[meta-intel] [PATCH v3] clang: build Intel common-clang and spirv
Anuj Mittal
anuj.mittal at intel.com
Fri May 3 01:04:05 PDT 2019
Common clang is a thin wrapper library around clang. Common clang has
OpenCL-oriented API and is capable to compile OpenCL C kernels to SPIR-V
modules.
This adds a bbappend to clang recipe from meta-clang to build the
necessary components and moves it to dynamic layers so it's built only
when clang-layer is included.
Signed-off-by: Anuj Mittal <anuj.mittal at intel.com>
---
conf/layer.conf | 4 +
.../recipes-devtools/clang/clang_%.bbappend | 13 +
...nge-type-of-block-pointer-for-OpenCL.patch | 156 +++
...CL-pass-to-handle-new-blocks-represn.patch | 1119 +++++++++++++++++
...001-dont-export-targets-for-binaries.patch | 66 +
.../files/0001-point-to-correct-clang.patch | 59 +
...-LLVM-IR-generated-for-OpenCL-blocks.patch | 294 +++++
...3-OpenCL-Fix-assertion-due-to-blocks.patch | 61 +
.../clang/llvm-project-source.bbappend | 16 +
9 files changed, 1788 insertions(+)
create mode 100644 dynamic-layers/clang-layer/recipes-devtools/clang/clang_%.bbappend
create mode 100644 dynamic-layers/clang-layer/recipes-devtools/clang/files/0001-OpenCL-Change-type-of-block-pointer-for-OpenCL.patch
create mode 100644 dynamic-layers/clang-layer/recipes-devtools/clang/files/0001-Update-LowerOpenCL-pass-to-handle-new-blocks-represn.patch
create mode 100644 dynamic-layers/clang-layer/recipes-devtools/clang/files/0001-dont-export-targets-for-binaries.patch
create mode 100644 dynamic-layers/clang-layer/recipes-devtools/clang/files/0001-point-to-correct-clang.patch
create mode 100644 dynamic-layers/clang-layer/recipes-devtools/clang/files/0002-OpenCL-Simplify-LLVM-IR-generated-for-OpenCL-blocks.patch
create mode 100644 dynamic-layers/clang-layer/recipes-devtools/clang/files/0003-OpenCL-Fix-assertion-due-to-blocks.patch
create mode 100644 dynamic-layers/clang-layer/recipes-devtools/clang/llvm-project-source.bbappend
diff --git a/conf/layer.conf b/conf/layer.conf
index 3e75b0fe..4cea5500 100644
--- a/conf/layer.conf
+++ b/conf/layer.conf
@@ -20,5 +20,9 @@ LAYERRECOMMENDS_intel = "dpdk intel-qat"
LAYERVERSION_intel = "5"
LAYERSERIES_COMPAT_intel = "thud warrior"
+BBFILES_DYNAMIC += " \
+ clang-layer:${LAYERDIR}/dynamic-layers/clang-layer/*/*/*.bb \
+ clang-layer:${LAYERDIR}/dynamic-layers/clang-layer/*/*/*.bbappend \
+"
require ${LAYERDIR}/conf/include/maintainers.inc
diff --git a/dynamic-layers/clang-layer/recipes-devtools/clang/clang_%.bbappend b/dynamic-layers/clang-layer/recipes-devtools/clang/clang_%.bbappend
new file mode 100644
index 00000000..f8d5a252
--- /dev/null
+++ b/dynamic-layers/clang-layer/recipes-devtools/clang/clang_%.bbappend
@@ -0,0 +1,13 @@
+FILESEXTRAPATHS_prepend_intel-x86-common := "${THISDIR}/files:"
+
+DEPENDS_append = " opencl-clang-native"
+LLVM_TARGETS_TO_BUILD = "X86"
+
+do_install_append_intel-x86-common() {
+ DESTDIR=${D} ninja -v install-cmake-exports
+}
+
+LIBCPLUSPLUS = ""
+
+# undefined reference to `__atomic_load' on i*86.
+COMPATIBLE_HOST = '(x86_64).*-linux'
diff --git a/dynamic-layers/clang-layer/recipes-devtools/clang/files/0001-OpenCL-Change-type-of-block-pointer-for-OpenCL.patch b/dynamic-layers/clang-layer/recipes-devtools/clang/files/0001-OpenCL-Change-type-of-block-pointer-for-OpenCL.patch
new file mode 100644
index 00000000..1c491402
--- /dev/null
+++ b/dynamic-layers/clang-layer/recipes-devtools/clang/files/0001-OpenCL-Change-type-of-block-pointer-for-OpenCL.patch
@@ -0,0 +1,156 @@
+From 39a3ac0065c23d1e2d55dfd8792cc28a146a4307 Mon Sep 17 00:00:00 2001
+From: Alexey Bader <alexey.bader at intel.com>
+Date: Tue, 19 Feb 2019 15:19:06 +0000
+Subject: [PATCH 1/2] [OpenCL] Change type of block pointer for OpenCL
+
+Summary:
+
+For some reason OpenCL blocks in LLVM IR are represented as function pointers.
+These pointers do not point to any real function and never get called. Actually
+they point to some structure, which in turn contains pointer to the real block
+invoke function.
+This patch changes represntation of OpenCL blocks in LLVM IR from function
+pointers to pointers to `%struct.__block_literal_generic`.
+Such representation allows to avoid unnecessary bitcasts and simplifies
+further processing (e.g. translation to SPIR-V ) of the module for targets
+which do not support function pointers.
+
+Patch by: Alexey Sotkin.
+
+Reviewers: Anastasia, yaxunl, svenvh
+
+Reviewed By: Anastasia
+
+Subscribers: alexbatashev, cfe-commits
+
+Tags: #clang
+
+Differential Revision: https://reviews.llvm.org/D58277
+
+git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@354337 91177308-0d34-0410-b5e6-96231b3b80d8
+
+Upstream-Status: Backport
+[https://github.com/llvm-mirror/clang/commit/283f308bdb5893bab1f36791711346e746045f94]
+Signed-off-by: Anuj Mittal <anuj.mittal at intel.com>
+---
+ lib/CodeGen/CodeGenTypes.cpp | 4 +++-
+ test/CodeGenOpenCL/blocks.cl | 18 ++++++++----------
+ test/CodeGenOpenCL/cl20-device-side-enqueue.cl | 18 +++++++++---------
+ 3 files changed, 20 insertions(+), 20 deletions(-)
+
+diff --git a/lib/CodeGen/CodeGenTypes.cpp b/lib/CodeGen/CodeGenTypes.cpp
+index 2acf1ac..93b3ebf 100644
+--- a/lib/CodeGen/CodeGenTypes.cpp
++++ b/lib/CodeGen/CodeGenTypes.cpp
+@@ -637,7 +637,9 @@ llvm::Type *CodeGenTypes::ConvertType(QualType T) {
+
+ case Type::BlockPointer: {
+ const QualType FTy = cast<BlockPointerType>(Ty)->getPointeeType();
+- llvm::Type *PointeeType = ConvertTypeForMem(FTy);
++ llvm::Type *PointeeType = CGM.getLangOpts().OpenCL
++ ? CGM.getGenericBlockLiteralType()
++ : ConvertTypeForMem(FTy);
+ unsigned AS = Context.getTargetAddressSpace(FTy);
+ ResultType = llvm::PointerType::get(PointeeType, AS);
+ break;
+diff --git a/test/CodeGenOpenCL/blocks.cl b/test/CodeGenOpenCL/blocks.cl
+index 675240c..19aacc3 100644
+--- a/test/CodeGenOpenCL/blocks.cl
++++ b/test/CodeGenOpenCL/blocks.cl
+@@ -35,11 +35,10 @@ void foo(){
+ // SPIR: %[[block_captured:.*]] = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 }>, <{ i32, i32, i8 addrspace(4)*, i32 }>* %[[block]], i32 0, i32 3
+ // 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_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: %[[blk_ptr:.*]] = bitcast <{ i32, i32, i8 addrspace(4)*, i32 }>* %[[block]] to %struct.__opencl_block_literal_generic*
++ // SPIR: %[[blk_gen_ptr:.*]] = addrspacecast %struct.__opencl_block_literal_generic* %[[blk_ptr]] to %struct.__opencl_block_literal_generic addrspace(4)*
++ // SPIR: store %struct.__opencl_block_literal_generic addrspace(4)* %[[blk_gen_ptr]], %struct.__opencl_block_literal_generic addrspace(4)** %[[block_B:.*]],
++ // SPIR: %[[block_literal:.*]] = load %struct.__opencl_block_literal_generic addrspace(4)*, %struct.__opencl_block_literal_generic addrspace(4)** %[[block_B]]
+ // 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]]
+@@ -50,11 +49,10 @@ void foo(){
+ // AMDGCN: %[[block_captured:.*]] = getelementptr inbounds <{ i32, i32, i8*, i32 }>, <{ i32, i32, i8*, i32 }> addrspace(5)* %[[block]], i32 0, i32 3
+ // 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_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: %[[blk_ptr:.*]] = bitcast <{ i32, i32, i8*, i32 }> addrspace(5)* %[[block]] to %struct.__opencl_block_literal_generic addrspace(5)*
++ // AMDGCN: %[[blk_gen_ptr:.*]] = addrspacecast %struct.__opencl_block_literal_generic addrspace(5)* %[[blk_ptr]] to %struct.__opencl_block_literal_generic*
++ // AMDGCN: store %struct.__opencl_block_literal_generic* %[[blk_gen_ptr]], %struct.__opencl_block_literal_generic* addrspace(5)* %[[block_B:.*]],
++ // AMDGCN: %[[block_literal:.*]] = load %struct.__opencl_block_literal_generic*, %struct.__opencl_block_literal_generic* addrspace(5)* %[[block_B]]
+ // 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]]
+diff --git a/test/CodeGenOpenCL/cl20-device-side-enqueue.cl b/test/CodeGenOpenCL/cl20-device-side-enqueue.cl
+index 4732194..8445016 100644
+--- a/test/CodeGenOpenCL/cl20-device-side-enqueue.cl
++++ b/test/CodeGenOpenCL/cl20-device-side-enqueue.cl
+@@ -11,7 +11,7 @@ typedef struct {int a;} ndrange_t;
+
+ // 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: @block_G = addrspace(1) constant %struct.__opencl_block_literal_generic addrspace(4)* addrspacecast (%struct.__opencl_block_literal_generic addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BL_GLOBAL]] to %struct.__opencl_block_literal_generic addrspace(1)*) to %struct.__opencl_block_literal_generic 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)*) }
+@@ -77,9 +77,9 @@ 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: 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 ()*
+- // COMMON: [[BL_I8:%[0-9]+]] = addrspacecast void ()* [[BL]] to i8 addrspace(4)*
++ // B32: [[BL:%[0-9]+]] = bitcast <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }>* %block to %struct.__opencl_block_literal_generic*
++ // B64: [[BL:%[0-9]+]] = bitcast <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32 addrspace(1)*, i32 }>* %block to %struct.__opencl_block_literal_generic*
++ // COMMON: [[BL_I8:%[0-9]+]] = addrspacecast %struct.__opencl_block_literal_generic* [[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)*),
+@@ -95,8 +95,8 @@ kernel void device_side_enqueue(global int *a, global int *b, int i) {
+ // 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_I8:%[0-9]+]] = addrspacecast void ()* [[BL]] to i8 addrspace(4)*
++ // COMMON: [[BL:%[0-9]+]] = bitcast <{ i32, i32, i8 addrspace(4)*, i32{{.*}}, i32{{.*}}, i32{{.*}} }>* %block3 to %struct.__opencl_block_literal_generic*
++ // COMMON: [[BL_I8:%[0-9]+]] = addrspacecast %struct.__opencl_block_literal_generic* [[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)*),
+@@ -300,13 +300,13 @@ kernel void device_side_enqueue(global int *a, global int *b, int i) {
+ // 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 %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)*), %struct.__opencl_block_literal_generic 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 %struct.__opencl_block_literal_generic addrspace(4)* addrspacecast (%struct.__opencl_block_literal_generic addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BLG9]] to %struct.__opencl_block_literal_generic addrspace(1)*) to %struct.__opencl_block_literal_generic addrspace(4)*), %struct.__opencl_block_literal_generic addrspace(4)** %block_B
+ void (^const block_B)(local void *) = ^(local void *a) {
+ return;
+ };
+@@ -346,7 +346,7 @@ kernel void device_side_enqueue(global int *a, global int *b, int i) {
+ // 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: [[BL_I8:%[0-9]+]] = addrspacecast %struct.__opencl_block_literal_generic* {{.*}} 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)*),
+--
+1.8.3.1
+
diff --git a/dynamic-layers/clang-layer/recipes-devtools/clang/files/0001-Update-LowerOpenCL-pass-to-handle-new-blocks-represn.patch b/dynamic-layers/clang-layer/recipes-devtools/clang/files/0001-Update-LowerOpenCL-pass-to-handle-new-blocks-represn.patch
new file mode 100644
index 00000000..4a528674
--- /dev/null
+++ b/dynamic-layers/clang-layer/recipes-devtools/clang/files/0001-Update-LowerOpenCL-pass-to-handle-new-blocks-represn.patch
@@ -0,0 +1,1119 @@
+From 9ce0fe02fd6cda5fb29fbb0d5037a1798a810b8a Mon Sep 17 00:00:00 2001
+From: Alexey Sotkin <alexey.sotkin at intel.com>
+Date: Thu, 21 Feb 2019 17:14:36 +0300
+Subject: [PATCH 1/3] Update LowerOpenCL pass to handle new blocks
+ represntation in LLVM IR
+
+Upstream-Status: Backport
+[https://github.com/KhronosGroup/SPIRV-LLVM-Translator/commit/bd6ddfaf7232cd81c7f2fe9877e66f286731bd8e]
+Signed-off-by: Anuj Mittal <anuj.mittal at intel.com>
+---
+ lib/SPIRV/SPIRVLowerOCLBlocks.cpp | 413 ++++++++----------------------
+ test/global_block.ll | 71 ++---
+ test/literal-struct.ll | 31 ++-
+ test/transcoding/block_w_struct_return.ll | 47 ++--
+ test/transcoding/enqueue_kernel.ll | 237 ++++++++++-------
+ 5 files changed, 317 insertions(+), 482 deletions(-)
+
+diff --git a/lib/SPIRV/SPIRVLowerOCLBlocks.cpp b/lib/SPIRV/SPIRVLowerOCLBlocks.cpp
+index 50e1838..b42a4ec 100644
+--- a/lib/SPIRV/SPIRVLowerOCLBlocks.cpp
++++ b/lib/SPIRV/SPIRVLowerOCLBlocks.cpp
+@@ -1,303 +1,110 @@
+-//===- SPIRVLowerOCLBlocks.cpp - OCL Utilities ----------------------------===//
+-//
+-// The LLVM/SPIRV Translator
+-//
+-// This file is distributed under the University of Illinois Open Source
+-// License. See LICENSE.TXT for details.
+-//
+-// Copyright (c) 2018 Intel Corporation. All rights reserved.
+-//
+-// Permission is hereby granted, free of charge, to any person obtaining a
+-// copy of this software and associated documentation files (the "Software"),
+-// to deal with the Software without restriction, including without limitation
+-// the rights to use, copy, modify, merge, publish, distribute, sublicense,
+-// and/or sell copies of the Software, and to permit persons to whom the
+-// Software is furnished to do so, subject to the following conditions:
+-//
+-// Redistributions of source code must retain the above copyright notice,
+-// this list of conditions and the following disclaimers.
