vhdl code for n*n matrix inversion

hello

I want to write a vhdl code for real or complex matrix inversion.

i don't know how i can send arrays of my matrix to output ports.

i need your helps for writting this code because i'm beginner.

tnx

this is my code but it have errors:

library ieee;

use ieee.std_logic_1164.all;

use ieee.numeric_std.all;

use work.newtype.all;

entity inverse is

port(input: in matrix_t;

--clk: in std_logic;

inversed_matrix: out matrix_t);

end inverse;

architecture arch of inverse is

function change (matrix: matrix_t) return matrix_t is

variable ret,test: matrix_t;

--variable a,b,c: integer ;

begin

--for i in matrix'range(1) loop

--for j in matrix'range(2) loop

test(0,0)<= matrix(1,1) * matrix(0,0);

test(0,1)<= matrix(0,1) * matrix(1,0);

test(1,1)<=test(0,0)-test(0,1);

ret (0,0):= (matrix (1,1))/test(1,1);

ret (1,1):= (matrix (0,0))/test(1,1);

ret (0,1):= -(matrix (0,1))/test(1,1);

ret (1,0):= -(matrix (1,0))/test(1,1);

--end loop;

--end loop;

return ret;

end ;

--end function;

begin

inversed_matrix <= change (input);

end arch;

help me plz