aboutsummaryrefslogtreecommitdiffstats
path: root/fpga/usrp3/lib/fifo/axi_packet_gate.v
diff options
context:
space:
mode:
authorMartin Braun <martin.braun@ettus.com>2020-01-23 16:10:22 -0800
committerMartin Braun <martin.braun@ettus.com>2020-01-28 09:35:36 -0800
commitbafa9d95453387814ef25e6b6256ba8db2df612f (patch)
tree39ba24b5b67072d354775272e687796bb511848d /fpga/usrp3/lib/fifo/axi_packet_gate.v
parent3075b981503002df3115d5f1d0b97d2619ba30f2 (diff)
downloaduhd-bafa9d95453387814ef25e6b6256ba8db2df612f.tar.gz
uhd-bafa9d95453387814ef25e6b6256ba8db2df612f.tar.bz2
uhd-bafa9d95453387814ef25e6b6256ba8db2df612f.zip
Merge FPGA repository back into UHD repository
The FPGA codebase was removed from the UHD repository in 2014 to reduce the size of the repository. However, over the last half-decade, the split between the repositories has proven more burdensome than it has been helpful. By merging the FPGA code back, it will be possible to create atomic commits that touch both FPGA and UHD codebases. Continuous integration testing is also simplified by merging the repositories, because it was previously difficult to automatically derive the correct UHD branch when testing a feature branch on the FPGA repository. This commit also updates the license files and paths therein. We are therefore merging the repositories again. Future development for FPGA code will happen in the same repository as the UHD host code and MPM code. == Original Codebase and Rebasing == The original FPGA repository will be hosted for the foreseeable future at its original local location: https://github.com/EttusResearch/fpga/ It can be used for bisecting, reference, and a more detailed history. The final commit from said repository to be merged here is 05003794e2da61cabf64dd278c45685a7abad7ec. This commit is tagged as v4.0.0.0-pre-uhd-merge. If you have changes in the FPGA repository that you want to rebase onto the UHD repository, simply run the following commands: - Create a directory to store patches (this should be an empty directory): mkdir ~/patches - Now make sure that your FPGA codebase is based on the same state as the code that was merged: cd src/fpga # Or wherever your FPGA code is stored git rebase v4.0.0.0-pre-uhd-merge Note: The rebase command may look slightly different depending on what exactly you're trying to rebase. - Create a patch set for your changes versus v4.0.0.0-pre-uhd-merge: git format-patch v4.0.0.0-pre-uhd-merge -o ~/patches Note: Make sure that only patches are stored in your output directory. It should otherwise be empty. Make sure that you picked the correct range of commits, and only commits you wanted to rebase were exported as patch files. - Go to the UHD repository and apply the patches: cd src/uhd # Or wherever your UHD repository is stored git am --directory fpga ~/patches/* rm -rf ~/patches # This is for cleanup == Contributors == The following people have contributed mainly to these files (this list is not complete): Co-authored-by: Alex Williams <alex.williams@ni.com> Co-authored-by: Andrej Rode <andrej.rode@ettus.com> Co-authored-by: Ashish Chaudhari <ashish@ettus.com> Co-authored-by: Ben Hilburn <ben.hilburn@ettus.com> Co-authored-by: