[SOLVED] How to have an undefined range in function?

[wtr]

Hello, pleases look at the following.

---

Code VHDL - [expand]
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
type byte_2d_matrix_t is array(natural range <>, natural range <>) of BYTE_T;
subtype  COL_RANGE is NATURAL range 0 TO 3;
subtype  ROW_RANGE is NATURAL range 0 TO 3;
 
    function slv2matrix2d (a : STD_LOGIC_VECTOR) return byte_2d_matrix_t is 
        -- sqr(a)
        variable b : byte_2d_matrix_t(ROW_RANGE, COL_RANGE);
        -- variable b : byte_2d_matrix_t(rr, col);
    begin
        for i in COL_RANGE loop
            for j in ROW_RANGE loop
                b(j, i) := a(
                                (BLK_WIDTH - (1+(i*DWORD_WIDTH) + (j*BYTE_WIDTH))) 
                                DOWNTO 
                                (BLK_WIDTH - ((i*DWORD_WIDTH) + ((j+1)*BYTE_WIDTH))) 
                            );
            end loop;
        end loop;
    
    return b;
    end function;

---

How can I facilitate the situation where I want to change this function to

function slv2matrix2d (a : STD_LOGIC_VECTOR, range_of_row, Range_of_col) return byte_2d_matrix_t

Whereby this means that BLK_WIDTH = a'length and DWORD_WIDTH=cellsize(byte)*numbers of col

I don't think range ascending or decending matters
Regards

 

[TrickyDicky]

Define all the parameters for the function for word matrix size:

---

Code VHDL - [expand]
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
function slv2matrix2d(a : in std_logic_vector; n_rows, n_cols : integer) return byte_2d_matrix_t is
  variable  ret : byte_2d_matrix_t(0 to n_rows-1, 0 to n_cols-1);
begin
  assert (a rem 8) = 0 and n_rows*n_cols*8 = a'length
    report "Specified paramters do name match input word length"
      severity failure;
 
  for i in 0 to n_rows-1 loop
    for j in 0 to n_cols-1 loop
      ret(i,j) = a( (i*n_cols + j)*8 + 7 downto (i*n_cols + j)*8);
    end loop;
  end loop
 
  return ret;
end function;

---

This way, it doesnt matter what the range of the output is, because this will already be known. the 0 to n_rows etc will just map onto whatever the output is:

signal my_byte2d : byte_2d_matrix_t(123 downto 77, 5 to 104732464);

my_byte2d <= slv2matrix2d( a, my_byte2d'length(1), my_byte2d'length(2));

 

[wtr]

Tricky, sometimes I wonder if you wrote the language.

 

[TrickyDicky]

no problem. Please see the revised version below, that will work with all versions of a, whether to or downto (and any length of BYTE_T)

---

Code VHDL - [expand]
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
function slv2matrix2d(a : in std_logic_vector; n_rows, n_cols : integer) return byte_2d_matrix_t is
  variable  ret : byte_2d_matrix_t(0 to n_rows-1, 0 to n_cols-1);
  alias a_a : std_logic_vector(a'length-1 downto 0) is a;
  
begin
  assert (a_a'length rem BYTE_T'length) = 0 and n_rows*n_cols*BYTE_T'length = a'length
    report "Specified paramters do name match input word length"
      severity failure;
 
  for i in 0 to n_rows-1 loop
    for j in 0 to n_cols-1 loop
      ret(i,j) = a_a( (i*n_cols + j)*BYTE_T'length + 7 downto (i*n_cols + j)*BYTE_T'length);
    end loop;
  end loop
 
  return ret;
end function;

---