FPGA/corundum
1
0
Fork 0
mirror of https://github.com/corundum/corundum.git synced 2025-01-16 08:12:53 +08:00
Code Issues Projects Releases Wiki Activity
corundum/fpga/common/rtl/tx_checksum.v
Alex Forencich fb0f6f67f7 Remove debug code
2021-11-06 16:14:32 -07:00

575 lines
19 KiB
Verilog
Raw Blame History

/*
Copyright 2019, The Regents of the University of California.
All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:
1. Redistributions of source code must retain the above copyright notice,
this list of conditions and the following disclaimer.
2. Redistributions in binary form must reproduce the above copyright notice,
this list of conditions and the following disclaimer in the documentation
and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE REGENTS OF THE UNIVERSITY OF CALIFORNIA ''AS
IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
DISCLAIMED. IN NO EVENT SHALL THE REGENTS OF THE UNIVERSITY OF CALIFORNIA OR
CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT
OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY
OF SUCH DAMAGE.
The views and conclusions contained in the software and documentation are those
of the authors and should not be interpreted as representing official policies,
either expressed or implied, of The Regents of the University of California.
*/
// Language: Verilog 2001
`resetall
`timescale 1ns / 1ps
`default_nettype none
/*
* Transmit checksum offload module
*/
module tx_checksum #
(
// Width of AXI stream interfaces in bits
parameter DATA_WIDTH = 256,
// AXI stream tkeep signal width (words per cycle)
parameter KEEP_WIDTH = (DATA_WIDTH/8),
// Propagate tid signal
parameter ID_ENABLE = 0,
// tid signal width
parameter ID_WIDTH = 8,
// Propagate tdest signal
parameter DEST_ENABLE = 0,
// tdest signal width
parameter DEST_WIDTH = 8,
// Propagate tuser signal
parameter USER_ENABLE = 1,
// tuser signal width
parameter USER_WIDTH = 1,
// Use checksum init value
parameter USE_INIT_VALUE = 0,
// Depth of data FIFO in words
parameter DATA_FIFO_DEPTH = 4096,
// Depth of checksum FIFO
parameter CHECKSUM_FIFO_DEPTH = 64
)
(
input wire clk,
input wire rst,
/*
* AXI input
*/
input wire [DATA_WIDTH-1:0] s_axis_tdata,
input wire [KEEP_WIDTH-1:0] s_axis_tkeep,
input wire s_axis_tvalid,
output wire s_axis_tready,
input wire s_axis_tlast,
input wire [ID_WIDTH-1:0] s_axis_tid,
input wire [DEST_WIDTH-1:0] s_axis_tdest,
input wire [USER_WIDTH-1:0] s_axis_tuser,
/*
* AXI output
*/
output wire [DATA_WIDTH-1:0] m_axis_tdata,
output wire [KEEP_WIDTH-1:0] m_axis_tkeep,
output wire m_axis_tvalid,
input wire m_axis_tready,
output wire m_axis_tlast,
output wire [ID_WIDTH-1:0] m_axis_tid,
