1 | // RUN: %clang_cc1 %s -cl-std=CL2.0 -emit-llvm -o - -O0 -triple spir-unknown-unknown | FileCheck -check-prefixes=COMMON,SPIR %s |
2 | // RUN: %clang_cc1 %s -cl-std=CL2.0 -emit-llvm -o - -O0 -triple amdgcn-amd-amdhsa | FileCheck -check-prefixes=COMMON,AMDGCN %s |
3 | // RUN: %clang_cc1 %s -cl-std=CL2.0 -emit-llvm -o - -O0 -debug-info-kind=limited -triple spir-unknown-unknown | FileCheck -check-prefixes=CHECK-DEBUG %s |
4 | // RUN: %clang_cc1 %s -cl-std=CL2.0 -emit-llvm -o - -O0 -debug-info-kind=limited -triple amdgcn-amd-amdhsa | FileCheck -check-prefixes=CHECK-DEBUG %s |
5 | |
6 | // SPIR: %struct.__opencl_block_literal_generic = type { i32, i32, i8 addrspace(4)* } |
7 | // AMDGCN: %struct.__opencl_block_literal_generic = type { i32, i32, i8* } |
8 | // SPIR: @__block_literal_global = internal addrspace(1) constant { i32, i32, i8 addrspace(4)* } { i32 12, i32 4, i8 addrspace(4)* addrspacecast (i8* bitcast (void (i8 addrspace(4)*, i8 addrspace(3)*)* @block_A_block_invoke to i8*) to i8 addrspace(4)*) } |
9 | // AMDGCN: @__block_literal_global = internal addrspace(1) constant { i32, i32, i8* } { i32 16, i32 8, i8* bitcast (void (i8*, i8 addrspace(3)*)* @block_A_block_invoke to i8*) } |
10 | // COMMON-NOT: .str |
11 | |
12 | // SPIR-LABEL: define internal {{.*}}void @block_A_block_invoke(i8 addrspace(4)* %.block_descriptor, i8 addrspace(3)* %a) |
13 | // AMDGCN-LABEL: define internal {{.*}}void @block_A_block_invoke(i8* %.block_descriptor, i8 addrspace(3)* %a) |
14 | void (^block_A)(local void *) = ^(local void *a) { |
15 | return; |
16 | }; |
17 | |
18 | // COMMON-LABEL: define {{.*}}void @foo() |
19 | void foo(){ |
20 | int i; |
21 | // COMMON-NOT: %block.isa |
22 | // COMMON-NOT: %block.flags |
23 | // COMMON-NOT: %block.reserved |
24 | // COMMON-NOT: %block.descriptor |
25 | // SPIR: %[[block_size:.*]] = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 }>, <{ i32, i32, i8 addrspace(4)*, i32 }>* %block, i32 0, i32 0 |
26 | // AMDGCN: %[[block_size:.*]] = getelementptr inbounds <{ i32, i32, i8*, i32 }>, <{ i32, i32, i8*, i32 }> addrspace(5)* %block, i32 0, i32 0 |
27 | // SPIR: store i32 16, i32* %[[block_size]] |
28 | // AMDGCN: store i32 20, i32 addrspace(5)* %[[block_size]] |
29 | // SPIR: %[[block_align:.*]] = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 }>, <{ i32, i32, i8 addrspace(4)*, i32 }>* %block, i32 0, i32 1 |
30 | // AMDGCN: %[[block_align:.*]] = getelementptr inbounds <{ i32, i32, i8*, i32 }>, <{ i32, i32, i8*, i32 }> addrspace(5)* %block, i32 0, i32 1 |
31 | // SPIR: store i32 4, i32* %[[block_align]] |
32 | // AMDGCN: store i32 8, i32 addrspace(5)* %[[block_align]] |
33 | // SPIR: %[[block_invoke:.*]] = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 }>, <{ i32, i32, i8 addrspace(4)*, i32 }>* %[[block:.*]], i32 0, i32 2 |
34 | // SPIR: store i8 addrspace(4)* addrspacecast (i8* bitcast (i32 (i8 addrspace(4)*)* @__foo_block_invoke to i8*) to i8 addrspace(4)*), i8 addrspace(4)** %[[block_invoke]] |
35 | // SPIR: %[[block_captured:.*]] = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 }>, <{ i32, i32, i8 addrspace(4)*, i32 }>* %[[block]], i32 0, i32 3 |
36 | // SPIR: %[[i_value:.*]] = load i32, i32* %i |
37 | // SPIR: store i32 %[[i_value]], i32* %[[block_captured]], |
38 | // SPIR: %[[blk_ptr:.*]] = bitcast <{ i32, i32, i8 addrspace(4)*, i32 }>* %[[block]] to %struct.__opencl_block_literal_generic* |
39 | // SPIR: %[[blk_gen_ptr:.*]] = addrspacecast %struct.__opencl_block_literal_generic* %[[blk_ptr]] to %struct.__opencl_block_literal_generic addrspace(4)* |
40 | // SPIR: store %struct.__opencl_block_literal_generic addrspace(4)* %[[blk_gen_ptr]], %struct.__opencl_block_literal_generic addrspace(4)** %[[block_B:.*]], |
41 | // SPIR: %[[block_literal:.*]] = load %struct.__opencl_block_literal_generic addrspace(4)*, %struct.__opencl_block_literal_generic addrspace(4)** %[[block_B]] |
