1
0
mirror of https://github.com/corundum/corundum.git synced 2025-01-16 08:12:53 +08:00
corundum/fpga/common/rtl/tx_scheduler_rr.v
2021-08-30 01:27:53 -07:00

890 lines
34 KiB
Verilog

/*
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
`timescale 1ns / 1ps
/*
* Transmit scheduler (round-robin)
*/
module tx_scheduler_rr #
(
// Width of AXI lite data bus in bits
parameter AXIL_DATA_WIDTH = 32,
// Width of AXI lite address bus in bits
parameter AXIL_ADDR_WIDTH = 16,
// Width of AXI lite wstrb (width of data bus in words)
parameter AXIL_STRB_WIDTH = (AXIL_DATA_WIDTH/8),
// Length field width
parameter LEN_WIDTH = 16,
// Transmit request tag field width
parameter REQ_TAG_WIDTH = 8,
// Number of outstanding operations
parameter OP_TABLE_SIZE = 16,
// Queue index width
parameter QUEUE_INDEX_WIDTH = 6,
// Pipeline stages
parameter PIPELINE = 2,
// Scheduler control input enable
parameter SCHED_CTRL_ENABLE = 0
)
(
input wire clk,
input wire rst,
/*
* Transmit request output (queue index)
*/
output wire [QUEUE_INDEX_WIDTH-1:0] m_axis_tx_req_queue,
output wire [REQ_TAG_WIDTH-1:0] m_axis_tx_req_tag,
output wire m_axis_tx_req_valid,
input wire m_axis_tx_req_ready,
/*
* Transmit request status input
*/
input wire [LEN_WIDTH-1:0] s_axis_tx_req_status_len,
input wire [REQ_TAG_WIDTH-1:0] s_axis_tx_req_status_tag,
input wire s_axis_tx_req_status_valid,
/*
* Doorbell input
*/
input wire [QUEUE_INDEX_WIDTH-1:0] s_axis_doorbell_queue,
input wire s_axis_doorbell_valid,
/*
* Scheduler control input
*/
input wire [QUEUE_INDEX_WIDTH-1:0] s_axis_sched_ctrl_queue,
input wire s_axis_sched_ctrl_enable,
input wire s_axis_sched_ctrl_valid,
output wire s_axis_sched_ctrl_ready,
/*
* AXI-Lite slave interface
*/
input wire [AXIL_ADDR_WIDTH-1:0] s_axil_awaddr,
input wire [2:0] s_axil_awprot,
input wire s_axil_awvalid,
output wire s_axil_awready,
input wire [AXIL_DATA_WIDTH-1:0] s_axil_wdata,
input wire [AXIL_STRB_WIDTH-1:0] s_axil_wstrb,
input wire s_axil_wvalid,
output wire s_axil_wready,
output wire [1:0] s_axil_bresp,
output wire s_axil_bvalid,
input wire s_axil_bready,
input wire [AXIL_ADDR_WIDTH-1:0] s_axil_araddr,
input wire [2:0] s_axil_arprot,
input wire s_axil_arvalid,
output wire s_axil_arready,
output wire [AXIL_DATA_WIDTH-1:0] s_axil_rdata,
output wire [1:0] s_axil_rresp,
output wire s_axil_rvalid,
input wire s_axil_rready,
/*
* Control
*/
input wire enable,
output wire active
);
parameter QUEUE_COUNT = 2**QUEUE_INDEX_WIDTH;
parameter CL_OP_TABLE_SIZE = $clog2(OP_TABLE_SIZE);
parameter QUEUE_RAM_BE_WIDTH = 2;
parameter QUEUE_RAM_WIDTH = QUEUE_RAM_BE_WIDTH*8;
// bus width assertions
initial begin
if (REQ_TAG_WIDTH < CL_OP_TABLE_SIZE) begin
$error("Error: REQ_TAG_WIDTH insufficient for OP_TABLE_SIZE (instance %m)");
$finish;
end
if (AXIL_DATA_WIDTH != 32) begin
$error("Error: AXI lite interface width must be 32 (instance %m)");
$finish;
end
if (AXIL_STRB_WIDTH * 8 != AXIL_DATA_WIDTH) begin
$error("Error: AXI lite interface requires byte (8-bit) granularity (instance %m)");
$finish;
end
if (AXIL_ADDR_WIDTH < QUEUE_INDEX_WIDTH+5) begin
$error("Error: AXI lite address width too narrow (instance %m)");
$finish;
end
if (PIPELINE < 2) begin
$error("Error: PIPELINE must be at least 2 (instance %m)");
$finish;
end
end
reg op_axil_write_pipe_hazard;
reg op_axil_read_pipe_hazard;
reg op_doorbell_pipe_hazard;
reg op_req_pipe_hazard;
reg op_complete_pipe_hazard;
reg op_ctrl_pipe_hazard;
reg op_internal_pipe_hazard;
reg stage_active;
reg [PIPELINE-1:0] op_axil_write_pipe_reg = {PIPELINE{1'b0}}, op_axil_write_pipe_next;
reg [PIPELINE-1:0] op_axil_read_pipe_reg = {PIPELINE{1'b0}}, op_axil_read_pipe_next;