+-// Redistributions in binary form must reproduce the above copyright notice,
+-// this list of conditions and the following disclaimers in the documentation
+-// and/or other materials provided with the distribution.
+-// Neither the names of Intel Corporation, nor the names of its
+-// contributors may be used to endorse or promote products derived from this
+-// Software without specific prior written permission.
+-// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+-// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+-// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+-// CONTRIBUTORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+-// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+-// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS WITH
+-// THE SOFTWARE.
+-//
+-//===----------------------------------------------------------------------===//
+-//
+-// SPIR-V specification doesn't allow function pointers, so SPIR-V translator
+-// is designed to fail if a value with function type (except calls) is occured.
+-// Currently there is only two cases, when function pointers are generating in
+-// LLVM IR in OpenCL - block calls and device side enqueue built-in calls.
+-//
+-// In both cases values with function type used as intermediate representation
+-// for block literal structure.
+-//
+-// This pass is designed to find such cases and simplify them to avoid any
+-// function pointer types occurrences in LLVM IR in 4 steps.
+-//
+-// 1. Find all function pointer allocas, like
+-// %block = alloca void () *
+-//
+-// Then find a single store to that alloca:
+-// %blockLit = alloca <{ i32, i32, ...}>, align 4
+-// %0 = bitcast <{ i32, i32, ... }>* %blockLit to void ()*
+-// > store void ()* %0, void ()** %block, align 4
+-//
+-// And replace the alloca users by new instructions which used stored value
+-// %blockLit itself instead of function pointer alloca %block.
+-//
+-// 2. Find consecutive casts from block literal type to i8 addrspace(4)*
+-// used function pointers as an intermediate type:
+-// %0 = bitcast <{ i32, i32 }> %block to void() *
+-// %1 = addrspacecast void() * %0 to i8 addrspace(4)*
+-// And simplify them:
+-// %2 = addrspacecast <{ i32, i32 }> %block to i8 addrspace(4)*
+-//
+-// 3. Find all unused instructions with function pointer type occured after
+-// pp.1-2 and remove them.
+-//
+-// 4. Find unused globals with function pointer type, like
+-// @block = constant void ()*
+-// bitcast ({ i32, i32 }* @__block_literal_global to void ()*
+-//
+-// And remove them.
+-//
+-//===----------------------------------------------------------------------===//
+-#define DEBUG_TYPE "spv-lower-ocl-blocks"
+-
+-#include "OCLUtil.h"
+-#include "SPIRVInternal.h"
+-
+-#include "llvm/ADT/SetVector.h"
+-#include "llvm/Analysis/ValueTracking.h"
+-#include "llvm/IR/GlobalVariable.h"
+-#include "llvm/IR/InstIterator.h"
+-#include "llvm/IR/Module.h"
+-#include "llvm/Pass.h"
+-#include "llvm/PassSupport.h"
+-#include "llvm/Support/Casting.h"
+-
+-using namespace llvm;
+-
+-namespace {
+-
+-static void
+-removeUnusedFunctionPtrInst(Instruction *I,
+- SmallSetVector<Instruction *, 16> &FuncPtrInsts) {
+- for (unsigned OpIdx = 0, Ops = I->getNumOperands(); OpIdx != Ops; ++OpIdx) {
+- Instruction *OpI = dyn_cast<Instruction>(I->getOperand(OpIdx));
+- I->setOperand(OpIdx, nullptr);
+- if (OpI && OpI != I && OpI->user_empty())
+- FuncPtrInsts.insert(OpI);
+- }
+- I->eraseFromParent();
+-}
+-
+-static bool isFuncPtrAlloca(const AllocaInst *AI) {
+- auto *ET = dyn_cast<PointerType>(AI->getAllocatedType());
+- return ET && ET->getElementType()->isFunctionTy();
+-}
+-
+-static bool hasFuncPtrType(const Value *V) {
+- auto *PT = dyn_cast<PointerType>(V->getType());
+- return PT && PT->getElementType()->isFunctionTy();
+-}
+-
+-static bool isFuncPtrInst(const Instruction *I) {
+- if (auto *AI = dyn_cast<AllocaInst>(I))
+- return isFuncPtrAlloca(AI);
+-
+- for (auto &Op : I->operands()) {
+- if (auto *AI = dyn_cast<AllocaInst>(Op))
+- return isFuncPtrAlloca(AI);
+-
+- auto *OpI = dyn_cast<Instruction>(&Op);
+- if (OpI && OpI != I && hasFuncPtrType(OpI))
+- return true;
+- }
+- return false;
+-}
+-
+-static StoreInst *findSingleStore(AllocaInst *AI) {
+- StoreInst *Store = nullptr;
+- for (auto *U : AI->users()) {
+- if (!isa<StoreInst>(U))
+- continue; // not a store
+- if (Store)
+- return nullptr; // there are more than one stores
+- Store = dyn_cast<StoreInst>(U);
+- }
+- return Store;
+-}
+-
+-static void fixFunctionPtrAllocaUsers(AllocaInst *AI) {
+- // Find and remove a single store to alloca
+- auto *SingleStore = findSingleStore(AI);
+- assert(SingleStore && "More than one store to the function pointer alloca");
+- auto *StoredVal = SingleStore->getValueOperand();
+- SingleStore->eraseFromParent();
+-
+- // Find loads from the alloca and replace thier users
+- for (auto *U : AI->users()) {
+- auto *LI = dyn_cast<LoadInst>(U);
+- if (!LI)
+- continue;
+-
+- for (auto *U : LI->users()) {
+- auto *UInst = cast<Instruction>(U);
+- auto *Cast = CastInst::CreatePointerBitCastOrAddrSpaceCast(
+- StoredVal, UInst->getType(), "", UInst);
+- UInst->replaceAllUsesWith(Cast);
+- }
+- }
+-}
+-
+-static int getBlockLiteralIdx(const Function &F) {
+- StringRef FName = F.getName();
+- if (isEnqueueKernelBI(FName))
+- return FName.contains("events") ? 7 : 4;
+- if (isKernelQueryBI(FName))
+- return FName.contains("for_ndrange") ? 2 : 1;
+- if (FName.startswith("__") && FName.contains("_block_invoke"))
+- return F.hasStructRetAttr() ? 1 : 0;
+-
+- return -1; // No block literal argument
+-}
+-
+-static bool hasBlockLiteralArg(const Function &F) {
+- return getBlockLiteralIdx(F) != -1;
+-}
+-
+-static bool simplifyFunctionPtrCasts(Function &F) {
+- bool Changed = false;
+- int BlockLiteralIdx = getBlockLiteralIdx(F);
+- for (auto *U : F.users()) {
+- auto *Call = dyn_cast<CallInst>(U);
+- if (!Call)
+- continue;
+- if (Call->getFunction()->getName() == F.getName().str() + "_kernel")
+- continue; // Skip block invoke function calls inside block invoke kernels
+-
+- const DataLayout &DL = F.getParent()->getDataLayout();
+- auto *BlockLiteral = Call->getOperand(BlockLiteralIdx);
+- auto *BlockLiteralVal = GetUnderlyingObject(BlockLiteral, DL);
+- if (isa<GlobalVariable>(BlockLiteralVal))
+- continue; // nothing to do with globals
+-
+- auto *BlockLiteralAlloca = cast<AllocaInst>(BlockLiteralVal);
+- assert(!BlockLiteralAlloca->getAllocatedType()->isFunctionTy() &&
+- "Function type shouldn't be there");
+-
+- auto *NewBlockLiteral = CastInst::CreatePointerBitCastOrAddrSpaceCast(
+- BlockLiteralAlloca, BlockLiteral->getType(), "", Call);
+- BlockLiteral->replaceAllUsesWith(NewBlockLiteral);
+- Changed |= true;
+- }
+- return Changed;
+-}
+-
+-static void
+-findFunctionPtrAllocas(Module &M,
+- SmallVectorImpl<AllocaInst *> &FuncPtrAllocas) {
+- for (auto &F : M) {
+- if (F.isDeclaration())
+- continue;
+- for (auto &I : instructions(F)) {
+- auto *AI = dyn_cast<AllocaInst>(&I);
+- if (!AI || !isFuncPtrAlloca(AI))
+- continue;
+- FuncPtrAllocas.push_back(AI);
+- }
+- }
+-}
+-
+-static void
+-findUnusedFunctionPtrInsts(Module &M,
+- SmallSetVector<Instruction *, 16> &FuncPtrInsts) {
+- for (auto &F : M) {
+- if (F.isDeclaration())
+- continue;
+- for (auto &I : instructions(F))
+- if (I.user_empty() && isFuncPtrInst(&I))
+- FuncPtrInsts.insert(&I);
+- }
+-}
+-
+-static void
+-findUnusedFunctionPtrGlbs(Module &M,
+- SmallVectorImpl<GlobalVariable *> &FuncPtrGlbs) {
+- for (auto &GV : M.globals()) {
+- if (!GV.user_empty())
+- continue;
+- auto *GVType = dyn_cast<PointerType>(GV.getType()->getElementType());
+- if (GVType && GVType->getElementType()->isFunctionTy())
+- FuncPtrGlbs.push_back(&GV);
+- }
+-}
+-
+-class SPIRVLowerOCLBlocks : public ModulePass {
+-
+-public:
+- SPIRVLowerOCLBlocks() : ModulePass(ID) {}
+-
+- bool runOnModule(Module &M) {
+- bool Changed = false;
+-
+- // 1. Find function pointer allocas and fix their users
+- SmallVector<AllocaInst *, 16> FuncPtrAllocas;
+- findFunctionPtrAllocas(M, FuncPtrAllocas);
+-
+- Changed |= !FuncPtrAllocas.empty();
+- for (auto *AI : FuncPtrAllocas)
+- fixFunctionPtrAllocaUsers(AI);
+-
+- // 2. Simplify consecutive casts which use function pointer types
+- for (auto &F : M)
+- if (hasBlockLiteralArg(F))
+- Changed |= simplifyFunctionPtrCasts(F);
+-
+- // 3. Cleanup unused instructions with function pointer type
+- // which are occured after pp. 1-2
+- SmallSetVector<Instruction *, 16> FuncPtrInsts;
+- findUnusedFunctionPtrInsts(M, FuncPtrInsts);
+-
+- Changed |= !FuncPtrInsts.empty();
+- while (!FuncPtrInsts.empty()) {
+- Instruction *I = FuncPtrInsts.pop_back_val();
+- removeUnusedFunctionPtrInst(I, FuncPtrInsts);
+- }
+-
+- // 4. Find and remove unused global variables with function pointer type
+- SmallVector<GlobalVariable *, 16> FuncPtrGlbs;
+- findUnusedFunctionPtrGlbs(M, FuncPtrGlbs);
+-
+- Changed |= !FuncPtrGlbs.empty();
+- for (auto *GV : FuncPtrGlbs)
+- GV->eraseFromParent();
+-
+- return Changed;
+- }
+-
+- static char ID;
+-}; // class SPIRVLowerOCLBlocks
+-
+-char SPIRVLowerOCLBlocks::ID = 0;
+-
+-} // namespace
+-
+-INITIALIZE_PASS(
+- SPIRVLowerOCLBlocks, "spv-lower-ocl-blocks",
+- "Remove function pointers occured in case of using OpenCL blocks", false,
+- false)
+-
+-llvm::ModulePass *llvm::createSPIRVLowerOCLBlocks() {
+- return new SPIRVLowerOCLBlocks();
+-}
++//===- SPIRVLowerOCLBlocks.cpp - OCL Utilities ----------------------------===//
++//
++// The LLVM/SPIRV Translator
++//
++// This file is distributed under the University of Illinois Open Source
++// License. See LICENSE.TXT for details.
++//
++// Copyright (c) 2018 Intel Corporation. All rights reserved.
++//
++// Permission is hereby granted, free of charge, to any person obtaining a
++// copy of this software and associated documentation files (the "Software"),
++// to deal with the Software without restriction, including without limitation
++// the rights to use, copy, modify, merge, publish, distribute, sublicense,
++// and/or sell copies of the Software, and to permit persons to whom the
++// Software is furnished to do so, subject to the following conditions:
++//
++// Redistributions of source code must retain the above copyright notice,
++// this list of conditions and the following disclaimers.
++// Redistributions in binary form must reproduce the above copyright notice,
++// this list of conditions and the following disclaimers in the documentation
++// and/or other materials provided with the distribution.
++// Neither the names of Intel Corporation, nor the names of its
++// contributors may be used to endorse or promote products derived from this
++// Software without specific prior written permission.
++// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
++// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
++// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
++// CONTRIBUTORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
++// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
++// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS WITH
++// THE SOFTWARE.
++//
++//===----------------------------------------------------------------------===//
++//
++// SPIR-V specification doesn't allow function pointers, so SPIR-V translator
++// is designed to fail if a value with function type (except calls) is occured.
++// Currently there is only two cases, when function pointers are generating in
++// LLVM IR in OpenCL - block calls and device side enqueue built-in calls.
++//
++// In both cases values with function type used as intermediate representation
++// for block literal structure.
++//
++// In LLVM IR produced by clang, blocks are represented with the following
++// structure:
++// %struct.__opencl_block_literal_generic = type { i32, i32, i8 addrspace(4)* }
++// Pointers to block invoke functions are stored in the third field. Clang
++// replaces inderect function calls in all cases except if block is passed as a
++// function argument. Note that it is somewhat unclear if the OpenCL C spec
++// should allow passing blocks as function argumernts. This pass is not supposed
++// to work correctly with such functions.
++// Clang though has to store function pointers to this structure. Purpose of
++// this pass is to replace store of function pointers(not allowed in SPIR-V)
++// with null pointers.
