 typedef struct {int a;} ndrange_t;
 
-// COMMON: %struct.__opencl_block_literal_generic = type { i32, i32, i8 addrspace(4)* }
-
 // For a block global variable, first emit the block literal as a global variable, then emit the block variable itself.
-// COMMON: [[BL_GLOBAL:@__block_literal_global[^ ]*]] = internal addrspace(1) constant { i32, i32, i8 addrspace(4)* } { i32 {{[0-9]+}}, i32 {{[0-9]+}}, i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*, i8 addrspace(3)*)* [[INV_G:@[^ ]+]] to i8*) to i8 addrspace(4)*) }
-// COMMON: @block_G =  addrspace(1) constant void (i8 addrspace(3)*) addrspace(4)* addrspacecast (void (i8 addrspace(3)*) addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BL_GLOBAL]] to void (i8 addrspace(3)*) addrspace(1)*) to void (i8 addrspace(3)*) addrspace(4)*)
+// COMMON: [[BL_GLOBAL:@__block_literal_global[^ ]*]] = internal addrspace(1) constant { i32, i32 } { i32 {{[0-9]+}}, i32 {{[0-9]+}} }
+// COMMON: @block_G =  addrspace(1) constant void (i8 addrspace(3)*) addrspace(4)* addrspacecast (void (i8 addrspace(3)*) addrspace(1)* bitcast ({ i32, i32 } addrspace(1)* [[BL_GLOBAL]] to void (i8 addrspace(3)*) addrspace(1)*) to void (i8 addrspace(3)*) addrspace(4)*)
 