Ciro Nishiguchi <ciro.nishiguchi@ni.com> Co-authored-by: Daniel Jepson <daniel.jepson@ni.com> Co-authored-by: Derek Kozel <derek.kozel@ettus.com> Co-authored-by: EJ Kreinar <ej@he360.com> Co-authored-by: Humberto Jimenez <humberto.jimenez@ni.com> Co-authored-by: Ian Buckley <ian.buckley@gmail.com> Co-authored-by: Jörg Hofrichter <joerg.hofrichter@ni.com> Co-authored-by: Jon Kiser <jon.kiser@ni.com> Co-authored-by: Josh Blum <josh@joshknows.com> Co-authored-by: Jonathon Pendlum <jonathan.pendlum@ettus.com> Co-authored-by: Martin Braun <martin.braun@ettus.com> Co-authored-by: Matt Ettus <matt@ettus.com> Co-authored-by: Michael West <michael.west@ettus.com> Co-authored-by: Moritz Fischer <moritz.fischer@ettus.com> Co-authored-by: Nick Foster <nick@ettus.com> Co-authored-by: Nicolas Cuervo <nicolas.cuervo@ettus.com> Co-authored-by: Paul Butler <paul.butler@ni.com> Co-authored-by: Paul David <paul.david@ettus.com> Co-authored-by: Ryan Marlow <ryan.marlow@ettus.com> Co-authored-by: Sugandha Gupta <sugandha.gupta@ettus.com> Co-authored-by: Sylvain Munaut <tnt@246tNt.com> Co-authored-by: Trung Tran <trung.tran@ettus.com> Co-authored-by: Vidush Vishwanath <vidush.vishwanath@ettus.com> Co-authored-by: Wade Fife <wade.fife@ettus.com>
Diffstat (limited to 'fpga/usrp3/lib/fifo/axi_packet_gate.v')
-rw-r--r--fpga/usrp3/lib/fifo/axi_packet_gate.v229
1 files changed, 229 insertions, 0 deletions
diff --git a/fpga/usrp3/lib/fifo/axi_packet_gate.v b/fpga/usrp3/lib/fifo/axi_packet_gate.v
new file mode 100644
index 000000000..771fb2200
--- /dev/null
+++ b/fpga/usrp3/lib/fifo/axi_packet_gate.v
@@ -0,0 +1,229 @@
+//
+// Copyright 2012 Ettus Research LLC
+// Copyright 2018 Ettus Research, a National Instruments Company
+//
+// SPDX-License-Identifier: LGPL-3.0-or-later
+//
+// Description:
+// Holds packets in a FIFO until they are complete. This allows buffering
+// slowly-built packets so they don't clog up downstream logic. If o_tready
+// is held high, this module guarantees that o_tvalid will not be deasserted
+// until a full packet is transferred. This module can also optionally drop
+// a packet if the i_terror bit is asserted along with i_tlast. This allows
+// discarding packet, say, if a CRC check fails.
+// NOTE:
+// - The maximum size of a packet that can pass through this module is
+// 2^SIZE lines. If a larger packet is sent, this module will lock up.
+// - Assuming that upstream is valid and downstream is ready, the maximum
+// in to out latency per packet is (2^SIZE + 2) clock cycles.
+// 2^SIZE because this module gates a packet, 1 cycle for the RAM read and
+// 1 more cycle for the output register. This is not guaranteed behavior though.
+// - The USE_AS_BUFF parameter can be used to treat this packet gate as
+// a multi-packet buffer. When USE_AS_BUFF=0, the max number of packets
+// (regardless of size) that the module can store is 2. When USE_AS_BUFF=1,
+// the entire storage of this module can be used to buffer packets but at
+// the cost of some additional RAM. Beware the sequence of (big packet,
+// small packet, small packet), as some outside buffering may be needed
+// to handle this case if USE_AS_BUFF=0.