++//
++//===----------------------------------------------------------------------===//
++#define DEBUG_TYPE "spv-lower-ocl-blocks"
++
++#include "SPIRVInternal.h"
++
++#include "llvm/IR/Module.h"
++#include "llvm/Pass.h"
++#include "llvm/Support/Regex.h"
++
++using namespace llvm;
++
++namespace {
++
++static bool isBlockInvoke(Function &F) {
++ static Regex BlockInvokeRegex("_block_invoke_?[0-9]*$");
++ return BlockInvokeRegex.match(F.getName());
++}
++
++class SPIRVLowerOCLBlocks : public ModulePass {
++
++public:
++ SPIRVLowerOCLBlocks() : ModulePass(ID) {}
++
++ bool runOnModule(Module &M) {
++ bool Changed = false;
++ for (Function &F : M) {
++ if (!isBlockInvoke(F))
++ continue;
++ for (User *U : F.users()) {
++ if (!isa<Constant>(U))
++ continue;
++ Constant *Null = Constant::getNullValue(U->getType());
++ if (U != Null) {
++ U->replaceAllUsesWith(Null);
++ Changed = true;
++ }
++ }
++ }
++ return Changed;
++ }
++
++ static char ID;
++};
++
++char SPIRVLowerOCLBlocks::ID = 0;
++
++} // namespace
++
++INITIALIZE_PASS(
++ SPIRVLowerOCLBlocks, "spv-lower-ocl-blocks",
++ "Remove function pointers occured in case of using OpenCL blocks", false,
++ false)
++
++llvm::ModulePass *llvm::createSPIRVLowerOCLBlocks() {
++ return new SPIRVLowerOCLBlocks();
++}
+diff --git a/test/global_block.ll b/test/global_block.ll
+index a9267d8..efb4cf3 100644
+--- a/test/global_block.ll
++++ b/test/global_block.ll
+@@ -16,7 +16,7 @@
+ ; RUN: llvm-spirv %t.bc -o %t.spv
+ ; RUN: llvm-spirv -r %t.spv -o - | llvm-dis | FileCheck %s --check-prefix=CHECK-LLVM
+
+-target datalayout = "e-p:32:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024-n8:16:32:64"
++target datalayout = "e-p:32:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024"
+ target triple = "spir-unknown-unknown"
+
+ ; CHECK-SPIRV: Name [[block_invoke:[0-9]+]] "_block_invoke"
+@@ -26,71 +26,56 @@ target triple = "spir-unknown-unknown"
+ ; CHECK-SPIRV: TypePointer [[int8Ptr:[0-9]+]] 8 [[int8]]
+ ; CHECK-SPIRV: TypeFunction [[block_invoke_type:[0-9]+]] [[int]] [[int8Ptr]] [[int]]
+
+-;; This variable is not needed in SPIRV
+-; CHECK-SPIRV-NOT: Name {{[0-9]+}} block_kernel.b1
+-; CHECK-LLVM-NOT: @block_kernel.b1
+- at block_kernel.b1 = internal addrspace(2) constant i32 (i32) addrspace(4)* addrspacecast (i32 (i32) addrspace(1)* bitcast ({ i32, i32 } addrspace(1)* @__block_literal_global to i32 (i32) addrspace(1)*) to i32 (i32) addrspace(4)*), align 8
++%struct.__opencl_block_literal_generic = type { i32, i32, i8 addrspace(4)* }
+
+- at __block_literal_global = internal addrspace(1) constant { i32, i32 } { i32 8, i32 4 }, align 4
++ at block_kernel.b1 = internal addrspace(2) constant %struct.__opencl_block_literal_generic addrspace(4)* addrspacecast (%struct.__opencl_block_literal_generic addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* @__block_literal_global to %struct.__opencl_block_literal_generic addrspace(1)*) to %struct.__opencl_block_literal_generic addrspace(4)*), align 4
++ at __block_literal_global = internal addrspace(1) constant { i32, i32, i8 addrspace(4)* } { i32 12, i32 4, i8 addrspace(4)* addrspacecast (i8* bitcast (i32 (i8 addrspace(4)*, i32)* @_block_invoke to i8*) to i8 addrspace(4)*) }, align 4
+
+-; Function Attrs: convergent nounwind
+-define spir_kernel void @block_kernel(i32 addrspace(1)* %res) #0 !kernel_arg_addr_space !4 !kernel_arg_access_qual !5 !kernel_arg_type !6 !kernel_arg_base_type !6 !kernel_arg_type_qual !7 {
++; Function Attrs: convergent noinline nounwind optnone
++define spir_kernel void @block_kernel(i32 addrspace(1)* %res) #0 !kernel_arg_addr_space !3 !kernel_arg_access_qual !4 !kernel_arg_type !5 !kernel_arg_base_type !5 !kernel_arg_type_qual !6 {
+ entry:
+- %res.addr = alloca i32 addrspace(1)*, align 8
+- store i32 addrspace(1)* %res, i32 addrspace(1)** %res.addr, align 8, !tbaa !10
+-
++ %res.addr = alloca i32 addrspace(1)*, align 4
++ store i32 addrspace(1)* %res, i32 addrspace(1)** %res.addr, align 4
+ ; CHECK-SPIRV: FunctionCall [[int]] {{[0-9]+}} [[block_invoke]] {{[0-9]+}} [[five]]
+ ; CHECK-LLVM: %call = call spir_func i32 @_block_invoke(i8 addrspace(4)* {{.*}}, i32 5)
+- %call = call spir_func i32 @_block_invoke(i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32 } addrspace(1)* @__block_literal_global to i8 addrspace(1)*) to i8 addrspace(4)*), i32 5) #2
+-
+- %0 = load i32 addrspace(1)*, i32 addrspace(1)** %res.addr, align 8, !tbaa !10
+- store i32 %call, i32 addrspace(1)* %0, align 4, !tbaa !14
++ %call = call spir_func i32 @_block_invoke(i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* @__block_literal_global to i8 addrspace(1)*) to i8 addrspace(4)*), i32 5) #2
++ %0 = load i32 addrspace(1)*, i32 addrspace(1)** %res.addr, align 4
++ store i32 %call, i32 addrspace(1)* %0, align 4
+ ret void
+ }
+
+-; CHECK-SPIRV: 5 Function [[int]] [[block_invoke]] 0 [[block_invoke_type]]
++; CHECK-SPIRV: 5 Function [[int]] [[block_invoke]] 2 [[block_invoke_type]]
+ ; CHECK-SPIRV-NEXT: 3 FunctionParameter [[int8Ptr]] {{[0-9]+}}
+ ; CHECK-SPIRV-NEXT: 3 FunctionParameter [[int]] {{[0-9]+}}
+ ; CHECK-LLVM: define internal spir_func i32 @_block_invoke(i8 addrspace(4)* {{.*}}, i32 %{{.*}})
+-; Function Attrs: convergent nounwind
++; Function Attrs: convergent noinline nounwind optnone
+ define internal spir_func i32 @_block_invoke(i8 addrspace(4)* %.block_descriptor, i32 %i) #1 {
+ entry:
+- %.block_descriptor.addr = alloca i8 addrspace(4)*, align 8
++ %.block_descriptor.addr = alloca i8 addrspace(4)*, align 4
+ %i.addr = alloca i32, align 4
+- store i8 addrspace(4)* %.block_descriptor, i8 addrspace(4)** %.block_descriptor.addr, align 8
+- %block = bitcast i8 addrspace(4)* %.block_descriptor to <{ i32, i32 }> addrspace(4)*
+- store i32 %i, i32* %i.addr, align 4, !tbaa !14
+- %0 = load i32, i32* %i.addr, align 4, !tbaa !14
++ %block.addr = alloca <{ i32, i32, i8 addrspace(4)* }> addrspace(4)*, align 4
++ store i8 addrspace(4)* %.block_descriptor, i8 addrspace(4)** %.block_descriptor.addr, align 4
++ %block = bitcast i8 addrspace(4)* %.block_descriptor to <{ i32, i32, i8 addrspace(4)* }> addrspace(4)*
++ store i32 %i, i32* %i.addr, align 4
++ store <{ i32, i32, i8 addrspace(4)* }> addrspace(4)* %block, <{ i32, i32, i8 addrspace(4)* }> addrspace(4)** %block.addr, align 4
++ %0 = load i32, i32* %i.addr, align 4
+ %add = add nsw i32 %0, 1
+ ret i32 %add
+ }
+
+-attributes #0 = { convergent nounwind "correctly-rounded-divide-sqrt-fp-math"="false" "disable-tail-calls"="false" "less-precise-fpmad"="false" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-jump-tables"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "uniform-work-group-size"="false" "unsafe-fp-math"="false" "use-soft-float"="false" }
+-attributes #1 = { convergent nounwind "correctly-rounded-divide-sqrt-fp-math"="false" "disable-tail-calls"="false" "less-precise-fpmad"="false" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-jump-tables"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
++attributes #0 = { convergent noinline nounwind optnone "correctly-rounded-divide-sqrt-fp-math"="false" "denorms-are-zero"="false" "disable-tail-calls"="false" "less-precise-fpmad"="false" "min-legal-vector-width"="0" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-jump-tables"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "uniform-work-group-size"="false" "unsafe-fp-math"="false" "use-soft-float"="false" }
++attributes #1 = { convergent noinline nounwind optnone "correctly-rounded-divide-sqrt-fp-math"="false" "denorms-are-zero"="false" "disable-tail-calls"="false" "less-precise-fpmad"="false" "min-legal-vector-width"="0" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-jump-tables"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
+ attributes #2 = { convergent }
+
+ !llvm.module.flags = !{!0}
+-!opencl.enable.FP_CONTRACT = !{}
+ !opencl.ocl.version = !{!1}
+ !opencl.spir.version = !{!1}
+-!opencl.used.extensions = !{!2}
+-!opencl.used.optional.core.features = !{!2}
+-!opencl.compiler.options = !{!2}
+-!llvm.ident = !{!3}
++!llvm.ident = !{!2}
+
+ !0 = !{i32 1, !"wchar_size", i32 4}
+ !1 = !{i32 2, i32 0}
+-!2 = !{}
+-!3 = !{!"clang version 7.0.0"}
+-!4 = !{i32 1}
+-!5 = !{!"none"}
+-!6 = !{!"int*"}
+-!7 = !{!""}
+-!8 = !{i1 false}
+-!9 = !{i32 0}
+-!10 = !{!11, !11, i64 0}
+-!11 = !{!"any pointer", !12, i64 0}
+-!12 = !{!"omnipotent char", !13, i64 0}
+-!13 = !{!"Simple C/C++ TBAA"}
+-!14 = !{!15, !15, i64 0}
+-!15 = !{!"int", !12, i64 0}
++!2 = !{!"clang version 9.0.0 (https://llvm.org/git/clang 04fb8964a801a5c5d7baa5a22272243a7d183896) (https://llvm.org/git/llvm 384f64397f6ad95a361b72d62c07d7bac9f24163)"}
++!3 = !{i32 1}
++!4 = !{!"none"}
++!5 = !{!"int*"}
++!6 = !{!""}
+diff --git a/test/literal-struct.ll b/test/literal-struct.ll
+index c52170a..52a731a 100644
+--- a/test/literal-struct.ll
++++ b/test/literal-struct.ll
+@@ -2,7 +2,7 @@
+ ; structs, i.e. structs whose type has no name. Typicaly clang generate such
+ ; structs if the kernel contains OpenCL 2.0 blocks. The IR was produced with
+ ; the following command:
+-; clang -cc1 -triple spir -cl-std=cl2.0 -O0 -finclude-default-header literal-struct.cl -emit-llvm -o test/literal-struct.ll
++; clang -cc1 -triple spir -cl-std=cl2.0 -O0 literal-struct.cl -emit-llvm -o test/literal-struct.ll
+
+ ; literal-struct.cl:
+ ; void foo()
+@@ -14,25 +14,28 @@
+ ; RUN: llvm-as < %s | llvm-spirv -spirv-text -o %t
+ ; RUN: FileCheck < %t %s
+
+-; CHECK-DAG: TypeInt [[Int:[0-9]+]] 32 0
+-; CHECK-DAG: TypeStruct [[StructType:[0-9]+]] [[Int]] [[Int]] {{$}}
++; CHECK: TypeInt [[Int:[0-9]+]] 32 0
++; CHECK: TypeInt [[Int8:[0-9]+]] 8 0
++; CHECK: TypePointer [[Int8Ptr:[0-9]+]] 8 [[Int8]]
++; CHECK: TypeStruct [[StructType:[0-9]+]] [[Int]] [[Int]] [[Int8Ptr]]
+
+ target datalayout = "e-p:32:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024"
+ target triple = "spir"
+
+- at __block_literal_global = internal addrspace(1) constant { i32, i32 } { i32 8, i32 4 }, align 4
++%struct.__opencl_block_literal_generic = type { i32, i32, i8 addrspace(4)* }
++
++ at __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)*)* @__foo_block_invoke to i8*) to i8 addrspace(4)*) }, align 4
+ ; CHECK: ConstantComposite [[StructType]]
+
+-; This is artificial case is added to cover ConstantNull instrucitions with TypeStruct.