 // For anonymous blocks without captures, emit block literals as global variable.
-// COMMON: [[BLG1:@__block_literal_global[^ ]*]] = internal addrspace(1) constant { i32, i32, i8 addrspace(4)* } { i32 {{[0-9]+}}, i32 {{[0-9]+}}, i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*, i8 addrspace(3)*)* {{@[^ ]+}} to i8*) to i8 addrspace(4)*) }
-// COMMON: [[BLG2:@__block_literal_global[^ ]*]] = internal addrspace(1) constant { i32, i32, i8 addrspace(4)* } { i32 {{[0-9]+}}, i32 {{[0-9]+}}, i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*, i8 addrspace(3)*)* {{@[^ ]+}} to i8*) to i8 addrspace(4)*) }
-// COMMON: [[BLG3:@__block_literal_global[^ ]*]] = internal addrspace(1) constant { i32, i32, i8 addrspace(4)* } { i32 {{[0-9]+}}, i32 {{[0-9]+}}, i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*, i8 addrspace(3)*)* {{@[^ ]+}} to i8*) to i8 addrspace(4)*) }
-// COMMON: [[BLG4:@__block_literal_global[^ ]*]] = internal addrspace(1) constant { i32, i32, i8 addrspace(4)* } { i32 {{[0-9]+}}, i32 {{[0-9]+}}, i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*, i8 addrspace(3)*)* {{@[^ ]+}} to i8*) to i8 addrspace(4)*) }
-// COMMON: [[BLG5:@__block_literal_global[^ ]*]] = internal addrspace(1) constant { i32, i32, i8 addrspace(4)* } { i32 {{[0-9]+}}, i32 {{[0-9]+}}, i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*, i8 addrspace(3)*)* {{@[^ ]+}} to i8*) to i8 addrspace(4)*) }
-// COMMON: [[BLG6:@__block_literal_global[^ ]*]] = internal addrspace(1) constant { i32, i32, i8 addrspace(4)* } { i32 {{[0-9]+}}, i32 {{[0-9]+}}, i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*, i8 addrspace(3)*, i8 addrspace(3)*, i8 addrspace(3)*)* {{@[^ ]+}} to i8*) to i8 addrspace(4)*) }
-// COMMON: [[BLG7:@__block_literal_global[^ ]*]] = internal addrspace(1) constant { i32, i32, i8 addrspace(4)* } { i32 {{[0-9]+}}, i32 {{[0-9]+}}, i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*, i8 addrspace(3)*)* {{@[^ ]+}} to i8*) to i8 addrspace(4)*) }
-// COMMON: [[BLG8:@__block_literal_global[^ ]*]] = internal addrspace(1) constant { i32, i32, i8 addrspace(4)* } { i32 {{[0-9]+}}, i32 {{[0-9]+}}, i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*)* [[INVG8:@[^ ]+]] to i8*) to i8 addrspace(4)*) }
-// COMMON: [[BLG9:@__block_literal_global[^ ]*]] = internal addrspace(1) constant { i32, i32, i8 addrspace(4)* } { i32 {{[0-9]+}}, i32 {{[0-9]+}}, i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*, i8 addrspace(3)*)* [[INVG9:@[^ ]+]] to i8*) to i8 addrspace(4)*) }
-// COMMON: [[BLG10:@__block_literal_global[^ ]*]] = internal addrspace(1) constant { i32, i32, i8 addrspace(4)* } { i32 {{[0-9]+}}, i32 {{[0-9]+}}, i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*)* {{@[^ ]+}} to i8*) to i8 addrspace(4)*) }
-// COMMON: [[BLG11:@__block_literal_global[^ ]*]] = internal addrspace(1) constant { i32, i32, i8 addrspace(4)* } { i32 {{[0-9]+}}, i32 {{[0-9]+}}, i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*)* {{@[^ ]+}} to i8*) to i8 addrspace(4)*) }
+// COMMON: [[BLG1:@__block_literal_global[^ ]*]] = internal addrspace(1) constant { i32, i32 } { i32 {{[0-9]+}}, i32 {{[0-9]+}} }
+// COMMON: [[BLG2:@__block_literal_global[^ ]*]] = internal addrspace(1) constant { i32, i32 } { i32 {{[0-9]+}}, i32 {{[0-9]+}} }
+// COMMON: [[BLG3:@__block_literal_global[^ ]*]] = internal addrspace(1) constant { i32, i32 } { i32 {{[0-9]+}}, i32 {{[0-9]+}} }
+// COMMON: [[BLG4:@__block_literal_global[^ ]*]] = internal addrspace(1) constant { i32, i32 } { i32 {{[0-9]+}}, i32 {{[0-9]+}} }
+// COMMON: [[BLG5:@__block_literal_global[^ ]*]] = internal addrspace(1) constant { i32, i32 } { i32 {{[0-9]+}}, i32 {{[0-9]+}} }
+// COMMON: [[BLG6:@__block_literal_global[^ ]*]] = internal addrspace(1) constant { i32, i32 } { i32 {{[0-9]+}}, i32 {{[0-9]+}} }
+// COMMON: [[BLG7:@__block_literal_global[^ ]*]] = internal addrspace(1) constant { i32, i32 } { i32 {{[0-9]+}}, i32 {{[0-9]+}} }
+// COMMON: [[BLG8:@__block_literal_global[^ ]*]] = internal addrspace(1) constant { i32, i32 } { i32 {{[0-9]+}}, i32 {{[0-9]+}} }
+// COMMON: [[BLG9:@__block_literal_global[^ ]*]] = internal addrspace(1) constant { i32, i32 } { i32 {{[0-9]+}}, i32 {{[0-9]+}} }
+// COMMON: [[BLG10:@__block_literal_global[^ ]*]] = internal addrspace(1) constant { i32, i32 } { i32 {{[0-9]+}}, i32 {{[0-9]+}} }
+// COMMON: [[BLG11:@__block_literal_global[^ ]*]] = internal addrspace(1) constant { i32, i32 } { i32 {{[0-9]+}}, i32 {{[0-9]+}} }
 
 // Emits block literal [[BL_GLOBAL]], invoke function [[INV_G]] and global block variable @block_G
-// COMMON: define internal spir_func void [[INV_G]](i8 addrspace(4)* %{{.*}}, i8 addrspace(3)* %{{.*}})
+// COMMON: define internal spir_func void [[INV_G:.*]](i8 addrspace(4)* %{{.*}}, i8 addrspace(3)* %{{.*}})
 const bl_t block_G = (bl_t) ^ (local void *a) {};
 
 void callee(int id, __global int *out) {
   out[id] = id;
 }
 
 // COMMON-LABEL: define spir_kernel void @device_side_enqueue(i32 addrspace(1)* %{{.*}}, i32 addrspace(1)* %b, i32 %i)
 kernel void device_side_enqueue(global int *a, global int *b, int i) {
kernel void device_side_enqueue(global int *a, global int *b, int i) {
   clk_event_t event_wait_list;
   // COMMON: %event_wait_list2 = alloca [1 x %opencl.clk_event_t*]
   clk_event_t event_wait_list2[] = {clk_event};
 