+
+module axi_packet_gate #(
+ parameter WIDTH = 64, // Width of datapath
+ parameter SIZE = 10, // log2 of the buffer size (must be >= MTU of packet)
+ parameter USE_AS_BUFF = 0, // Allow the packet gate to be used as a buffer (uses more RAM)
+ parameter MIN_PKT_SIZE= 0 // log2 of minimum valid packet size (rounded down, used to reduce addr fifo size)
+) (
+ input wire clk,
+ input wire reset,
+ input wire clear,
+ input wire [WIDTH-1:0] i_tdata,
+ input wire i_tlast,
+ input wire i_terror,
+ input wire i_tvalid,
+ output wire i_tready,
+ output reg [WIDTH-1:0] o_tdata = {WIDTH{1'b0}},
+ output reg o_tlast = 1'b0,
+ output reg o_tvalid = 1'b0,
+ input wire o_tready
+);
+
+ localparam [SIZE-1:0] ADDR_ZERO = {SIZE{1'b0}};
+ localparam [SIZE-1:0] ADDR_ONE = {{(SIZE-1){1'b0}}, 1'b1};
+
+ // -------------------------------------------
+ // RAM block that will hold pkts
+ // -------------------------------------------
+ wire wr_en, rd_en;
+ wire [WIDTH:0] wr_data, rd_data;
+ reg [SIZE-1:0] wr_addr = ADDR_ZERO, rd_addr = ADDR_ZERO;
+
+ // Threshold to explicitly instantiate LUTRAM
+ localparam LUTRAM_THRESH = 5;
+
+ // We need to instantiate a simple dual-port RAM here so
+ // we use the ram_2port module with one read port and one
+ // write port and "NO-CHANGE" mode.
+ ram_2port #(
+ .DWIDTH (WIDTH+1), .AWIDTH(SIZE),
+ .RW_MODE("NO-CHANGE"), .OUT_REG(0),
+ .RAM_TYPE(SIZE <= LUTRAM_THRESH ? "LUTRAM" : "AUTOMATIC")
+ ) ram_i (
+ .clka (clk), .ena(1'b1), .wea(wr_en),
+ .addra(wr_addr), .dia(wr_data), .doa(),
+ .clkb (clk), .enb(rd_en), .web(1'b0),
+ .addrb(rd_addr), .dib({WIDTH{1'b0}}), .dob(rd_data)
+ );
+
+ // FIFO empty/full logic. The condition for both
+ // empty and full is when rd_addr == wr_addr. However,
+ // it matters if we approach that case from the low side
+ // or the high side. So keep track of the almost empty/full
+ // state for determine if the next transaction will cause
+ // the FIFO to be truly empty or full.
+ reg ram_full = 1'b0, ram_empty = 1'b1;
+ wire almost_full = (wr_addr == rd_addr - ADDR_ONE);
+ wire almost_empty = (wr_addr == rd_addr + ADDR_ONE);
+
+ always @(posedge clk) begin
+ if (reset | clear) begin
+ ram_full <= 1'b0;
+ end else begin
+ if (almost_full) begin
+ if (wr_en & ~rd_en)
+ ram_full <= 1'b1;
+ end else begin
+ if (~wr_en & rd_en)
+ ram_full <= 1'b0;
+ end
+ end
+ end
+
+ always @(posedge clk) begin
+ if (reset | clear) begin
+ ram_empty <= 1'b1;
+ end else begin
+ if (almost_empty) begin
+ if (rd_en & ~wr_en)
+ ram_empty <= 1'b1;
+ end else begin
+ if (~rd_en & wr_en)
+ ram_empty <= 1'b0;
+ end
+ end
+ end
+
+ // -------------------------------------------
+ // Address FIFO
+ // -------------------------------------------
+ // The address FIFO will hold the write address
+ // for the last line in a non-errant packet
+
+ wire [SIZE-1:0] afifo_i_tdata, afifo_o_tdata, afifo_p_tdata;
+ wire afifo_i_tvalid, afifo_i_tready;
+ wire afifo_o_tvalid, afifo_o_tready;
+ wire afifo_p_tvalid, afifo_p_tready;
+
+ axi_fifo #(.WIDTH(SIZE), .SIZE(USE_AS_BUFF==1 ? SIZE-MIN_PKT_SIZE : 1)) addr_fifo_i (
+ .clk(clk), .reset(reset), .clear(clear),
+ .i_tdata(afifo_i_tdata), .i_tvalid(afifo_i_tvalid), .i_tready(afifo_i_tready),