reg [PIPELINE-1:0] op_doorbell_pipe_reg = {PIPELINE{1'b0}}, op_doorbell_pipe_next;
reg [PIPELINE-1:0] op_req_pipe_reg = {PIPELINE{1'b0}}, op_req_pipe_next;
reg [PIPELINE-1:0] op_complete_pipe_reg = {PIPELINE{1'b0}}, op_complete_pipe_next;
reg [PIPELINE-1:0] op_ctrl_pipe_reg = {PIPELINE{1'b0}}, op_ctrl_pipe_next;
reg [PIPELINE-1:0] op_internal_pipe_reg = {PIPELINE{1'b0}}, op_internal_pipe_next;
reg [QUEUE_INDEX_WIDTH-1:0] queue_ram_addr_pipeline_reg[PIPELINE-1:0], queue_ram_addr_pipeline_next[PIPELINE-1:0];
reg [AXIL_DATA_WIDTH-1:0] write_data_pipeline_reg[PIPELINE-1:0], write_data_pipeline_next[PIPELINE-1:0];
reg [AXIL_STRB_WIDTH-1:0] write_strobe_pipeline_reg[PIPELINE-1:0], write_strobe_pipeline_next[PIPELINE-1:0];
reg [REQ_TAG_WIDTH-1:0] req_tag_pipeline_reg[PIPELINE-1:0], req_tag_pipeline_next[PIPELINE-1:0];
reg [CL_OP_TABLE_SIZE-1:0] op_index_pipeline_reg[PIPELINE-1:0], op_index_pipeline_next[PIPELINE-1:0];
reg [QUEUE_INDEX_WIDTH-1:0] m_axis_tx_req_queue_reg = {QUEUE_INDEX_WIDTH{1'b0}}, m_axis_tx_req_queue_next;
reg [REQ_TAG_WIDTH-1:0] m_axis_tx_req_tag_reg = {REQ_TAG_WIDTH{1'b0}}, m_axis_tx_req_tag_next;
reg m_axis_tx_req_valid_reg = 1'b0, m_axis_tx_req_valid_next;
reg s_axis_sched_ctrl_ready_reg = 1'b0, s_axis_sched_ctrl_ready_next;
reg s_axil_awready_reg = 0, s_axil_awready_next;
reg s_axil_wready_reg = 0, s_axil_wready_next;
reg s_axil_bvalid_reg = 0, s_axil_bvalid_next;
reg s_axil_arready_reg = 0, s_axil_arready_next;
reg [AXIL_DATA_WIDTH-1:0] s_axil_rdata_reg = 0, s_axil_rdata_next;
reg s_axil_rvalid_reg = 0, s_axil_rvalid_next;
reg [QUEUE_RAM_WIDTH-1:0] queue_ram[QUEUE_COUNT-1:0];
reg [QUEUE_INDEX_WIDTH-1:0] queue_ram_read_ptr;
reg [QUEUE_INDEX_WIDTH-1:0] queue_ram_write_ptr;
reg [QUEUE_RAM_WIDTH-1:0] queue_ram_write_data;
reg queue_ram_wr_en;
reg [QUEUE_RAM_BE_WIDTH-1:0] queue_ram_be;
reg [QUEUE_RAM_WIDTH-1:0] queue_ram_read_data_reg = 0;
reg [QUEUE_RAM_WIDTH-1:0] queue_ram_read_data_pipeline_reg[PIPELINE-1:1];
wire queue_ram_read_data_enabled = queue_ram_read_data_pipeline_reg[PIPELINE-1][0];
wire queue_ram_read_data_global_enable = queue_ram_read_data_pipeline_reg[PIPELINE-1][1];
wire queue_ram_read_data_sched_enable = queue_ram_read_data_pipeline_reg[PIPELINE-1][2];
wire queue_ram_read_data_active = queue_ram_read_data_pipeline_reg[PIPELINE-1][6];
wire queue_ram_read_data_scheduled = queue_ram_read_data_pipeline_reg[PIPELINE-1][7];
wire [CL_OP_TABLE_SIZE-1:0] queue_ram_read_data_op_tail_index = queue_ram_read_data_pipeline_reg[PIPELINE-1][15:8];
reg [OP_TABLE_SIZE-1:0] op_table_active = 0;
reg [QUEUE_INDEX_WIDTH-1:0] op_table_queue[OP_TABLE_SIZE-1:0];
reg op_table_doorbell[OP_TABLE_SIZE-1:0];
reg op_table_is_head[OP_TABLE_SIZE-1:0];
reg [CL_OP_TABLE_SIZE-1:0] op_table_next_index[OP_TABLE_SIZE-1:0];
reg [CL_OP_TABLE_SIZE-1:0] op_table_prev_index[OP_TABLE_SIZE-1:0];
wire [CL_OP_TABLE_SIZE-1:0] op_table_start_ptr;
wire op_table_start_ptr_valid;
reg [QUEUE_INDEX_WIDTH-1:0] op_table_start_queue;
reg op_table_start_en;
reg [CL_OP_TABLE_SIZE-1:0] op_table_doorbell_ptr;
reg op_table_doorbell_en;
reg [CL_OP_TABLE_SIZE-1:0] op_table_release_ptr;
reg op_table_release_en;
reg [CL_OP_TABLE_SIZE-1:0] op_table_update_next_ptr;
reg [CL_OP_TABLE_SIZE-1:0] op_table_update_next_index;
reg op_table_update_next_en;
reg [CL_OP_TABLE_SIZE-1:0] op_table_update_prev_ptr;
reg [CL_OP_TABLE_SIZE-1:0] op_table_update_prev_index;
reg op_table_update_prev_is_head;
reg op_table_update_prev_en;
reg [CL_OP_TABLE_SIZE+1-1:0] finish_fifo_wr_ptr_reg = 0, finish_fifo_wr_ptr_next;
reg [CL_OP_TABLE_SIZE+1-1:0] finish_fifo_rd_ptr_reg = 0, finish_fifo_rd_ptr_next;
reg [REQ_TAG_WIDTH-1:0] finish_fifo_tag[(2**CL_OP_TABLE_SIZE)-1:0];
