Found a strange occurance with this register I coded. I'm very new to VHDL, but I was taught when writing a value to output ports like data_out you should always use a "middleman" signal to transfer your value. Here I tried to use the signal "data" to transfer the signal, but this implementation resulted in a delay before data_out changes (when ld goes high). Taking out data completely and coding how I would in a C program removes this delay and the register works perfectly. Any idea on why this is and why I shouldn't do this?
Broken code:
entity register is
generic (
DATA_WIDTH : natural := 12);
port (
data_in : in std_logic_vector(DATA_WIDTH-1 downto 0);
ld : in std_logic;
clk : in std_logic;
rst_L : in std_logic;
data_out : out std_logic_vector(DATA_WIDTH-1 downto 0));
end entity register;
architecture bhv of register is
signal data : std_logic_vector(DATA_WIDTH-1 downto 0);
begin -- bhv
REG : process (clk, rst_L, ld, data_in)
begin -- process REG
if rst_L = '0' then
data <= (others => '0');
else
if clk'event and clk = '1' then
if ld = '1' then
data <= data_in;
end if;
end if;
end if;
data_out <= data;
end process REG;
end architecture bhv;
Changes for process that made it work:
REG : process (clk, rst_L, ld, data_in)
begin -- process REG
if rst_L = '0' then
data <= (others => '0');
else
if clk'event and clk = '1' then
if ld = '1' then
data_out <= data_in;
end if;
end if;
end if;
end process REG;
Just wanted to know what I did wrong and why I even have to use a signal to transfer the value if the code works fine. Thanks!
The problem in the broken process code is that data signal is not updated for
read until after a delta delay, thus the data_out update by
data_out <= data assign is postponed until next execution of the process code, thereby giving a delay in simulation.
Note that the ld and data_in in the sensitivity list of the initial process are not required, since use of these are guarded by rising clk.
Update of the code can be:
reg : process (clk, rst_L)
begin -- process REG
if rst_L = '0' then
data <= (others => '0');
else
if clk'event and clk = '1' then -- May be written as rising_edge(clk) instead
if ld = '1' then
data <= data_in;
end if;
end if;
end if;
end process REG;
data_out <= data;
It may be useful to take a look at
VHDL's crown jewel for some information about processes and delta cycles in VHDL.
Note that register is a VHDL reserved word, so it can't be used as identifier
for the entity.
Unexpected delays with register VHDL
The problem with the original process is that data is not in the sensitivity list, delaying the assignment to data_out until the next time the process executes instead of the next simulation data cycle. The process will execute on any transaction of a signal in the sensitivity list - originally as you indicated ld.
Adding data to the sensitivity list of your original process:
REG : process (clk, rst_L, data)
begin -- process REG
if rst_L = '0' then
data <= (others => '0');
else
if clk'event and clk = '1' then
if ld = '1' then
data <= data_in;
end if;
end if;
end if;
data_out <= data;
end process REG;
Will allow the assignment to data_out to occur in a delta simulation cycle instead of waiting until a transition on a signal currently in the sensitivity list.
I removed the extraneous signals from the sensitivity list for the same reason Morten did in his answer. These were causing execution of the process at other times, but not when a transaction occurred on data.
And yes, as long as you don't have the data_out signal in an expression (e.g. on the right hand side of an assignment statement) in your architecture you don't need the intermediary variable data and your simulation will go a bit faster.
Your changed process is the most efficient implementation once the sensitivity list is pared down.
Simply adding data to the sensitivity list would have caused the original process to simulate correctly, feel free to try it.
Related
I have some Filter coefficients in BRAM those coefficients need to be written into an array to perform convolution. I have created an array using type and assigned it to a signal. That signal I have port mapped to DATA_OUT of BRAM. it's giving an error "expecting STD_LOGIC_VECTOR"
I've tried writing the data in an array with for loop. It results in an error "indices is not a STD_LOGIC_VECTOR"
my type declaration
TYPE coeff_pipe IS ARRAY(0 TO 15) OF std_logic_vector(7 downto 0);
Signal coeff:coeff_pipe;
my for loop is like this
for i in 0 to loop
coeff(i) <= data_out_BRAM(i); end loop;
Help me with suitable changes in my code to make it work
You mix the for loop behavior in a coding language (C) and in a hardware descritpion language (VHDL). In a coding language, if you write a for loop, processor will execute the content of the loop several times in row sequentially (one after one). In an HDL, for loop is used to instanciate several times the same circuit with different inputs/outputs. There is no time notion in a for loop.
In your case, you have to use a sequential process and increment your BRAM address :
process(clk, rst)
begin
if rst = '1' then
addr_BRAM <= (others => '0');
addr_BRAM_d <= (others => '0');
ram_init_en <= '1';
ram_init_en_d <= '0';
coeff <= (others => (others => '0'));
elsif rising_edge(clk) then
addr_BRAM_d <= addr_BRAM ; -- Delay of 1 clk cycle
ram_init_en_d <= ram_init_en; -- Delay of 1 clk cycle
-- Init done
if addr_BRAM = x"1111" then
ram_init_en <= '0';
end if;
-- Increment BRAM address
if ram_init_en = '1' then
addr_BRAM <= std_logic_vector(unsigned(addr_BRAM) + 1);
end if;
-- Get data one cycle after set address because a BRAM doesn't answer instant, it answers in one clk cycle.
if ram_init_en_d = '1' then
coeff(to_integer(unsigned(addr_BRAM_d))) <= data_out_BRAM;
end if;
end if;
end process;
I am unfortunately new to VHDL but not new to software development. What is the equivalency to functions in VHDL? Specifically, in the code below I need to debounce four push buttons instead of one. Obviously repeating my process code four times and suffixing each of my signals with a number to make them unique for the four instances is not the professional nor correct way of doing this. How do I collapse all this down into one process "function" to which I can "pass" the signals so I can excise all this duplicate code?