+ .o_tdata(afifo_p_tdata), .o_tvalid(afifo_p_tvalid), .o_tready(afifo_p_tready),
+ .space(), .occupied()
+ );
+
+ axi_fifo #(.WIDTH(SIZE), .SIZE(1)) addr_fifo_pipe_i (
+ .clk(clk), .reset(reset), .clear(clear),
+ .i_tdata(afifo_p_tdata), .i_tvalid(afifo_p_tvalid), .i_tready(afifo_p_tready),
+ .o_tdata(afifo_o_tdata), .o_tvalid(afifo_o_tvalid), .o_tready(afifo_o_tready),
+ .space(), .occupied()
+ );
+
+ // -------------------------------------------
+ // Write state machine
+ // -------------------------------------------
+ reg [SIZE-1:0] wr_head_addr = ADDR_ZERO;
+
+ assign i_tready = ~ram_full & afifo_i_tready;
+ assign wr_en = i_tvalid & i_tready;
+ assign wr_data = {i_tlast, i_tdata};
+
+ always @(posedge clk) begin
+ if (reset | clear) begin
+ wr_addr <= ADDR_ZERO;
+ wr_head_addr <= ADDR_ZERO;
+ end else begin
+ if (wr_en) begin
+ if (i_tlast) begin
+ if (i_terror) begin
+ // Incoming packet had an error. Rewind the write
+ // pointer and pretend that a packet never came in.
+ wr_addr <= wr_head_addr;
+ end else begin
+ // Incoming packet had no error, advance wr_addr and
+ // wr_head_addr for the next packet.
+ wr_addr <= wr_addr + ADDR_ONE;
+ wr_head_addr <= wr_addr + ADDR_ONE;
+ end
+ end else begin
+ // Packet is still in progress, only update wr_addr
+ wr_addr <= wr_addr + ADDR_ONE;
+ end
+ end
+ end
+ end
+
+ // Push the write address to the address FIFO if
+ // - It is the last one in the packet
+ // - The packet has no errors
+ assign afifo_i_tdata = wr_addr;
+ assign afifo_i_tvalid = ~ram_full & i_tvalid & i_tlast & ~i_terror;
+
+ // -------------------------------------------
+ // Read state machine
+ // -------------------------------------------
+ reg rd_data_valid = 1'b0;
+ wire update_out_reg;
+ // Data can be read if there is a valid last address in the
+ // address FIFO (signifying the end of an input packet) and
+ // if there is data available in RAM
+ wire ready_to_read = (~ram_empty) & afifo_o_tvalid;
+ // Pop from address FIFO once we have see the end of the pkt
+ assign afifo_o_tready = rd_en & (afifo_o_tdata == rd_addr);
+
+ // Read from RAM if
+ // - A full packet has been written AND
+ // - Output data is not valid OR is currently being transferred
+ assign rd_en = ready_to_read & (update_out_reg | ~rd_data_valid);
+
+ always @(posedge clk) begin
+ if (reset | clear) begin
+ rd_data_valid <= 1'b0;
+ rd_addr <= ADDR_ZERO;
+ end else begin
+ if (update_out_reg | ~rd_data_valid) begin
+ // Output data is not valid OR is currently being transferred
+ if (ready_to_read) begin
+ rd_data_valid <= 1'b1;
+ rd_addr <= rd_addr + ADDR_ONE;
+ end else begin
+ rd_data_valid <= 1'b0; // Don't read
+ end
+ end
+ end
+ end
+
+ // Instantiate an output register to break critical paths starting
+ // at the RAM module. When ram_2port is inferred as BRAM, the tools
+ // should absorb this register into the BRAM block without using
+ // SLICE resources.
+ always @(posedge clk) begin
+ if (reset | clear) begin
+ o_tvalid <= 1'b0;
+ end else if (update_out_reg) begin
+ o_tvalid <= rd_data_valid;
+ {o_tlast, o_tdata} <= rd_data;
+ end
+ end
+ // Update the output reg only *after* the downstream
+ // block has consumed the current value
+ assign update_out_reg = o_tready | ~o_tvalid;
+
+endmodule