From 85167c8b9a85f4d05dc72513f69f3643928ddd0e Mon Sep 17 00:00:00 2001 From: Joey Gouly Date: Wed, 10 Aug 2016 15:57:02 +0000 Subject: [PATCH] [OpenCL] Change block descriptor address space to constant. The block descriptor is a GlobalVariable in the LLVM IR, so it shouldn't be in the private address space. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@278234 91177308-0d34-0410-b5e6-96231b3b80d8 --- lib/CodeGen/CGBlocks.cpp | 12 ++++++++++-- test/CodeGenOpenCL/cl20-device-side-enqueue.cl | 14 +++++++------- 2 files changed, 17 insertions(+), 9 deletions(-) diff --git a/lib/CodeGen/CGBlocks.cpp b/lib/CodeGen/CGBlocks.cpp index 5a41f361d9..e3658ab9b7 100644 --- a/lib/CodeGen/CGBlocks.cpp +++ b/lib/CodeGen/CGBlocks.cpp @@ -125,10 +125,15 @@ static llvm::Constant *buildBlockDescriptor(CodeGenModule &CGM, llvm::Constant *init = llvm::ConstantStruct::getAnon(elements); + unsigned AddrSpace = 0; + if (C.getLangOpts().OpenCL) + AddrSpace = C.getTargetAddressSpace(LangAS::opencl_constant); llvm::GlobalVariable *global = new llvm::GlobalVariable(CGM.getModule(), init->getType(), true, llvm::GlobalValue::InternalLinkage, - init, "__block_descriptor_tmp"); + init, "__block_descriptor_tmp", nullptr, + llvm::GlobalValue::NotThreadLocal, + AddrSpace); return llvm::ConstantExpr::getBitCast(global, CGM.getBlockDescriptorType()); } @@ -927,7 +932,10 @@ llvm::Type *CodeGenModule::getBlockDescriptorType() { UnsignedLongTy, UnsignedLongTy, nullptr); // Now form a pointer to that. - BlockDescriptorType = llvm::PointerType::getUnqual(BlockDescriptorType); + unsigned AddrSpace = 0; + if (getLangOpts().OpenCL) + AddrSpace = getContext().getTargetAddressSpace(LangAS::opencl_constant); + BlockDescriptorType = llvm::PointerType::get(BlockDescriptorType, AddrSpace); return BlockDescriptorType; } diff --git a/test/CodeGenOpenCL/cl20-device-side-enqueue.cl b/test/CodeGenOpenCL/cl20-device-side-enqueue.cl index 08b24bdc5b..61418d86b0 100644 --- a/test/CodeGenOpenCL/cl20-device-side-enqueue.cl +++ b/test/CodeGenOpenCL/cl20-device-side-enqueue.cl @@ -5,7 +5,7 @@ typedef void (^bl_t)(local void *); const bl_t block_G = (bl_t) ^ (local void *a) {}; kernel void device_side_enqueue(global int *a, global int *b, int i) { - // CHECK: %default_queue = alloca %opencl.queue_t* + // CHECK: %deafault_queue = alloca %opencl.queue_t* queue_t default_queue; // CHECK: %flags = alloca i32 unsigned flags = 0; @@ -21,7 +21,7 @@ kernel void device_side_enqueue(global int *a, global int *b, int i) { // CHECK: [[DEF_Q:%[0-9]+]] = load %opencl.queue_t*, %opencl.queue_t** %default_queue // CHECK: [[FLAGS:%[0-9]+]] = load i32, i32* %flags // CHECK: [[NDR:%[0-9]+]] = load %opencl.ndrange_t*, %opencl.ndrange_t** %ndrange - // CHECK: [[BL:%[0-9]+]] = bitcast <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, i32{{.*}}, i32{{.*}}, i32{{.*}} }>* %block to void ()* + // CHECK: [[BL:%[0-9]+]] = bitcast <{ i8*, i32, i32, i8*, %struct.__block_descriptor addrspace(3)*, i32{{.