   // Emits block literal on stack and block kernel [[INVLK1]].
   // 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: store i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*)* [[INVL1:@__device_side_enqueue_block_invoke[^ ]*]] to i8*) to i8 addrspace(4)*), i8 addrspace(4)** %block.invoke
-  // B32: [[BL:%[0-9]+]] = bitcast <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }>* %block to void ()*
-  // B64: [[BL:%[0-9]+]] = bitcast <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32 addrspace(1)*, i32 }>* %block to void ()*
+  // B32: [[BL:%[0-9]+]] = bitcast <{ i32, i32, i32 addrspace(1)*, i32, i32 addrspace(1)* }>* %block to void ()*
+  // B64: [[BL:%[0-9]+]] = bitcast <{ i32, i32, i32 addrspace(1)*, i32 addrspace(1)*, i32 }>* %block to void ()*
   // COMMON: [[BL_I8:%[0-9]+]] = addrspacecast void ()* [[BL]] to i8 addrspace(4)*
   // COMMON-LABEL: call i32 @__enqueue_kernel_basic(
   // COMMON-SAME: %opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %struct.ndrange_t* byval [[NDR]]{{([0-9]+)?}},
   // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8* bitcast ({{.*}} [[INVLK1:[^ ]+_kernel]] to i8*) to i8 addrspace(4)*),
   // COMMON-SAME: i8 addrspace(4)* [[BL_I8]])
   enqueue_kernel(default_queue, flags, ndrange,
                  ^(void) {
                    a[i] = b[i];
                  });
 
   // Emits block literal on stack and block kernel [[INVLK2]].
   // COMMON: [[DEF_Q:%[0-9]+]] = load %opencl.queue_t{{.*}}*, %opencl.queue_t{{.*}}** %default_queue
   // COMMON: [[FLAGS:%[0-9]+]] = load i32, i32* %flags
   // 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: store i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*)* [[INVL2:@__device_side_enqueue_block_invoke[^ ]*]] to i8*) to i8 addrspace(4)*), i8 addrspace(4)** %block.invoke
-  // COMMON: [[BL:%[0-9]+]] = bitcast <{ i32, i32, i8 addrspace(4)*, i32{{.*}}, i32{{.*}}, i32{{.*}} }>* %block3 to void ()*
+  // COMMON: [[BL:%[0-9]+]] = bitcast <{ i32, i32, i32{{.*}}, i32{{.*}}, i32{{.*}} }>* %block3 to void ()*
   // COMMON: [[BL_I8:%[0-9]+]] = addrspacecast void ()* [[BL]] to i8 addrspace(4)*
   // COMMON-LABEL: call i32 @__enqueue_kernel_basic_events
   // COMMON-SAME: (%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]],
   // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8* bitcast ({{.*}} [[INVLK2:[^ ]+_kernel]] to i8*) to i8 addrspace(4)*),
   // COMMON-SAME: i8 addrspace(4)* [[BL_I8]])
 
   enqueue_kernel(default_queue, flags, ndrange, 2, &event_wait_list, &clk_event,
                  ^(void) {
                    a[i] = b[i];
                  });
 
   // Emits global block literal [[BLG1]] and block kernel [[INVGK1]].
   // COMMON: [[DEF_Q:%[0-9]+]] = load %opencl.queue_t{{.*}}*, %opencl.queue_t{{.*}}** %default_queue
   // COMMON: [[FLAGS:%[0-9]+]] = load i32, i32* %flags
   // B32: %[[TMP:.*]] = alloca [1 x i32]
   // B32: %[[TMP1:.*]] = getelementptr [1 x i32], [1 x i32]* %[[TMP]], i32 0, i32 0
   // B32: store i32 256, i32* %[[TMP1]], align 4
   // B64: %[[TMP:.*]] = alloca [1 x i64]
   // B64: %[[TMP1:.*]] = getelementptr [1 x i64], [1 x i64]* %[[TMP]], i32 0, i32 0
   // B64: store i64 256, i64* %[[TMP1]], align 8
   // COMMON-LABEL: call i32 @__enqueue_kernel_vaargs(
   // COMMON-SAME: %opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %struct.ndrange_t* [[NDR]]{{([0-9]+)?}},
   // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8* bitcast ({{.*}} [[INVGK1:[^ ]+_kernel]] to i8*) to i8 addrspace(4)*),
-  // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BLG1]] to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1,
+  // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32 } addrspace(1)* [[BLG1]] to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1,
   // B32-SAME: i32* %[[TMP1]])
   // B64-SAME: i64* %[[TMP1]])
   enqueue_kernel(default_queue, flags, ndrange,
                  ^(local void *p) {
                    return;
                  },
                  256);
   char c;
   // Emits global block literal [[BLG2]] and block kernel [[INVGK2]].
   // COMMON: [[DEF_Q:%[0-9]+]] = load %opencl.queue_t{{.*}}*, %opencl.queue_t{{.*}}** %default_queue
   // COMMON: [[FLAGS:%[0-9]+]] = load i32, i32* %flags
   // B32: %[[TMP:.*]] = alloca [1 x i32]
   // B32: %[[TMP1:.*]] = getelementptr [1 x i32], [1 x i32]* %[[TMP]], i32 0, i32 0
   // B32: store i32 %{{.*}}, i32* %[[TMP1]], align 4
   // B64: %[[TMP:.*]] = alloca [1 x i64]
   // B64: %[[TMP1:.*]] = getelementptr [1 x i64], [1 x i64]* %[[TMP]], i32 0, i32 0
   // B64: store i64 %{{.*}}, i64* %[[TMP1]], align 8
   // COMMON-LABEL: call i32 @__enqueue_kernel_vaargs(
   // COMMON-SAME: %opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %struct.ndrange_t* [[NDR]]{{([0-9]+)?}},
   // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8* bitcast ({{.*}} [[INVGK2:[^ ]+_kernel]] to i8*) to i8 addrspace(4)*),
-  // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BLG2]] to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1,
+  // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32 } addrspace(1)* [[BLG2]] to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1,
   // B32-SAME: i32* %[[TMP1]])
   // B64-SAME: i64* %[[TMP1]])
   enqueue_kernel(default_queue, flags, ndrange,
                  ^(local void *p) {
                    return;
                  },
                  c);
 