output wire [DEST_WIDTH-1:0] m_axis_tdest,
output wire [USER_WIDTH-1:0] m_axis_tuser,
/*
* Control
*/
input wire s_axis_cmd_csum_enable,
input wire [7:0] s_axis_cmd_csum_start,
input wire [7:0] s_axis_cmd_csum_offset,
input wire [15:0] s_axis_cmd_csum_init,
input wire s_axis_cmd_valid,
output wire s_axis_cmd_ready
);
parameter LEVELS = $clog2(DATA_WIDTH/8);
// bus width assertions
initial begin
if (KEEP_WIDTH * 8 != DATA_WIDTH) begin
$error("Error: AXI stream interface requires byte (8-bit) granularity (instance %m)");
$finish;
end
end
reg transfer_in_reg = 1'b0;
reg [15:0] csum_in_csum_reg = 0;
reg [7:0] csum_in_offset_reg = 0;
reg csum_in_enable_reg = 1'b0;
reg csum_in_valid_reg = 1'b0;
wire csum_in_ready;
wire [15:0] csum_out_csum;
wire [7:0] csum_out_offset;
wire csum_out_enable;
wire csum_out_valid;
reg csum_out_ready = 1'b0;
reg [KEEP_WIDTH-1:0] mask_reg = 0;
reg first_cycle_reg = 1'b0;
reg [7:0] input_offset_reg = 0;
reg [DATA_WIDTH-1:0] s_axis_tdata_masked;
reg frame_reg = 1'b0, frame_next;
reg [15:0] csum_reg = 16'd0, csum_next;
reg [7:0] csum_offset_reg = 8'd0, csum_offset_next;
reg csum_enable_reg = 1'b0, csum_enable_next;
reg csum_split_reg = 1'b0, csum_split_next;
reg [DATA_WIDTH-1:0] sum_reg[LEVELS-2:0];
reg [LEVELS-2:0] sum_valid_reg = 0;
reg [LEVELS-2:0] sum_odd_reg = 0;
reg [LEVELS-2:0] sum_last_reg = 0;
reg [LEVELS-2:0] sum_enable_reg = 0;
reg [7:0] sum_offset_reg[LEVELS-2:0];
reg [15:0] sum_init_reg[LEVELS-2:0];
reg [LEVELS-2:0] sum_init_valid_reg = 0;
reg [16+LEVELS-1:0] sum_acc_temp = 0;
reg [15:0] sum_acc_reg = 0;
// internal datapath
reg [DATA_WIDTH-1:0] m_axis_tdata_int;
reg [KEEP_WIDTH-1:0] m_axis_tkeep_int;
reg m_axis_tvalid_int;
reg m_axis_tready_int_reg = 1'b0;
reg m_axis_tlast_int;
reg [ID_WIDTH-1:0] m_axis_tid_int;
reg [DEST_WIDTH-1:0] m_axis_tdest_int;
reg [USER_WIDTH-1:0] m_axis_tuser_int;
wire m_axis_tready_int_early;
wire [DATA_WIDTH-1:0] data_in_axis_tdata;
wire [KEEP_WIDTH-1:0] data_in_axis_tkeep;
wire data_in_axis_tvalid;
wire data_in_axis_tready;
wire data_in_axis_tlast;
wire [ID_WIDTH-1:0] data_in_axis_tid;
wire [DEST_WIDTH-1:0] data_in_axis_tdest;
wire [USER_WIDTH-1:0] data_in_axis_tuser;
wire [DATA_WIDTH-1:0] data_out_axis_tdata;
wire [KEEP_WIDTH-1:0] data_out_axis_tkeep;
wire data_out_axis_tvalid;
reg data_out_axis_tready;
wire data_out_axis_tlast;
wire [ID_WIDTH-1:0] data_out_axis_tid;
wire [DEST_WIDTH-1:0] data_out_axis_tdest;
wire [USER_WIDTH-1:0] data_out_axis_tuser;
assign s_axis_tready = data_in_axis_tready && csum_in_ready && transfer_in_reg;
assign s_axis_cmd_ready = csum_in_ready && !transfer_in_reg;
// data FIFO
assign data_in_axis_tdata = s_axis_tdata;