reg finish_fifo_status[(2**CL_OP_TABLE_SIZE)-1:0];
reg finish_fifo_we;
reg [REQ_TAG_WIDTH-1:0] finish_fifo_wr_tag;
reg finish_fifo_wr_status;
reg [CL_OP_TABLE_SIZE-1:0] finish_ptr_reg = {CL_OP_TABLE_SIZE{1'b0}}, finish_ptr_next;
reg finish_status_reg = 1'b0, finish_status_next;
reg finish_valid_reg = 1'b0, finish_valid_next;
reg init_reg = 1'b0, init_next;
reg [QUEUE_INDEX_WIDTH-1:0] init_index_reg = 0, init_index_next;
reg [QUEUE_INDEX_WIDTH:0] active_queue_count_reg = 0, active_queue_count_next;
assign m_axis_tx_req_queue = m_axis_tx_req_queue_reg;
assign m_axis_tx_req_tag = m_axis_tx_req_tag_reg;
assign m_axis_tx_req_valid = m_axis_tx_req_valid_reg;
assign s_axis_sched_ctrl_ready = s_axis_sched_ctrl_ready_reg;
assign s_axil_awready = s_axil_awready_reg;
assign s_axil_wready = s_axil_wready_reg;
assign s_axil_bresp = 2'b00;
assign s_axil_bvalid = s_axil_bvalid_reg;
assign s_axil_arready = s_axil_arready_reg;
assign s_axil_rdata = s_axil_rdata_reg;
assign s_axil_rresp = 2'b00;
assign s_axil_rvalid = s_axil_rvalid_reg;
assign active = active_queue_count_reg != 0;
wire [QUEUE_INDEX_WIDTH-1:0] s_axil_awaddr_queue = s_axil_awaddr >> 2;
wire [QUEUE_INDEX_WIDTH-1:0] s_axil_araddr_queue = s_axil_araddr >> 2;
wire queue_tail_active = op_table_active[queue_ram_read_data_op_tail_index] && op_table_queue[queue_ram_read_data_op_tail_index] == queue_ram_addr_pipeline_reg[PIPELINE-1];
wire [QUEUE_INDEX_WIDTH-1:0] axis_doorbell_fifo_queue;
wire axis_doorbell_fifo_valid;
reg axis_doorbell_fifo_ready;
axis_fifo #(
.DEPTH(256),
.DATA_WIDTH(QUEUE_INDEX_WIDTH),
.KEEP_ENABLE(0),
.KEEP_WIDTH(1),
.LAST_ENABLE(0),
.ID_ENABLE(0),
.DEST_ENABLE(0),
.USER_ENABLE(0),
.FRAME_FIFO(0)
)
doorbell_fifo (
.clk(clk),
.rst(rst),
// AXI input
.s_axis_tdata(s_axis_doorbell_queue),
.s_axis_tkeep(0),
.s_axis_tvalid(s_axis_doorbell_valid),
.s_axis_tready(),
.s_axis_tlast(0),
.s_axis_tid(0),
.s_axis_tdest(0),
.s_axis_tuser(0),
// AXI output
.m_axis_tdata(axis_doorbell_fifo_queue),
.m_axis_tkeep(),
.m_axis_tvalid(axis_doorbell_fifo_valid),
.m_axis_tready(axis_doorbell_fifo_ready),
.m_axis_tlast(),
.m_axis_tid(),
.m_axis_tdest(),
.m_axis_tuser(),
// Status
.status_overflow(),
.status_bad_frame(),
.status_good_frame()
);
reg [QUEUE_INDEX_WIDTH-1:0] axis_scheduler_fifo_in_queue;
reg axis_scheduler_fifo_in_valid;
wire axis_scheduler_fifo_in_ready;
wire [QUEUE_INDEX_WIDTH-1:0] axis_scheduler_fifo_out_queue;
wire axis_scheduler_fifo_out_valid;
reg axis_scheduler_fifo_out_ready;
axis_fifo #(
.DEPTH(2**QUEUE_INDEX_WIDTH),
.DATA_WIDTH(QUEUE_INDEX_WIDTH),
.KEEP_ENABLE(0),
.KEEP_WIDTH(1),
.LAST_ENABLE(0),
.ID_ENABLE(0),
.DEST_ENABLE(0),
.USER_ENABLE(0),
.FRAME_FIFO(0)
)
rr_fifo (
.clk(clk),
.rst(rst),
// AXI input
.s_axis_tdata(axis_scheduler_fifo_in_queue),
.s_axis_tkeep(0),
.s_axis_tvalid(axis_scheduler_fifo_in_valid),
.s_axis_tready(axis_scheduler_fifo_in_ready),
.s_axis_tlast(0),
.s_axis_tid(0),
.s_axis_tdest(0),
.s_axis_tuser(0),
// AXI output
.m_axis_tdata(axis_scheduler_fifo_out_queue),
.m_axis_tkeep(),
.m_axis_tvalid(axis_scheduler_fifo_out_valid),
.m_axis_tready(axis_scheduler_fifo_out_ready),
.m_axis_tlast(),
.m_axis_tid(),
.m_axis_tdest(),
.m_axis_tuser(),
// Status
.status_overflow(),
.status_bad_frame(),
.status_good_frame()
);
priority_encoder #(
.WIDTH(OP_TABLE_SIZE),
.LSB_HIGH_PRIORITY(1)
)
op_table_start_enc_inst (
.input_unencoded(~op_table_active),
.output_valid(op_table_start_ptr_valid),
.output_encoded(op_table_start_ptr),
.output_unencoded()
);
integer i;
initial begin
for (i = 0; i < QUEUE_COUNT; i = i + 1) begin
queue_ram[i] = 0;
end
for (i = 0; i < PIPELINE; i = i + 1) begin
queue_ram_addr_pipeline_reg[i] = 0;