   // Emits global block literal [[BLG3]] and block kernel [[INVGK3]].
   // COMMON: [[DEF_Q:%[0-9]+]] = load %opencl.queue_t{{.*}}*, %opencl.queue_t{{.*}}** %default_queue
   // COMMON: [[FLAGS:%[0-9]+]] = load i32, i32* %flags
   // 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: %[[TMP:.*]] = alloca [1 x i32]
   // B32: %[[TMP1:.*]] = getelementptr [1 x i32], [1 x i32]* %[[TMP]], i32 0, i32 0
   // B32: store i32 256, i32* %[[TMP1]], align 4
   // B64: %[[TMP:.*]] = alloca [1 x i64]
   // B64: %[[TMP1:.*]] = getelementptr [1 x i64], [1 x i64]* %[[TMP]], i32 0, i32 0
   // B64: store i64 256, i64* %[[TMP1]], align 8
   // COMMON-LABEL: call i32 @__enqueue_kernel_events_vaargs
   // COMMON-SAME: (%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]],  %struct.ndrange_t* {{.*}}, i32 2, %opencl.clk_event_t{{.*}} [[WAIT_EVNT]], %opencl.clk_event_t{{.*}} [[EVNT]],
   // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8* bitcast ({{.*}} [[INVGK3:[^ ]+_kernel]] to i8*) to i8 addrspace(4)*),
-  // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BLG3]] to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1,
+  // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32 } addrspace(1)* [[BLG3]] to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1,
   // B32-SAME: i32* %[[TMP1]])
   // B64-SAME: i64* %[[TMP1]])
   enqueue_kernel(default_queue, flags, ndrange, 2, event_wait_list2, &clk_event,
                  ^(local void *p) {
                    return;
                  },
                  256);
 
   // Emits global block literal [[BLG4]] and block kernel [[INVGK4]].
   // COMMON: [[DEF_Q:%[0-9]+]] = load %opencl.queue_t{{.*}}*, %opencl.queue_t{{.*}}** %default_queue
   // COMMON: [[FLAGS:%[0-9]+]] = load i32, i32* %flags
   // 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: %[[TMP:.*]] = alloca [1 x i32]
   // B32: %[[TMP1:.*]] = getelementptr [1 x i32], [1 x i32]* %[[TMP]], i32 0, i32 0
   // B32: store i32 %{{.*}}, i32* %[[TMP1]], align 4
   // B64: %[[TMP:.*]] = alloca [1 x i64]
   // B64: %[[TMP1:.*]] = getelementptr [1 x i64], [1 x i64]* %[[TMP]], i32 0, i32 0
   // B64: store i64 %{{.*}}, i64* %[[TMP1]], align 8
   // COMMON-LABEL: call i32 @__enqueue_kernel_events_vaargs
   // COMMON-SAME: (%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]],
   // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8* bitcast ({{.*}} [[INVGK4:[^ ]+_kernel]] to i8*) to i8 addrspace(4)*),
-  // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BLG4]] to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1,
+  // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32 } addrspace(1)* [[BLG4]] to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1,
   // B32-SAME: i32* %[[TMP1]])
   // B64-SAME: i64* %[[TMP1]])
   enqueue_kernel(default_queue, flags, ndrange, 2, event_wait_list2, &clk_event,
                  ^(local void *p) {
                    return;
                  },
                  c);
 