----------------------------------------------------------------------------------
-- Debounced pushbutton examples
----------------------------------------------------------------------------------
library IEEE;
use IEEE.STD_LOGIC_1164.ALL;
use IEEE.STD_LOGIC_UNSIGNED.ALL;
entity pushbutton is
generic(
counter_size : integer := 19 -- counter size (19 bits gives 10.5ms with 50MHz clock)
);
port(
CLK : in std_logic; -- input clock
BTN : in std_logic_vector(0 to 3); -- input buttons
AN : out std_logic_vector(0 to 3); -- 7-segment digit anodes ports
LED : out std_logic_vector(0 to 3) -- LEDs
);
end pushbutton;
architecture pb of pushbutton is
signal flipflops0 : std_logic_vector(1 downto 0); -- input flip flops
signal flipflops1 : std_logic_vector(1 downto 0);
signal flipflops2 : std_logic_vector(1 downto 0);
signal flipflops3 : std_logic_vector(1 downto 0);
signal counter_set0 : std_logic; -- sync reset to zero
signal counter_set1 : std_logic;
signal counter_set2 : std_logic;
signal counter_set3 : std_logic;
signal counter_out0 : std_logic_vector(counter_size downto 0) := (others => '0'); -- counter output
signal counter_out1 : std_logic_vector(counter_size downto 0) := (others => '0');
signal counter_out2 : std_logic_vector(counter_size downto 0) := (others => '0');
signal counter_out3 : std_logic_vector(counter_size downto 0) := (others => '0');
signal button0 : std_logic; -- debounce input
signal button1 : std_logic;
signal button2 : std_logic;
signal button3 : std_logic;
signal result0 : std_logic; -- debounced signal
signal result1 : std_logic;
signal result2 : std_logic;
signal result3 : std_logic;
begin
-- Make sure Mercury BaseBoard 7-Seg Display is disabled (anodes are pulled high)
AN <= (others => '1');
-- Feed buttons into debouncers
button0 <= BTN(0);
button1 <= BTN(1);
button2 <= BTN(2);
button3 <= BTN(3);
-- Start or reset the counter at the right time
counter_set0 <= flipflops0(0) xor flipflops0(1);
counter_set1 <= flipflops1(0) xor flipflops1(1);
counter_set2 <= flipflops2(0) xor flipflops2(1);
counter_set3 <= flipflops3(0) xor flipflops3(1);
-- Feed LEDs from the debounce circuitry
LED(0) <= result0;
LED(1) <= result1;
LED(2) <= result2;
LED(3) <= result3;
-- Debounce circuit 0
process (CLK)
begin
if (CLK'EVENT and CLK = '1') then
flipflops0(0) <= button0;
flipflops0(1) <= flipflops0(0);
if (counter_set0 = '1') then -- reset counter because input is changing
counter_out0 <= (others => '0');
elsif (counter_out0(counter_size) = '0') then -- stable input time is not yet met
counter_out0 <= counter_out0 + 1;
else -- stable input time is met
result0 <= flipflops0(1);
end if;
end if;
end process;
-- Debounce circuit 1
process (CLK)
begin
if (CLK'EVENT and CLK = '1') then
flipflops1(0) <= button1;
flipflops1(1) <= flipflops1(0);
if (counter_set1 = '1') then -- reset counter because input is changing
counter_out1 <= (others => '0');
elsif (counter_out1(counter_size) = '0') then -- stable input time is not yet met
counter_out1 <= counter_out1 + 1;
else -- stable input time is met
result1 <= flipflops1(1);
end if;
end if;
end process;
-- Debounce circuit 2
process (CLK)
begin
if (CLK'EVENT and CLK = '1') then
flipflops2(0) <= button2;
flipflops2(1) <= flipflops2(0);
if (counter_set2 = '1') then -- reset counter because input is changing
counter_out2 <= (others => '0');
elsif (counter_out2(counter_size) = '0') then -- stable input time is not yet met
counter_out2 <= counter_out2 + 1;
else -- stable input time is met
result2 <= flipflops2(1);
end if;
end if;
end process;
-- Debounce circuit 3
process (CLK)
begin
if (CLK'EVENT and CLK = '1') then
flipflops3(0) <= button3;
flipflops3(1) <= flipflops3(0);
if (counter_set3 = '1') then -- reset counter because input is changing
counter_out3 <= (others => '0');
elsif (counter_out3(counter_size) = '0') then -- stable input time is not yet met
counter_out3 <= counter_out3 + 1;
else -- stable input time is met
result3 <= flipflops3(1);
end if;
end if;
end process;
end pb;
VHDL has functions but function calls are expressions and not statements or expression statements as in some programming languages. A function call always returns a value of a type and an expression can't represent a portion of a design hierarchy.
Consider the other subprogram class procedures which are statements instead.