write_data_pipeline_reg[i] = 0;
write_strobe_pipeline_reg[i] = 0;
req_tag_pipeline_reg[i] = 0;
end
for (i = 0; i < OP_TABLE_SIZE; i = i + 1) begin
op_table_queue[i] = 0;
op_table_next_index[i] = 0;
op_table_prev_index[i] = 0;
op_table_doorbell[i] = 0;
op_table_is_head[i] = 0;
end
end
integer j;
always @* begin
op_axil_write_pipe_next = {op_axil_write_pipe_reg, 1'b0};
op_axil_read_pipe_next = {op_axil_read_pipe_reg, 1'b0};
op_doorbell_pipe_next = {op_doorbell_pipe_reg, 1'b0};
op_req_pipe_next = {op_req_pipe_reg, 1'b0};
op_complete_pipe_next = {op_complete_pipe_reg, 1'b0};
op_ctrl_pipe_next = {op_ctrl_pipe_reg, 1'b0};
op_internal_pipe_next = {op_internal_pipe_reg, 1'b0};
queue_ram_addr_pipeline_next[0] = 0;
write_data_pipeline_next[0] = 0;
write_strobe_pipeline_next[0] = 0;
req_tag_pipeline_next[0] = 0;
op_index_pipeline_next[0] = 0;
for (j = 1; j < PIPELINE; j = j + 1) begin
queue_ram_addr_pipeline_next[j] = queue_ram_addr_pipeline_reg[j-1];
write_data_pipeline_next[j] = write_data_pipeline_reg[j-1];
write_strobe_pipeline_next[j] = write_strobe_pipeline_reg[j-1];
req_tag_pipeline_next[j] = req_tag_pipeline_reg[j-1];
op_index_pipeline_next[j] = op_index_pipeline_reg[j-1];
end
m_axis_tx_req_queue_next = m_axis_tx_req_queue_reg;
m_axis_tx_req_tag_next = m_axis_tx_req_tag_reg;
m_axis_tx_req_valid_next = m_axis_tx_req_valid_reg && !m_axis_tx_req_ready;
s_axis_sched_ctrl_ready_next = 1'b0;
s_axil_awready_next = 1'b0;
s_axil_wready_next = 1'b0;
s_axil_bvalid_next = s_axil_bvalid_reg && !s_axil_bready;
s_axil_arready_next = 1'b0;
s_axil_rdata_next = s_axil_rdata_reg;
s_axil_rvalid_next = s_axil_rvalid_reg && !s_axil_rready;
queue_ram_read_ptr = 0;
queue_ram_write_ptr = queue_ram_addr_pipeline_reg[PIPELINE-1];
queue_ram_write_data = queue_ram_read_data_pipeline_reg[PIPELINE-1];
queue_ram_wr_en = 0;
queue_ram_be = 0;
op_table_start_queue = queue_ram_addr_pipeline_reg[PIPELINE-1];
op_table_start_en = 1'b0;
op_table_doorbell_ptr = queue_ram_read_data_op_tail_index;
op_table_doorbell_en = 1'b0;
op_table_release_ptr = op_index_pipeline_reg[PIPELINE-1];
op_table_release_en = 1'b0;
op_table_update_next_ptr = queue_ram_read_data_op_tail_index;
op_table_update_next_index = op_index_pipeline_reg[PIPELINE-1];
op_table_update_next_en = 1'b0;
op_table_update_prev_ptr = op_index_pipeline_reg[PIPELINE-1];
op_table_update_prev_index = queue_ram_read_data_op_tail_index;
op_table_update_prev_is_head = !(queue_tail_active && op_index_pipeline_reg[PIPELINE-1] != queue_ram_read_data_op_tail_index);
op_table_update_prev_en = 1'b0;
finish_fifo_rd_ptr_next = finish_fifo_rd_ptr_reg;
finish_fifo_wr_ptr_next = finish_fifo_wr_ptr_reg;
finish_fifo_we = 1'b0;
finish_fifo_wr_tag = s_axis_tx_req_status_tag;
finish_fifo_wr_status = s_axis_tx_req_status_len != 0;
finish_ptr_next = finish_ptr_reg;
finish_status_next = finish_status_reg;
finish_valid_next = finish_valid_reg;
init_next = init_reg;
init_index_next = init_index_reg;
active_queue_count_next = active_queue_count_reg;
axis_doorbell_fifo_ready = 1'b0;
axis_scheduler_fifo_in_queue = queue_ram_addr_pipeline_reg[PIPELINE-1];
axis_scheduler_fifo_in_valid = 1'b0;
axis_scheduler_fifo_out_ready = 1'b0;
op_axil_write_pipe_hazard = 1'b0;
op_axil_read_pipe_hazard = 1'b0;
op_doorbell_pipe_hazard = 1'b0;
op_req_pipe_hazard = 1'b0;
op_complete_pipe_hazard = 1'b0;
op_ctrl_pipe_hazard = 1'b0;
op_internal_pipe_hazard = 1'b0;
stage_active = 1'b0;
for (j = 0; j < PIPELINE; j = j + 1) begin
stage_active = op_axil_write_pipe_reg[j] || op_axil_read_pipe_reg[j] || op_doorbell_pipe_reg[j] || op_req_pipe_reg[j] || op_complete_pipe_reg[j] || op_ctrl_pipe_reg[j] || op_internal_pipe_reg[j];