   long l;
   // Emits global block literal [[BLG5]] and block kernel [[INVGK5]].
   // COMMON: [[DEF_Q:%[0-9]+]] = load %opencl.queue_t{{.*}}*, %opencl.queue_t{{.*}}** %default_queue
   // COMMON: [[FLAGS:%[0-9]+]] = load i32, i32* %flags
   // B32: %[[TMP:.*]] = alloca [1 x i32]
   // B32: %[[TMP1:.*]] = getelementptr [1 x i32], [1 x i32]* %[[TMP]], i32 0, i32 0
   // B32: store i32 %{{.*}}, i32* %[[TMP1]], align 4
   // B64: %[[TMP:.*]] = alloca [1 x i64]
   // B64: %[[TMP1:.*]] = getelementptr [1 x i64], [1 x i64]* %[[TMP]], i32 0, i32 0
   // B64: store i64 %{{.*}}, i64* %[[TMP1]], align 8
   // COMMON-LABEL: call i32 @__enqueue_kernel_vaargs
   // COMMON-SAME: (%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %struct.ndrange_t* [[NDR]]{{([0-9]+)?}},
   // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8* bitcast ({{.*}} [[INVGK5:[^ ]+_kernel]] to i8*) to i8 addrspace(4)*),
-  // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BLG5]] to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1,
+  // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32 } addrspace(1)* [[BLG5]] to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1,
   // B32-SAME: i32* %[[TMP1]])
   // B64-SAME: i64* %[[TMP1]])
   enqueue_kernel(default_queue, flags, ndrange,
                  ^(local void *p) {
                    return;
                  },
                  l);
 
kernel void device_side_enqueue(global int *a, global int *b, int i) {
   // B64: store i64 1, i64* %[[TMP1]], align 8
   // B64: %[[TMP2:.*]] = getelementptr [3 x i64], [3 x i64]* %[[TMP]], i32 0, i32 1
   // B64: store i64 2, i64* %[[TMP2]], align 8
   // B64: %[[TMP3:.*]] = getelementptr [3 x i64], [3 x i64]* %[[TMP]], i32 0, i32 2
   // B64: store i64 4, i64* %[[TMP3]], align 8
   // COMMON-LABEL: call i32 @__enqueue_kernel_vaargs
   // COMMON-SAME: (%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %struct.ndrange_t* [[NDR]]{{([0-9]+)?}},
   // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8* bitcast ({{.*}} [[INVGK6:[^ ]+_kernel]] to i8*) to i8 addrspace(4)*),
-  // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BLG6]] to i8 addrspace(1)*) to i8 addrspace(4)*), i32 3,
+  // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32 } addrspace(1)* [[BLG6]] to i8 addrspace(1)*) to i8 addrspace(4)*), i32 3,
   // B32-SAME: i32* %[[TMP1]])
   // B64-SAME: i64* %[[TMP1]])
   enqueue_kernel(default_queue, flags, ndrange,
                  ^(local void *p1, local void *p2, local void *p3) {
                    return;
                  },
                  1, 2, 4);
 
   // Emits global block literal [[BLG7]] and block kernel [[INVGK7]].
   // COMMON: [[DEF_Q:%[0-9]+]] = load %opencl.queue_t*, %opencl.queue_t** %default_queue
   // COMMON: [[FLAGS:%[0-9]+]] = load i32, i32* %flags
   // B32: %[[TMP:.*]] = alloca [1 x i32]
   // B32: %[[TMP1:.*]] = getelementptr [1 x i32], [1 x i32]* %[[TMP]], i32 0, i32 0
   // B32: store i32 0, i32* %[[TMP1]], align 4
   // B64: %[[TMP:.*]] = alloca [1 x i64]
   // B64: %[[TMP1:.*]] = getelementptr [1 x i64], [1 x i64]* %[[TMP]], i32 0, i32 0
   // B64: store i64 4294967296, i64* %[[TMP1]], align 8
   // COMMON-LABEL: call i32 @__enqueue_kernel_vaargs
   // COMMON-SAME: (%opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %struct.ndrange_t* [[NDR]]{{([0-9]+)?}},
   // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8* bitcast ({{.*}} [[INVGK7:[^ ]+_kernel]] to i8*) to i8 addrspace(4)*),
-  // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BLG7]] to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1,
+  // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32 } addrspace(1)* [[BLG7]] to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1,
   // B32-SAME: i32* %[[TMP1]])
   // B64-SAME: i64* %[[TMP1]])
   enqueue_kernel(default_queue, flags, ndrange,
                  ^(local void *p) {
                    return;
                  },
                  4294967296L);
 