The debouncer processes and associated declarations can also be simplified without using a procedure:
library ieee;
use ieee.std_logic_1164.all;
use ieee.numeric_std.all;
entity pushbutton is
generic (
counter_size: integer := 19 -- The left bound of debounce counters
);
port (
clk: in std_logic;
btn: in std_logic_vector(0 to 3);
an: out std_logic_vector(0 to 3);
led: out std_logic_vector(0 to 3)
);
end entity pushbutton;
architecture pb1 of pushbutton is
-- There are two flip flops for each of four buttons:
subtype buttons is std_logic_vector(0 to 3);
type flip_flops is array (0 to 1) of buttons;
signal flipflops: flip_flops;
signal counter_set: std_logic_vector(0 to 3);
use ieee.numeric_std.all;
type counter is array (0 to 3) of
unsigned(counter_size downto 0);
signal counter_out: counter := (others => (others => '0'));
begin
an <= (others => '1');
counter_set <= flipflops(0) xor flipflops(1);
DEBOUNCE:
process (clk)
begin
if rising_edge (clk) then
flipflops(0) <= btn;
flipflops(1) <= flipflops(0);
for i in 0 to 3 loop
if counter_set(i) = '1' then
counter_out(i) <= (others => '0');
elsif counter_out(i)(counter_size) = '0' then
counter_out(i) <= counter_out(i) + 1;
else
led(i) <= flipflops(1)(i);
end if;
end loop;
end if;
end process;
end architecture pb1;
Moving part of the design specification into a procedure:
architecture pb2 of pushbutton is
-- There are two flip flops for each of four buttons:
subtype buttons is std_logic_vector(0 to 3);
type flip_flops is array (0 to 1) of buttons;
signal flipflops: flip_flops;
signal counter_set: std_logic_vector(0 to 3);
use ieee.numeric_std.all;
type counter is array (0 to 3) of
unsigned(counter_size downto 0);
signal counter_out: counter := (others => (others => '0'));
procedure debounce (
-- Can eliminate formals of mode IN within the scope of their declaration:
-- signal counter_set: in std_logic_vector (0 to 3);
-- signal flipflops: in flip_flops;
signal counter_out: inout counter;
signal led: out std_logic_vector(0 to 3)
) is
begin
for i in 0 to 3 loop
if counter_set(i) = '1' then
counter_out(i) <= (others => '0');
elsif counter_out(i)(counter_size) = '0' then
counter_out(i) <= counter_out(i) + 1;
else
led(i) <= flipflops(1)(i);
end if;
end loop;
end procedure;
begin
an <= (others => '1');
counter_set <= flipflops(0) xor flipflops(1);
DEBOUNCER:
process (clk)
begin
if rising_edge (clk) then
flipflops(0) <= btn;
flipflops(1) <= flipflops(0);
-- debounce(counter_set, flipflops, counter_out, led);
debounce (counter_out, led);
end if;
end process;
end architecture pb2;
Here the procedure serves as a collection of sequential statements and doesn't save any lines of code.
Sequential procedure calls can be useful to hide repetitious clutter. The clutter has been removed already by consolidating declarations and using the loop statement. There's a balancing act between the design entry effort, code maintenance effort and user readability, which can also be affected by coding style. Coding style is also affected by RTL constructs implying hardware.
Moving the clock evaluation into a procedure would require the procedure call be be a concurrent statement, similar to an instantiation, which you already have. It doesn't seem worthwhile here should you consolidate signals declared as block declarative items in the architecture body or when using a loop statement.
Note that result and button declarations have been eliminated. Also the use of package numeric_std and type unsigned for the counters prevents inadvertent assignment to other objects with the same subtype. The counter values are treated as unsigned numbers while counter_set for instance is not.
Also there's an independent counter for each input being debounced just as in the original. Without independent counters some events might be lost for independent inputs when a single counter is repetitively cleared.
This code hasn't been validated by simulation, lacking a testbench. With the entity both architectures analyze and elaborate.
There doesn't appear to be anything here other than sequential statements now found in a for loop that would benefit from a function call. Because a function call returns a value the type of that value would either need to be a composite (here a record type) or be split into separate function calls for each assignment target.
There's also the generate statement which can elaborate zero or more copies of declarations and concurrent statements (here a process) as block statements with block declarative items. Any signal used only in an elaborated block can be a block declarative item.
architecture pb3 of pushbutton is
begin
DEBOUNCERS:
for i in btn'range generate
signal flipflops: std_logic_vector (0 to 1);
signal counter_set: std_logic;
signal counter_out: unsigned (counter_size downto 0) :=
(others => '0');
begin
counter_set <= flipflops(0) xor flipflops(1);
DEBOUNCE:
process (clk)
begin
if rising_edge (clk) then
flipflops(0) <= btn(i);
flipflops(1) <= flipflops(0);
if counter_set = '1' then
counter_out <= (others => '0');
elsif counter_out(counter_size) = '0' then
counter_out <= counter_out + 1;
else
led(i) <= flipflops(1);
end if;
end if;
end process;
end generate;
end architecture pb3;
Addendum
The OP pointed out an error made in the above code due to a lack of simulation and complexity hidden by abstraction when synthesizing architecture pb2. While the time for the debounce counter was given at 10.5 ms (50 MHz clock) the name of the generic (counter_size) is also actually the left bound of the counter (given as an unsigned binary counter using type unsigned).
The mistake (two flip flops in the synchronizer for each of four buttons) and simply acceding to the OP's naming convention with respect to the counter has been corrected in the above code.
The OP's synthesis error in the comment relates to the requirement there be a matching element for each element on the left hand or right hand of an aassignment statement.