+- at __block_literal_global.1 = internal addrspace(1) constant { i32, i32 } zeroinitializer, align 4
++ at __block_literal_global.1 = internal addrspace(1) constant { i32, i32, i8 addrspace(4)* } zeroinitializer, align 4
+ ; CHECK: ConstantNull [[StructType]]
+
+ ; Function Attrs: convergent noinline nounwind optnone
+ define spir_func void @foo() #0 {
+ entry:
+- %myBlock = alloca void () addrspace(4)*, align 4
+- store void () addrspace(4)* addrspacecast (void () addrspace(1)* bitcast ({ i32, i32 } addrspace(1)* @__block_literal_global to void () addrspace(1)*) to void () addrspace(4)*), void () addrspace(4)** %myBlock, align 4
+- call spir_func void @__foo_block_invoke(i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32 } addrspace(1)* @__block_literal_global to i8 addrspace(1)*) to i8 addrspace(4)*)) #1
++ %myBlock = alloca %struct.__opencl_block_literal_generic addrspace(4)*, align 4
++ store %struct.__opencl_block_literal_generic addrspace(4)* addrspacecast (%struct.__opencl_block_literal_generic addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* @__block_literal_global to %struct.__opencl_block_literal_generic addrspace(1)*) to %struct.__opencl_block_literal_generic addrspace(4)*), %struct.__opencl_block_literal_generic addrspace(4)** %myBlock, align 4
++ call spir_func void @__foo_block_invoke(i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* @__block_literal_global to i8 addrspace(1)*) to i8 addrspace(4)*)) #1
+ ret void
+ }
+
+@@ -40,14 +43,14 @@ entry:
+ define internal spir_func void @__foo_block_invoke(i8 addrspace(4)* %.block_descriptor) #0 {
+ entry:
+ %.block_descriptor.addr = alloca i8 addrspace(4)*, align 4
+- %block.addr = alloca <{ i32, i32 }> addrspace(4)*, align 4
++ %block.addr = alloca <{ i32, i32, i8 addrspace(4)* }> addrspace(4)*, align 4
+ store i8 addrspace(4)* %.block_descriptor, i8 addrspace(4)** %.block_descriptor.addr, align 4
+- %block = bitcast i8 addrspace(4)* %.block_descriptor to <{ i32, i32 }> addrspace(4)*
+- store <{ i32, i32 }> addrspace(4)* %block, <{ i32, i32 }> addrspace(4)** %block.addr, align 4
++ %block = bitcast i8 addrspace(4)* %.block_descriptor to <{ i32, i32, i8 addrspace(4)* }> addrspace(4)*
++ store <{ i32, i32, i8 addrspace(4)* }> addrspace(4)* %block, <{ i32, i32, i8 addrspace(4)* }> addrspace(4)** %block.addr, align 4
+ ret void
+ }
+
+-attributes #0 = { convergent noinline nounwind optnone "correctly-rounded-divide-sqrt-fp-math"="false" "denorms-are-zero"="false" "disable-tail-calls"="false" "less-precise-fpmad"="false" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-jump-tables"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
++attributes #0 = { convergent noinline nounwind optnone "correctly-rounded-divide-sqrt-fp-math"="false" "denorms-are-zero"="false" "disable-tail-calls"="false" "less-precise-fpmad"="false" "min-legal-vector-width"="0" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-jump-tables"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
+ attributes #1 = { convergent }
+
+ !llvm.module.flags = !{!0}
+@@ -57,4 +60,4 @@ attributes #1 = { convergent }
+
+ !0 = !{i32 1, !"wchar_size", i32 4}
+ !1 = !{i32 2, i32 0}
+-!2 = !{!"clang version 8.0.0 "}
++!2 = !{!"clang version 9.0.0 (https://llvm.org/git/clang 04fb8964a801a5c5d7baa5a22272243a7d183896) (https://llvm.org/git/llvm 384f64397f6ad95a361b72d62c07d7bac9f24163)"}
+diff --git a/test/transcoding/block_w_struct_return.ll b/test/transcoding/block_w_struct_return.ll
+index 76e29f0..df89b13 100644
+--- a/test/transcoding/block_w_struct_return.ll
++++ b/test/transcoding/block_w_struct_return.ll
+@@ -16,6 +16,8 @@
+ ; res[tid] = kernelBlock(aa).a - 6;
+ ; }
+
++; clang -cc1 -triple spir -cl-std=cl2.0 -disable-llvm-passes -finclude-default-header block_w_struct_return.cl -emit-llvm -o test/transcoding/block_w_struct_return.ll
++
+ ; RUN: llvm-as %s -o %t.bc
+ ; RUN: llvm-spirv %t.bc -spirv-text -o %t.spv.txt
+ ; RUN: FileCheck < %t.spv.txt %s --check-prefix=CHECK-SPIRV
+@@ -27,12 +29,14 @@
+ ; CHECK-SPIRV: Name [[BlockInv:[0-9]+]] "__block_ret_struct_block_invoke"
+
+ ; CHECK-SPIRV: 4 TypeInt [[IntTy:[0-9]+]] 32
++; CHECK-SPIRV: 4 TypeInt [[Int8Ty:[0-9]+]] 8
++; CHECK-SPIRV: 4 TypePointer [[Int8Ptr:[0-9]+]] 8 [[Int8Ty]]
+ ; CHECK-SPIRV: 3 TypeStruct [[StructTy:[0-9]+]] [[IntTy]]
+ ; CHECK-SPIRV: 4 TypePointer [[StructPtrTy:[0-9]+]] 7 [[StructTy]]
+
+ ; CHECK-SPIRV: 4 Variable [[StructPtrTy]] [[StructArg:[0-9]+]] 7
+ ; CHECK-SPIRV: 4 Variable [[StructPtrTy]] [[StructRet:[0-9]+]] 7
+-; CHECK-SPIRV: 4 PtrCastToGeneric {{[0-9]+}} [[BlockLit:[0-9]+]] {{[0-9]+}}
++; CHECK-SPIRV: 4 PtrCastToGeneric [[Int8Ptr]] [[BlockLit:[0-9]+]] {{[0-9]+}}
+ ; CHECK-SPIRV: 7 FunctionCall {{[0-9]+}} {{[0-9]+}} [[BlockInv]] [[StructRet]] [[BlockLit]] [[StructArg]]
+
+ ; CHECK-LLVM: %[[StructA:.*]] = type { i32 }
+@@ -41,20 +45,21 @@
+ target datalayout = "e-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024"
+ target triple = "spir64-unknown-unknown"
+
++%struct.__opencl_block_literal_generic = type { i32, i32, i8 addrspace(4)* }
+ %struct.A = type { i32 }
+
+- at __block_literal_global = internal addrspace(1) constant { i32, i32 } { i32 8, i32 4 }, align 4
++ at __block_literal_global = internal addrspace(1) constant { i32, i32, i8 addrspace(4)* } { i32 16, i32 8, i8 addrspace(4)* addrspacecast (i8* bitcast (void (%struct.A*, i8 addrspace(4)*, %struct.A*)* @__block_ret_struct_block_invoke to i8*) to i8 addrspace(4)*) }, align 8
+
+ ; Function Attrs: convergent noinline nounwind optnone
+-define spir_kernel void @block_ret_struct(i32 addrspace(1)* %res) #0 !kernel_arg_addr_space !4 !kernel_arg_access_qual !5 !kernel_arg_type !6 !kernel_arg_base_type !6 !kernel_arg_type_qual !7 !kernel_arg_host_accessible !8 !kernel_arg_pipe_depth !9 !kernel_arg_pipe_io !7 !kernel_arg_buffer_location !7 {
++define spir_kernel void @block_ret_struct(i32 addrspace(1)* %res) #0 !kernel_arg_addr_space !3 !kernel_arg_access_qual !4 !kernel_arg_type !5 !kernel_arg_base_type !5 !kernel_arg_type_qual !6 {
+ entry:
+ %res.addr = alloca i32 addrspace(1)*, align 8
+- %kernelBlock = alloca void (%struct.A*, %struct.A*) addrspace(4)*, align 8
++ %kernelBlock = alloca %struct.__opencl_block_literal_generic addrspace(4)*, align 8
+ %tid = alloca i64, align 8
+ %aa = alloca %struct.A, align 4
+ %tmp = alloca %struct.A, align 4
+ store i32 addrspace(1)* %res, i32 addrspace(1)** %res.addr, align 8
+- store void (%struct.A*, %struct.A*) addrspace(4)* addrspacecast (void (%struct.A*, %struct.A*) addrspace(1)* bitcast ({ i32, i32 } addrspace(1)* @__block_literal_global to void (%struct.A*, %struct.A*) addrspace(1)*) to void (%struct.A*, %struct.A*) addrspace(4)*), void (%struct.A*, %struct.A*) addrspace(4)** %kernelBlock, align 8
++ store %struct.__opencl_block_literal_generic addrspace(4)* addrspacecast (%struct.__opencl_block_literal_generic addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* @__block_literal_global to %struct.__opencl_block_literal_generic addrspace(1)*) to %struct.__opencl_block_literal_generic addrspace(4)*), %struct.__opencl_block_literal_generic addrspace(4)** %kernelBlock, align 8
+ %call = call spir_func i64 @_Z13get_global_idj(i32 0) #4
+ store i64 %call, i64* %tid, align 8
+ %0 = load i32 addrspace(1)*, i32 addrspace(1)** %res.addr, align 8
+@@ -63,7 +68,7 @@ entry:
+ store i32 -1, i32 addrspace(1)* %arrayidx, align 4
+ %a = getelementptr inbounds %struct.A, %struct.A* %aa, i32 0, i32 0
+ store i32 5, i32* %a, align 4
+- call spir_func void @__block_ret_struct_block_invoke(%struct.A* sret %tmp, i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32 } addrspace(1)* @__block_literal_global to i8 addrspace(1)*) to i8 addrspace(4)*), %struct.A* byval align 4 %aa) #5
++ call spir_func void @__block_ret_struct_block_invoke(%struct.A* sret %tmp, i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* @__block_literal_global to i8 addrspace(1)*) to i8 addrspace(4)*), %struct.A* byval align 4 %aa) #5
+ %a1 = getelementptr inbounds %struct.A, %struct.A* %tmp, i32 0, i32 0
+ %2 = load i32, i32* %a1, align 4
+ %sub = sub nsw i32 %2, 6
+@@ -78,10 +83,10 @@ entry:
+ define internal spir_func void @__block_ret_struct_block_invoke(%struct.A* noalias sret %agg.result, i8 addrspace(4)* %.block_descriptor, %struct.A* byval align 4 %a) #1 {
+ entry:
+ %.block_descriptor.addr = alloca i8 addrspace(4)*, align 8
+- %block.addr = alloca <{ i32, i32 }> addrspace(4)*, align 8
++ %block.addr = alloca <{ i32, i32, i8 addrspace(4)* }> addrspace(4)*, align 8
+ store i8 addrspace(4)* %.block_descriptor, i8 addrspace(4)** %.block_descriptor.addr, align 8
+- %block = bitcast i8 addrspace(4)* %.block_descriptor to <{ i32, i32 }> addrspace(4)*
+- store <{ i32, i32 }> addrspace(4)* %block, <{ i32, i32 }> addrspace(4)** %block.addr, align 8
++ %block = bitcast i8 addrspace(4)* %.block_descriptor to <{ i32, i32, i8 addrspace(4)* }> addrspace(4)*
++ store <{ i32, i32, i8 addrspace(4)* }> addrspace(4)* %block, <{ i32, i32, i8 addrspace(4)* }> addrspace(4)** %block.addr, align 8
+ %a1 = getelementptr inbounds %struct.A, %struct.A* %a, i32 0, i32 0
+ store i32 6, i32* %a1, align 4
+ %0 = bitcast %struct.A* %agg.result to i8*
+@@ -96,30 +101,22 @@ declare void @llvm.memcpy.p0i8.p0i8.i64(i8* nocapture writeonly, i8* nocapture r
+ ; Function Attrs: convergent nounwind readnone
+ declare spir_func i64 @_Z13get_global_idj(i32) #3
+
+-attributes #0 = { convergent noinline nounwind optnone "correctly-rounded-divide-sqrt-fp-math"="false" "denorms-are-zero"="false" "disable-tail-calls"="false" "less-precise-fpmad"="false" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-jump-tables"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "uniform-work-group-size"="false" "unsafe-fp-math"="false" "use-soft-float"="false" }
+-attributes #1 = { convergent noinline nounwind optnone "correctly-rounded-divide-sqrt-fp-math"="false" "denorms-are-zero"="false" "disable-tail-calls"="false" "less-precise-fpmad"="false" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-jump-tables"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
++attributes #0 = { convergent noinline nounwind optnone "correctly-rounded-divide-sqrt-fp-math"="false" "denorms-are-zero"="false" "disable-tail-calls"="false" "less-precise-fpmad"="false" "min-legal-vector-width"="0" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-jump-tables"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "uniform-work-group-size"="false" "unsafe-fp-math"="false" "use-soft-float"="false" }
++attributes #1 = { convergent noinline nounwind optnone "correctly-rounded-divide-sqrt-fp-math"="false" "denorms-are-zero"="false" "disable-tail-calls"="false" "less-precise-fpmad"="false" "min-legal-vector-width"="0" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-jump-tables"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
+ attributes #2 = { argmemonly nounwind }
+ attributes #3 = { convergent nounwind readnone "correctly-rounded-divide-sqrt-fp-math"="false" "denorms-are-zero"="false" "disable-tail-calls"="false" "less-precise-fpmad"="false" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
+ attributes #4 = { convergent nounwind readnone }
+ attributes #5 = { convergent }
+
+ !llvm.module.flags = !{!0}
+-!opencl.enable.FP_CONTRACT = !{}
+ !opencl.ocl.version = !{!1}
+ !opencl.spir.version = !{!1}
+-!opencl.used.extensions = !{!2}
+-!opencl.used.optional.core.features = !{!2}
+-!opencl.compiler.options = !{!2}
+-!llvm.ident = !{!3}
++!llvm.ident = !{!2}
+
+ !0 = !{i32 1, !"wchar_size", i32 4}
+ !1 = !{i32 2, i32 0}
+-!2 = !{}
+-!3 = !{!"clang version 7.0.0"}
+-!4 = !{i32 1}
+-!5 = !{!"none"}
+-!6 = !{!"int*"}
+-!7 = !{!""}
+-!8 = !{i1 false}
+-!9 = !{i32 0}
+-
++!2 = !{!"clang version 9.0.0 (https://llvm.org/git/clang 04fb8964a801a5c5d7baa5a22272243a7d183896) (https://llvm.org/git/llvm 384f64397f6ad95a361b72d62c07d7bac9f24163)"}
++!3 = !{i32 1}
++!4 = !{!"none"}
++!5 = !{!"int*"}
++!6 = !{!""}
+diff --git a/test/transcoding/enqueue_kernel.ll b/test/transcoding/enqueue_kernel.ll
+index 0d29c71..435871d 100644
+--- a/test/transcoding/enqueue_kernel.ll
++++ b/test/transcoding/enqueue_kernel.ll
+@@ -51,11 +51,12 @@
+ ; ModuleID = 'enqueue_kernel.cl'
+ source_filename = "enqueue_kernel.cl"
+ target datalayout = "e-p:32:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024"
+-target triple = "spir-unknown-unknown"
++target triple = "spir"
+
+ %opencl.queue_t = type opaque
+ %struct.ndrange_t = type { i32 }
+ %opencl.clk_event_t = type opaque
++%struct.__opencl_block_literal_generic = type { i32, i32, i8 addrspace(4)* }
+
+ ; CHECK-SPIRV: EntryPoint {{[0-9]+}} [[BlockKer1:[0-9]+]] "__device_side_enqueue_block_invoke_kernel"