op_axil_write_pipe_hazard = op_axil_write_pipe_hazard || (stage_active && queue_ram_addr_pipeline_reg[j] == s_axil_awaddr_queue);
op_axil_read_pipe_hazard = op_axil_read_pipe_hazard || (stage_active && queue_ram_addr_pipeline_reg[j] == s_axil_araddr_queue);
op_doorbell_pipe_hazard = op_doorbell_pipe_hazard || (stage_active && queue_ram_addr_pipeline_reg[j] == axis_doorbell_fifo_queue);
op_req_pipe_hazard = op_req_pipe_hazard || (stage_active && queue_ram_addr_pipeline_reg[j] == axis_scheduler_fifo_out_queue);
op_complete_pipe_hazard = op_complete_pipe_hazard || (stage_active && queue_ram_addr_pipeline_reg[j] == op_table_queue[finish_ptr_reg]);
op_ctrl_pipe_hazard = op_ctrl_pipe_hazard || (stage_active && queue_ram_addr_pipeline_reg[j] == s_axis_sched_ctrl_queue);
op_internal_pipe_hazard = op_internal_pipe_hazard || (stage_active && queue_ram_addr_pipeline_reg[j] == init_index_reg);
end
// pipeline stage 0 - receive request
if (!init_reg && !op_internal_pipe_hazard) begin
// init queue states
op_internal_pipe_next[0] = 1'b1;
init_index_next = init_index_reg + 1;
queue_ram_read_ptr = init_index_reg;
queue_ram_addr_pipeline_next[0] = init_index_reg;
if (init_index_reg == {QUEUE_INDEX_WIDTH{1'b1}}) begin
init_next = 1'b1;
end
end else if (s_axil_awvalid && s_axil_wvalid && (!s_axil_bvalid || s_axil_bready) && !op_axil_write_pipe_reg[0] && !op_axil_write_pipe_hazard) begin
// AXIL write
op_axil_write_pipe_next[0] = 1'b1;
s_axil_awready_next = 1'b1;
s_axil_wready_next = 1'b1;
write_data_pipeline_next[0] = s_axil_wdata;
write_strobe_pipeline_next[0] = s_axil_wstrb;
queue_ram_read_ptr = s_axil_awaddr_queue;
queue_ram_addr_pipeline_next[0] = s_axil_awaddr_queue;
end else if (s_axil_arvalid && (!s_axil_rvalid || s_axil_rready) && !op_axil_read_pipe_reg[0] && !op_axil_read_pipe_hazard) begin
// AXIL read
op_axil_read_pipe_next[0] = 1'b1;
s_axil_arready_next = 1'b1;
queue_ram_read_ptr = s_axil_araddr_queue;
queue_ram_addr_pipeline_next[0] = s_axil_araddr_queue;
end else if (axis_doorbell_fifo_valid && !op_doorbell_pipe_hazard) begin
// handle doorbell
op_doorbell_pipe_next[0] = 1'b1;
axis_doorbell_fifo_ready = 1'b1;
queue_ram_read_ptr = axis_doorbell_fifo_queue;
queue_ram_addr_pipeline_next[0] = axis_doorbell_fifo_queue;
end else if (finish_valid_reg && !op_complete_pipe_reg[0] && !op_complete_pipe_hazard) begin
// transmit complete
op_complete_pipe_next[0] = 1'b1;
write_data_pipeline_next[0][0] = finish_status_reg || op_table_doorbell[finish_ptr_reg];
op_index_pipeline_next[0] = finish_ptr_reg;
finish_valid_next = 1'b0;
queue_ram_read_ptr = op_table_queue[finish_ptr_reg];
queue_ram_addr_pipeline_next[0] = op_table_queue[finish_ptr_reg];
end else if (SCHED_CTRL_ENABLE && s_axis_sched_ctrl_valid && !op_ctrl_pipe_reg[0] && !op_ctrl_pipe_hazard) begin
// Scheduler control
op_ctrl_pipe_next[0] = 1'b1;
s_axis_sched_ctrl_ready_next = 1'b1;
write_data_pipeline_next[0] = s_axis_sched_ctrl_enable;
queue_ram_read_ptr = s_axis_sched_ctrl_queue;
queue_ram_addr_pipeline_next[0] = s_axis_sched_ctrl_queue;
end else if (enable && op_table_start_ptr_valid && axis_scheduler_fifo_out_valid && (!m_axis_tx_req_valid || m_axis_tx_req_ready) && !op_req_pipe_reg && !op_req_pipe_hazard) begin
// transmit request
op_req_pipe_next[0] = 1'b1;
op_table_start_en = 1'b1;
op_table_start_queue = axis_scheduler_fifo_out_queue;
op_index_pipeline_next[0] = op_table_start_ptr;
axis_scheduler_fifo_out_ready = 1'b1;
queue_ram_read_ptr = axis_scheduler_fifo_out_queue;
queue_ram_addr_pipeline_next[0] = axis_scheduler_fifo_out_queue;
end
// read complete, perform operation
if (op_internal_pipe_reg[PIPELINE-1]) begin
// internal operation
// init queue state