Without synthesizing the code (which the OP did) the error can't be found without simulation. Because the only thing necessary to find the particular error assigning flipflops(0) is the clock a simple testbench can be written:
use ieee.std_logic_1164.all;
entity pushbutton_tb is
end entity;
architecture fum of pushbutton_tb is
signal clk: std_logic := '0';
signal btn: std_logic_vector (0 to 3);
signal an: std_logic_vector(0 to 3);
signal led: std_logic_vector(0 to 3);
begin
CLOCK:
process
begin
wait for 0.5 ms;
clk <= not clk;
if now > 50 ms then
wait;
end if;
end process;
DUT:
entity work.pushbutton (pb2)
generic map (
counter_size => 4 -- FOR SIMULATION
)
port map (
clk => clk,
btn => btn,
an => an,
led => led
);
STIMULUS:
process
begin
btn <= (others => '0');
wait for 20 ms;
btn(0) <= '1';
wait for 2 ms;
btn(1) <= '1';
wait for 3 ms;
btn(2) <= '1';
wait for 6 ms;
btn(3) <= '1';
wait;
end process;
end architecture;
The corrected code and a testbench to demonstrate there are no matching element errors in assignment during simulation.
Simulation was provided for both architectures with identical results.
The generic was used to reduce the size of the debounce counters using a 1 millisecond clock in the testbench (to avoid simulation time with 50 MHz clock events that don't add to the narrative).
Here's the output of the first architecture's simulation:
The caution here is that designs should be simulated. There's a class of VHDL semantic error conditions that are checked only at runtime (or in synthesis).
Added abstraction for reducing 'uniquified' code otherwise identically performing can introduce such errors.
The generate statement wouldn't have that issue using names in a design hierarchy:
The concurrent statements and declarations found in a generate statement are replicated in any generated block statements implied by the generate statement. Each block statement represents a portion of a design hierarchy.
There's been a trade off between design complexity and waveform display organization for debugging.
A design description depending on hiding repetitious detail should be simulated anyway. Here there are two references to the generate parameter i used in selected names, susceptible to the same errors as ranges should parameter substitution be overlooked.
A multiple bit debouncing circuit might look like this:
library IEEE;
use IEEE.std_logic_1164.all;
use IEEE.numeric_std.all;
use work.Utilities.all;
entity Debouncer is
generic (
CLOCK_PERIOD_NS : positive := 10;
DEBOUNCE_TIME_MS : positive := 3;
BITS : positive
);
port (
Clock : in std_logic;
Input : in std_logic_vector(BITS - 1 downto 0);
Output : out std_logic_vector(BITS - 1 downto 0) := (others => '0')
);
end entity;
architecture rtl of Debouncer is
begin
genBits: for i in Input'range generate
constant DEBOUNCE_COUNTER_MAX : positive := (DEBOUNCE_TIME_MS * 1000000) / CLOCK_PERIOD_NS;
constant DEBOUNCE_COUNTER_BITS : positive := log2(DEBOUNCE_COUNTER_MAX);
signal DebounceCounter : signed(DEBOUNCE_COUNTER_BITS downto 0) := to_signed(DEBOUNCE_COUNTER_MAX - 3, DEBOUNCE_COUNTER_BITS + 1);
begin
process (Clock)
begin
if rising_edge(Clock) then
-- restart counter, whenever Input(i) was unstable within DEBOUNCE_TIME_MS
if (Input(i) /= Output(i)) then
DebounceCounter <= DebounceCounter - 1;
else
DebounceCounter <= to_signed(DEBOUNCE_COUNTER_MAX - 3, DebounceCounter'length);
end if;
-- latch input bit, if input was stable for DEBOUNCE_TIME_MS
if (DebounceCounter(DebounceCounter'high) = '1') then
Output(i) <= Input(i);
end if;
end if;
end process;
end generate;
end architecture;
In stead of a counter size, it expects the user to provide a frequency (as period in nanoseconds) and a debounce time (in milliseconds).
The referenced package implements a log2 function.
Have JUST started learning how to use this tool so if my question seems silly i apologize in advance. I have searched the error in numerous forums (already answered posts , not mine) and couldn't understand what i was doing wrong so here is my question:
My Behavioral Code:
----------------------------------------------------------------------------- -----
-- Company:
-- Engineer:
--
-- Create Date: 01:47:22 07/07/2015
-- Design Name:
-- Module Name: Module_1 - Behavioral
-- Project Name:
-- Target Devices:
-- Tool versions:
-- Description:
--
-- Dependencies:
--
-- Revision:
-- Revision 0.01 - File Created
-- Additional Comments:
--
----------------------------------------------------------------------------- -----
library IEEE;
use IEEE.STD_LOGIC_1164.ALL;
use IEEE.STD_LOGIC_ARITH.ALL;
use IEEE.STD_LOGIC_UNSIGNED.ALL;
-- Uncomment the following library declaration if using
-- arithmetic functions with Signed or Unsigned valuessss
--use IEEE.NUMERIC_STD.ALL;
-- Uncomment the following library declaration if instantiating
-- any Xilinx primitives in this code.