+ ; CHECK-SPIRV: EntryPoint {{[0-9]+}} [[BlockKer2:[0-9]+]] "__device_side_enqueue_block_invoke_2_kernel"
+@@ -66,89 +67,123 @@ target triple = "spir-unknown-unknown"
+
+ ; CHECK-SPIRV: TypeInt [[Int32Ty:[0-9]+]] 32
+ ; CHECK-SPIRV: TypeInt [[Int8Ty:[0-9]+]] 8
+-; CHECK-SPIRV: Constant [[Int32Ty]] [[ConstInt8:[0-9]+]] 8
+ ; CHECK-SPIRV: Constant [[Int32Ty]] [[ConstInt0:[0-9]+]] 0
+-; CHECK-SPIRV: Constant [[Int32Ty]] [[ConstInt17:[0-9]+]] 17
++; CHECK-SPIRV: Constant [[Int32Ty]] [[ConstInt17:[0-9]+]] 21
+ ; CHECK-SPIRV: Constant [[Int32Ty]] [[ConstInt2:[0-9]+]] 2
+-; CHECK-SPIRV: Constant [[Int32Ty]] [[ConstInt20:[0-9]+]] 20
+-; CHECK-SPIRV: TypeVoid [[VoidTy:[0-9]+]]
++; CHECK-SPIRV: Constant [[Int32Ty]] [[ConstInt8:[0-9]+]] 8
++; CHECK-SPIRV: Constant [[Int32Ty]] [[ConstInt20:[0-9]+]] 24
+
+ ; CHECK-SPIRV: TypePointer {{[0-9]+}} 7 {{[0-9]+}}
++; CHECK-SPIRV: TypePointer [[Int8PtrGenTy:[0-9]+]] 8 [[Int8Ty]]
++; CHECK-SPIRV: TypeVoid [[VoidTy:[0-9]+]]
+ ; CHECK-SPIRV: TypePointer [[Int32LocPtrTy:[0-9]+]] 7 [[Int32Ty]]
+ ; CHECK-SPIRV: TypeDeviceEvent [[EventTy:[0-9]+]]
+-; CHECK-SPIRV: TypePointer [[Int8PtrGenTy:[0-9]+]] 8 [[Int8Ty]]
+ ; CHECK-SPIRV: TypePointer [[EventPtrTy:[0-9]+]] 8 [[EventTy]]
+ ; CHECK-SPIRV: TypeFunction [[BlockTy1:[0-9]+]] [[VoidTy]] [[Int8PtrGenTy]]
+ ; CHECK-SPIRV: TypeFunction [[BlockTy2:[0-9]+]] [[VoidTy]] [[Int8PtrGenTy]]
+ ; CHECK-SPIRV: TypeFunction [[BlockTy3:[0-9]+]] [[VoidTy]] [[Int8PtrGenTy]]
+ ; CHECK-SPIRV: ConstantNull [[EventPtrTy]] [[EventNull:[0-9]+]]
+
+-; CHECK-LLVM: [[BlockTy1:%[0-9]+]] = type { i32, i32 }
+-; CHECK-LLVM: [[BlockTy2:%[0-9]+]] = type <{ i32, i32, i32 addrspace(1)*, i32, i8 }>
+-; CHECK-LLVM: [[BlockTy3:%[0-9]+]] = type <{ i32, i32, i32 addrspace(1)*, i32, i32 addrspace(1)* }>
+-; CHECK-LLVM: [[BlockTy4:%[0-9]+]] = type <{ i32, i32 }>
++; CHECK-LLVM: [[BlockTy1:%[0-9]+]] = type { i32, i32, i8 addrspace(4)* }
++; CHECK-LLVM: [[BlockTy2:%[0-9]+]] = type <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i8 }>
++; CHECK-LLVM: [[BlockTy3:%[0-9]+]] = type <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }>
++; CHECK-LLVM: [[BlockTy4:%[0-9]+]] = type <{ i32, i32, i8 addrspace(4)* }>
+
+-; CHECK-LLVM: @__block_literal_global = internal addrspace(1) constant [[BlockTy1]] { i32 8, i32 4 }, align 4
+-; CHECK-LLVM: @__block_literal_global.1 = internal addrspace(1) constant [[BlockTy1]] { i32 8, i32 4 }, align 4
++; CHECK-LLVM: @__block_literal_global = internal addrspace(1) constant [[BlockTy1]] { i32 12, i32 4, i8 addrspace(4)* addrspacecast (i8* null to i8 addrspace(4)*) }, align 4
++; CHECK-LLVM: @__block_literal_global.1 = internal addrspace(1) constant [[BlockTy1]] { i32 12, i32 4, i8 addrspace(4)* addrspacecast (i8* null to i8 addrspace(4)*) }, align 4
+
+- at __block_literal_global = internal addrspace(1) constant { i32, i32 } { i32 8, i32 4 }, align 4
+- at __block_literal_global.1 = internal addrspace(1) constant { i32, i32 } { i32 8, i32 4 }, align 4
++ at __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)*)* @__device_side_enqueue_block_invoke_3 to i8*) to i8 addrspace(4)*) }, align 4
++ at __block_literal_global.1 = 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)*, i8 addrspace(3)*, i8 addrspace(3)*)* @__device_side_enqueue_block_invoke_4 to i8*) to i8 addrspace(4)*) }, align 4
+
+ ; Function Attrs: convergent noinline nounwind optnone
+-define spir_kernel void @device_side_enqueue(i32 addrspace(1)* %a, i32 addrspace(1)* %b, i32 %i, i8 signext %c0) #0 !kernel_arg_addr_space !4 !kernel_arg_access_qual !5 !kernel_arg_type !6 !kernel_arg_base_type !6 !kernel_arg_type_qual !7 {
++define spir_kernel void @device_side_enqueue(i32 addrspace(1)* %a, i32 addrspace(1)* %b, i32 %i, i8 signext %c0) #0 !kernel_arg_addr_space !3 !kernel_arg_access_qual !4 !kernel_arg_type !5 !kernel_arg_base_type !5 !kernel_arg_type_qual !6 {
+ entry:
++ %a.addr = alloca i32 addrspace(1)*, align 4
++ %b.addr = alloca i32 addrspace(1)*, align 4
++ %i.addr = alloca i32, align 4
++ %c0.addr = alloca i8, align 1
+ %default_queue = alloca %opencl.queue_t*, align 4
+ %flags = alloca i32, align 4
+ %ndrange = alloca %struct.ndrange_t, align 4
+ %clk_event = alloca %opencl.clk_event_t*, align 4
+ %event_wait_list = alloca %opencl.clk_event_t*, align 4
+ %event_wait_list2 = alloca [1 x %opencl.clk_event_t*], align 4
+- %block = alloca <{ i32, i32, i32 addrspace(1)*, i32, i8 }>, align 4
+- %block3 = alloca <{ i32, i32, i32 addrspace(1)*, i32, i32 addrspace(1)* }>, align 4
++ %block = alloca <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i8 }>, align 4
++ %tmp = alloca %struct.ndrange_t, align 4
++ %block3 = alloca <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }>, align 4
++ %tmp4 = alloca %struct.ndrange_t, align 4
+ %c = alloca i8, align 1
++ %tmp11 = alloca %struct.ndrange_t, align 4
++ %block_sizes = alloca [1 x i32], align 4
++ %tmp12 = alloca %struct.ndrange_t, align 4
++ %block_sizes13 = alloca [3 x i32], align 4
++ store i32 addrspace(1)* %a, i32 addrspace(1)** %a.addr, align 4
++ store i32 addrspace(1)* %b, i32 addrspace(1)** %b.addr, align 4
++ store i32 %i, i32* %i.addr, align 4
++ store i8 %c0, i8* %c0.addr, align 1
+ store i32 0, i32* %flags, align 4
+ %arrayinit.begin = getelementptr inbounds [1 x %opencl.clk_event_t*], [1 x %opencl.clk_event_t*]* %event_wait_list2, i32 0, i32 0
+ %0 = load %opencl.clk_event_t*, %opencl.clk_event_t** %clk_event, align 4
+ store %opencl.clk_event_t* %0, %opencl.clk_event_t** %arrayinit.begin, align 4
+ %1 = load %opencl.queue_t*, %opencl.queue_t** %default_queue, align 4
+ %2 = load i32, i32* %flags, align 4
+- %block.size = getelementptr inbounds <{ i32, i32, i32 addrspace(1)*, i32, i8 }>, <{ i32, i32, i32 addrspace(1)*, i32, i8 }>* %block, i32 0, i32 0
+- store i32 17, i32* %block.size, align 4
+- %block.align = getelementptr inbounds <{ i32, i32, i32 addrspace(1)*, i32, i8 }>, <{ i32, i32, i32 addrspace(1)*, i32, i8 }>* %block, i32 0, i32 1
++ %3 = bitcast %struct.ndrange_t* %tmp to i8*
++ %4 = bitcast %struct.ndrange_t* %ndrange to i8*
++ call void @llvm.memcpy.p0i8.p0i8.i32(i8* align 4 %3, i8* align 4 %4, i32 4, i1 false)
++ %block.size = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i8 }>, <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i8 }>* %block, i32 0, i32 0
++ store i32 21, i32* %block.size, align 4
++ %block.align = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i8 }>, <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i8 }>* %block, i32 0, i32 1
+ store i32 4, i32* %block.align, align 4
+- %block.captured = getelementptr inbounds <{ i32, i32, i32 addrspace(1)*, i32, i8 }>, <{ i32, i32, i32 addrspace(1)*, i32, i8 }>* %block, i32 0, i32 2
+- store i32 addrspace(1)* %a, i32 addrspace(1)** %block.captured, align 4
+- %block.captured1 = getelementptr inbounds <{ i32, i32, i32 addrspace(1)*, i32, i8 }>, <{ i32, i32, i32 addrspace(1)*, i32, i8 }>* %block, i32 0, i32 3
+- store i32 %i, i32* %block.captured1, align 4
+- %block.captured2 = getelementptr inbounds <{ i32, i32, i32 addrspace(1)*, i32, i8 }>, <{ i32, i32, i32 addrspace(1)*, i32, i8 }>* %block, i32 0, i32 4
+- store i8 %c0, i8* %block.captured2, align 4
+- %3 = bitcast <{ i32, i32, i32 addrspace(1)*, i32, i8 }>* %block to void ()*
+- %4 = addrspacecast void ()* %3 to i8 addrspace(4)*
++ %block.invoke = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i8 }>, <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i8 }>* %block, i32 0, i32 2
++ store i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*)* @__device_side_enqueue_block_invoke to i8*) to i8 addrspace(4)*), i8 addrspace(4)** %block.invoke, align 4
++ %block.captured = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i8 }>, <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i8 }>* %block, i32 0, i32 3
++ %5 = load i32 addrspace(1)*, i32 addrspace(1)** %a.addr, align 4
++ store i32 addrspace(1)* %5, i32 addrspace(1)** %block.captured, align 4
++ %block.captured1 = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i8 }>, <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i8 }>* %block, i32 0, i32 4
++ %6 = load i32, i32* %i.addr, align 4
++ store i32 %6, i32* %block.captured1, align 4
++ %block.captured2 = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i8 }>, <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i8 }>* %block, i32 0, i32 5
++ %7 = load i8, i8* %c0.addr, align 1
++ store i8 %7, i8* %block.captured2, align 4
++ %8 = bitcast <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i8 }>* %block to %struct.__opencl_block_literal_generic*
++ %9 = addrspacecast %struct.__opencl_block_literal_generic* %8 to i8 addrspace(4)*
+
+ ; CHECK-SPIRV: PtrCastToGeneric [[Int8PtrGenTy]] [[BlockLit1:[0-9]+]]
+ ; CHECK-SPIRV: EnqueueKernel [[Int32Ty]] {{[0-9]+}} {{[0-9]+}} {{[0-9]+}} {{[0-9]+}}
+ ; [[ConstInt0]] [[EventNull]] [[EventNull]]
+ ; [[BlockKer1]] [[BlockLit1]] [[ConstInt17]] [[ConstInt8]]
+
+-; CHECK-LLVM: [[Block2:%[0-9]+]] = addrspacecast [[BlockTy2]]* %block to i8 addrspace(4)*
++; CHECK-LLVM: [[Block2:%[0-9]+]] = bitcast [[BlockTy2]]* %block to %struct.__opencl_block_literal_generic*
++; CHECK-LLVM: [[Block2Ptr:%[0-9]+]] = addrspacecast %struct.__opencl_block_literal_generic* [[Block2]] to i8 addrspace(4)*
+ ; CHECK-LLVM: [[BlockInv2:%[0-9]+]] = addrspacecast void (i8 addrspace(4)*)* @__device_side_enqueue_block_invoke_kernel to i8 addrspace(4)*
+-; CHECK-LLVM: call i32 @__enqueue_kernel_basic_events(%opencl.queue_t* {{.*}}, i32 {{.*}}, %struct.ndrange_t* {{.*}}, i32 0, %opencl.clk_event_t* addrspace(4)* null, %opencl.clk_event_t* addrspace(4)* null, i8 addrspace(4)* [[BlockInv2]], i8 addrspace(4)* [[Block2]])
+-
+- %5 = call i32 @__enqueue_kernel_basic(%opencl.queue_t* %1, i32 %2, %struct.ndrange_t* byval %ndrange, i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*)* @__device_side_enqueue_block_invoke_kernel to i8*) to i8 addrspace(4)*), i8 addrspace(4)* %4)
+- %6 = addrspacecast %opencl.clk_event_t** %event_wait_list to %opencl.clk_event_t* addrspace(4)*
+- %7 = addrspacecast %opencl.clk_event_t** %clk_event to %opencl.clk_event_t* addrspace(4)*
+- %block.size5 = getelementptr inbounds <{ i32, i32, i32 addrspace(1)*, i32, i32 addrspace(1)* }>, <{ i32, i32, i32 addrspace(1)*, i32, i32 addrspace(1)* }>* %block3, i32 0, i32 0
+- store i32 20, i32* %block.size5, align 4
+- %block.align6 = getelementptr inbounds <{ i32, i32, i32 addrspace(1)*, i32, i32 addrspace(1)* }>, <{ i32, i32, i32 addrspace(1)*, i32, i32 addrspace(1)* }>* %block3, i32 0, i32 1
++; CHECK-LLVM: call i32 @__enqueue_kernel_basic_events(%opencl.queue_t* {{.*}}, i32 {{.*}}, %struct.ndrange_t* {{.*}}, i32 0, %opencl.clk_event_t* addrspace(4)* null, %opencl.clk_event_t* addrspace(4)* null, i8 addrspace(4)* [[BlockInv2]], i8 addrspace(4)* [[Block2Ptr]])
++
++ %10 = call i32 @__enqueue_kernel_basic(%opencl.queue_t* %1, i32 %2, %struct.ndrange_t* byval %tmp, i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*)* @__device_side_enqueue_block_invoke_kernel to i8*) to i8 addrspace(4)*), i8 addrspace(4)* %9)
++ %11 = load %opencl.queue_t*, %opencl.queue_t** %default_queue, align 4
++ %12 = load i32, i32* %flags, align 4
++ %13 = bitcast %struct.ndrange_t* %tmp4 to i8*
++ %14 = bitcast %struct.ndrange_t* %ndrange to i8*
++ call void @llvm.memcpy.p0i8.p0i8.i32(i8* align 4 %13, i8* align 4 %14, i32 4, i1 false)
++ %15 = addrspacecast %opencl.clk_event_t** %event_wait_list to %opencl.clk_event_t* addrspace(4)*
++ %16 = addrspacecast %opencl.clk_event_t** %clk_event to %opencl.clk_event_t* addrspace(4)*
++ %block.size5 = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }>, <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }>* %block3, i32 0, i32 0
++ store i32 24, i32* %block.size5, align 4
++ %block.align6 = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }>, <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }>* %block3, i32 0, i32 1
+ store i32 4, i32* %block.align6, align 4
+- %block.captured7 = getelementptr inbounds <{ i32, i32, i32 addrspace(1)*, i32, i32 addrspace(1)* }>, <{ i32, i32, i32 addrspace(1)*, i32, i32 addrspace(1)* }>* %block3, i32 0, i32 2
+- store i32 addrspace(1)* %a, i32 addrspace(1)** %block.captured7, align 4
+- %block.captured8 = getelementptr inbounds <{ i32, i32, i32 addrspace(1)*, i32, i32 addrspace(1)* }>, <{ i32, i32, i32 addrspace(1)*, i32, i32 addrspace(1)* }>* %block3, i32 0, i32 3
+- store i32 %i, i32* %block.captured8, align 4
+- %block.captured9 = getelementptr inbounds <{ i32, i32, i32 addrspace(1)*, i32, i32 addrspace(1)* }>, <{ i32, i32, i32 addrspace(1)*, i32, i32 addrspace(1)* }>* %block3, i32 0, i32 4
+- store i32 addrspace(1)* %b, i32 addrspace(1)** %block.captured9, align 4
