VGA RLT代码封装成AXI Stream

kenson · 发表于 2019-4-12 17:48

本帖最后由 kenson 于 2019-4-12 17:59 编辑

本编文章将对VGA的RTL代码，封装成AXI Stream,并且在vivado 里用TPG进行测试

本篇文章的VGA RTL代码在【ZYNQ-7000开发之一】基础上修改，封装好的VGA Stream可以方便我们实现视频图像处理

TGP 等的规范说明可以到官网下载最新版本

本文所使用的开发板是Miz702(兼容zedboard)
7 C7 y! |; n1 I i+ y" J9 iPC 开发环境版本：Vivado 2015.2 Xilinx SDK 2015.2 " F* y X8 X: f; Y0 f
其它：VGA显示器

AXI Stream原理

首先这里列出axi stream的信号，红框里的是要用到的
0 J: @: x9 b+ u1 J& W- e1 ^5 L2 ` VGA RLT代码封装成AXI Stream-1.jpg
: O* g4 `4 l% v* Y) {* y" M+ c9 }2 s7 n核心的信号是，TVALID,TREADY,TLAST,TUSER

TVALID和TREADY握手信号

VGA RLT代码封装成AXI Stream-2.jpg & |2 f: I) w/ Z5 |
在TVALID和TREADY同时有效的时候，数据才有效。对于TPG来说，大部分时间TVALID是有效的，TREADY由我们自己控制。 $ p7 T/ S7 j8 P; D# }2 ^
TUSER(SOF，Start Of Frame),代表每一帧的开始。TLAST(EOL,End Of Line),代表每一行的结束。
v9 s J' S1 a6 s5 i' s- ?# f$ P这里给出TPG的时序图，比较简洁，要理解之后才好把RTL封装成AXI Stream 5 P7 i6 Y) J f
VGA RLT代码封装成AXI Stream-3.jpg
5 \8 p6 ?0 ~' K8 f4 i" Q0 [ VGA RLT代码封装成AXI Stream-4.jpg

封装AXI Stream打开vivado，建立一个工程（选择zed）选择Tools->Create and Package IP

VGA RLT代码封装成AXI Stream-5.jpg

点击NEXT,选择Create AXI4,如图所示

VGA RLT代码封装成AXI Stream-6.jpg

输入好name和路径后，点击NEXT

VGA RLT代码封装成AXI Stream-7.jpg

按照如下配置

VGA RLT代码封装成AXI Stream-8.jpg

把VGA_AXI_Stream_v1_0.v文件里的代码，修改如下`timescale 1 ns / 1 ps

module myip_v1_0 #
(
      // Users to add parameters here

      // User parameters ends
      // Do not modify the parameters beyond this line

      // Parameters of Axi Slave Bus Interface S00_AXIS_VGA
      parameter integer C_S00_AXIS_VGA_TDATA_WIDTH = 32
)
(
      // Users to add ports here
      input wire          clk_25mhz,
      output  wire          hsync,
      output  wire          vsync,
      output  wire [11:0] rgb,
      output  reg          LED,
      // User ports ends
      // Do not modify the ports beyond this line

      // Ports of Axi Slave Bus Interface S00_AXIS_VGA
      input wire  s00_axis_vga_aclk,
      input wire  s00_axis_vga_aresetn,
      output wire  s00_axis_vga_tready,
      input wire [C_S00_AXIS_VGA_TDATA_WIDTH-1 : 0] s00_axis_vga_tdata,
      input wire [(C_S00_AXIS_VGA_TDATA_WIDTH/8)-1 : 0] s00_axis_vga_tstrb,
      input wire  s00_axis_vga_tlast,
      input wire  s00_axis_vga_tvalid,
      input wire  s00_axis_vga_tuser
);
// Instantiation of Axi Bus Interface S00_AXIS_VGA
myip_v1_0_S00_AXIS_VGA # (
      .C_S_AXIS_TDATA_WIDTH(C_S00_AXIS_VGA_TDATA_WIDTH)
) myip_v1_0_S00_AXIS_VGA_inst (
      .S_AXIS_ACLK(s00_axis_vga_aclk),
      .S_AXIS_ARESETN(s00_axis_vga_aresetn),
      .S_AXIS_TREADY(s00_axis_vga_tready),
      .S_AXIS_TDATA(s00_axis_vga_tdata),
      .S_AXIS_TSTRB(s00_axis_vga_tstrb),
      .S_AXIS_TLAST(s00_axis_vga_tlast),
      .S_AXIS_TVALID(s00_axis_vga_tvalid),
      .S_AXIS_TUSER(s00_axis_vga_tuser),
      .clk          (clk_25mhz),
      .rst_n       (s00_axis_vga_aresetn),
      .video_en    (video_en),
      .hsync       (hsync),
      .vsync       (vsync),
      .pixel_x       (pixel_x),
      .pixel_y       (pixel_y)
);

