如何在SystemC中使用结构作为信号类型？

Question

I'm trying to simulate in SystemC a block which adds the red component of two pixels P1 and P2 , and always keeps the green and blue components of pixel P1 . I have declared the pixel as an struct and its overload function in the following way:

struct pixel {
   sc_uint<8> r;
   sc_uint<8> g;
   sc_uint<8> b;

   pixel( sc_uint<8> _r = 0, sc_uint<8> _g = 0, sc_uint<8> _b = 0): r(_r), g(_g), b(_b) { }

   bool operator == (const pixel &other) {
      return (r == other.r) && (g == other.g) && (b == other.b);
   }

   // Displaying
   friend ostream& operator << ( ostream& o, const pixel& P ) {
      o << "{" << P.r << "," << P.g << "," << P.b << "}" ;
      return o;
   }
};

//Overload function
void sc_trace( sc_trace_file* _f, const pixel& _foo, const std::string& _s ) {
   sc_trace( _f, _foo.r, _s + "_r" );
   sc_trace( _f, _foo.g, _s + "_g" );
   sc_trace( _f, _foo.b, _s + "_b" );
}

Then I have coded the adder module considering that the signals sc_in are of type pixel , as follows:

SC_MODULE(adder){
    sc_in<pixel> pin1;
    sc_in<pixel> pin2;
    sc_out<pixel> pout;

    SC_CTOR(adder){

        SC_METHOD(addpixel);
        sensitive << pin1 << pin2;
    }

    void addpixel(){
        sc_uint<8> ir;
        sc_uint<8> ig;
        sc_uint<8> ib;

        ir = pin1.r + pin2.r;
        ig = pin1.g;
        ib = pin1.b;

        pout = pixels(ir,ig,ib);

        cout << " P1 = " << pin1 << endl;
        cout << " P2 = " << pin2 << endl;

    }
};

I get the following compiling errors:

test.cpp:46:13: error: ‘class sc_core::sc_in<pixel>’ has no member named ‘r’
   ir = pin1.r + pin2.r;
             ^
test.cpp:46:22: error: ‘class sc_core::sc_in<pixel>’ has no member named ‘r’
   ir = pin1.r + pin2.r;
                      ^
test.cpp:47:13: error: ‘class sc_core::sc_in<pixel>’ has no member named ‘g’
   ig = pin1.g;
             ^
test.cpp:48:13: error: ‘class sc_core::sc_in<pixel>’ has no member named ‘b’
   ib = pin1.b;
             ^
<builtin>: recipe for target 'test' failed

I would like to know how could the method addpixel access to each component RGB of the pixels and make the operation. If I delete the error lines I arrive to display in the terminal the values of pixels P1 and P2 .

Answer 1

sc_in<pixel> is not pixel . I guess you should fetch the value via sc_in::read() like this:

void addpixel(){
    sc_uint<8> ir;
    sc_uint<8> ig;
    sc_uint<8> ib;
    pixel pin1_value = pin1.read();
    pixel pin2_value = pin2.read();

    ir = pin1_value.r + pin2_value.r;
    ig = pin1_value.g;
    ib = pin1_value.b;

    pout = pixels(ir,ig,ib);

    cout << " P1 = " << pin1_value << endl;
    cout << " P2 = " << pin2_value << endl;

}