*}}, i32{{.*}}, i32{{.*}} }>* %block to void ()* // CHECK: [[BL_I8:%[0-9]+]] = bitcast void ()* [[BL]] to i8* // CHECK: call i32 @__enqueue_kernel_basic(%opencl.queue_t* [[DEF_Q]], i32 [[FLAGS]], %opencl.ndrange_t* [[NDR]], i8* [[BL_I8]]) enqueue_kernel(default_queue, flags, ndrange, @@ -32,7 +32,7 @@ kernel void device_side_enqueue(global int *a, global int *b, int i) { // CHECK: [[DEF_Q:%[0-9]+]] = load %opencl.queue_t*, %opencl.queue_t** %default_queue // CHECK: [[FLAGS:%[0-9]+]] = load i32, i32* %flags // CHECK: [[NDR:%[0-9]+]] = load %opencl.ndrange_t*, %opencl.ndrange_t** %ndrange - // CHECK: [[BL:%[0-9]+]] = bitcast <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, i32{{.*}}, i32{{.*}}, i32{{.*}} }>* %block3 to void ()* + // CHECK: [[BL:%[0-9]+]] = bitcast <{ i8*, i32, i32, i8*, %struct.__block_descriptor addrspace(3)*, i32{{.*}}, i32{{.*}}, i32{{.*}} }>* %block3 to void ()* // CHECK: [[BL_I8:%[0-9]+]] = bitcast void ()* [[BL]] to i8* // CHECK: call i32 @__enqueue_kernel_basic_events(%opencl.queue_t* [[DEF_Q]], i32 [[FLAGS]], %opencl.ndrange_t* [[NDR]], i32 2, %opencl.clk_event_t** %event_wait_list, %opencl.clk_event_t** %clk_event, i8* [[BL_I8]]) enqueue_kernel(default_queue, flags, ndrange, 2, &event_wait_list, &clk_event, @@ -43,7 +43,7 @@ kernel void device_side_enqueue(global int *a, global int *b, int i) { // CHECK: [[DEF_Q:%[0-9]+]] = load %opencl.queue_t*, %opencl.queue_t** %default_queue // CHECK: [[FLAGS:%[0-9]+]] = load i32, i32* %flags // CHECK: [[NDR:%[0-9]+]] = load %opencl.ndrange_t*, %opencl.ndrange_t** %ndrange - // CHECK: call i32 (%opencl.queue_t*, i32, %opencl.ndrange_t*, i8*, i32, ...) @__enqueue_kernel_vaargs(%opencl.queue_t* [[DEF_Q]], i32 [[FLAGS]], %opencl.ndrange_t* [[NDR]], i8* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor* }* @__block_literal_global{{(.[0-9]+)?}} to i8*), i32 1, i32 256) + // CHECK: call i32 (%opencl.queue_t*, i32, %opencl.ndrange_t*, i8*, i32, ...) @__enqueue_kernel_vaargs(%opencl.queue_t* [[DEF_Q]], i32 [[FLAGS]], %opencl.ndrange_t* [[NDR]], i8* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(3)* }* @__block_literal_global{{(.[0-9]+)?}} to i8*), i32 1, i32 256) enqueue_kernel(default_queue, flags, ndrange, ^(local void *p) { return; @@ -54,7 +54,7 @@ kernel void device_side_enqueue(global int *a, global int *b, int i) { // CHECK: [[FLAGS:%[0-9]+]] = load i32, i32* %flags // CHECK: [[NDR:%[0-9]+]] = load %opencl.ndrange_t*, %opencl.ndrange_t** %ndrange // CHECK: [[SIZE:%[0-9]+]] = zext i8 {{%[0-9]+}} to i32 - // CHECK: call i32 (%opencl.queue_t*, i32, %opencl.ndrange_t*, i8*, i32, ...) @__enqueue_kernel_vaargs(%opencl.queue_t* [[DEF_Q]], i32 [[FLAGS]], %opencl.ndrange_t* [[NDR]], i8* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor* }* @__block_literal_global{{(.[0-9]+)?}} to i8*), i32 1, i32 [[SIZE]]) + // CHECK: call i32 (%opencl.queue_t*, i32, %opencl.ndrange_t*, i8*, i32, ...) @__enqueue_kernel_vaargs(%opencl.queue_t* [[DEF_Q]], i32 [[FLAGS]], %opencl.ndrange_t* [[NDR]], i8* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(3)* }* @__block_literal_global{{(.