queue_ram_write_ptr = queue_ram_addr_pipeline_reg[PIPELINE-1];
queue_ram_write_data[0] = 1'b0; // queue enabled
if (SCHED_CTRL_ENABLE) begin
queue_ram_write_data[1] = 1'b0; // queue global enable
queue_ram_write_data[2] = 1'b0; // queue sched enable
end
queue_ram_write_data[6] = 1'b0; // queue active
queue_ram_write_data[7] = 1'b0; // queue scheduled
queue_ram_be[0] = 1'b1;
queue_ram_wr_en = 1'b1;
end else if (op_doorbell_pipe_reg[PIPELINE-1]) begin
// handle doorbell
// mark queue active
queue_ram_write_ptr = queue_ram_addr_pipeline_reg[PIPELINE-1];
queue_ram_write_data[6] = 1'b1; // queue active
queue_ram_be[0] = 1'b1;
queue_ram_wr_en = 1'b1;
// schedule queue if necessary
if (queue_ram_read_data_enabled && (!SCHED_CTRL_ENABLE || queue_ram_read_data_global_enable || queue_ram_read_data_sched_enable) && !queue_ram_read_data_scheduled) begin
queue_ram_write_data[7] = 1'b1; // queue scheduled
axis_scheduler_fifo_in_queue = queue_ram_addr_pipeline_reg[PIPELINE-1];
axis_scheduler_fifo_in_valid = 1'b1;
active_queue_count_next = active_queue_count_reg + 1;
end
if (queue_tail_active) begin
// record doorbell in table so we don't lose it
op_table_doorbell_ptr = queue_ram_read_data_op_tail_index;
op_table_doorbell_en = 1'b1;
end
end else if (op_req_pipe_reg[PIPELINE-1]) begin
// transmit request
m_axis_tx_req_queue_next = queue_ram_addr_pipeline_reg[PIPELINE-1];
m_axis_tx_req_tag_next = op_index_pipeline_reg[PIPELINE-1];
axis_scheduler_fifo_in_queue = queue_ram_addr_pipeline_reg[PIPELINE-1];
// update state
queue_ram_write_ptr = queue_ram_addr_pipeline_reg[PIPELINE-1];
queue_ram_write_data[15:8] = op_index_pipeline_reg[PIPELINE-1]; // tail index
queue_ram_be[0] = 1'b1;
queue_ram_wr_en = 1'b1;
op_table_update_prev_ptr = op_index_pipeline_reg[PIPELINE-1];
op_table_update_prev_index = queue_ram_read_data_op_tail_index;
op_table_update_prev_is_head = !(queue_tail_active && op_index_pipeline_reg[PIPELINE-1] != queue_ram_read_data_op_tail_index);
op_table_update_next_ptr = queue_ram_read_data_op_tail_index;
op_table_update_next_index = op_index_pipeline_reg[PIPELINE-1];
if (queue_ram_read_data_enabled && (!SCHED_CTRL_ENABLE || queue_ram_read_data_global_enable || queue_ram_read_data_sched_enable) && queue_ram_read_data_active && queue_ram_read_data_scheduled) begin
// queue enabled, active, and scheduled
// issue transmit request
m_axis_tx_req_valid_next = 1'b1;
// reschedule
axis_scheduler_fifo_in_valid = 1'b1;
// update state
queue_ram_write_data[7] = 1'b1; // queue scheduled
queue_ram_be[1] = 1'b1; // tail index
op_table_update_prev_en = 1'b1;
op_table_update_next_en = queue_tail_active && op_index_pipeline_reg[PIPELINE-1] != queue_ram_read_data_op_tail_index;
end else begin
// queue not enabled, not active, or not scheduled
// deschedule queue
op_table_release_ptr = op_index_pipeline_reg[PIPELINE-1];
op_table_release_en = 1'b1;
// update state
queue_ram_write_data[7] = 1'b0; // queue scheduled
if (queue_ram_read_data_scheduled) begin
active_queue_count_next = active_queue_count_reg - 1;
end
end
end else if (op_complete_pipe_reg[PIPELINE-1]) begin
// tx complete
// update state
queue_ram_write_ptr = queue_ram_addr_pipeline_reg[PIPELINE-1];
queue_ram_be[0] = 1'b1;
queue_ram_wr_en = 1'b1;
op_table_update_prev_ptr = op_table_next_index[op_index_pipeline_reg[PIPELINE-1]];
op_table_update_prev_index = op_table_prev_index[op_index_pipeline_reg[PIPELINE-1]];
op_table_update_prev_is_head = op_table_is_head[op_index_pipeline_reg[PIPELINE-1]];
op_table_update_prev_en = op_index_pipeline_reg[PIPELINE-1] != queue_ram_read_data_op_tail_index; // our next pointer only valid if we're not the tail