+- %8 = bitcast <{ i32, i32, i32 addrspace(1)*, i32, i32 addrspace(1)* }>* %block3 to void ()*
+- %9 = addrspacecast void ()* %8 to i8 addrspace(4)*
++ %block.invoke7 = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }>, <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }>* %block3, i32 0, i32 2
++ store i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*)* @__device_side_enqueue_block_invoke_2 to i8*) to i8 addrspace(4)*), i8 addrspace(4)** %block.invoke7, align 4
++ %block.captured8 = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }>, <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }>* %block3, i32 0, i32 3
++ %17 = load i32 addrspace(1)*, i32 addrspace(1)** %a.addr, align 4
++ store i32 addrspace(1)* %17, i32 addrspace(1)** %block.captured8, align 4
++ %block.captured9 = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }>, <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }>* %block3, i32 0, i32 4
++ %18 = load i32, i32* %i.addr, align 4
++ store i32 %18, i32* %block.captured9, align 4
++ %block.captured10 = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }>, <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }>* %block3, i32 0, i32 5
++ %19 = load i32 addrspace(1)*, i32 addrspace(1)** %b.addr, align 4
++ store i32 addrspace(1)* %19, i32 addrspace(1)** %block.captured10, align 4
++ %20 = bitcast <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }>* %block3 to %struct.__opencl_block_literal_generic*
++ %21 = addrspacecast %struct.__opencl_block_literal_generic* %20 to i8 addrspace(4)*
++
+
+ ; CHECK-SPIRV: PtrCastToGeneric [[EventPtrTy]] [[Event1:[0-9]+]]
+ ; CHECK-SPIRV: PtrCastToGeneric [[EventPtrTy]] [[Event2:[0-9]+]]
+@@ -158,16 +193,24 @@ entry:
+ ; [[ConstInt2]] [[Event1]] [[Event2]]
+ ; [[BlockKer2]] [[BlockLit2]] [[ConstInt20]] [[ConstInt8]]
+
+-; CHECK-LLVM: [[Block3:%[0-9]+]] = addrspacecast [[BlockTy3]]* %block3 to i8 addrspace(4)*
++; CHECK-LLVM: [[Block3:%[0-9]+]] = bitcast [[BlockTy3]]* %block3 to %struct.__opencl_block_literal_generic*
++; CHECK-LLVM: [[Block3Ptr:%[0-9]+]] = addrspacecast %struct.__opencl_block_literal_generic* [[Block3]] to i8 addrspace(4)
+ ; CHECK-LLVM: [[BlockInv3:%[0-9]+]] = addrspacecast void (i8 addrspace(4)*)* @__device_side_enqueue_block_invoke_2_kernel to i8 addrspace(4)*
+-; CHECK-LLVM: call i32 @__enqueue_kernel_basic_events(%opencl.queue_t* {{.*}}, i32 {{.*}}, %struct.ndrange_t* {{.*}}, i32 2, %opencl.clk_event_t* addrspace(4)* {{.*}}, %opencl.clk_event_t* addrspace(4)* {{.*}}, i8 addrspace(4)* [[BlockInv3]], i8 addrspace(4)* [[Block3]])
+-
+- %10 = call i32 @__enqueue_kernel_basic_events(%opencl.queue_t* %1, i32 %2, %struct.ndrange_t* %ndrange, i32 2, %opencl.clk_event_t* addrspace(4)* %6, %opencl.clk_event_t* addrspace(4)* %7, i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*)* @__device_side_enqueue_block_invoke_2_kernel to i8*) to i8 addrspace(4)*), i8 addrspace(4)* %9)
+- %11 = alloca [1 x i32]
+- %12 = getelementptr [1 x i32], [1 x i32]* %11, i32 0, i32 0
+- %13 = load i8, i8* %c, align 1
+- %14 = zext i8 %13 to i32
+- store i32 %14, i32* %12, align 4
++; CHECK-LLVM: call i32 @__enqueue_kernel_basic_events(%opencl.queue_t* {{.*}}, i32 {{.*}}, %struct.ndrange_t* {{.*}}, i32 2, %opencl.clk_event_t* addrspace(4)* {{.*}}, %opencl.clk_event_t* addrspace(4)* {{.*}}, i8 addrspace(4)* [[BlockInv3]], i8 addrspace(4)* [[Block3Ptr]])
++
++ %22 = call i32 @__enqueue_kernel_basic_events(%opencl.queue_t* %11, i32 %12, %struct.ndrange_t* %tmp4, i32 2, %opencl.clk_event_t* addrspace(4)* %15, %opencl.clk_event_t* addrspace(4)* %16, i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*)* @__device_side_enqueue_block_invoke_2_kernel to i8*) to i8 addrspace(4)*), i8 addrspace(4)* %21)
++ %23 = load %opencl.queue_t*, %opencl.queue_t** %default_queue, align 4
++ %24 = load i32, i32* %flags, align 4
++ %25 = bitcast %struct.ndrange_t* %tmp11 to i8*
++ %26 = bitcast %struct.ndrange_t* %ndrange to i8*
++ call void @llvm.memcpy.p0i8.p0i8.i32(i8* align 4 %25, i8* align 4 %26, i32 4, i1 false)
++ %arraydecay = getelementptr inbounds [1 x %opencl.clk_event_t*], [1 x %opencl.clk_event_t*]* %event_wait_list2, i32 0, i32 0
++ %27 = addrspacecast %opencl.clk_event_t** %arraydecay to %opencl.clk_event_t* addrspace(4)*
++ %28 = addrspacecast %opencl.clk_event_t** %clk_event to %opencl.clk_event_t* addrspace(4)*
++ %29 = getelementptr [1 x i32], [1 x i32]* %block_sizes, i32 0, i32 0
++ %30 = load i8, i8* %c, align 1
++ %31 = zext i8 %30 to i32
++ store i32 %31, i32* %29, align 4
+
+ ; CHECK-SPIRV: PtrAccessChain [[Int32LocPtrTy]] [[LocalBuf31:[0-9]+]]
+ ; CHECK-SPIRV: Bitcast {{[0-9]+}} [[BlockLit3Tmp:[0-9]+]] [[BlockGlb1:[0-9]+]]
+@@ -182,14 +225,18 @@ entry:
+ ; CHECK-LLVM: [[BlockInv0:%[0-9]+]] = addrspacecast void (i8 addrspace(4)*, i8 addrspace(3)*)* @__device_side_enqueue_block_invoke_3_kernel to i8 addrspace(4)*
+ ; CHECK-LLVM: call i32 @__enqueue_kernel_events_varargs(%opencl.queue_t* {{.*}}, i32 {{.*}}, %struct.ndrange_t* {{.*}}, i32 2, %opencl.clk_event_t* addrspace(4)* {{.*}}, %opencl.clk_event_t* addrspace(4)* {{.*}}, i8 addrspace(4)* [[BlockInv0]], i8 addrspace(4)* [[Block0]], i32 1, i32* {{.*}})
+
+- %15 = call i32 @__enqueue_kernel_events_varargs(%opencl.queue_t* %1, i32 %2, %struct.ndrange_t* %ndrange, i32 2, %opencl.clk_event_t* addrspace(4)* %6, %opencl.clk_event_t* addrspace(4)* %7, i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*, i8 addrspace(3)*)* @__device_side_enqueue_block_invoke_3_kernel to i8*) to i8 addrspace(4)*), i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32 } addrspace(1)* @__block_literal_global to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1, i32* %12)
+- %16 = alloca [3 x i32]
+- %17 = getelementptr [3 x i32], [3 x i32]* %16, i32 0, i32 0
+- store i32 1, i32* %17, align 4
+- %18 = getelementptr [3 x i32], [3 x i32]* %16, i32 0, i32 1
+- store i32 2, i32* %18, align 4
+- %19 = getelementptr [3 x i32], [3 x i32]* %16, i32 0, i32 2
+- store i32 4, i32* %19, align 4
++ %32 = call i32 @__enqueue_kernel_events_varargs(%opencl.queue_t* %23, i32 %24, %struct.ndrange_t* %tmp11, i32 2, %opencl.clk_event_t* addrspace(4)* %27, %opencl.clk_event_t* addrspace(4)* %28, i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*, i8 addrspace(3)*)* @__device_side_enqueue_block_invoke_3_kernel to i8*) to i8 addrspace(4)*), i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* @__block_literal_global to i8 addrspace(1)*) to i8 addrspace(4)*), i32 1, i32* %29)
++ %33 = load %opencl.queue_t*, %opencl.queue_t** %default_queue, align 4
++ %34 = load i32, i32* %flags, align 4
++ %35 = bitcast %struct.ndrange_t* %tmp12 to i8*
++ %36 = bitcast %struct.ndrange_t* %ndrange to i8*
++ call void @llvm.memcpy.p0i8.p0i8.i32(i8* align 4 %35, i8* align 4 %36, i32 4, i1 false)
++ %37 = getelementptr [3 x i32], [3 x i32]* %block_sizes13, i32 0, i32 0
++ store i32 1, i32* %37, align 4
++ %38 = getelementptr [3 x i32], [3 x i32]* %block_sizes13, i32 0, i32 1
++ store i32 2, i32* %38, align 4
++ %39 = getelementptr [3 x i32], [3 x i32]* %block_sizes13, i32 0, i32 2
++ store i32 4, i32* %39, align 4
+
+ ; CHECK-SPIRV: PtrAccessChain [[Int32LocPtrTy]] [[LocalBuf41:[0-9]+]]
+ ; CHECK-SPIRV: PtrAccessChain [[Int32LocPtrTy]] [[LocalBuf42:[0-9]+]]
+@@ -206,24 +253,27 @@ entry:
+ ; CHECK-LLVM: [[BlockInv1:%[0-9]+]] = addrspacecast void (i8 addrspace(4)*, i8 addrspace(3)*, i8 addrspace(3)*, i8 addrspace(3)*)* @__device_side_enqueue_block_invoke_4_kernel to i8 addrspace(4)*
+ ; CHECK-LLVM: call i32 @__enqueue_kernel_events_varargs(%opencl.queue_t* {{.*}}, i32 {{.*}}, %struct.ndrange_t* {{.*}}, i32 0, %opencl.clk_event_t* addrspace(4)* null, %opencl.clk_event_t* addrspace(4)* null, i8 addrspace(4)* [[BlockInv1]], i8 addrspace(4)* [[Block1]], i32 3, i32* {{.*}})
+
+- %20 = call i32 @__enqueue_kernel_varargs(%opencl.queue_t* %1, i32 %2, %struct.ndrange_t* %ndrange, i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*, i8 addrspace(3)*, i8 addrspace(3)*, i8 addrspace(3)*)* @__device_side_enqueue_block_invoke_4_kernel to i8*) to i8 addrspace(4)*), i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32 } addrspace(1)* @__block_literal_global.1 to i8 addrspace(1)*) to i8 addrspace(4)*), i32 3, i32* %17)
++ %40 = call i32 @__enqueue_kernel_varargs(%opencl.queue_t* %33, i32 %34, %struct.ndrange_t* %tmp12, i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*, i8 addrspace(3)*, i8 addrspace(3)*, i8 addrspace(3)*)* @__device_side_enqueue_block_invoke_4_kernel to i8*) to i8 addrspace(4)*), i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* @__block_literal_global.1 to i8 addrspace(1)*) to i8 addrspace(4)*), i32 3, i32* %37)
+ ret void
+ }
+
++; Function Attrs: argmemonly nounwind
++declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture writeonly, i8* nocapture readonly, i32, i1) #1
++
+ ; Function Attrs: convergent noinline nounwind optnone
+ define internal spir_func void @__device_side_enqueue_block_invoke(i8 addrspace(4)* %.block_descriptor) #2 {
+ entry:
+ %.block_descriptor.addr = alloca i8 addrspace(4)*, align 4
+- %block.addr = alloca <{ i32, i32, i32 addrspace(1)*, i32, i8 }> addrspace(4)*, align 4
++ %block.addr = alloca <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i8 }> addrspace(4)*, align 4
+ store i8 addrspace(4)* %.block_descriptor, i8 addrspace(4)** %.block_descriptor.addr, align 4
+- %block = bitcast i8 addrspace(4)* %.block_descriptor to <{ i32, i32, i32 addrspace(1)*, i32, i8 }> addrspace(4)*
+- store <{ i32, i32, i32 addrspace(1)*, i32, i8 }> addrspace(4)* %block, <{ i32, i32, i32 addrspace(1)*, i32, i8 }> addrspace(4)** %block.addr, align 4
+- %block.capture.addr = getelementptr inbounds <{ i32, i32, i32 addrspace(1)*, i32, i8 }>, <{ i32, i32, i32 addrspace(1)*, i32, i8 }> addrspace(4)* %block, i32 0, i32 4
++ %block = bitcast i8 addrspace(4)* %.block_descriptor to <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i8 }> addrspace(4)*
++ store <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i8 }> addrspace(4)* %block, <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i8 }> addrspace(4)** %block.addr, align 4
++ %block.capture.addr = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i8 }>, <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i8 }> addrspace(4)* %block, i32 0, i32 5
+ %0 = load i8, i8 addrspace(4)* %block.capture.addr, align 4
+ %conv = sext i8 %0 to i32
+- %block.capture.addr1 = getelementptr inbounds <{ i32, i32, i32 addrspace(1)*, i32, i8 }>, <{ i32, i32, i32 addrspace(1)*, i32, i8 }> addrspace(4)* %block, i32 0, i32 2
++ %block.capture.addr1 = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i8 }>, <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i8 }> addrspace(4)* %block, i32 0, i32 3
+ %1 = load i32 addrspace(1)*, i32 addrspace(1)* addrspace(4)* %block.capture.addr1, align 4
+- %block.capture.addr2 = getelementptr inbounds <{ i32, i32, i32 addrspace(1)*, i32, i8 }>, <{ i32, i32, i32 addrspace(1)*, i32, i8 }> addrspace(4)* %block, i32 0, i32 3
++ %block.capture.addr2 = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i8 }>, <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i8 }> addrspace(4)* %block, i32 0, i32 4
+ %2 = load i32, i32 addrspace(4)* %block.capture.addr2, align 4
+ %arrayidx = getelementptr inbounds i32, i32 addrspace(1)* %1, i32 %2
+ store i32 %conv, i32 addrspace(1)* %arrayidx, align 4
+@@ -243,19 +293,19 @@ declare i32 @__enqueue_kernel_basic(%opencl.queue_t*, i32, %struct.ndrange_t*, i
+ define internal spir_func void @__device_side_enqueue_block_invoke_2(i8 addrspace(4)* %.block_descriptor) #2 {
+ entry:
+ %.block_descriptor.addr = alloca i8 addrspace(4)*, align 4
+- %block.addr = alloca <{ i32, i32, i32 addrspace(1)*, i32, i32 addrspace(1)* }> addrspace(4)*, align 4
++ %block.addr = alloca <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }> addrspace(4)*, align 4
+ store i8 addrspace(4)* %.block_descriptor, i8 addrspace(4)** %.block_descriptor.addr, align 4
+- %block = bitcast i8 addrspace(4)* %.block_descriptor to <{ i32, i32, i32 addrspace(1)*, i32, i32 addrspace(1)* }> addrspace(4)*
+- store <{ i32, i32, i32 addrspace(1)*, i32, i32 addrspace(1)* }> addrspace(4)* %block, <{ i32, i32, i32 addrspace(1)*, i32, i32 addrspace(1)* }> addrspace(4)** %block.addr, align 4
+- %block.capture.addr = getelementptr inbounds <{ i32, i32, i32 addrspace(1)*, i32, i32 addrspace(1)* }>, <{ i32, i32, i32 addrspace(1)*, i32, i32 addrspace(1)* }> addrspace(4)* %block, i32 0, i32 4
++ %block = bitcast i8 addrspace(4)* %.block_descriptor to <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }> addrspace(4)*
++ store <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }> addrspace(4)* %block, <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }> addrspace(4)** %block.addr, align 4
