replacing whens with some kind of for-loop,

[moberg2236]

To be able to make the design i'm working on mor general so it will be easier to change,change the number of inports for example.
Therefore i would like to replace the code below with a for loop.



time_out <= inports(0) when id_in_sig = "0000" else
inports(1) when id_in_sig = "0001" else
inports(2) when id_in_sig = "0010" else
inports(3) when id_in_sig = "0011" else
inports(4) when id_in_sig = "0100" else
inports(5) when id_in_sig = "0101" else
inports(7) when id_in_sig = "0111" else
inports(7) when id_in_sig = "0111" else
inports(8) when id_in_sig = "1000" else
inports(9) when id_in_sig = "1001" else
inports(10) when id_in_sig = "1010" else
inports(11) when id_in_sig = "1011" else
inports(12) when id_in_sig = "1100" else
inports(13) when id_in_sig = "1101" else
inports(14) when id_in_sig = "1110" else
inports(15) when id_in_sig = "1111";

With different variations, i have tested the following for loop


snurra: for N in 0 to interrupt_in_size generate
time_out<= inports(N) when id_in_sig = conv_std_logic_vector(N, id_size);

end generate;

The problem is now that i cant find a way to make it work in any way, as you see, the for loop will result in that time_out will get assigned multiple values, but i only want it to be set to one value when a certain id_in_sig is set,

For your information, inports is an array of size 16 of std_logic_vectors (15 downto 0)

Time out is a std_logic_vector(15 downto 0)
id_in_sig is a 4 bit id.

Making a for-loop works for me in every other case, except for this one where i want 1 signal to have 16 possible values depending on the id_sig

This is not a problem to make my code work, but since figuring this one out, probably could help speeding up my coding i would appreciate any help

 

[eliascm]

Replacing "whens"

The code below does essentially what you are trying to do. The function num_bits is in a package I wrote (vhdl_lib_2.arithmetic_pkg).

--=====================================================================
library ieee;
use ieee.std_logic_1164.all;
--=====================================================================
package m_to_n_mux_pkg is

constant mux_in_width : positive := 256; -- m
constant mux_out_width : positive := 9; -- n

type vec_array is array(mux_in_width - 1 downto 0) of std_logic_vector(mux_out_width - 1 downto 0);

component m_to_n_mux
port( sel : in std_logic_vector;
mux_in : in vec_array;
mux_out : out std_logic_vector(mux_out_width - 1 downto 0)
);
end component;

end m_to_n_mux_pkg;
--=====================================================================
library ieee;
use ieee.std_logic_1164.all;
use ieee.numeric_std.all;
library vhdl_lib_2;
use vhdl_lib_2.arithmetic_pkg.all;
use work.m_to_n_mux_pkg.all;
--=====================================================================
--The num_bits function returns the minimum number of bits required to represent a specified number
-- of states using binary encoding. For example: num_bits(5) = 3, num_bits(8) = 3, num_bits(9) = 4
-----------------------------------------------------------------------
entity m_to_n_mux is
port( sel : in std_logic_vector(num_bits(mux_in_width) - 1 downt to 0);
mux_in : in vec_array;
mux_out : out std_logic_vector(mux_out_width - 1 downto 0)
);
end m_to_n_mux;

architecture arch_mux of m_to_n_mux is

begin
pmux : process(mux_in, sel)
begin
mux_out <= mux_in(to_integer(unsigned(sel)));
end process;
end arch_mux;