assign data_in_axis_tkeep = s_axis_tkeep;
assign data_in_axis_tvalid = s_axis_tvalid && csum_in_ready && transfer_in_reg;
assign data_in_axis_tlast = s_axis_tlast;
assign data_in_axis_tid = s_axis_tid;
assign data_in_axis_tdest = s_axis_tdest;
assign data_in_axis_tuser = s_axis_tuser;
axis_fifo #(
.DEPTH(DATA_FIFO_DEPTH),
.DATA_WIDTH(DATA_WIDTH),
.KEEP_ENABLE(1),
.KEEP_WIDTH(KEEP_WIDTH),
.LAST_ENABLE(1),
.ID_ENABLE(ID_ENABLE),
.ID_WIDTH(ID_WIDTH),
.DEST_ENABLE(DEST_ENABLE),
.DEST_WIDTH(DEST_WIDTH),
.USER_ENABLE(USER_ENABLE),
.USER_WIDTH(USER_WIDTH),
.FRAME_FIFO(0)
)
data_fifo (
.clk(clk),
.rst(rst),
// AXI input
.s_axis_tdata(data_in_axis_tdata),
.s_axis_tkeep(data_in_axis_tkeep),
.s_axis_tvalid(data_in_axis_tvalid),
.s_axis_tready(data_in_axis_tready),
.s_axis_tlast(data_in_axis_tlast),
.s_axis_tid(data_in_axis_tid),
.s_axis_tdest(data_in_axis_tdest),
.s_axis_tuser(data_in_axis_tuser),
// AXI output
.m_axis_tdata(data_out_axis_tdata),
.m_axis_tkeep(data_out_axis_tkeep),
.m_axis_tvalid(data_out_axis_tvalid),
.m_axis_tready(data_out_axis_tready),
.m_axis_tlast(data_out_axis_tlast),
.m_axis_tid(data_out_axis_tid),
.m_axis_tdest(data_out_axis_tdest),
.m_axis_tuser(data_out_axis_tuser),
// Status
.status_overflow(),
.status_bad_frame(),
.status_good_frame()
);
// checksum FIFO
axis_fifo #(
.DEPTH(CHECKSUM_FIFO_DEPTH),
.DATA_WIDTH(16+8+1),
.KEEP_ENABLE(0),
.LAST_ENABLE(0),
.ID_ENABLE(0),
.DEST_ENABLE(0),
.USER_ENABLE(0),
.FRAME_FIFO(0)
)
csum_fifo (
.clk(clk),
.rst(rst),
// AXI input
.s_axis_tdata({csum_in_csum_reg, csum_in_offset_reg, csum_in_enable_reg}),
.s_axis_tkeep(0),
.s_axis_tvalid(csum_in_valid_reg),
.s_axis_tready(csum_in_ready),
.s_axis_tlast(0),
.s_axis_tid(0),
.s_axis_tdest(0),
.s_axis_tuser(0),
// AXI output
.m_axis_tdata({csum_out_csum, csum_out_offset, csum_out_enable}),
.m_axis_tkeep(),
.m_axis_tvalid(csum_out_valid),
.m_axis_tready(csum_out_ready),
.m_axis_tlast(),
.m_axis_tid(),
.m_axis_tdest(),
.m_axis_tuser(),
// Status
.status_overflow(),
.status_bad_frame(),
.status_good_frame()
);
// Mask input data
integer j;
always @* begin
for (j = 0; j < KEEP_WIDTH; j = j + 1) begin
s_axis_tdata_masked[j*8 +: 8] = (s_axis_tkeep[j] && mask_reg[j]) ? s_axis_tdata[j*8 +: 8] : 8'd0;
end
end
// Compute checksum
integer i;
always @(posedge clk) begin
sum_valid_reg[0] <= sum_valid_reg[0] && !csum_in_ready;
if (s_axis_tvalid && s_axis_tready) begin
for (i = 0; i < DATA_WIDTH/8/4; i = i + 1) begin
sum_reg[0][i*17 +: 17] <= {s_axis_tdata_masked[(4*i+0)*8 +: 8], s_axis_tdata_masked[(4*i+1)*8 +: 8]} + {s_axis_tdata_masked[(4*i+2)*8 +: 8], s_axis_tdata_masked[(4*i+3)*8 +: 8]};
end
sum_valid_reg[0] <= 1'b1;