++ %block.capture.addr = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }>, <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }> addrspace(4)* %block, i32 0, i32 5
+ %0 = load i32 addrspace(1)*, i32 addrspace(1)* addrspace(4)* %block.capture.addr, align 4
+- %block.capture.addr1 = getelementptr inbounds <{ i32, i32, i32 addrspace(1)*, i32, i32 addrspace(1)* }>, <{ i32, i32, i32 addrspace(1)*, i32, i32 addrspace(1)* }> addrspace(4)* %block, i32 0, i32 3
++ %block.capture.addr1 = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }>, <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }> addrspace(4)* %block, i32 0, i32 4
+ %1 = load i32, i32 addrspace(4)* %block.capture.addr1, align 4
+ %arrayidx = getelementptr inbounds i32, i32 addrspace(1)* %0, i32 %1
+ %2 = load i32, i32 addrspace(1)* %arrayidx, align 4
+- %block.capture.addr2 = getelementptr inbounds <{ i32, i32, i32 addrspace(1)*, i32, i32 addrspace(1)* }>, <{ i32, i32, i32 addrspace(1)*, i32, i32 addrspace(1)* }> addrspace(4)* %block, i32 0, i32 2
++ %block.capture.addr2 = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }>, <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }> addrspace(4)* %block, i32 0, i32 3
+ %3 = load i32 addrspace(1)*, i32 addrspace(1)* addrspace(4)* %block.capture.addr2, align 4
+- %block.capture.addr3 = getelementptr inbounds <{ i32, i32, i32 addrspace(1)*, i32, i32 addrspace(1)* }>, <{ i32, i32, i32 addrspace(1)*, i32, i32 addrspace(1)* }> addrspace(4)* %block, i32 0, i32 3
++ %block.capture.addr3 = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }>, <{ i32, i32, i8 addrspace(4)*, i32 addrspace(1)*, i32, i32 addrspace(1)* }> addrspace(4)* %block, i32 0, i32 4
+ %4 = load i32, i32 addrspace(4)* %block.capture.addr3, align 4
+ %arrayidx4 = getelementptr inbounds i32, i32 addrspace(1)* %3, i32 %4
+ store i32 %2, i32 addrspace(1)* %arrayidx4, align 4
+@@ -276,11 +326,11 @@ define internal spir_func void @__device_side_enqueue_block_invoke_3(i8 addrspac
+ entry:
+ %.block_descriptor.addr = alloca i8 addrspace(4)*, align 4
+ %p.addr = alloca i8 addrspace(3)*, align 4
+- %block.addr = alloca <{ i32, i32 }> addrspace(4)*, align 4
++ %block.addr = alloca <{ i32, i32, i8 addrspace(4)* }> addrspace(4)*, align 4
+ store i8 addrspace(4)* %.block_descriptor, i8 addrspace(4)** %.block_descriptor.addr, align 4
+- %block = bitcast i8 addrspace(4)* %.block_descriptor to <{ i32, i32 }> addrspace(4)*
++ %block = bitcast i8 addrspace(4)* %.block_descriptor to <{ i32, i32, i8 addrspace(4)* }> addrspace(4)*
+ store i8 addrspace(3)* %p, i8 addrspace(3)** %p.addr, align 4
+- store <{ i32, i32 }> addrspace(4)* %block, <{ i32, i32 }> addrspace(4)** %block.addr, align 4
++ store <{ i32, i32, i8 addrspace(4)* }> addrspace(4)* %block, <{ i32, i32, i8 addrspace(4)* }> addrspace(4)** %block.addr, align 4
+ ret void
+ }
+
+@@ -300,13 +350,13 @@ entry:
+ %p1.addr = alloca i8 addrspace(3)*, align 4
+ %p2.addr = alloca i8 addrspace(3)*, align 4
+ %p3.addr = alloca i8 addrspace(3)*, align 4
+- %block.addr = alloca <{ i32, i32 }> addrspace(4)*, align 4
++ %block.addr = alloca <{ i32, i32, i8 addrspace(4)* }> addrspace(4)*, align 4
+ store i8 addrspace(4)* %.block_descriptor, i8 addrspace(4)** %.block_descriptor.addr, align 4
+- %block = bitcast i8 addrspace(4)* %.block_descriptor to <{ i32, i32 }> addrspace(4)*
++ %block = bitcast i8 addrspace(4)* %.block_descriptor to <{ i32, i32, i8 addrspace(4)* }> addrspace(4)*
+ store i8 addrspace(3)* %p1, i8 addrspace(3)** %p1.addr, align 4
+ store i8 addrspace(3)* %p2, i8 addrspace(3)** %p2.addr, align 4
+ store i8 addrspace(3)* %p3, i8 addrspace(3)** %p3.addr, align 4
+- store <{ i32, i32 }> addrspace(4)* %block, <{ i32, i32 }> addrspace(4)** %block.addr, align 4
++ store <{ i32, i32, i8 addrspace(4)* }> addrspace(4)* %block, <{ i32, i32, i8 addrspace(4)* }> addrspace(4)** %block.addr, align 4
+ ret void
+ }
+
+@@ -329,27 +379,20 @@ declare i32 @__enqueue_kernel_varargs(%opencl.queue_t*, i32, %struct.ndrange_t*,
+ ; CHECK-LLVM-DAG: define spir_kernel void @__device_side_enqueue_block_invoke_3_kernel(i8 addrspace(4)*, i8 addrspace(3)*)
+ ; CHECK-LLVM-DAG: define spir_kernel void @__device_side_enqueue_block_invoke_4_kernel(i8 addrspace(4)*, i8 addrspace(3)*, i8 addrspace(3)*, i8 addrspace(3)*)
+
+-attributes #0 = { convergent noinline nounwind optnone "correctly-rounded-divide-sqrt-fp-math"="false" "disable-tail-calls"="false" "less-precise-fpmad"="false" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-jump-tables"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "uniform-work-group-size"="false" "unsafe-fp-math"="false" "use-soft-float"="false" }
++attributes #0 = { convergent noinline nounwind optnone "correctly-rounded-divide-sqrt-fp-math"="false" "denorms-are-zero"="false" "disable-tail-calls"="false" "less-precise-fpmad"="false" "min-legal-vector-width"="0" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-jump-tables"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "uniform-work-group-size"="false" "unsafe-fp-math"="false" "use-soft-float"="false" }
+ attributes #1 = { argmemonly nounwind }
+-attributes #2 = { convergent noinline nounwind optnone "correctly-rounded-divide-sqrt-fp-math"="false" "disable-tail-calls"="false" "less-precise-fpmad"="false" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-jump-tables"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
++attributes #2 = { convergent noinline nounwind optnone "correctly-rounded-divide-sqrt-fp-math"="false" "denorms-are-zero"="false" "disable-tail-calls"="false" "less-precise-fpmad"="false" "min-legal-vector-width"="0" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-jump-tables"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
+ attributes #3 = { nounwind }
+
+ !llvm.module.flags = !{!0}
+-!opencl.enable.FP_CONTRACT = !{}
+ !opencl.ocl.version = !{!1}
+ !opencl.spir.version = !{!1}
+-!opencl.used.extensions = !{!2}
+-!opencl.used.optional.core.features = !{!2}
+-!opencl.compiler.options = !{!2}
+-!llvm.ident = !{!3}
++!llvm.ident = !{!2}
+
+ !0 = !{i32 1, !"wchar_size", i32 4}
+ !1 = !{i32 2, i32 0}
+-!2 = !{}
+-!3 = !{!"clang version 7.0.0"}
+-!4 = !{i32 1, i32 1, i32 0, i32 0}
+-!5 = !{!"none", !"none", !"none", !"none"}
+-!6 = !{!"int*", !"int*", !"int", !"char"}
+-!7 = !{!"", !"", !"", !""}
+-!8 = !{i1 false, i1 false, i1 false, i1 false}
+-!9 = !{i32 0, i32 0, i32 0, i32 0}
++!2 = !{!"clang version 9.0.0 (https://llvm.org/git/clang 04fb8964a801a5c5d7baa5a22272243a7d183896) (https://llvm.org/git/llvm 384f64397f6ad95a361b72d62c07d7bac9f24163)"}
++!3 = !{i32 1, i32 1, i32 0, i32 0}
++!4 = !{!"none", !"none", !"none", !"none"}
++!5 = !{!"int*", !"int*", !"int", !"char"}
++!6 = !{!"", !"", !"", !""}
+--
+1.8.3.1
+
diff --git a/dynamic-layers/clang-layer/recipes-devtools/clang/files/0001-dont-export-targets-for-binaries.patch b/dynamic-layers/clang-layer/recipes-devtools/clang/files/0001-dont-export-targets-for-binaries.patch
new file mode 100644
index 00000000..9d25bbad
--- /dev/null
+++ b/dynamic-layers/clang-layer/recipes-devtools/clang/files/0001-dont-export-targets-for-binaries.patch
@@ -0,0 +1,66 @@
+From 7bbd0058362ac3bb5edd7a82d43e1785810776b3 Mon Sep 17 00:00:00 2001
+From: Anuj Mittal <anuj.mittal at intel.com>
+Date: Fri, 29 Mar 2019 08:56:53 +0800
+Subject: [PATCH] dont export targets for binaries
+
+The projects using LLVM cmake modules look for target binaries in
+sysroot as a result which isn't desirable in this case and isn't needed
+either.
+
+Upstream-Status: Inappropriate [cross-compile specific]
+
+Signed-off-by: Anuj Mittal <anuj.mittal at intel.com>
+---
+ llvm/cmake/modules/AddLLVM.cmake | 9 ---------
+ llvm/cmake/modules/TableGen.cmake | 6 ------
+ 2 files changed, 15 deletions(-)
+
+diff --git a/llvm/cmake/modules/AddLLVM.cmake b/llvm/cmake/modules/AddLLVM.cmake
+index 0df6845..b79f4fa 100644
+--- a/llvm/cmake/modules/AddLLVM.cmake
++++ b/llvm/cmake/modules/AddLLVM.cmake
+@@ -866,12 +866,6 @@ macro(add_llvm_tool name)
+
+ if ( ${name} IN_LIST LLVM_TOOLCHAIN_TOOLS OR NOT LLVM_INSTALL_TOOLCHAIN_ONLY)
+ if( LLVM_BUILD_TOOLS )
+- if(${name} IN_LIST LLVM_DISTRIBUTION_COMPONENTS OR
+- NOT LLVM_DISTRIBUTION_COMPONENTS)
+- set(export_to_llvmexports EXPORT LLVMExports)
+- set_property(GLOBAL PROPERTY LLVM_HAS_EXPORTS True)
+- endif()
+-
+ install(TARGETS ${name}
+ ${export_to_llvmexports}
+ RUNTIME DESTINATION ${LLVM_TOOLS_INSTALL_DIR}
+@@ -884,9 +878,6 @@ macro(add_llvm_tool name)
+ endif()
+ endif()
+ endif()
+- if( LLVM_BUILD_TOOLS )
+- set_property(GLOBAL APPEND PROPERTY LLVM_EXPORTS ${name})
+- endif()
+ set_target_properties(${name} PROPERTIES FOLDER "Tools")
+ endmacro(add_llvm_tool name)
+
+diff --git a/llvm/cmake/modules/TableGen.cmake b/llvm/cmake/modules/TableGen.cmake
+index 3c84ae7..141219f 100644
+--- a/llvm/cmake/modules/TableGen.cmake
++++ b/llvm/cmake/modules/TableGen.cmake
+@@ -164,14 +164,8 @@ macro(add_tablegen target project)
+ endif()
+
+ if (${project} STREQUAL LLVM AND NOT LLVM_INSTALL_TOOLCHAIN_ONLY)
+- if(${target} IN_LIST LLVM_DISTRIBUTION_COMPONENTS OR
+- NOT LLVM_DISTRIBUTION_COMPONENTS)
+- set(export_to_llvmexports EXPORT LLVMExports)
+- endif()
+-
+ install(TARGETS ${target}
+ ${export_to_llvmexports}
+ RUNTIME DESTINATION ${LLVM_TOOLS_INSTALL_DIR})
+ endif()
+- set_property(GLOBAL APPEND PROPERTY LLVM_EXPORTS ${target})
+ endmacro()
+--
+2.7.4
+
diff --git a/dynamic-layers/clang-layer/recipes-devtools/clang/files/0001-point-to-correct-clang.patch b/dynamic-layers/clang-layer/recipes-devtools/clang/files/0001-point-to-correct-clang.patch
new file mode 100644
index 00000000..0dfc537b
--- /dev/null
+++ b/dynamic-layers/clang-layer/recipes-devtools/clang/files/0001-point-to-correct-clang.patch
@@ -0,0 +1,59 @@
+From 6c33fb58869ffb17106047c45ab8d3856966eaf7 Mon Sep 17 00:00:00 2001
+From: Anuj Mittal <anuj.mittal at intel.com>
+Date: Tue, 26 Mar 2019 14:11:29 +0800
+Subject: [PATCH] point to correct clang project and tblgen
+
+Point to correct path for clang project as per the way we unpack. Also
+let llvm-tblgen path be passed from recipe itself.
+
+Also since we're going to do the patching ourselves, no need to look for
+git through cmake.
+
+Upstream-Status: Inappropriate [OE specific]
+---
+ CMakeLists.txt | 8 ++++----
+ 1 file changed, 4 insertions(+), 4 deletions(-)
+
+diff --git a/CMakeLists.txt b/CMakeLists.txt
+index 174133b..c769f08 100644
+--- a/CMakeLists.txt
++++ b/CMakeLists.txt
+@@ -53,7 +53,7 @@ endif(CMAKE_SOURCE_DIR STREQUAL CMAKE_CURRENT_SOURCE_DIR)
+ include(AddLLVM)
+ include(TableGen)
+
+-find_package(Git REQUIRED)
++#find_package(Git REQUIRED)
+
+ if (NOT WIN32)
+ add_subdirectory( linux_linker )
+@@ -80,7 +80,7 @@ set(TARGET_NAME ${COMMON_CLANG_LIBRARY_NAME}${BUILD_PLATFORM} )
+
+ if(NOT USE_PREBUILT_LLVM)
+ set(TARGET_BRANCH "ocl-open-80")
+- set(CLANG_SOURCE_DIR ${LLVM_SOURCE_DIR}/tools/clang)
++ set(CLANG_SOURCE_DIR ${LLVM_SOURCE_DIR}/../clang)
+ set(CLANG_BASE_REVISION a03da8be08a208122e292016cb6cea1f30229677)
+
+ set(SPIRV_SOURCE_DIR ${LLVM_SOURCE_DIR}/projects/llvm-spirv)
+@@ -102,7 +102,7 @@ endif(NOT USE_PREBUILT_LLVM)
+ set (COMPILE_OPTIONS_TD opencl_clang_options.td)
+ set (COMPILE_OPTIONS_INC opencl_clang_options.inc)
+
+-set(LLVM_TABLEGEN_EXE "llvm-tblgen")
++#set(LLVM_TABLEGEN_EXE "llvm-tblgen")
+ set(LLVM_TARGET_DEFINITIONS ${COMPILE_OPTIONS_TD})
+ if(USE_PREBUILT_LLVM)
+ set(TABLEGEN_ADDITIONAL -I ${LLVM_INCLUDE_DIRS})
+@@ -153,7 +153,7 @@ endif()
+
+ if(NOT USE_PREBUILT_LLVM)
+ set(CLANG_BINARY_DIR ${LLVM_BINARY_DIR}/tools/clang/)
+- set(CLANG_SOURCE_DIR ${LLVM_MAIN_SRC_DIR}/tools/clang/)
++ set(CLANG_SOURCE_DIR ${LLVM_MAIN_SRC_DIR}/../clang/)
+ include_directories(
+ ${CLANG_BINARY_DIR}/include # for tablegened includes
+ ${CLANG_SOURCE_DIR}/include # for basic headers
+--
+2.19.1
+
diff --git a/dynamic-layers/clang-layer/recipes-devtools/clang/files/0002-OpenCL-Simplify-LLVM-IR-generated-for-OpenCL-blocks.patch b/dynamic-layers/clang-layer/recipes-devtools/clang/files/0002-OpenCL-Simplify-LLVM-IR-generated-for-OpenCL-blocks.patch
new file mode 100644
index 00000000..2e935a13
--- /dev/null
+++ b/dynamic-layers/clang-layer/recipes-devtools/clang/files/0002-OpenCL-Simplify-LLVM-IR-generated-for-OpenCL-blocks.patch
@@ -0,0 +1,294 @@
+From c94ec28600255098ffb9d73d1b386a7c8a535590 Mon Sep 17 00:00:00 2001
+From: Andrew Savonichev <andrew.savonichev at intel.com>
+Date: Thu, 21 Feb 2019 11:02:10 +0000
+Subject: [PATCH 2/2] [OpenCL] Simplify LLVM IR generated for OpenCL blocks
+
+Summary:
+Emit direct call of block invoke functions when possible, i.e. in case the
+block is not passed as a function argument.