42 | // SPIR: %[[blk_gen_ptr:.*]] = bitcast %struct.__opencl_block_literal_generic addrspace(4)* %[[block_literal]] to i8 addrspace(4)* |
43 | // SPIR: call {{.*}}i32 @__foo_block_invoke(i8 addrspace(4)* %[[blk_gen_ptr]]) |
44 | // AMDGCN: %[[block_invoke:.*]] = getelementptr inbounds <{ i32, i32, i8*, i32 }>, <{ i32, i32, i8*, i32 }> addrspace(5)* %[[block:.*]], i32 0, i32 2 |
45 | // AMDGCN: store i8* bitcast (i32 (i8*)* @__foo_block_invoke to i8*), i8* addrspace(5)* %[[block_invoke]] |
46 | // AMDGCN: %[[block_captured:.*]] = getelementptr inbounds <{ i32, i32, i8*, i32 }>, <{ i32, i32, i8*, i32 }> addrspace(5)* %[[block]], i32 0, i32 3 |
47 | // AMDGCN: %[[i_value:.*]] = load i32, i32 addrspace(5)* %i |
48 | // AMDGCN: store i32 %[[i_value]], i32 addrspace(5)* %[[block_captured]], |
49 | // AMDGCN: %[[blk_ptr:.*]] = bitcast <{ i32, i32, i8*, i32 }> addrspace(5)* %[[block]] to %struct.__opencl_block_literal_generic addrspace(5)* |
50 | // AMDGCN: %[[blk_gen_ptr:.*]] = addrspacecast %struct.__opencl_block_literal_generic addrspace(5)* %[[blk_ptr]] to %struct.__opencl_block_literal_generic* |
51 | // AMDGCN: store %struct.__opencl_block_literal_generic* %[[blk_gen_ptr]], %struct.__opencl_block_literal_generic* addrspace(5)* %[[block_B:.*]], |
52 | // AMDGCN: %[[block_literal:.*]] = load %struct.__opencl_block_literal_generic*, %struct.__opencl_block_literal_generic* addrspace(5)* %[[block_B]] |
53 | // AMDGCN: %[[blk_gen_ptr:.*]] = bitcast %struct.__opencl_block_literal_generic* %[[block_literal]] to i8* |
54 | // AMDGCN: call {{.*}}i32 @__foo_block_invoke(i8* %[[blk_gen_ptr]]) |
55 | |
56 | int (^ block_B)(void) = ^{ |
57 | return i; |
58 | }; |
59 | block_B(); |
60 | } |
61 | |
62 | // SPIR-LABEL: define internal {{.*}}i32 @__foo_block_invoke(i8 addrspace(4)* %.block_descriptor) |
63 | // SPIR: %[[block:.*]] = bitcast i8 addrspace(4)* %.block_descriptor to <{ i32, i32, i8 addrspace(4)*, i32 }> addrspace(4)* |
64 | // SPIR: %[[block_capture_addr:.*]] = getelementptr inbounds <{ i32, i32, i8 addrspace(4)*, i32 }>, <{ i32, i32, i8 addrspace(4)*, i32 }> addrspace(4)* %[[block]], i32 0, i32 3 |
65 | // SPIR: %[[block_capture:.*]] = load i32, i32 addrspace(4)* %[[block_capture_addr]] |
66 | // AMDGCN-LABEL: define internal {{.*}}i32 @__foo_block_invoke(i8* %.block_descriptor) |
67 | // AMDGCN: %[[block:.*]] = bitcast i8* %.block_descriptor to <{ i32, i32, i8*, i32 }>* |
68 | // AMDGCN: %[[block_capture_addr:.*]] = getelementptr inbounds <{ i32, i32, i8*, i32 }>, <{ i32, i32, i8*, i32 }>* %[[block]], i32 0, i32 3 |
69 | // AMDGCN: %[[block_capture:.*]] = load i32, i32* %[[block_capture_addr]] |
70 | |
71 | // COMMON-NOT: define{{.*}}@__foo_block_invoke_kernel |
72 | |
73 | // Test that we support block arguments. |
74 | // COMMON-LABEL: define {{.*}} @blockArgFunc |
75 | int blockArgFunc(int (^ bl)(void)) { |
76 | return bl(); |
77 | } |
78 | |
79 | // COMMON-LABEL: define {{.*}} @get21 |
80 | // COMMON: define {{.*}} @__get21_block_invoke |
81 | // COMMON: ret i32 21 |
82 | int get21() { |
83 | return blockArgFunc(^{return 21;}); |
84 | } |
85 | |
86 | // COMMON-LABEL: define {{.*}} @get42 |
87 | // COMMON: define {{.*}} @__get42_block_invoke |
88 | // COMMON: ret i32 42 |
89 | int get42() { |
90 | return blockArgFunc(^{return 42;}); |
91 | } |
92 | |
93 | // COMMON-LABEL: define {{.*}}@call_block |
94 | // call {{.*}}@__call_block_block_invoke |
95 | int call_block() { |
96 | return ^int(int num) { return num; } (11); |
97 | } |
98 | |
99 | // CHECK-DEBUG: !DIDerivedType(tag: DW_TAG_member, name: "__size" |
100 | // CHECK-DEBUG: !DIDerivedType(tag: DW_TAG_member, name: "__align" |
101 | |
102 | // CHECK-DEBUG-NOT: !DIDerivedType(tag: DW_TAG_member, name: "__isa" |
103 | // CHECK-DEBUG-NOT: !DIDerivedType(tag: DW_TAG_member, name: "__flags" |
104 | // CHECK-DEBUG-NOT: !DIDerivedType(tag: DW_TAG_member, name: "__reserved" |
105 | // CHECK-DEBUG-NOT: !DIDerivedType(tag: DW_TAG_member, name: "__FuncPtr" |
106 | |