clang/test/CodeGenOpenCL/amdgcn-automatic-variable.cl

   1 // NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
   2 // RUN: %clang_cc1 -O0 -cl-std=CL1.2 -triple amdgcn-amd-amdhsa -emit-llvm %s -o - | FileCheck -check-prefixes=CL12 %s
   3 // RUN: %clang_cc1 -O0 -cl-std=CL2.0 -triple amdgcn-amd-amdhsa -emit-llvm %s -o - | FileCheck -check-prefixes=CL20 %s
   4
   5 // CL12-LABEL: define dso_local void @func1(
   6 // CL12-SAME: ptr addrspace(5) noundef [[X:%.*]]) #[[ATTR0:[0-9]+]] {
   7 // CL12-NEXT:  [[ENTRY:.*:]]
   8 // CL12-NEXT:    [[X_ADDR:%.*]] = alloca ptr addrspace(5), align 4, addrspace(5)
   9 // CL12-NEXT:    store ptr addrspace(5) [[X]], ptr addrspace(5) [[X_ADDR]], align 4
  10 // CL12-NEXT:    [[TMP0:%.*]] = load ptr addrspace(5), ptr addrspace(5) [[X_ADDR]], align 4
  11 // CL12-NEXT:    store i32 1, ptr addrspace(5) [[TMP0]], align 4
  12 // CL12-NEXT:    ret void
  13 //
  14 // CL20-LABEL: define dso_local void @func1(
  15 // CL20-SAME: ptr noundef [[X:%.*]]) #[[ATTR0:[0-9]+]] {
  16 // CL20-NEXT:  [[ENTRY:.*:]]
  17 // CL20-NEXT:    [[X_ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
  18 // CL20-NEXT:    [[X_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[X_ADDR]] to ptr
  19 // CL20-NEXT:    store ptr [[X]], ptr [[X_ADDR_ASCAST]], align 8
  20 // CL20-NEXT:    [[TMP0:%.*]] = load ptr, ptr [[X_ADDR_ASCAST]], align 8
  21 // CL20-NEXT:    store i32 1, ptr [[TMP0]], align 4
  22 // CL20-NEXT:    ret void
  23 //
  24 void func1(int *x) {
  25   *x = 1;
  26 }
  27
  28 // CL12-LABEL: define dso_local void @func2(
  29 // CL12-SAME: ) #[[ATTR0]] {
  30 // CL12-NEXT:  [[ENTRY:.*:]]
  31 // CL12-NEXT:    [[LV1:%.*]] = alloca i32, align 4, addrspace(5)
  32 // CL12-NEXT:    [[LV2:%.*]] = alloca i32, align 4, addrspace(5)
  33 // CL12-NEXT:    [[LA:%.*]] = alloca [100 x i32], align 4, addrspace(5)
  34 // CL12-NEXT:    [[LP1:%.*]] = alloca ptr addrspace(5), align 4, addrspace(5)
  35 // CL12-NEXT:    [[LP2:%.*]] = alloca ptr addrspace(5), align 4, addrspace(5)
  36 // CL12-NEXT:    [[LVC:%.*]] = alloca i32, align 4, addrspace(5)
  37 // CL12-NEXT:    store i32 1, ptr addrspace(5) [[LV1]], align 4
  38 // CL12-NEXT:    store i32 2, ptr addrspace(5) [[LV2]], align 4
  39 // CL12-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds [100 x i32], ptr addrspace(5) [[LA]], i64 0, i64 0
  40 // CL12-NEXT:    store i32 3, ptr addrspace(5) [[ARRAYIDX]], align 4
  41 // CL12-NEXT:    store ptr addrspace(5) [[LV1]], ptr addrspace(5) [[LP1]], align 4
  42 // CL12-NEXT:    [[ARRAYDECAY:%.*]] = getelementptr inbounds [100 x i32], ptr addrspace(5) [[LA]], i64 0, i64 0
  43 // CL12-NEXT:    store ptr addrspace(5) [[ARRAYDECAY]], ptr addrspace(5) [[LP2]], align 4
  44 // CL12-NEXT:    call void @func1(ptr addrspace(5) noundef [[LV1]]) #[[ATTR2:[0-9]+]]
  45 // CL12-NEXT:    store i32 4, ptr addrspace(5) [[LVC]], align 4
  46 // CL12-NEXT:    store i32 4, ptr addrspace(5) [[LV1]], align 4
  47 // CL12-NEXT:    ret void
  48 //