--library UNISIM;
--use UNISIM.VComponents.all;
entity Module_1 is
port (A,B,WE,reset : in std_logic;
clk : in std_logic;
DIN : in signed(3 downto 0);
FULL,EMPTY,ERROR : out std_logic:= '0';
PARKFREE : out signed(3 downto 0)
);
end Module_1;
architecture Behavioral of Module_1 is
signal current_state,next_state:std_ulogic_vector(1 downto 0);
signal empty_bf, full_bf :std_ulogic;
signal enter, reset_b : std_ulogic := '0' ;
constant s0: std_ulogic_vector (1 downto 0):="00";
constant s1: std_ulogic_vector (1 downto 0):="10";
constant s2: std_ulogic_vector (1 downto 0):="11";
constant s3: std_ulogic_vector (1 downto 0):="01";
signal park_counter,buffr: signed(3 downto 0):="0000";
signal PARKTOTAL,free_park_counter: signed(3 downto 0):= "1111";
begin
p1: process (clk,reset,reset_b)
begin
if (reset = '1') then
current_state <= s0;
elsif clk'event and clk = '1' then
current_state <= next_state;
end if;
end process p1;
p2: process (current_state,A,B)
begin
next_state <= current_state;
case current_state is
when s0 =>
if A = '1' then
enter <= '1';
next_state <= s1;
elsif B = '1' then
next_state <= s3;
end if;
when s1 =>
if A = '0' then
enter <= '0';
next_state <= s0;
elsif B = '1' then
next_state <= s2;
end if;
when s2 =>
if A = '0' then
next_state <= s3;
elsif B = '0' then
next_state <= s1;
end if;
when s3 =>
if B = '0' then
enter <= '0';
next_state <= s0;
elsif A = '1' then
next_state <= s2;
end if;
when others =>
end case;
end process p2;
p3: process(current_state,A,B)
begin
case current_state is
when s1 =>
if enter = '0' and A = '0' and empty_bf = '0' then
park_counter <= park_counter - 1;
free_park_counter <= free_park_counter + 1;
ERROR <= '0';
end if;
when s3 =>
if enter = '1' and B = '0' and full_bf = '0' then
park_counter <= park_counter + 1;
free_park_counter <= free_park_counter - 1;
ERROR <= '0';
end if;
when others =>
end case;
end process p3;
max: process(WE)
begin
if clk'event and clk = '1' and WE = '1' then
PARKTOTAL <= DIN ;
buffr <= DIN ;
if (free_park_counter < buffr - park_counter) then
ERROR <= '1';
reset_b <= '1';
else free_park_counter <= buffr - park_counter;
end if;
end if;
end process max;
incr: process(free_park_counter,DIN)
begin
PARKFREE <= free_park_counter;
if (free_park_counter = 15) then
EMPTY <= '1';
empty_bf <= '1';
else EMPTY <= '0';
empty_bf <= '0';
end if;
if (free_park_counter = 0) then
FULL <= '1';
full_bf <= '1';
else FULL <= '0';
full_bf <= '0';
end if;
end process incr;
end Behavioral;
My Testbench
----------------------------------------------------------------------------- ---
-- Company:
-- Engineer:
--
-- Create Date: 02:17:07 07/11/2015
-- Design Name:
-- Module Name: D:/Users/ErgasiaFPGA/Testbench.vhd
-- Project Name: ErgasiaFPGA
-- Target Device:
-- Tool versions:
-- Description:
--
-- VHDL Test Bench Created by ISE for module: Module_1
--
-- Dependencies:
--
-- Revision:
-- Revision 0.01 - File Created
-- Additional Comments:
--
-- Notes:
-- This testbench has been automatically generated using types std_logic and
-- std_logic_vector for the ports of the unit under test. Xilinx recommends
-- that these types always be used for the top-level I/O of a design in order
-- to guarantee that the testbench will bind correctly to the post-implementation
-- simulation model.
--------------------------------------------------------------------------------
LIBRARY ieee;
USE ieee.std_logic_1164.ALL;
use IEEE.STD_LOGIC_ARITH.ALL;
use IEEE.STD_LOGIC_UNSIGNED.ALL;
-- Uncomment the following library declaration if using
-- arithmetic functions with Signed or Unsigned values
--USE ieee.numeric_std.ALL;
ENTITY Testbench IS
END Testbench;
ARCHITECTURE behavior OF Testbench IS
-- Component Declaration for the Unit Under Test (UUT)
COMPONENT Module_1
PORT(
A : IN std_logic;
B : IN std_logic;
WE : IN std_logic;
reset : IN std_logic;
clk : IN std_logic;
DIN : IN signed(3 downto 0);
FULL : OUT std_logic;
EMPTY : OUT std_logic;
ERROR : OUT std_logic;
PARKFREE : OUT signed(3 downto 0)
);
END COMPONENT;
--Inputs
signal A : std_logic := '0';
signal B : std_logic := '0';
signal WE : std_logic := '0';
signal reset : std_logic := '0';
signal clk : std_logic := '0';
signal DIN : signed(3 downto 0) := (others => '0');
--Outputs
signal FULL : std_logic;
signal EMPTY : std_logic;
signal ERROR : std_logic;
signal PARKFREE : signed(3 downto 0);
-- Clock period definitions
constant clk_period : time := 10 ns;
BEGIN
-- Instantiate the Unit Under Test (UUT)
uut: Module_1 PORT MAP (
A => A,
B => B,
WE => WE,
reset => reset,
clk => clk,
DIN => DIN,
FULL => FULL,
EMPTY => EMPTY,
ERROR => ERROR,
PARKFREE => PARKFREE
);
-- Clock process definitions
clk_process :process
begin
clk <= '0';
wait for clk_period/2;
clk <= '1';
wait for clk_period/2;
end process;
-- Stimulus process
stim_proc: process
begin
-- hold reset state for 100 ns.
reset <= '1' ;
wait for 100 ns;
reset <= '0' ;
wait for clk_period*10;
-- insert stimulus here
A <= '1' ;
wait for clk_period*5;
B <= '1' ;
wait for clk_period*5;
A <= '0' ;
wait for clk_period*5;
B <= '0' ;
wait for clk_period*5;
B <= '1' ;
wait for clk_period*5;
A <= '1' ;
wait for clk_period*5;
B <= '0' ;
wait for clk_period*5;
A <= '0' ;
wait;
end process;
END;
I posted the whole code just in case i'm missing something in some part of it that i wouldn't think about. So , when i ISim it , with any "succesful" trigger of p3...