sum_last_reg[0] <= s_axis_tlast;
sum_init_valid_reg[0] <= first_cycle_reg;
first_cycle_reg <= 1'b0;
if (s_axis_tlast) begin
transfer_in_reg <= 1'b0;
end
if (input_offset_reg > 0) begin
if (input_offset_reg >= KEEP_WIDTH) begin
mask_reg <= 0;
input_offset_reg <= input_offset_reg - KEEP_WIDTH;
end else begin
mask_reg <= {KEEP_WIDTH{1'b1}} << input_offset_reg;
input_offset_reg <= 0;
end
end else begin
mask_reg <= {KEEP_WIDTH{1'b1}};
end
end
if (s_axis_cmd_valid && s_axis_cmd_ready) begin
transfer_in_reg <= 1'b1;
sum_odd_reg[0] <= s_axis_cmd_csum_start[0];
sum_enable_reg[0] <= s_axis_cmd_csum_enable;
sum_offset_reg[0] <= s_axis_cmd_csum_offset;
sum_init_reg[0] <= s_axis_cmd_csum_init;
first_cycle_reg <= 1'b1;
if (s_axis_cmd_csum_start >= KEEP_WIDTH) begin
mask_reg <= 0;
input_offset_reg <= s_axis_cmd_csum_start - KEEP_WIDTH;
end else begin
mask_reg <= {KEEP_WIDTH{1'b1}} << s_axis_cmd_csum_start;
input_offset_reg <= 0;
end
end
if (rst) begin
transfer_in_reg <= 1'b0;
sum_valid_reg[0] <= 1'b0;
end
end
generate
genvar l;
for (l = 1; l < LEVELS-1; l = l + 1) begin
always @(posedge clk) begin
sum_valid_reg[l] <= sum_valid_reg[l] && !csum_in_ready;
if (sum_valid_reg[l-1] && csum_in_ready) begin
for (i = 0; i < DATA_WIDTH/8/4/2**l; i = i + 1) begin
sum_reg[l][i*(17+l) +: (17+l)] <= sum_reg[l-1][(i*2+0)*(17+l-1) +: (17+l-1)] + sum_reg[l-1][(i*2+1)*(17+l-1) +: (17+l-1)];
end
sum_valid_reg[l] <= 1'b1;
sum_odd_reg[l] <= sum_odd_reg[l-1];
sum_last_reg[l] <= sum_last_reg[l-1];
sum_enable_reg[l] <= sum_enable_reg[l-1];
sum_offset_reg[l] <= sum_offset_reg[l-1];
sum_init_reg[l] <= sum_init_reg[l-1];
sum_init_valid_reg[l] <= sum_init_valid_reg[l-1];
end
if (rst) begin
sum_valid_reg[l] <= 1'b0;
end
end
end
endgenerate
always @(posedge clk) begin
csum_in_valid_reg <= 1'b0;
if (sum_valid_reg[LEVELS-2] && csum_in_ready) begin
sum_acc_temp = sum_reg[LEVELS-2][16+LEVELS-1-1:0] + (sum_init_valid_reg[LEVELS-2] && USE_INIT_VALUE ? sum_init_reg[LEVELS-2] : sum_acc_reg);
sum_acc_temp = sum_acc_temp[15:0] + (sum_acc_temp >> 16);
sum_acc_temp = sum_acc_temp[15:0] + sum_acc_temp[16];
if (sum_last_reg[LEVELS-2]) begin
if (sum_odd_reg[LEVELS-2]) begin
csum_in_csum_reg[7:0] <= ~sum_acc_temp[15:8];
csum_in_csum_reg[15:8] <= ~sum_acc_temp[7:0];
end else begin
csum_in_csum_reg[7:0] <= ~sum_acc_temp[7:0];
csum_in_csum_reg[15:8] <= ~sum_acc_temp[15:8];
end
csum_in_offset_reg <= sum_offset_reg[LEVELS-2];
csum_in_enable_reg <= sum_enable_reg[LEVELS-2];
csum_in_valid_reg <= 1'b1;
sum_acc_reg <= 0;
end else begin
sum_acc_reg <= sum_acc_temp;
end
end
if (rst) begin
csum_in_valid_reg <= 1'b0;
end
end
// Insert checksum
always @* begin