  49 // CL20-LABEL: define dso_local void @func2(
  50 // CL20-SAME: ) #[[ATTR0]] {
  51 // CL20-NEXT:  [[ENTRY:.*:]]
  52 // CL20-NEXT:    [[LV1:%.*]] = alloca i32, align 4, addrspace(5)
  53 // CL20-NEXT:    [[LV2:%.*]] = alloca i32, align 4, addrspace(5)
  54 // CL20-NEXT:    [[LA:%.*]] = alloca [100 x i32], align 4, addrspace(5)
  55 // CL20-NEXT:    [[LP1:%.*]] = alloca ptr, align 8, addrspace(5)
  56 // CL20-NEXT:    [[LP2:%.*]] = alloca ptr, align 8, addrspace(5)
  57 // CL20-NEXT:    [[LVC:%.*]] = alloca i32, align 4, addrspace(5)
  58 // CL20-NEXT:    [[LV1_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[LV1]] to ptr
  59 // CL20-NEXT:    [[LV2_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[LV2]] to ptr
  60 // CL20-NEXT:    [[LA_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[LA]] to ptr
  61 // CL20-NEXT:    [[LP1_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[LP1]] to ptr
  62 // CL20-NEXT:    [[LP2_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[LP2]] to ptr
  63 // CL20-NEXT:    [[LVC_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[LVC]] to ptr
  64 // CL20-NEXT:    store i32 1, ptr [[LV1_ASCAST]], align 4
  65 // CL20-NEXT:    store i32 2, ptr [[LV2_ASCAST]], align 4
  66 // CL20-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds [100 x i32], ptr [[LA_ASCAST]], i64 0, i64 0
  67 // CL20-NEXT:    store i32 3, ptr [[ARRAYIDX]], align 4
  68 // CL20-NEXT:    store ptr [[LV1_ASCAST]], ptr [[LP1_ASCAST]], align 8
  69 // CL20-NEXT:    [[ARRAYDECAY:%.*]] = getelementptr inbounds [100 x i32], ptr [[LA_ASCAST]], i64 0, i64 0
  70 // CL20-NEXT:    store ptr [[ARRAYDECAY]], ptr [[LP2_ASCAST]], align 8
  71 // CL20-NEXT:    call void @func1(ptr noundef [[LV1_ASCAST]]) #[[ATTR2:[0-9]+]]
  72 // CL20-NEXT:    store i32 4, ptr [[LVC_ASCAST]], align 4
  73 // CL20-NEXT:    store i32 4, ptr [[LV1_ASCAST]], align 4
  74 // CL20-NEXT:    ret void
  75 //
  76 void func2(void) {
  77   int lv1;
  78   lv1 = 1;
  79   int lv2 = 2;
  80
  81   int la[100];
  82   la[0] = 3;
  83
  84   int *lp1 = &lv1;
  85
  86   int *lp2 = la;
  87
  88   func1(&lv1);
  89
  90   const int lvc = 4;
  91   lv1 = lvc;
  92 }
  93
  94 // CL12-LABEL: define dso_local void @func3(
  95 // CL12-SAME: ) #[[ATTR0]] {
  96 // CL12-NEXT:  [[ENTRY:.*:]]
  97 // CL12-NEXT:    [[A:%.*]] = alloca [16 x [1 x float]], align 4, addrspace(5)
  98 // CL12-NEXT:    call void @llvm.memset.p5.i64(ptr addrspace(5) align 4 [[A]], i8 0, i64 64, i1 false)
  99 // CL12-NEXT:    ret void
 100 //
 101 // CL20-LABEL: define dso_local void @func3(
 102 // CL20-SAME: ) #[[ATTR0]] {
 103 // CL20-NEXT:  [[ENTRY:.*:]]
 104 // CL20-NEXT:    [[A:%.*]] = alloca [16 x [1 x float]], align 4, addrspace(5)
 105 // CL20-NEXT:    [[A_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[A]] to ptr
 106 // CL20-NEXT:    call void @llvm.memset.p0.i64(ptr align 4 [[A_ASCAST]], i8 0, i64 64, i1 false)
 107 // CL20-NEXT:    ret void
 108 //
 109 void func3(void) {
 110   float a[16][1] = {{0.}};
 111 }