op_table_update_next_ptr = op_table_prev_index[op_index_pipeline_reg[PIPELINE-1]];
op_table_update_next_index = op_table_next_index[op_index_pipeline_reg[PIPELINE-1]];
op_table_update_next_en = !op_table_is_head[op_index_pipeline_reg[PIPELINE-1]]; // our prev index only valid if we're not the head element
op_table_doorbell_ptr = op_table_prev_index[op_index_pipeline_reg[PIPELINE-1]];
op_table_doorbell_en = !op_table_is_head[op_index_pipeline_reg[PIPELINE-1]] && op_table_doorbell[op_index_pipeline_reg[PIPELINE-1]];;
op_table_release_ptr = op_index_pipeline_reg[PIPELINE-1];
op_table_release_en = 1'b1;
if (write_data_pipeline_reg[PIPELINE-1][0]) begin
queue_ram_write_data[6] = 1'b1; // queue active
// schedule if disabled
if ((!SCHED_CTRL_ENABLE || write_data_pipeline_reg[PIPELINE-1][1] || queue_ram_read_data_sched_enable) && !queue_ram_read_data_scheduled) begin
queue_ram_write_data[7] = 1'b1; // queue scheduled
axis_scheduler_fifo_in_queue = queue_ram_addr_pipeline_reg[PIPELINE-1];
axis_scheduler_fifo_in_valid = 1'b1;
active_queue_count_next = active_queue_count_reg + 1;
end
end else begin
queue_ram_write_data[6] = 1'b0; // queue active
end
end else if (SCHED_CTRL_ENABLE && op_ctrl_pipe_reg[PIPELINE-1]) begin
// Scheduler control
queue_ram_write_ptr = queue_ram_addr_pipeline_reg[PIPELINE-1];
queue_ram_wr_en = 1'b1;
queue_ram_write_data[2] = write_data_pipeline_reg[PIPELINE-1][0]; // queue sched enable
queue_ram_be[0] = 1'b1;
// schedule if disabled
if (queue_ram_read_data_enabled && queue_ram_read_data_active && (queue_ram_read_data_global_enable || write_data_pipeline_reg[PIPELINE-1][0]) && !queue_ram_read_data_scheduled) begin
queue_ram_write_data[7] = 1'b1; // queue scheduled
axis_scheduler_fifo_in_queue = queue_ram_addr_pipeline_reg[PIPELINE-1];
axis_scheduler_fifo_in_valid = 1'b1;
active_queue_count_next = active_queue_count_reg + 1;
end
end else if (op_axil_write_pipe_reg[PIPELINE-1]) begin
// AXIL write
s_axil_bvalid_next = 1'b1;
queue_ram_write_ptr = queue_ram_addr_pipeline_reg[PIPELINE-1];
queue_ram_wr_en = 1'b1;
queue_ram_write_data[0] = write_data_pipeline_reg[PIPELINE-1][0]; // queue enabled
queue_ram_write_data[1] = write_data_pipeline_reg[PIPELINE-1][1]; // queue global enable
queue_ram_be[0] = write_strobe_pipeline_reg[PIPELINE-1][0];
// schedule if disabled
if (write_data_pipeline_reg[PIPELINE-1][0] && queue_ram_read_data_active && (!SCHED_CTRL_ENABLE || write_data_pipeline_reg[PIPELINE-1][1] || queue_ram_read_data_sched_enable) && !queue_ram_read_data_scheduled) begin
queue_ram_write_data[7] = 1'b1; // queue scheduled
axis_scheduler_fifo_in_queue = queue_ram_addr_pipeline_reg[PIPELINE-1];
axis_scheduler_fifo_in_valid = 1'b1;
active_queue_count_next = active_queue_count_reg + 1;
end
end else if (op_axil_read_pipe_reg[PIPELINE-1]) begin
// AXIL read
s_axil_rvalid_next = 1'b1;
s_axil_rdata_next = 0;
s_axil_rdata_next[0] = queue_ram_read_data_enabled;
if (SCHED_CTRL_ENABLE) begin
s_axil_rdata_next[1] = queue_ram_read_data_global_enable;
s_axil_rdata_next[2] = queue_ram_read_data_sched_enable;
end
s_axil_rdata_next[16] = queue_ram_read_data_active;
s_axil_rdata_next[24] = queue_ram_read_data_scheduled;
end
// finish transmit operation
if (s_axis_tx_req_status_valid) begin
finish_fifo_we = 1'b1;
finish_fifo_wr_tag = s_axis_tx_req_status_tag;
finish_fifo_wr_status = s_axis_tx_req_status_len != 0;
finish_fifo_wr_ptr_next = finish_fifo_wr_ptr_reg + 1;
end
if (!finish_valid_reg && finish_fifo_wr_ptr_reg != finish_fifo_rd_ptr_reg) begin
finish_ptr_next = finish_fifo_tag[finish_fifo_rd_ptr_reg[CL_OP_TABLE_SIZE-1:0]];
finish_status_next = finish_fifo_status[finish_fifo_rd_ptr_reg[CL_OP_TABLE_SIZE-1:0]];
finish_valid_next = 1'b1;