data_out_axis_tready = m_axis_tready_int_reg && frame_reg;
csum_out_ready = 1'b0;
frame_next = frame_reg;
csum_next = csum_reg;
csum_offset_next = csum_offset_reg;
csum_enable_next = csum_enable_reg;
csum_split_next = csum_split_reg;
m_axis_tdata_int = data_out_axis_tdata;
m_axis_tkeep_int = data_out_axis_tkeep;
m_axis_tvalid_int = data_out_axis_tvalid && data_out_axis_tready;
m_axis_tlast_int = data_out_axis_tlast;
m_axis_tid_int = data_out_axis_tid;
m_axis_tdest_int = data_out_axis_tdest;
m_axis_tuser_int = data_out_axis_tuser;
if (frame_reg) begin
if (data_out_axis_tvalid && data_out_axis_tready) begin
if (data_out_axis_tlast) begin
frame_next = 1'b0;
end
if (csum_enable_reg) begin
if (csum_offset_reg >= KEEP_WIDTH) begin
csum_offset_next = csum_offset_reg - KEEP_WIDTH;
end else if (csum_split_reg) begin
// other byte of split checksum
m_axis_tdata_int[0 +: 8] = csum_reg[7:0];
csum_enable_next = 1'b0;
end else if (csum_offset_reg == KEEP_WIDTH-1) begin
// split across two cycles
m_axis_tdata_int[DATA_WIDTH-8 +: 8] = csum_reg[15:8];
csum_split_next = 1'b1;
end else begin
m_axis_tdata_int[csum_offset_reg*8 +: 8] = csum_reg[15:8];
m_axis_tdata_int[(csum_offset_reg+1)*8 +: 8] = csum_reg[7:0];
csum_enable_next = 1'b0;
end
end
end
end else begin
csum_out_ready = 1'b1;
csum_next = csum_out_csum;
csum_offset_next = csum_out_offset;
csum_enable_next = csum_out_enable;
csum_split_next = 1'b0;
if (csum_out_valid) begin
frame_next = 1'b1;
end
end
end
always @(posedge clk) begin
frame_reg <= frame_next;
csum_reg <= csum_next;
csum_offset_reg <= csum_offset_next;
csum_enable_reg <= csum_enable_next;
csum_split_reg <= csum_split_next;
if (rst) begin
frame_reg <= 1'b0;
csum_enable_reg <= 1'b0;
end
end
// output datapath logic
reg [DATA_WIDTH-1:0] m_axis_tdata_reg = {DATA_WIDTH{1'b0}};
reg [KEEP_WIDTH-1:0] m_axis_tkeep_reg = {KEEP_WIDTH{1'b0}};
reg m_axis_tvalid_reg = 1'b0, m_axis_tvalid_next;
reg m_axis_tlast_reg = 1'b0;
reg [ID_WIDTH-1:0] m_axis_tid_reg = {ID_WIDTH{1'b0}};
reg [DEST_WIDTH-1:0] m_axis_tdest_reg = {DEST_WIDTH{1'b0}};
reg [USER_WIDTH-1:0] m_axis_tuser_reg = {USER_WIDTH{1'b0}};
reg [DATA_WIDTH-1:0] temp_m_axis_tdata_reg = {DATA_WIDTH{1'b0}};
reg [KEEP_WIDTH-1:0] temp_m_axis_tkeep_reg = {KEEP_WIDTH{1'b0}};
reg temp_m_axis_tvalid_reg = 1'b0, temp_m_axis_tvalid_next;
reg temp_m_axis_tlast_reg = 1'b0;
reg [ID_WIDTH-1:0] temp_m_axis_tid_reg = {ID_WIDTH{1'b0}};
reg [DEST_WIDTH-1:0] temp_m_axis_tdest_reg = {DEST_WIDTH{1'b0}};
reg [USER_WIDTH-1:0] temp_m_axis_tuser_reg = {USER_WIDTH{1'b0}};
// datapath control
reg store_axis_int_to_output;
reg store_axis_int_to_temp;
reg store_axis_temp_to_output;
assign m_axis_tdata = m_axis_tdata_reg;