Referencing it again here:
p3: process(current_state,A,B)
begin
case current_state is
when s1 =>
if enter = '0' and A = '0' and empty_bf = '0' then
park_counter <= park_counter - 1;
free_park_counter <= free_park_counter + 1;
ERROR <= '0';
end if;
when s3 =>
if enter = '1' and B = '0' and full_bf = '0' then
park_counter <= park_counter + 1;
free_park_counter <= free_park_counter - 1;
ERROR <= '0';
end if;
when others =>
end case;
end process p3;
...the ISim says that in this part
"There is an 'U'|'X'|'W'|'Z'|'-' in an arithmetic operand, the result will be 'X'(es)."
and proceeds to make Xs of some of the values after that part , although all of the signals have been initialized (at least the ones in this part)
The "park_counter <= park_counter + 1;" part works correctly in the simulation but the "free_park_counter <= free_park_counter -1;" doesn't. This completely baffles me as they are declared as the same type and are both initialized the same way , even with different values.
So what am i missing or even doing blatantly wrong? Any help will be incredibly appreciated. Only looking for the error , if you could please contain optimizations since i'm looking to learn through trial and error and thought and would like to struggle to make it better myself
In addition , please be patient with my responses since i log on 2 to 3 times per day. Thanks in advance
Your design is non-workable per Brian's answer. Your testbench causes the messages when going from s3 or s1 to s0 before the clock edge. free_park_counter goes to 'U's. (Once it gets U's it won't loop further, no events occur without a signal value change).
Your counters should be clocked to prevent combinatorial looping, plus they likely won't synthesize a clock usefully due to uneven combinatorial delays. Sensitivity lists should likewise be complete, if for no other reason than the intent is to make simulation match the synthesized result.
Looking at the result of your testbench:
(clickable)
We can compare that with the messages from the arithmetic operators found in the Synopsys package std_logic_arith:
../../../src/synopsys/std_logic_arith.vhdl:315:20:#350ns:(assertion warning): There is an 'U'|'X'|'W'|'Z'|'-' in an arithmetic operand, the result will be 'X'(es).
../../../src/synopsys/std_logic_arith.vhdl:315:20:#350ns:(assertion warning): There is an 'U'|'X'|'W'|'Z'|'-' in an arithmetic operand, the result will be 'X'(es).
../../../src/synopsys/std_logic_arith.vhdl:315:20:#550ns:(assertion warning): There is an 'U'|'X'|'W'|'Z'|'-' in an arithmetic operand, the result will be 'X'(es).
The signals displayed in the waveform are selected in order of importance and appearance a first pass selection and we immediately see we also get 'U's on free_park_counter as well as ERROR.
ERROR catches attention because you hadn't mentioned it previously. When asking 'where to the 'U' come from ?' it becomes apparent the issue is there are drivers on ERROR and free_park_counter in both processes p3 and max. The messages are a side effect.
Each process assigning a signal provides a driver. Signals with multiple drivers are either resolved or result in an error for non-resolved types.
The resolved value of free_park_counter with one or more elements having a metavalue will cause the diagnostic messages produced by package std_logic_arith. The 'U's in the waveform are caused by the resolution of the two drivers.
The difficulty your audience had in noticing the two drivers may be in part due to your strong insistence on focusing on process p3, which is not well specified. The title and focus of your question also seems a bit unclear. Without a Minimal Complete and Verifiable example there was also bound to be less scrutiny.
You might expect as a minimum to consolidate all the assignments to ERROR and free_park_counter into a single process. ERROR should likely be registered, and I'd expect something named park_counter would likely want to be registered, too.
There is some confusion in the question title : declaring a signal and setting its value are entirely separate.
Initialising a signal (in the declaration) will influence its value, but not fully determine it. If the initialisation and another driving value are different, the result probably will be 'X'. Likewise if the signal is driven from different processes which disagree on its value.
Now, you are using a multiple-process form of state machine, where the operations are split between clocked and combinational processes. These are recommended by more than one textbook. This is unfortunate because they are notoriously difficult to get right, and for example, a moment's inspection will show that the sensitivity list on process P3 is wrong.
Fixing P3's sensitivity list may not affect the problem, because P3 also drives its own inputs in what is known as a combinational loop. Consider that, if the process wakes up several times because of glitches on the combinational inputs in its sensitivity list, the additions will take place several times...
Rewriting these three processes in the form of a single clocked process P1, (which is, unfortunately, not well taught in several textbooks) will avoid all of these difficulties.
In ISim, if you browse the tree menu on the left you are able to add to then signals window any internal signal you want. Add all of them, rerun the simulation and look for the signals that have 'U'|'X'|'W'|'Z'|'-' values. This should give us a lead to track down the problem.