+Also doing some refactoring of `CodeGenFunction::EmitBlockCallExpr()`
+
+Reviewers: Anastasia, yaxunl, svenvh
+
+Reviewed By: Anastasia
+
+Subscribers: cfe-commits
+
+Tags: #clang
+
+Differential Revision: https://reviews.llvm.org/D58388
+
+git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@354568 91177308-0d34-0410-b5e6-96231b3b80d8
+
+Upstream-Status: Backport
+[https://github.com/llvm-mirror/clang/commit/eae71f8d05ce550c4e2595c9b7082cc2c7882c58]
+Signed-off-by: Anuj Mittal <anuj.mittal at intel.com>
+---
+ lib/CodeGen/CGBlocks.cpp | 77 +++++++++++++-------------
+ lib/CodeGen/CGOpenCLRuntime.cpp | 30 +++++++---
+ lib/CodeGen/CGOpenCLRuntime.h | 4 ++
+ test/CodeGenOpenCL/blocks.cl | 10 +---
+ test/CodeGenOpenCL/cl20-device-side-enqueue.cl | 34 +++++++++---
+ 5 files changed, 91 insertions(+), 64 deletions(-)
+
+diff --git a/lib/CodeGen/CGBlocks.cpp b/lib/CodeGen/CGBlocks.cpp
+index fa3c3ee..10a0238 100644
+--- a/lib/CodeGen/CGBlocks.cpp
++++ b/lib/CodeGen/CGBlocks.cpp
+@@ -1261,52 +1261,49 @@ RValue CodeGenFunction::EmitBlockCallExpr(const CallExpr *E,
+ ReturnValueSlot ReturnValue) {
+ const BlockPointerType *BPT =
+ E->getCallee()->getType()->getAs<BlockPointerType>();
+-
+ llvm::Value *BlockPtr = EmitScalarExpr(E->getCallee());
+-
+- // 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);
+-
+- // 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);
+-
+- // Add the block literal.
++ llvm::Type *GenBlockTy = CGM.getGenericBlockLiteralType();
++ llvm::Value *Func = nullptr;
++ QualType FnType = BPT->getPointeeType();
++ ASTContext &Ctx = getContext();
+ 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();
++ // For OpenCL, BlockPtr is already casted to generic block literal.
++
++ // First argument of a block call is a generic block literal casted to
++ // generic void pointer, i.e. i8 addrspace(4)*
++ llvm::Value *BlockDescriptor = Builder.CreatePointerCast(
++ BlockPtr, CGM.getOpenCLRuntime().getGenericVoidPointerType());
++ QualType VoidPtrQualTy = Ctx.getPointerType(
++ Ctx.getAddrSpaceQualType(Ctx.VoidTy, LangAS::opencl_generic));
++ Args.add(RValue::get(BlockDescriptor), VoidPtrQualTy);
++ // And the rest of the arguments.
++ EmitCallArgs(Args, FnType->getAs<FunctionProtoType>(), E->arguments());
++
++ // We *can* call the block directly unless it is a function argument.
++ if (!isa<ParmVarDecl>(E->getCalleeDecl()))
++ Func = CGM.getOpenCLRuntime().getInvokeFunction(E->getCallee());
++ else {
++ llvm::Value *FuncPtr = Builder.CreateStructGEP(GenBlockTy, BlockPtr, 2);
++ Func = Builder.CreateAlignedLoad(FuncPtr, getPointerAlign());
++ }
++ } else {
++ // Bitcast the block literal to a generic block literal.
++ BlockPtr = Builder.CreatePointerCast(
++ BlockPtr, llvm::PointerType::get(GenBlockTy, 0), "block.literal");
++ // Get pointer to the block invoke function
++ llvm::Value *FuncPtr = Builder.CreateStructGEP(GenBlockTy, BlockPtr, 3);
+
+- // And the rest of the arguments.
+- EmitCallArgs(Args, FnType->getAs<FunctionProtoType>(), E->arguments());
++ // First argument is a block literal casted to a void pointer
++ BlockPtr = Builder.CreatePointerCast(BlockPtr, VoidPtrTy);
++ Args.add(RValue::get(BlockPtr), Ctx.VoidPtrTy);
++ // And the rest of the arguments.
++ EmitCallArgs(Args, FnType->getAs<FunctionProtoType>(), E->arguments());
+
+- // Load the function.
+- llvm::Value *Func = Builder.CreateAlignedLoad(FuncPtr, getPointerAlign());
++ // Load the function.
++ Func = Builder.CreateAlignedLoad(FuncPtr, getPointerAlign());
++ }
+
+ const FunctionType *FuncTy = FnType->castAs<FunctionType>();
+ const CGFunctionInfo &FnInfo =
+diff --git a/lib/CodeGen/CGOpenCLRuntime.cpp b/lib/CodeGen/CGOpenCLRuntime.cpp
+index 7f6f595..75003e5 100644
+--- a/lib/CodeGen/CGOpenCLRuntime.cpp
++++ b/lib/CodeGen/CGOpenCLRuntime.cpp
+@@ -123,6 +123,23 @@ llvm::PointerType *CGOpenCLRuntime::getGenericVoidPointerType() {
+ 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) {
++ const Expr *Prev = nullptr; // to make sure we do not stuck in infinite loop.
++ while(!isa<BlockExpr>(E) && E != Prev) {
++ Prev = E;
++ E = E->IgnoreCasts();
++ if (auto DR = dyn_cast<DeclRefExpr>(E)) {
++ E = cast<VarDecl>(DR->getDecl())->getInit();
++ }
++ }
++ return cast<BlockExpr>(E);
++}
++
+ /// Record emitted llvm invoke function and llvm block literal for the
+ /// corresponding block expression.
+ void CGOpenCLRuntime::recordBlockInfo(const BlockExpr *E,
+@@ -137,20 +154,17 @@ void CGOpenCLRuntime::recordBlockInfo(const BlockExpr *E,
+ 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");
+diff --git a/lib/CodeGen/CGOpenCLRuntime.h b/lib/CodeGen/CGOpenCLRuntime.h
+index 750721f..4effc7e 100644
+--- a/lib/CodeGen/CGOpenCLRuntime.h
++++ b/lib/CodeGen/CGOpenCLRuntime.h
+@@ -92,6 +92,10 @@ public:
+ /// \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);
+ };
+
+ }
+diff --git a/test/CodeGenOpenCL/blocks.cl b/test/CodeGenOpenCL/blocks.cl
+index 19aacc3..ab5a2c6 100644
+--- a/test/CodeGenOpenCL/blocks.cl
++++ b/test/CodeGenOpenCL/blocks.cl
+@@ -39,11 +39,8 @@ void foo(){
+ // SPIR: %[[blk_gen_ptr:.*]] = addrspacecast %struct.__opencl_block_literal_generic* %[[blk_ptr]] to %struct.__opencl_block_literal_generic addrspace(4)*
+ // SPIR: store %struct.__opencl_block_literal_generic addrspace(4)* %[[blk_gen_ptr]], %struct.__opencl_block_literal_generic addrspace(4)** %[[block_B:.*]],
+ // SPIR: %[[block_literal:.*]] = load %struct.__opencl_block_literal_generic addrspace(4)*, %struct.__opencl_block_literal_generic addrspace(4)** %[[block_B]]
+- // 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]])
++ // SPIR: call {{.*}}i32 @__foo_block_invoke(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
+@@ -53,11 +50,8 @@ void foo(){
+ // AMDGCN: %[[blk_gen_ptr:.*]] = addrspacecast %struct.__opencl_block_literal_generic addrspace(5)* %[[blk_ptr]] to %struct.__opencl_block_literal_generic*
+ // AMDGCN: store %struct.__opencl_block_literal_generic* %[[blk_gen_ptr]], %struct.__opencl_block_literal_generic* addrspace(5)* %[[block_B:.*]],
+ // AMDGCN: %[[block_literal:.*]] = load %struct.__opencl_block_literal_generic*, %struct.__opencl_block_literal_generic* addrspace(5)* %[[block_B]]
+- // 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: call {{.*}}i32 @__foo_block_invoke(i8* %[[blk_gen_ptr]])
+
+ int (^ block_B)(void) = ^{
+ return i;
+diff --git a/test/CodeGenOpenCL/cl20-device-side-enqueue.cl b/test/CodeGenOpenCL/cl20-device-side-enqueue.cl
+index 8445016..1566912 100644
+--- a/test/CodeGenOpenCL/cl20-device-side-enqueue.cl
++++ b/test/CodeGenOpenCL/cl20-device-side-enqueue.cl
+@@ -312,9 +312,7 @@ kernel void device_side_enqueue(global int *a, global int *b, int i) {
+ };
+
+ // 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 @__device_side_enqueue_block_invoke_11(i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } 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]].
+@@ -333,15 +331,35 @@ kernel void device_side_enqueue(global int *a, global int *b, int i) {
+ 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 @__device_side_enqueue_block_invoke_11(i8 addrspace(4)* addrspacecast (i8 addrspace(1)* bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BLG8]] to i8 addrspace(1)*) to i8 addrspace(4)*))
+ block_A();
+
++ // Make sure that block invoke function is resolved correctly after sequence of assignements.
++ // COMMON: store %struct.__opencl_block_literal_generic addrspace(4)*
++ // COMMON-SAME: addrspacecast (%struct.__opencl_block_literal_generic addrspace(1)*
++ // COMMON-SAME: bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BL_GLOBAL]] to %struct.__opencl_block_literal_generic addrspace(1)*)
++ // COMMON-SAME: to %struct.__opencl_block_literal_generic addrspace(4)*),
++ // COMMON-SAME: %struct.__opencl_block_literal_generic addrspace(4)** %b1,
++ bl_t b1 = block_G;
++ // COMMON: store %struct.__opencl_block_literal_generic addrspace(4)*
++ // COMMON-SAME: addrspacecast (%struct.__opencl_block_literal_generic addrspace(1)*
++ // COMMON-SAME: bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BL_GLOBAL]] to %struct.__opencl_block_literal_generic addrspace(1)*)
++ // COMMON-SAME: to %struct.__opencl_block_literal_generic addrspace(4)*),
++ // COMMON-SAME: %struct.__opencl_block_literal_generic addrspace(4)** %b2,
++ bl_t b2 = b1;
++ // COMMON: call spir_func void @block_G_block_invoke(i8 addrspace(4)* addrspacecast (i8 addrspace(1)*
++ // COMMON-SAME: bitcast ({ i32, i32, i8 addrspace(4)* } addrspace(1)* [[BL_GLOBAL]] to i8 addrspace(1)*)
++ // COOMON-SAME: to i8 addrspace(4)*), i8 addrspace(3)* null)
++ b2(0);
++ // 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_size_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)*))
++ size = get_kernel_preferred_work_group_size_multiple(b2);
++
+ 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
+@@ -404,8 +422,8 @@ kernel void device_side_enqueue(global int *a, global int *b, int i) {
+ // 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_kernel void [[INVGK8]](i8 addrspace(4)*{{.*}})
++// COMMON: define internal spir_kernel void [[INV_G_K]](i8 addrspace(4)*{{.*}}, i8 addrspace(3)*{{.*}})
+ // 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)*{{.*}})
+--
+1.8.3.1
+
diff --git a/dynamic-layers/clang-layer/recipes-devtools/clang/files/0003-OpenCL-Fix-assertion-due-to-blocks.patch b/dynamic-layers/clang-layer/recipes-devtools/clang/files/0003-OpenCL-Fix-assertion-due-to-blocks.patch
new file mode 100644
index 00000000..510c7c6e
--- /dev/null
+++ b/dynamic-layers/clang-layer/recipes-devtools/clang/files/0003-OpenCL-Fix-assertion-due-to-blocks.patch
@@ -0,0 +1,61 @@
+From 29e2813a2ab7d5569860bb07892dfef7b5374d96 Mon Sep 17 00:00:00 2001
+From: Yaxun Liu <Yaxun.Liu at amd.com>
+Date: Tue, 26 Feb 2019 16:20:41 +0000
+Subject: [PATCH] [OpenCL] Fix assertion due to blocks
+
+A recent change caused assertion in CodeGenFunction::EmitBlockCallExpr when a block is called.
+
+There is code
+
+ Func = CGM.getOpenCLRuntime().getInvokeFunction(E->getCallee());
+getCalleeDecl calls Expr::getReferencedDeclOfCallee, which does not handle
+BlockExpr and returns nullptr, which causes isa to assert.
+
+This patch fixes that.
+
+Differential Revision: https://reviews.llvm.org/D58658
+
+
+git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@354893 91177308-0d34-0410-b5e6-96231b3b80d8
+
+Upstream-Status: Backport
+[https://github.com/llvm-mirror/clang/commit/29e2813a2ab7d5569860bb07892dfef7b5374d96]
+Signed-off-by: Anuj Mittal <anuj.mittal at intel.com>
+---
+ lib/AST/Expr.cpp | 2 ++
+ test/CodeGenOpenCL/blocks.cl | 6 ++++++
+ 2 files changed, 8 insertions(+)
+
+diff --git a/lib/AST/Expr.cpp b/lib/AST/Expr.cpp
+index aef1eab..85690c7 100644
+--- a/lib/AST/Expr.cpp
++++ b/lib/AST/Expr.cpp
+@@ -1358,6 +1358,8 @@ Decl *Expr::getReferencedDeclOfCallee() {
+ return DRE->getDecl();
+ if (MemberExpr *ME = dyn_cast<MemberExpr>(CEE))
+ return ME->getMemberDecl();
++ if (auto *BE = dyn_cast<BlockExpr>(CEE))
++ return BE->getBlockDecl();
+
+ return nullptr;
+ }
+diff --git a/test/CodeGenOpenCL/blocks.cl b/test/CodeGenOpenCL/blocks.cl
+index ab5a2c6..c3e2685 100644
+--- a/test/CodeGenOpenCL/blocks.cl
++++ b/test/CodeGenOpenCL/blocks.cl
+@@ -90,6 +90,12 @@ int get42() {
+ return blockArgFunc(^{return 42;});
+ }
+
++// COMMON-LABEL: define {{.*}}@call_block
++// call {{.*}}@__call_block_block_invoke
++int call_block() {
++ return ^int(int num) { return num; } (11);
++}
++
+ // CHECK-DEBUG: !DIDerivedType(tag: DW_TAG_member, name: "__size"
+ // CHECK-DEBUG: !DIDerivedType(tag: DW_TAG_member, name: "__align"
+
+--
+1.8.3.1
+
diff --git a/dynamic-layers/clang-layer/recipes-devtools/clang/llvm-project-source.bbappend b/dynamic-layers/clang-layer/recipes-devtools/clang/llvm-project-source.bbappend
new file mode 100644
index 00000000..f536f0f2
--- /dev/null
+++ b/dynamic-layers/clang-layer/recipes-devtools/clang/llvm-project-source.bbappend
@@ -0,0 +1,16 @@
+FILESEXTRAPATHS_prepend_intel-x86-common := "${THISDIR}/files:"
+
+SRC_URI_append_intel-x86-common = " \
+ git://github.com/intel/opencl-clang.git;protocol=https;branch=ocl-open-80;destsuffix=git/llvm/projects/opencl-clang;name=opencl-clang \
+ git://github.com/KhronosGroup/SPIRV-LLVM-Translator.git;protocol=https;branch=llvm_release_80;destsuffix=git/llvm/projects/llvm-spirv;name=spirv \
+ file://0001-point-to-correct-clang.patch;patchdir=llvm/projects/opencl-clang \
+ file://0001-OpenCL-Change-type-of-block-pointer-for-OpenCL.patch;patchdir=clang \
+ file://0002-OpenCL-Simplify-LLVM-IR-generated-for-OpenCL-blocks.patch;patchdir=clang \
+ file://0003-OpenCL-Fix-assertion-due-to-blocks.patch;patchdir=clang \
+ file://0001-dont-export-targets-for-binaries.patch \
+ file://0001-Update-LowerOpenCL-pass-to-handle-new-blocks-represn.patch;patchdir=llvm/projects/llvm-spirv \
+ "
+
+SRCREV_opencl-clang = "daf5e4dd718477ae8cf89a283c653939d9182f15"
+SRCREV_spirv = "bd0f28fb92061d49c0f120b4dac3fd8956006745"
+
--
2.20.1
More information about the meta-intel
mailing list