原创 同步FIFO之VHDL描述4

       同步FIFO之VHDL描述4

       今天准备完成同步FIFO其他模块的代码，主要包括读指针管理和产生电路，FIFO状态判断电路，以及双端口RAM的VHDL描述。先是读指针管理和产生电路：

-----------------------------------------------------------------------------------------------------------
-- Designer  : skycanny
-- Date   : 2007-2-3
-- Description : read_pointer is created

library ieee;
use ieee.std_logic_1164.all;
use ieee.std_logic_unsigned.all;
use ieee.std_logic_arith.all;

entity read_pointer is
 generic
 (
  depth : positive 
 );
 port
 (
  clk  : in std_logic;
  rst  : in std_logic;
  rq  : in std_logic;
  empty : in std_logic;
  rd_pt : out std_logic_vector(depth - 1 downto 0)  
 );
end entity read_pointer;

architecture RTL of read_pointer is
signal rd_pt_t : std_logic_vector(depth - 1 downto 0);  -- read pointer counter

begin
 process(rst, clk)
 begin
  if rst = '0' then
   rd_pt_t <= (others => '0');
  elsif clk'event and clk = '1' then
   if rq = '0' and empty = '0' then
    rd_pt_t <= rd_pt_t + 1;
   end if;
 end process;
 rd_pt <= rd_pt_t;
end RTL;
------------------------------------------------------------------------------------------------------------

<?xml:namespace prefix = o ns = "urn:schemas-microsoft-com:office:office" />

刚才想了一下，读写指针产生电路必须要对FIFO的状态进行判断，所以又对上面的代码进行了一点修改（上面的代码是修改过的），判断FIFO的状态，这一点在刚开始的时候给疏忽了，幸好刚才给发现了。同样昨天写的写指针产生电路没有判断FIFO的full状态，也要进行修改，还是在昨天的基础上修改吧，就不在这里罗嗦了，下面就是FIFO状态判断电路的VHDL描述。

-----------------------------------------------------------------------------------------------------------

-- Designer            :      skycanny

-- Date                  :      2007-2-3

-- Description :      read_pointer is created

library ieee;

use ieee.std_logic_1164.all;

use ieee.std_logic_unsigned.all;

use ieee.std_logic_arith.all;

entity judge_status is

       generic

       (

              depth      :      positive

       );

       port

       (

              clk          :      in     std_logic;

              rst          :      in     std_logic;

              wr_pt      :      in     std_logic_vector(depth - 1 downto 0);

              rd_pt       :      in     std_logic_vector(depth - 1 downto 0);

              empty     :      out   std_logic;

              full   :      out   std_logic

       );

end entity judge_status;

architecture RTL of judge_status is

begin

       process(rst, clk)

       begin

              if rst = '0 then

                     empty <= '1';

              elsif clk'event and clk = '1' then

                     if wr_pt = rd_pt then

                            empty <= '1';

                     else

                            empty <= '0';

                     end if;

              end if;

       end process;

       process(rst, clk)

       begin

              if rst = '0' then

                     full <= '0';

              elsif clk 'event and clk = '1' then

                     if wr_pt > rd_pt then

                            if (rd_pt + depth) = wr_pt then

                                   full <= '1';

                            else

                                   full <= '0';

                            end if;

                     else

                            if (wr_pt + 1 ) = rd_pt then

                                   full <= '1';

                            else

                                   full <= '0';

                            end if;

                     end if;

              end if;

       end process;

end RTL;

-----------------------------------------------------------------------------------------------------------

    广告时间欢迎访问skycanny的笔记(副站) 。现在就开始双口RAM的VHDL描述吧，这个可是很重要的。下面的代码在Altera的CycloneII系列的器件上通过了后仿真。

-----------------------------------------------------------------------------------------------------------

-- Designer            :      skycanny

-- Date                  :      2007-2-3

-- Description :      This VHDL file is designed to generate a dual port ram   

-- Device        :      Cyclone II

library ieee;

use ieee.std_logic_1164.all;

use ieee.std_logic_unsigned.all;

use ieee.std_logic_arith.all;

entity dualram is

       generic

       (

              width      :      positive   := 16;

              depth      :      positive   := 8

       );

       port

       (

              ------------------------- port a is only for writing -------------------------------

              clka :      in     std_logic;

              wr          :      in     std_logic;

              addra      :      in     std_logic_vector(depth - 1 downto 0);

              datain      :      in     std_logic_vector(width - 1 downto 0);

              ------------------------------------------------------------------------------------

              ------------------------- port b is only for reading -------------------------------

              clkb :      in     std_logic;

              rd           :      in     std_logic;

              addrb      :      in     std_logic_vector(depth - 1 downto 0);

              dataout    :      out   std_logic_vector(width - 1 downto 0)         

              ------------------------------------------------------------------------------------

       );

end entity dualram;

architecture Behavioral of dualram is

type ram is array(2 ** depth - 1 downto 0) of std_logic_vector(width - 1 downto 0);

signal      dualram          :      ram;

begin

       process(clka, clkb)

       begin

              if clka'event and clka = '1' then

                     if wr = '0' then

                            dualram(conv_integer(addra)) <= datain;

                     end if;

              end if;

       end process;

       process(clkb)

       begin

              if clkb'event and clkb = '1' then

                     if rd = '0' then

                            dataout <= dualram(conv_integer(addrb));

                     end if;

              end if;

       end process;

end Behavioral;

-----------------------------------------------------------------------------------------------------------

抓个后仿真的图形吧，需要testbench的投票留言。另外今天还发现前仿真的时候可以用generic传递参数，而后仿真的时候就不能用generic传递参数了，Modelsim会报错“No default binding for component at”。后来想了一下确实应该是这样的，因为后仿真的时候电路都已经布局布线完成了，还有什么参数需要generic传递呀？

今天就到这里吧，大家赶快给我投票啊，吼吼！欢迎访问skycanny的笔记(副站)。