If you are really new to VHDL, this answer of mine should help you undersand some of the basic concepts of this description language :)
VHDL - iSIM output uninitialised, doesn't change states
Another advice that I learned the hard way, but you can think about it after we solved this problem: textbooks and even Xilinx describe how to implement finite state machines with two or even three distinct processes. This comes from an educational approach where FSM are splitted in synchronous logic and asynchronous logic. In practice, this is doing more harm than good: most of the FSM can be described with a single synchronous process. Google it (or if you are interested we can talk about it) and try it, you will get the hang of it really quickly and it will really simplify the code (you won't even need two separate signals for the states anymore!).
New to VHDL and trying to implement a small cache.
Part of my cache.vhd
entity cache is
Port ( clock : in STD_LOGIC;
rw_sel : in STD_LOGIC; --'1' to read from cache ,
--'0' to write to cache
ad_sel: in STD_LOGIC; --'0' to send address, '1' to send data
data_in :in STD_LOGIC_VECTOR (7 downto 0);
--send data or address controled by ad_sel
...
end cache;
architecture Behavioral of cache is
....
signal block_addr_n : integer range 0 to 15;
-- cache block address
signal block_cell_addr_n: integer range 0 to 1;
-- byte-in-cache address
begin
process(clock, init)
begin
case init is
when '0' =>
for i in 0 to 15 loop
cache_memory(i)<="1111111111111111";
tag_memory(i)<="11";
end loop;
when others =>null;
end case;
case ad_sel is
when '0' =>
address<=data_in(6 downto 0);
when '1' => data_cpu_wr<=data_in;
when others =>null;
end case;
block_addr_n <= conv_integer(unsigned(address(4 downto 1)));
block_cell_addr_n<=conv_integer(address(0));
case rw_sel is
....
part of my testbench file :
....
--Inputs
signal clock : std_logic := '0';
signal rw_sel : std_logic := '1';
signal ad_sel : std_logic := '0';
signal init: std_logic:='1';
signal data_in : std_logic_vector(7 downto 0) := (others => '0');
....
stim_proc: process
begin
-- initialize
init<='0';
wait for 100 ns;
-- read address "0101111"(MSB cut), expecting miss
cache_mem_data<="1100110011001100";
ad_sel<='0';
rw_sel<='1';
data_in<="00101111";
clock <= not clock;
wait for 100 ns;
-- write to address "0101111" but written in "0000000" which is default
data_in<="10101010";
ad_sel<='1';
rw_sel<='0';
clock<=not clock;
wait for 100 ns;
data_in<="00101111";
ad_sel<='0';
rw_sel<='1';
clock<=not clock;
wait;
end process;
END;
And what I got in the ISim window is
Why doesn't the block_addr_n and block_cell_addr_n be assigned in the second 100 ns since the signal address be assigned at that time directly? I think this is the reason that causes my program into unexpected results, since the cache block 0 is written while I just set address 0 as default but pass in the address "0101111" after the first 100 ns.
Any ideas?
The sensitivity list does not contain all the signals that are used in the process, since the address is a signal (assigned in the same process, but that does not matter) but the signal is not in the sensitivity list of the process. So the process is not reexecuted by the simulator when the value of address changes. Process parts:
process(clock, init)
begin
...
address<=data_in(6 downto 0);
...
block_addr_n <= conv_integer(unsigned(address(4 downto 1)));
block_cell_addr_n<=conv_integer(address(0));
If you are using VHDL-2008, then you can make the sensitivity list automatically with process (all), but I don't think Xilinx has implemented VHDL-2008 yet, so you must probably make the sensitivity list yourself by manually including all used signals in the sensitivity list.
Btw, consider making the processes either purely combinatorial (gates), or clocked (flip-flops or RAMs), so that combinatorial process should not have for example clock in the sensitivity list.
I think that your actual problem might be a misunderstanding about how to write a clocked process.
The way your process is written now it will generate purely combinational logic and latches.
What you want here is registers, RAM and combinational logic.
A clocked process would be written like this:
process(clock, init) --A clocked process shall have only the clock and reset signal in the sensitivity list. This is correct.
begin
if init = '0' then --"init" is used as a reset signal
for i in 0 to 15 loop
--This will reset the cache memory. It works, but
--doing it this way prevents the synthesis tool from infering a RAM block.
--To allow RAM inference you can write only one memory location per clock cycle.
cache_memory(i)<="1111111111111111";
tag_memory(i)<="11";
end loop;
elsif rising_edge( clock ) then
case ad_sel is
when '0' =>
address<=data_in(6 downto 0);
when '1' => data_cpu_wr<=data_in;
when others =>null;
end case;
block_addr_n <= conv_integer(unsigned(address(4 downto 1)));
block_cell_addr_n<=conv_integer(address(0));
case rw_sel is
...
end if;
Note that this will not work straight away. There are issues in your testbench as well. For example, your init signal is low indefinitely.
You may also want to use "if" statements rather than "case" for simple conditions. Example:
if ad_sel='0' then
address<=data_in(6 downto 0);
else
data_cpu_wr<=data_in;
end if;
I write simple vhdl code for Uart receiver.
Simulation (iSIM) is fine but when implemented I have wrong reading behavior.