// Add user logic here

wire [9:0] pixel_x;
wire [9:0] pixel_y;
wire          clk_25mhz;
reg    [11:0]  rgb_reg;

//显示静态图像640*480
reg       [23:0] cnt;
always @(posedge clk_25mhz or negedge s00_axis_vga_aresetn)
      if(!s00_axis_vga_aresetn)
         begin
            cnt <= 0;
            led <= 0;
         end
      else
         begin
            cnt <= cnt + 1'b1;
            if(cnt == 24'd12500000)
                  begin
                     cnt <= 24'b0;
                     led <= ~led;
                  end
         end
always @ (posedge clk_25mhz or negedge s00_axis_vga_aresetn)
      if(!s00_axis_vga_aresetn)
         begin
            rgb_reg <= 12'b0;
         end
      else if(s00_axis_vga_tvalid == 1'b1 && s00_axis_vga_tready == 1'b1)
         begin             //显示图像
            rgb_reg[3:0] <= s00_axis_vga_tdata[7:4];
            rgb_reg[7:4] <= s00_axis_vga_tdata[15:12];
            rgb_reg[11:8] <= s00_axis_vga_tdata[23:20];
         end
/*
always @ (posedge clk_25mhz or negedge s00_axis_vga_aresetn)
      if(!s00_axis_vga_aresetn)
         begin
            address_sig <= 19'b0;
         end
      else
         begin
            if(pixel_x>=0 && pixel_x<= 639 && pixel_y>=0 && pixel_y<=479)
                  address_sig = (pixel_x + 640*pixel_y);
         end
*/
//////////////////////////////////////////////////////////////
assign rgb = (video_en == 1'b1) ? rgb_reg:12'b0;

// User logic ends
  endmodule
把VGA_AXI_Stream_v1_0_S00_AXI_VgA.v文件里的代码，修改如下

// Users to add ports here
      input wire          clk,
      input wire          rst_n,
      output  wire          video_en,
      output  reg          hsync,
      output  reg          vsync,
      output  wire [9:0] pixel_x,
      output  wire [9:0] pixel_y,
      // User ports ends
      // Do not modify the ports beyond this line

      // AXI4Stream sink: Clock
      input wire  S_AXIS_ACLK,
      // AXI4Stream sink: Reset
      input wire  S_AXIS_ARESETN,
      // Ready to accept data in
      output wire  S_AXIS_TREADY,
      // Data in
      input wire [C_S_AXIS_TDATA_WIDTH-1 : 0] S_AXIS_TDATA,
      // Byte qualifier
      input wire [(C_S_AXIS_TDATA_WIDTH/8)-1 : 0] S_AXIS_TSTRB,
      // Indicates boundary of last packet
      input wire  S_AXIS_TLAST,
      // Data is in valid
      input wire  S_AXIS_TVALID,
      input wire S_AXIS_TUSER
);
// function called clogb2 that returns an integer which has the
// value of the ceiling of the log base 2.
function integer clogb2 (input integer bit_depth);
   begin
      for(clogb2=0; bit_depth>0; clogb2=clogb2+1)
      bit_depth = bit_depth >> 1;
   end
endfunction

// Total number of input data.
localparam NUMBER_OF_INPUT_WORDS  = 8;
// bit_num gives the minimum number of bits needed to address 'NUMBER_OF_INPUT_WORDS' size of FIFO.
localparam bit_num  = clogb2(NUMBER_OF_INPUT_WORDS-1);
// Define the states of state machine
// The control state machine oversees the writing of input streaming data to the FIFO,
// and outputs the streaming data from the FIFO
parameter [1:0] IDLE = 1'b0,       // This is the initial/idle state

                  WRITE_FIFO  = 1'b1; // In this state FIFO is written with the
                                    // input stream data S_AXIS_TDATA
wire axis_tready;
// State variable
reg mst_exec_state;
// FIFO implementation signals
genvar byte_index;
// FIFO write enable
wire fifo_wren;
// FIFO full flag
reg fifo_full_flag;
// FIFO write pointer
reg [bit_num-1:0] write_pointer;
// sink has accepted all the streaming data and stored in FIFO
   reg writes_done;
// I/O Connections assignments

assign S_AXIS_TREADY = axis_tready;
// Control state machine implementation
always @(posedge S_AXIS_ACLK)
begin
   if (!S_AXIS_ARESETN)
   // Synchronous reset (active low)
      begin
      mst_exec_state <= IDLE;
      end
   else
      case (mst_exec_state)
      IDLE:
         // The sink starts accepting tdata when
         // there tvalid is asserted to mark the
         // presence of valid streaming data
            if (S_AXIS_TVALID)
            begin
               mst_exec_state <= WRITE_FIFO;
            end
            else
            begin
               mst_exec_state <= IDLE;
            end
      WRITE_FIFO:
         // When the sink has accepted all the streaming input data,
         // the interface swiches functionality to a streaming master
         if (writes_done)
            begin
            mst_exec_state <= IDLE;
            end
         else
            begin
            // The sink accepts and stores tdata
            // into FIFO
            mst_exec_state <= WRITE_FIFO;
            end