[0-9]+)?}} to i8*), i32 1, i32 [[SIZE]]) enqueue_kernel(default_queue, flags, ndrange, ^(local void *p) { return; @@ -65,7 +65,7 @@ kernel void device_side_enqueue(global int *a, global int *b, int i) { // CHECK: [[FLAGS:%[0-9]+]] = load i32, i32* %flags // CHECK: [[NDR:%[0-9]+]] = load %opencl.ndrange_t*, %opencl.ndrange_t** %ndrange // CHECK: [[AD:%arraydecay[0-9]*]] = getelementptr inbounds [1 x %opencl.clk_event_t*], [1 x %opencl.clk_event_t*]* %event_wait_list2, i32 0, i32 0 - // CHECK: call i32 (%opencl.queue_t*, i32, %opencl.ndrange_t*, i32, %opencl.clk_event_t**, %opencl.clk_event_t**, i8*, i32, ...) @__enqueue_kernel_events_vaargs(%opencl.queue_t* [[DEF_Q]], i32 [[FLAGS]], %opencl.ndrange_t* [[NDR]], i32 2, %opencl.clk_event_t** [[AD]], %opencl.clk_event_t** %clk_event, i8* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor* }* @__block_literal_global{{(.[0-9]+)?}} to i8*), i32 1, i32 256) + // CHECK: call i32 (%opencl.queue_t*, i32, %opencl.ndrange_t*, i32, %opencl.clk_event_t**, %opencl.clk_event_t**, i8*, i32, ...) @__enqueue_kernel_events_vaargs(%opencl.queue_t* [[DEF_Q]], i32 [[FLAGS]], %opencl.ndrange_t* [[NDR]], i32 2, %opencl.clk_event_t** [[AD]], %opencl.clk_event_t** %clk_event, i8* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(3)* }* @__block_literal_global{{(.[0-9]+)?}} to i8*), i32 1, i32 256) enqueue_kernel(default_queue, flags, ndrange, 2, event_wait_list2, &clk_event, ^(local void *p) { return; @@ -77,7 +77,7 @@ kernel void device_side_enqueue(global int *a, global int *b, int i) { // CHECK: [[NDR:%[0-9]+]] = load %opencl.ndrange_t*, %opencl.ndrange_t** %ndrange // CHECK: [[AD:%arraydecay[0-9]*]] = getelementptr inbounds [1 x %opencl.clk_event_t*], [1 x %opencl.clk_event_t*]* %event_wait_list2, i32 0, i32 0 // CHECK: [[SIZE:%[0-9]+]] = zext i8 {{%[0-9]+}} to i32 - // CHECK: call i32 (%opencl.queue_t*, i32, %opencl.ndrange_t*, i32, %opencl.clk_event_t**, %opencl.clk_event_t**, i8*, i32, ...) @__enqueue_kernel_events_vaargs(%opencl.queue_t* [[DEF_Q]], i32 [[FLAGS]], %opencl.ndrange_t* [[NDR]], i32 2, %opencl.clk_event_t** [[AD]], %opencl.clk_event_t** %clk_event, i8* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor* }* @__block_literal_global{{(.[0-9]+)?}} to i8*), i32 1, i32 [[SIZE]]) + // CHECK: call i32 (%opencl.queue_t*, i32, %opencl.ndrange_t*, i32, %opencl.clk_event_t**, %opencl.clk_event_t**, i8*, i32, ...) @__enqueue_kernel_events_vaargs(%opencl.queue_t* [[DEF_Q]], i32 [[FLAGS]], %opencl.ndrange_t* [[NDR]], i32 2, %opencl.clk_event_t** [[AD]], %opencl.clk_event_t** %clk_event, i8* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor addrspace(3)* }* @__block_literal_global{{(.[0-9]+)?}} to i8*), i32 1, i32 [[SIZE]]) enqueue_kernel(default_queue, flags, ndrange, 2, event_wait_list2, &clk_event, ^(local void *p) { return; -- 2.40.0