assign m_axis_tkeep = m_axis_tkeep_reg;
assign m_axis_tvalid = m_axis_tvalid_reg;
assign m_axis_tlast = m_axis_tlast_reg;
assign m_axis_tid = ID_ENABLE ? m_axis_tid_reg : {ID_WIDTH{1'b0}};
assign m_axis_tdest = DEST_ENABLE ? m_axis_tdest_reg : {DEST_WIDTH{1'b0}};
assign m_axis_tuser = USER_ENABLE ? m_axis_tuser_reg : {USER_WIDTH{1'b0}};
// enable ready input next cycle if output is ready or the temp reg will not be filled on the next cycle (output reg empty or no input)
assign m_axis_tready_int_early = m_axis_tready || (!temp_m_axis_tvalid_reg && (!m_axis_tvalid_reg || !m_axis_tvalid_int));
always @* begin
// transfer sink ready state to source
m_axis_tvalid_next = m_axis_tvalid_reg;
temp_m_axis_tvalid_next = temp_m_axis_tvalid_reg;
store_axis_int_to_output = 1'b0;
store_axis_int_to_temp = 1'b0;
store_axis_temp_to_output = 1'b0;
if (m_axis_tready_int_reg) begin
// input is ready
if (m_axis_tready || !m_axis_tvalid_reg) begin
// output is ready or currently not valid, transfer data to output
m_axis_tvalid_next = m_axis_tvalid_int;
store_axis_int_to_output = 1'b1;
end else begin
// output is not ready, store input in temp
temp_m_axis_tvalid_next = m_axis_tvalid_int;
store_axis_int_to_temp = 1'b1;
end
end else if (m_axis_tready) begin
// input is not ready, but output is ready
m_axis_tvalid_next = temp_m_axis_tvalid_reg;
temp_m_axis_tvalid_next = 1'b0;
store_axis_temp_to_output = 1'b1;
end
end
always @(posedge clk) begin
if (rst) begin
m_axis_tvalid_reg <= 1'b0;
m_axis_tready_int_reg <= 1'b0;
temp_m_axis_tvalid_reg <= 1'b0;
end else begin
m_axis_tvalid_reg <= m_axis_tvalid_next;
m_axis_tready_int_reg <= m_axis_tready_int_early;
temp_m_axis_tvalid_reg <= temp_m_axis_tvalid_next;
end
// datapath
if (store_axis_int_to_output) begin
m_axis_tdata_reg <= m_axis_tdata_int;
m_axis_tkeep_reg <= m_axis_tkeep_int;
m_axis_tlast_reg <= m_axis_tlast_int;
m_axis_tid_reg <= m_axis_tid_int;
m_axis_tdest_reg <= m_axis_tdest_int;
m_axis_tuser_reg <= m_axis_tuser_int;
end else if (store_axis_temp_to_output) begin
m_axis_tdata_reg <= temp_m_axis_tdata_reg;
m_axis_tkeep_reg <= temp_m_axis_tkeep_reg;
m_axis_tlast_reg <= temp_m_axis_tlast_reg;
m_axis_tid_reg <= temp_m_axis_tid_reg;
m_axis_tdest_reg <= temp_m_axis_tdest_reg;
m_axis_tuser_reg <= temp_m_axis_tuser_reg;
end
if (store_axis_int_to_temp) begin
temp_m_axis_tdata_reg <= m_axis_tdata_int;
temp_m_axis_tkeep_reg <= m_axis_tkeep_int;
temp_m_axis_tlast_reg <= m_axis_tlast_int;
temp_m_axis_tid_reg <= m_axis_tid_int;
temp_m_axis_tdest_reg <= m_axis_tdest_int;
temp_m_axis_tuser_reg <= m_axis_tuser_int;
end
end
endmodule
`resetall
Reference in New Issue View Git Blame Copy Permalink