   // Emits global block literal [[BLG8]] and invoke function [[INVG8]].
   // The full type of these expressions are long (and repeated elsewhere), so we
   // capture it as part of the regex for convenience and clarity.
-  // COMMON: store void () addrspace(4)* addrspacecast (void () addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BLG8]] to void () addrspace(1)*) to void () addrspace(4)*), void () addrspace(4)** %block_A
+  // COMMON: store void () addrspace(4)* addrspacecast (void () addrspace(1)* bitcast ({ i32, i32 } addrspace(1)* [[BLG8]] to void () addrspace(1)*) to void () addrspace(4)*), void () addrspace(4)** %block_A
   void (^const block_A)(void) = ^{
     return;
   };
 
   // Emits global block literal [[BLG9]] and invoke function [[INVG9]].
-  // COMMON: store void (i8 addrspace(3)*) addrspace(4)* addrspacecast (void (i8 addrspace(3)*) addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BLG9]] to void (i8 addrspace(3)*) addrspace(1)*) to void (i8 addrspace(3)*) addrspace(4)*), void (i8 addrspace(3)*) addrspace(4)** %block_B
+  // COMMON: store void (i8 addrspace(3)*) addrspace(4)* addrspacecast (void (i8 addrspace(3)*) addrspace(1)* bitcast ({ i32, i32 } addrspace(1)* [[BLG9]] to void (i8 addrspace(3)*) addrspace(1)*) to void (i8 addrspace(3)*) addrspace(4)*), void (i8 addrspace(3)*) addrspace(4)** %block_B
   void (^const block_B)(local void *) = ^(local void *a) {
     return;
   };
 
   // Uses global block literal [[BLG8]] and invoke function [[INVG8]].
-  // COMMON: [[r1:%.*]] = load i8 addrspace(4)*, i8 addrspace(4)* addrspace(4)* getelementptr inbounds (%struct.__opencl_block_literal_generic, %struct.__opencl_block_literal_generic addrspace(4)* addrspacecast (%struct.__opencl_block_literal_generic addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BLG8]] to %struct.__opencl_block_literal_generic addrspace(1)*) to %struct.__opencl_block_literal_generic addrspace(4)*), i32 0, i32 2)
-  // COMMON: [[r2:%.*]] = addrspacecast i8 addrspace(4)* [[r1]] to void (i8 addrspace(4)*)*
-  // COMMON: call spir_func void [[r2]](i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BLG8]] to i8 addrspace(1)*) to i8 addrspace(4)*))
+  // COMMON: call spir_func void [[INVG8:.*]](i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32 } addrspace(1)* [[BLG8]] to i8 addrspace(1)*) to i8 addrspace(4)*))
   block_A();
 
   // Emits global block literal [[BLG8]] and block kernel [[INVGK8]]. [[INVGK8]] calls [[INVG8]].
   // COMMON: [[DEF_Q:%[0-9]+]] = load %opencl.queue_t{{.*}}*, %opencl.queue_t{{.*}}** %default_queue
   // COMMON: [[FLAGS:%[0-9]+]] = load i32, i32* %flags
   // COMMON-LABEL: call i32 @__enqueue_kernel_basic(
   // COMMON-SAME: %opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %struct.ndrange_t* byval [[NDR]]{{([0-9]+)?}},
   // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8* bitcast ({{.*}} [[INVGK8:[^ ]+_kernel]] to i8*) to i8 addrspace(4)*),
-  // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BLG8]] to i8 addrspace(1)*) to i8 addrspace(4)*))
+  // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32 } addrspace(1)* [[BLG8]] to i8 addrspace(1)*) to i8 addrspace(4)*))
   enqueue_kernel(default_queue, flags, ndrange, block_A);
 
   // Uses block kernel [[INVGK8]] and global block literal [[BLG8]].
   // COMMON: call i32 @__get_kernel_work_group_size_impl(
   // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8* bitcast ({{.*}} [[INVGK8]] to i8*) to i8 addrspace(4)*),
-  // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BLG8]] to i8 addrspace(1)*) to i8 addrspace(4)*))
+  // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32 } addrspace(1)* [[BLG8]] to i8 addrspace(1)*) to i8 addrspace(4)*))
   unsigned size = get_kernel_work_group_size(block_A);
 
