VHDL aggregates assignment

Is it possible to use aggregates to perform the inverse of this
concatenation operation:

signal d : std_logic_vector(11 downto 0);
signal a1, b1, c1 : std_logic_vector(3 downto 0);
d <= a1 & b1 & c1;

With something like this:

signal a2, b2, c2 : std_logic_vector(3 downto 0);
(a2, b2, c2) <= d;

This syntax doesn't work, as the signals are vectors, not std_logic. I
probably just don't have the syntax right. I'd like to use this method
as it is less error prone than the following, especially if lengths are
changed later:

a2 <= d(11 downto 8);
b2 <= d(7 downto 4);
c2 <= d(3 downto 0);

Thanks,
Brian

 

[Mike Treseler]

bkuschak@gmail.com wrote:

Is it possible to use aggregates to perform the inverse of this
concatenation operation
....
With something like this:

signal a2, b2, c2 : std_logic_vector(3 downto 0);
(a2, b2, c2) <= d;

Yes, but it's probably more trouble than you expect.
You have to declare some types and subtypes.
See below:
(aa,bb,cc) := twelve_bit;

-- Mike Treseler

___________________________________________________________

library ieee;
use ieee.std_logic_1164.all;
-- use ieee.numeric_std.all;
-- Mon Sep 12 15:27:54 2005 Mike Treseler
entity aggregate is
end entity aggregate;

architecture play of aggregate is
begin
what : process is
variable a, b, c, d : std_ulogic;
subtype u3_t is std_logic_vector(2 downto 0);
subtype u4_t is std_logic_vector(3 downto 0);
type u4x3_t is array (0 to 2) of u4_t;
variable twelve_bit : u4x3_t := ("0000","1111","1010");
variable four_bit, aa,bb,cc : u4_t;
variable three_bit : u3_t;
begin -- process what
-- assign 4 bits to aggregate
-- vector type qualification required on right.
(a, b, c, d) := std_logic_vector'("0011");
-- assign four bit vector from aggregate
four_bit := (a, b, c, d);
assert four_bit = "0011" report "not 3";
-- reverse 4 bit values using two aggregates
-- vector type qualification required on right.
(d, c, b, a) := std_logic_vector'(a, b, c, d);
-- assign reversed values to vector
four_bit := (a, b, c, d);
assert four_bit = "1100" report "not 12";
three_bit := (a, b, c);
-- assign three small vectors from type u4x3_t
(aa,bb,cc) := twelve_bit;
assert cc = "1010" report "cc error";
assert bb = "1111" report "bb error";
assert aa = "0000" report "aa error";
report("Expect no assertions above.");
wait;
end process what;
end architecture play;
-------------------------------------------------------------------------------
--# vsim -c aggregate
--# Loading /steptoe/usr1/modeltech/linux/../std.standard
--# Loading /steptoe/usr1/modeltech/linux/../ieee.std_logic_1164(body)
--# Loading /steptoe/usr1/modeltech/linux/../ieee.numeric_std(body)
--# Loading play.aggregate(play)

--VSIM 1> run 1
--# ** Note: Expect no assertions above.
--# Time: 0 ns Iteration: 0 Instance: /aggregate
--VSIM 2>

 

Yeah, I figured it would require something like that.
Thanks,
Brian