1// RUN: %clang_cc1 %s -cl-std=CL2.0 -ffake-address-space-map -O0 -emit-llvm -o - | FileCheck %s 2 3typedef void (^bl_t)(local void *); 4 5const bl_t block_G = (bl_t) ^ (local void *a) {}; 6 7kernel void device_side_enqueue(global int *a, global int *b, int i) { 8 // CHECK: %default_queue = alloca %opencl.queue_t* 9 queue_t default_queue; 10 // CHECK: %flags = alloca i32 11 unsigned flags = 0; 12 // CHECK: %ndrange = alloca %opencl.ndrange_t* 13 ndrange_t ndrange; 14 // CHECK: %clk_event = alloca %opencl.clk_event_t* 15 clk_event_t clk_event; 16 // CHECK: %event_wait_list = alloca %opencl.clk_event_t* 17 clk_event_t event_wait_list; 18 // CHECK: %event_wait_list2 = alloca [1 x %opencl.clk_event_t*] 19 clk_event_t event_wait_list2[] = {clk_event}; 20 21 // CHECK: [[DEF_Q:%[0-9]+]] = load %opencl.queue_t*, %opencl.queue_t** %default_queue 22 // CHECK: [[FLAGS:%[0-9]+]] = load i32, i32* %flags 23 // CHECK: [[NDR:%[0-9]+]] = load %opencl.ndrange_t*, %opencl.ndrange_t** %ndrange 24 // CHECK: [[BL:%[0-9]+]] = bitcast <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, i32{{.*}}, i32{{.*}}, i32{{.*}} }>* %block to void ()* 25 // CHECK: [[BL_I8:%[0-9]+]] = bitcast void ()* [[BL]] to i8* 26 // CHECK: call i32 @__enqueue_kernel_basic(%opencl.queue_t* [[DEF_Q]], i32 [[FLAGS]], %opencl.ndrange_t* [[NDR]], i8* [[BL_I8]]) 27 enqueue_kernel(default_queue, flags, ndrange, 28 ^(void) { 29 a[i] = b[i]; 30 }); 31 32 // CHECK: [[DEF_Q:%[0-9]+]] = load %opencl.queue_t*, %opencl.queue_t** %default_queue 33 // CHECK: [[FLAGS:%[0-9]+]] = load i32, i32* %flags 34 // CHECK: [[NDR:%[0-9]+]] = load %opencl.ndrange_t*, %opencl.ndrange_t** %ndrange 35 // CHECK: [[BL:%[0-9]+]] = bitcast <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, i32{{.*}}, i32{{.*}}, i32{{.*}} }>* %block3 to void ()* 36 // CHECK: [[BL_I8:%[0-9]+]] = bitcast void ()* [[BL]] to i8* 37 // 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]]) 38 enqueue_kernel(default_queue, flags, ndrange, 2, &event_wait_list, &clk_event, 39 ^(void) { 40 a[i] = b[i]; 41 }); 42 43 // CHECK: [[DEF_Q:%[0-9]+]] = load %opencl.queue_t*, %opencl.queue_t** %default_queue 44 // CHECK: [[FLAGS:%[0-9]+]] = load i32, i32* %flags 45 // CHECK: [[NDR:%[0-9]+]] = load %opencl.ndrange_t*, %opencl.ndrange_t** %ndrange 46 // 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) 47 enqueue_kernel(default_queue, flags, ndrange, 48 ^(local void *p) { 49 return; 50 }, 51 256); 52 char c; 53 // CHECK: [[DEF_Q:%[0-9]+]] = load %opencl.queue_t*, %opencl.queue_t** %default_queue 54 // CHECK: [[FLAGS:%[0-9]+]] = load i32, i32* %flags 55 // CHECK: [[NDR:%[0-9]+]] = load %opencl.ndrange_t*, %opencl.ndrange_t** %ndrange 56 // CHECK: [[SIZE:%[0-9]+]] = zext i8 {{%[0-9]+}} to i32 57 // 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]]) 58 enqueue_kernel(default_queue, flags, ndrange, 59 ^(local void *p) { 60 return; 61 }, 62 c); 63 64 // CHECK: [[DEF_Q:%[0-9]+]] = load %opencl.queue_t*, %opencl.queue_t** %default_queue 65 // CHECK: [[FLAGS:%[0-9]+]] = load i32, i32* %flags 66 // CHECK: [[NDR:%[0-9]+]] = load %opencl.ndrange_t*, %opencl.ndrange_t** %ndrange 67 // 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 68 // 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) 69 enqueue_kernel(default_queue, flags, ndrange, 2, event_wait_list2, &clk_event, 70 ^(local void *p) { 71 return; 72 }, 73 256); 74 75 // CHECK: [[DEF_Q:%[0-9]+]] = load %opencl.queue_t*, %opencl.queue_t** %default_queue 76 // CHECK: [[FLAGS:%[0-9]+]] = load i32, i32* %flags 77 // CHECK: [[NDR:%[0-9]+]] = load %opencl.ndrange_t*, %opencl.ndrange_t** %ndrange 78 // 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 79 // CHECK: [[SIZE:%[0-9]+]] = zext i8 {{%[0-9]+}} to i32 80 // 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]]) 81 enqueue_kernel(default_queue, flags, ndrange, 2, event_wait_list2, &clk_event, 82 ^(local void *p) { 83 return; 84 }, 85 c); 86 87 void (^const block_A)(void) = ^{ 88 return; 89 }; 90 void (^const block_B)(local void *) = ^(local void *a) { 91 return; 92 }; 93 94 // CHECK: [[BL:%[0-9]+]] = load void ()*, void ()** %block_A 95 // CHECK: [[BL_I8:%[0-9]+]] = bitcast void ()* [[BL]] to i8* 96 // CHECK: call i32 @__get_kernel_work_group_size_impl(i8* [[BL_I8]]) 97 unsigned size = get_kernel_work_group_size(block_A); 98 // CHECK: [[BL:%[0-9]+]] = load void (i8 addrspace(2)*)*, void (i8 addrspace(2)*)** %block_B 99 // CHECK: [[BL_I8:%[0-9]+]] = bitcast void (i8 addrspace(2)*)* [[BL]] to i8* 100 // CHECK: call i32 @__get_kernel_work_group_size_impl(i8* [[BL_I8]]) 101 size = get_kernel_work_group_size(block_B); 102 // CHECK: [[BL:%[0-9]+]] = load void ()*, void ()** %block_A 103 // CHECK: [[BL_I8:%[0-9]+]] = bitcast void ()* [[BL]] to i8* 104 // CHECK: call i32 @__get_kernel_preferred_work_group_multiple_impl(i8* [[BL_I8]]) 105 size = get_kernel_preferred_work_group_size_multiple(block_A); 106 // CHECK: [[BL:%[0-9]+]] = load void (i8 addrspace(2)*)*, void (i8 addrspace(2)*)* addrspace(1)* @block_G 107 // CHECK: [[BL_I8:%[0-9]+]] = bitcast void (i8 addrspace(2)*)* [[BL]] to i8* 108 // CHECK: call i32 @__get_kernel_preferred_work_group_multiple_impl(i8* [[BL_I8]]) 109 size = get_kernel_preferred_work_group_size_multiple(block_G); 110} 111