      endcase
end
// AXI Streaming Sink
//
// The example design sink is always ready to accept the S_AXIS_TDATA  until
// the FIFO is not filled with NUMBER_OF_INPUT_WORDS number of input words.
//assign axis_tready = ((mst_exec_state == WRITE_FIFO) && (write_pointer <= NUMBER_OF_INPUT_WORDS-1)

assign axis_tready =( (pixel_x>=0 && pixel_x<= 639 && pixel_y>=0 && pixel_y<=479) );

always@(posedge S_AXIS_ACLK)
begin
   if(!S_AXIS_ARESETN)
      begin
      write_pointer <= 0;
      writes_done <= 1'b0;
      end
   else
      if (write_pointer <= NUMBER_OF_INPUT_WORDS-1)
      begin
         if (fifo_wren)
            begin
            // write pointer is incremented after every write to the FIFO
            // when FIFO write signal is enabled.
            write_pointer <= write_pointer + 1;
            writes_done <= 1'b0;
            end
            if ((write_pointer == NUMBER_OF_INPUT_WORDS-1)|| S_AXIS_TLAST)
            begin
               // reads_done is asserted when NUMBER_OF_INPUT_WORDS numbers of streaming data
               // has been written to the FIFO which is also marked by S_AXIS_TLAST(kept for optional usage).
               writes_done <= 1'b1;
            end
      end
end

// FIFO write enable generation
assign fifo_wren = S_AXIS_TVALID && axis_tready;

// FIFO Implementation
generate
   for(byte_index=0; byte_index<= (C_S_AXIS_TDATA_WIDTH/8-1); byte_index=byte_index+1)
   begin:FIFO_GEN

      reg  [(C_S_AXIS_TDATA_WIDTH/4)-1:0] stream_data_fifo [0 : NUMBER_OF_INPUT_WORDS-1];

      // Streaming input data is stored in FIFO

      always @( posedge S_AXIS_ACLK )
      begin
      if (fifo_wren)// && S_AXIS_TSTRB[byte_index])
         begin
            stream_data_fifo[write_pointer] <= S_AXIS_TDATA[(byte_index*8+7) -: 8];
         end
      end
   end
endgenerate

// Add user logic here

reg    [9:0] pixel_cnt;
reg    [9:0] line_cnt;
reg          v_video_en;
reg          h_video_en;

always @(posedge clk or negedge rst_n)
   if(!rst_n)
         begin
            pixel_cnt <= 10'b0;
         end
   else
         begin
            pixel_cnt <= pixel_cnt + 1'b1;
                  if(pixel_cnt == 10'd799 || S_AXIS_TUSER == 1'b1)
                  begin
                        pixel_cnt <= 10'b0;
                  end
         end

always @(posedge clk or negedge rst_n)
   if(!rst_n)
         begin
            h_video_en <= 1'b1;
                     hsync <= 1'b1;
         end
   else
         begin
            case (pixel_cnt)
               10'd0:
                     begin
                        h_video_en <= 1'b1;
                     end
               10'd639:
                     begin
                        h_video_en <= 1'b0;
                     end
               10'd655:
                     begin
                        hsync <= 1'b0;
                     end
               10'd751:
                     begin
                        hsync <= 1'b1;
                     end
            endcase
         end

always @(posedge clk or negedge rst_n)
   if(!rst_n)
         begin
            line_cnt <= 10'b0;
         end
   else
         begin
            if(pixel_cnt <= S_AXIS_TLAST) //pixel_cnt == 10'd799
               begin
                     line_cnt <= line_cnt + 1'b1;
               end

               if(line_cnt == 10'd524)
               begin
                     line_cnt <= 10'b0;
               end
               if(S_AXIS_TUSER == 1'b1)
               begin
                  line_cnt <= 10'b0;
               end
         end

always @(posedge clk or negedge rst_n)
   if(!rst_n)
         begin
            v_video_en <= 1'b1;
            vsync <= 1'b1;
         end
   else
         begin
            case(line_cnt)
               10'd0:
                     begin
                        v_video_en <= 1'b1;
                     end
               10'd479:
                     begin
                        v_video_en <= 1'b0;
                     end
               10'd489:
                     begin
                        vsync <= 1'b0;
                     end
               10'd491:
                     begin
                        vsync <= 1'b1;
                     end
            endcase
         end

assign pixel_x = pixel_cnt;
assign pixel_y = line_cnt;

assign video_en = ((h_video_en == 1'b1) &&  (v_video_en == 1'b1));
// User logic ends

endmodule

kenson · 发表于 2019-4-12 18:01

https://blog.csdn.net/rzjmpb/article/details/50609314

		自动登录	找回密码
密码			请使用微信账号登录和注册会员

VGA RLT代码封装成AXI Stream

浏览过的版块