   // Uses global block literal [[BLG8]] and invoke function [[INVG8]]. Make sure no redundant block literal and invoke functions are emitted.
-  // COMMON: [[r1:%.*]] = load i8 addrspace(4)*, i8 addrspace(4)* addrspace(4)* getelementptr inbounds (%struct.__opencl_block_literal_generic, %struct.__opencl_block_literal_generic addrspace(4)* addrspacecast (%struct.__opencl_block_literal_generic addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BLG8]] to %struct.__opencl_block_literal_generic addrspace(1)*) to %struct.__opencl_block_literal_generic addrspace(4)*), i32 0, i32 2)
-  // COMMON: [[r2:%.*]] = addrspacecast i8 addrspace(4)* [[r1]] to void (i8 addrspace(4)*)*
-  // COMMON: call spir_func void [[r2]](i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BLG8]] to i8 addrspace(1)*) to i8 addrspace(4)*))
+  // COMMON: call spir_func void [[INVG8]](i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32 } addrspace(1)* [[BLG8]] to i8 addrspace(1)*) to i8 addrspace(4)*))
   block_A();
 
   void (^block_C)(void) = ^{
     callee(i, a);
   };
 
   // Emits block literal on stack and block kernel [[INVLK3]].
-  // COMMON: store i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*)* [[INVL3:@__device_side_enqueue_block_invoke[^ ]*]] to i8*) to i8 addrspace(4)*), i8 addrspace(4)** %block.invoke
   // COMMON: [[DEF_Q:%[0-9]+]] = load %opencl.queue_t{{.*}}*, %opencl.queue_t{{.*}}** %default_queue
   // COMMON: [[FLAGS:%[0-9]+]] = load i32, i32* %flags
   // COMMON: [[BL_I8:%[0-9]+]] = addrspacecast void ()* {{.*}} to i8 addrspace(4)*
   // COMMON-LABEL: call i32 @__enqueue_kernel_basic(
   // COMMON-SAME: %opencl.queue_t{{.*}}* [[DEF_Q]], i32 [[FLAGS]], %struct.ndrange_t* byval [[NDR]]{{([0-9]+)?}},
   // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8* bitcast ({{.*}} [[INVLK3:[^ ]+_kernel]] to i8*) to i8 addrspace(4)*),
   // COMMON-SAME: i8 addrspace(4)* [[BL_I8]])
   enqueue_kernel(default_queue, flags, ndrange, block_C);
 
-  // Emits global block literal [[BLG9]] and block kernel [[INVGK9]]. [[INVGK9]] calls [[INV9]].
+  // Emits global block literal [[BLG9]] and block kernel [[INVGK9]]. [[INVGK9]] calls [[INVG9]].
   // COMMON: call i32 @__get_kernel_work_group_size_impl(
   // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8* bitcast ({{.*}} [[INVGK9:[^ ]+_kernel]] to i8*) to i8 addrspace(4)*),
-  // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BLG9]] to i8 addrspace(1)*) to i8 addrspace(4)*))
+  // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32 } addrspace(1)* [[BLG9]] to i8 addrspace(1)*) to i8 addrspace(4)*))
   size = get_kernel_work_group_size(block_B);
 
   // Uses global block literal [[BLG8]] and block kernel [[INVGK8]]. Make sure no redundant block literal ind invoke functions are emitted.
   // COMMON: call i32 @__get_kernel_preferred_work_group_multiple_impl(
   // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8* bitcast ({{.*}} [[INVGK8]] to i8*) to i8 addrspace(4)*),
-  // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BLG8]] to i8 addrspace(1)*) to i8 addrspace(4)*))
+  // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32 } addrspace(1)* [[BLG8]] to i8 addrspace(1)*) to i8 addrspace(4)*))
   size = get_kernel_preferred_work_group_size_multiple(block_A);
 
   // Uses global block literal [[BL_GLOBAL]] and block kernel [[INV_G_K]]. [[INV_G_K]] calls [[INV_G]].
   // COMMON: call i32 @__get_kernel_preferred_work_group_multiple_impl(
   // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8* bitcast ({{.*}} [[INV_G_K:[^ ]+_kernel]] to i8*) to i8 addrspace(4)*),
-  // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BL_GLOBAL]] to i8 addrspace(1)*) to i8 addrspace(4)*))
+  // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32 } addrspace(1)* [[BL_GLOBAL]] to i8 addrspace(1)*) to i8 addrspace(4)*))
   size = get_kernel_preferred_work_group_size_multiple(block_G);
 