When synthetized ISE tell me there are latches on state machine end on data_fill(x).
have you any suggestion.
thanks in advance
gian
here the code
library ieee;
use IEEE.STD_LOGIC_1164.all;
use IEEE.STD_LOGIC_UNSIGNED.all;
entity rx_uart is
port (
clk : in STD_LOGIC;
rst : in STD_LOGIC;
rx_data : in STD_LOGIC;
data_out: out STD_LOGIC_VECTOR(7 downto 0)
);
end rx_uart;
architecture fizzim of rx_uart is
-- state bits
subtype state_type is STD_LOGIC_VECTOR(2 downto 0);
constant idle: state_type:="000"; -- receive_en=0 load_en=0 cnt_en=0
constant receive: state_type:="101"; -- receive_en=1 load_en=0 cnt_en=1
constant stop_load: state_type:="010"; -- receive_en=0 load_en=1 cnt_en=0
signal state,nextstate: state_type;
signal cnt_en_internal: STD_LOGIC;
signal load_en_internal: STD_LOGIC;
signal receive_en_internal: STD_LOGIC;
signal count : integer range 0 to 54686:=0;
signal cnt_en : STD_LOGIC;
signal load_en : STD_LOGIC;
signal receive_en : STD_LOGIC;
signal data_fill : STD_LOGIC_VECTOR(9 downto 0);
-- comb always block
begin
COUNTER_EN : process(clk,rst,cnt_en) begin
if (rst ='1') then
count <= 0;
elsif rising_edge(clk) then
if (cnt_en ='1') then
count <= count+1;
else
count <= 0;
end if;
end if;
end process;
LOADER: process(clk,rst,load_en) begin
if (rst='1') then
data_out <= (others =>'0');
elsif (rising_edge(clk) and load_en='1')then
data_out <= data_fill(8 downto 1);
end if;
end process;
ASSIGNATION : process(clk,rst,receive_en) begin
if (rst ='1') then
data_fill <= (others =>'0');
elsif (receive_en='1') then
case count is
when 7812 =>
data_fill(1) <= rx_data;
when 13020 =>
data_fill(2) <= rx_data;
when 18228 =>
data_fill(3) <= rx_data;
when 23436 =>
data_fill(4) <= rx_data;
when 28664 =>
data_fill(5) <= rx_data;
when 33852 =>
data_fill(6) <= rx_data;
when 39060 =>
data_fill(7) <= rx_data;
when 44268 =>
data_fill(8) <= rx_data;
when 49476 =>
data_fill(9) <= rx_data;
when others =>
data_fill(0) <= '0';
end case;
end if;
end process;
COMB: process(state,clk,count,rst,rx_data) begin
case state is
when idle =>
if (rx_data='0') then
nextstate <= receive;
elsif (rx_data='1') then
nextstate <= idle;
end if;
when receive =>
if (count<=54685) then
nextstate <= receive;
elsif (count>54685) then
nextstate <= stop_load;
end if;
when stop_load =>
nextstate <= idle;
when others =>
end case;
end process;
-- Assign reg'd outputs to state bits
cnt_en_internal <= state(0);
load_en_internal <= state(1);
receive_en_internal <= state(2);
-- Port renames for vhdl
cnt_en <= cnt_en_internal;
load_en <= load_en_internal;
receive_en <= receive_en_internal;
-- sequential always block
FF: process(clk,rst,nextstate) begin
if (rst='1') then
state <= idle;
elsif (rising_edge(clk)) then
state <= nextstate;
end if;
end process;
end fizzim;
You need to understand when latches are generated. Latches are generated when you have an incomplete assignment in combinational logic. Your case statements are combinational. You do not completely assign all possibilities of your data signal and your state machine. When you have an incomplete assignment in a combinational process you will always generate a latch. Latches are bad!
Use your when others properly to assign all of your signals under all conditions. Your data_fill signal will always generate a latch because you don't handle all conditions for data 0:9 on all cases.
Read more about how to avoid latches in VHDL
Edit: It seems too that you don't consistently create sequential logic in VHDL. You need to be creating a clocked process or remove clk from your sensitivity list in a combinational process.
your design has several bad code sites besides your latch problem. I'll number them for better reference.
(1)
Xilinx XST won't recognize the implemented state machine as a FSM. See your XST report or the *.syr file. There won't be a FSM section. If XST finds no FSM it's not able to choose "the best" state encoding and optimize your FSM.
You should use an enum as a state-type and also initialize your state signal:
type t_state is (st_idle, st_receive, st_stop);
signal state : t_state := st_idle;
signal nextstate : t_state;
Your FSM process also needs default assignments (see Russell's explanation) like
nextstate <= state;
(2)
asynchronous resets are no good design practice and complicate timing closure calculations.
(3)
you are handling raw input signals from outside you FPGA without any timing information. to prevent meta stability problems place two D-flip-flops on the rx_data path. you should also add the async-reg and no-srl-extract attribute to these 2 registers to prevent XST optimizations
SIGNAL I_async : STD_LOGIC := '0';
SIGNAL I_sync : STD_LOGIC := '0';
-- Mark register "I_async" as asynchronous
ATTRIBUTE ASYNC_REG OF I_async : SIGNAL IS "TRUE";
-- Prevent XST from translating two FFs into SRL plus FF
ATTRIBUTE SHREG_EXTRACT OF I_async : SIGNAL IS "NO";
ATTRIBUTE SHREG_EXTRACT OF I_sync : SIGNAL IS "NO";
[...]
BEGIN
[...]
I_async <= In WHEN rising_edge(Clock);
I_sync <= I_async WHEN rising_edge(Clock);
Let 'In' be your asynchronous input. Now you can use I_sync in your clock domain. I choose to describe these 2 registers in a one-liner, you can also use a classic process :) The suffix '_async' allows you to define a matching timing rule in your *.xcf and *.ucf files.