finish_fifo_rd_ptr_next = finish_fifo_rd_ptr_reg + 1;
end
end
always @(posedge clk) begin
if (rst) begin
op_axil_write_pipe_reg <= {PIPELINE{1'b0}};
op_axil_read_pipe_reg <= {PIPELINE{1'b0}};
op_doorbell_pipe_reg <= {PIPELINE{1'b0}};
op_req_pipe_reg <= {PIPELINE{1'b0}};
op_complete_pipe_reg <= {PIPELINE{1'b0}};
op_ctrl_pipe_reg <= {PIPELINE{1'b0}};
op_internal_pipe_reg <= {PIPELINE{1'b0}};
finish_fifo_rd_ptr_reg <= {CL_OP_TABLE_SIZE+1{1'b0}};
finish_fifo_wr_ptr_reg <= {CL_OP_TABLE_SIZE+1{1'b0}};
finish_valid_reg <= 1'b0;
m_axis_tx_req_valid_reg <= 1'b0;
s_axis_sched_ctrl_ready_reg <= 1'b0;
s_axil_awready_reg <= 1'b0;
s_axil_wready_reg <= 1'b0;
s_axil_bvalid_reg <= 1'b0;
s_axil_arready_reg <= 1'b0;
s_axil_rvalid_reg <= 1'b0;
init_reg <= 1'b0;
init_index_reg <= 0;
active_queue_count_reg <= 0;
op_table_active <= 0;
end else begin
op_axil_write_pipe_reg <= op_axil_write_pipe_next;
op_axil_read_pipe_reg <= op_axil_read_pipe_next;
op_doorbell_pipe_reg <= op_doorbell_pipe_next;
op_req_pipe_reg <= op_req_pipe_next;
op_complete_pipe_reg <= op_complete_pipe_next;
op_ctrl_pipe_reg <= op_ctrl_pipe_next;
op_internal_pipe_reg <= op_internal_pipe_next;
finish_fifo_rd_ptr_reg <= finish_fifo_rd_ptr_next;
finish_fifo_wr_ptr_reg <= finish_fifo_wr_ptr_next;
finish_valid_reg <= finish_valid_next;
m_axis_tx_req_valid_reg <= m_axis_tx_req_valid_next;
s_axis_sched_ctrl_ready_reg <= s_axis_sched_ctrl_ready_next;
s_axil_awready_reg <= s_axil_awready_next;
s_axil_wready_reg <= s_axil_wready_next;
s_axil_bvalid_reg <= s_axil_bvalid_next;
s_axil_arready_reg <= s_axil_arready_next;
s_axil_rvalid_reg <= s_axil_rvalid_next;
init_reg <= init_next;
init_index_reg <= init_index_next;
active_queue_count_reg <= active_queue_count_next;
if (op_table_start_en) begin
op_table_active[op_table_start_ptr] <= 1'b1;
end
if (op_table_release_en) begin
op_table_active[op_table_release_ptr] <= 1'b0;
end
end
for (i = 0; i < PIPELINE; i = i + 1) begin
queue_ram_addr_pipeline_reg[i] <= queue_ram_addr_pipeline_next[i];
write_data_pipeline_reg[i] <= write_data_pipeline_next[i];
write_strobe_pipeline_reg[i] <= write_strobe_pipeline_next[i];
req_tag_pipeline_reg[i] <= req_tag_pipeline_next[i];
op_index_pipeline_reg[i] <= op_index_pipeline_next[i];
end
finish_ptr_reg <= finish_ptr_next;
finish_status_reg <= finish_status_next;
m_axis_tx_req_queue_reg <= m_axis_tx_req_queue_next;
m_axis_tx_req_tag_reg <= m_axis_tx_req_tag_next;
s_axil_rdata_reg <= s_axil_rdata_next;
if (queue_ram_wr_en) begin
for (i = 0; i < QUEUE_RAM_BE_WIDTH; i = i + 1) begin
if (queue_ram_be[i]) begin
queue_ram[queue_ram_write_ptr][i*8 +: 8] <= queue_ram_write_data[i*8 +: 8];
end
end
end
queue_ram_read_data_reg <= queue_ram[queue_ram_read_ptr];
queue_ram_read_data_pipeline_reg[1] <= queue_ram_read_data_reg;
for (i = 2; i < PIPELINE; i = i + 1) begin
queue_ram_read_data_pipeline_reg[i] <= queue_ram_read_data_pipeline_reg[i-1];
end
if (op_table_start_en) begin
op_table_queue[op_table_start_ptr] <= op_table_start_queue;
op_table_doorbell[op_table_start_ptr] <= 1'b0;
end
if (op_table_doorbell_en) begin
op_table_doorbell[op_table_doorbell_ptr] <= 1'b1;
end
if (op_table_update_next_en) begin
op_table_next_index[op_table_update_next_ptr] <= op_table_update_next_index;
end
if (op_table_update_prev_en) begin
op_table_prev_index[op_table_update_prev_ptr] <= op_table_update_prev_index;
op_table_is_head[op_table_update_prev_ptr] <= op_table_update_prev_is_head;
end
if (finish_fifo_we) begin
finish_fifo_tag[finish_fifo_wr_ptr_reg[CL_OP_TABLE_SIZE-1:0]] <= finish_fifo_wr_tag;
finish_fifo_status[finish_fifo_wr_ptr_reg[CL_OP_TABLE_SIZE-1:0]] <= finish_fifo_wr_status;
end
end
endmodule