   // Emits global block literal [[BLG10]] and block kernel [[INVGK10]].
   // COMMON: call i32 @__get_kernel_max_sub_group_size_for_ndrange_impl(%struct.ndrange_t* {{[^,]+}},
   // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8* bitcast ({{.*}} [[INVGK10:[^ ]+_kernel]] to i8*) to i8 addrspace(4)*),
-  // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BLG10]] to i8 addrspace(1)*) to i8 addrspace(4)*))
+  // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32 } addrspace(1)* [[BLG10]] to i8 addrspace(1)*) to i8 addrspace(4)*))
   size = get_kernel_max_sub_group_size_for_ndrange(ndrange, ^(){});
 
   // Emits global block literal [[BLG11]] and block kernel [[INVGK11]].
   // COMMON: call i32 @__get_kernel_sub_group_count_for_ndrange_impl(%struct.ndrange_t* {{[^,]+}},
   // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8* bitcast ({{.*}} [[INVGK11:[^ ]+_kernel]] to i8*) to i8 addrspace(4)*),
-  // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BLG11]] to i8 addrspace(1)*) to i8 addrspace(4)*))
+  // COMMON-SAME: i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32 } addrspace(1)* [[BLG11]] to i8 addrspace(1)*) to i8 addrspace(4)*))
   size = get_kernel_sub_group_count_for_ndrange(ndrange, ^(){});
 }
 
 // COMMON: define internal spir_kernel void [[INVLK1]](i8 addrspace(4)*) #{{[0-9]+}} {
 // COMMON: entry:
 // COMMON:  call void @__device_side_enqueue_block_invoke(i8 addrspace(4)* %0)
 // COMMON:  ret void
 // COMMON: }
 // COMMON: define internal spir_kernel void [[INVLK2]](i8 addrspace(4)*{{.*}})
 // COMMON: define internal spir_kernel void [[INVGK1]](i8 addrspace(4)*{{.*}}, i8 addrspace(3)*{{.*}})
 // COMMON: define internal spir_kernel void [[INVGK2]](i8 addrspace(4)*{{.*}}, i8 addrspace(3)*{{.*}})
 // COMMON: define internal spir_kernel void [[INVGK3]](i8 addrspace(4)*{{.*}}, i8 addrspace(3)*{{.*}})
 // COMMON: define internal spir_kernel void [[INVGK4]](i8 addrspace(4)*{{.*}}, i8 addrspace(3)*{{.*}})
 // COMMON: define internal spir_kernel void [[INVGK5]](i8 addrspace(4)*{{.*}}, i8 addrspace(3)*{{.*}})
 // COMMON: define internal spir_kernel void [[INVGK6]](i8 addrspace(4)*, i8 addrspace(3)*, i8 addrspace(3)*, i8 addrspace(3)*) #{{[0-9]+}} {
 // COMMON: entry:
 // COMMON:  call void @__device_side_enqueue_block_invoke_8(i8 addrspace(4)* %0, i8 addrspace(3)* %1, i8 addrspace(3)* %2, i8 addrspace(3)* %3)
 // COMMON:  ret void
 // COMMON: }
 // COMMON: define internal spir_kernel void [[INVGK7]](i8 addrspace(4)*{{.*}}, i8 addrspace(3)*{{.*}})
 // COMMON: define internal spir_func void [[INVG8]](i8 addrspace(4)*{{.*}})
-// COMMON: define internal spir_func void [[INVG9]](i8 addrspace(4)*{{.*}}, i8 addrspace(3)* %{{.*}})
+// COMMON: define internal spir_func void [[INVG9:.*]](i8 addrspace(4)*{{.*}}, i8 addrspace(3)* %{{.*}})
 // COMMON: define internal spir_kernel void [[INVGK8]](i8 addrspace(4)*{{.*}})
 // COMMON: define internal spir_kernel void [[INVLK3]](i8 addrspace(4)*{{.*}})
 // COMMON: define internal spir_kernel void [[INVGK9]](i8 addrspace(4)*{{.*}}, i8 addrspace(3)*{{.*}})
 // COMMON: define internal spir_kernel void [[INV_G_K]](i8 addrspace(4)*{{.*}}, i8 addrspace(3)*{{.*}})
 // COMMON: define internal spir_kernel void [[INVGK10]](i8 addrspace(4)*{{.*}})
 // COMMON: define internal spir_kernel void [[INVGK11]](i8 addrspace(4)*{{.*}})