test/tint/access/ptr.wgsl.expected.ir.dxc.hlsl

struct S {
  int a;
  int b;
};

struct main_inputs {
  uint tint_local_index : SV_GroupIndex;
};


RWByteAddressBuffer s : register(u0);
groupshared int g1;
int accept_value(int val) {
  return val;
}

int accept_ptr_deref_call_func(inout int val) {
  int v = val;
  return (v + accept_value(val));
}

int accept_ptr_deref_pass_through(inout int val) {
  int v_1 = val;
  return (v_1 + accept_ptr_deref_call_func(val));
}

int accept_ptr_to_struct_and_access(inout S val) {
  return (val.a + val.b);
}

int accept_ptr_to_struct_access_pass_ptr(inout S val) {
  val.a = int(2);
  return val.a;
}

int tint_f32_to_i32(float value) {
  return (((value <= 2147483520.0f)) ? ((((value >= -2147483648.0f)) ? (int(value)) : (int(-2147483648)))) : (int(2147483647)));
}

int accept_ptr_vec_access_elements(inout float3 v1) {
  v1[0u] = cross(v1, v1)[0u];
  return tint_f32_to_i32(v1.x);
}

int call_builtin_with_mod_scope_ptr() {
  int v_2 = int(0);
  InterlockedOr(g1, int(0), v_2);
  return v_2;
}

void main_inner(uint tint_local_index) {
  if ((tint_local_index == 0u)) {
    int v_3 = int(0);
    InterlockedExchange(g1, int(0), v_3);
  }
  GroupMemoryBarrierWithGroupSync();
  int v1 = int(0);
  S v2 = (S)0;
  float3 v4 = (0.0f).xxx;
  int v_4 = int(0);
  InterlockedOr(g1, int(0), v_4);
  int t1 = v_4;
  int v_5 = accept_ptr_deref_pass_through(v1);
  int v_6 = (v_5 + accept_ptr_to_struct_and_access(v2));
  int v_7 = (v_6 + accept_ptr_to_struct_and_access(v2));
  int v_8 = (v_7 + accept_ptr_vec_access_elements(v4));
  int v_9 = (v_8 + accept_ptr_to_struct_access_pass_ptr(v2));
  s.Store(0u, asuint(((v_9 + call_builtin_with_mod_scope_ptr()) + t1)));
}

[numthreads(1, 1, 1)]
void main(main_inputs inputs) {
  main_inner(inputs.tint_local_index);
}