aoeldemann / netfpga_sume_loopback

NetFPGA-SUME network interface loopback design

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

NetFPGA SUME Network Interface Loopback Design

A hardware design for the NetFPGA SUME, which loops-back the AXI-Stream RX and TX interfaces of the Ethernet MACs. Comes in two flavors: cut-through and store-and-forward.

Cut-Through Design

Location: projects/cut_through/

The AXI-Stream RX interfaces are directly connected to the TX interfaces.

Store-and-Forward Design

Location: projects/store_and_forward

Frames received on the AXI-Stream interfaces from the MAC are placed in a small FIFO and only transmitted back for tranmission on the network when comeplely received. Frames, which are invalid (e.g., too small, bad FCS), are discarded.

Loopback Connections

  • Port 0 RX -> Port 1 TX
  • Port 1 RX -> Port 0 TX
  • Port 2 RX -> Port 3 TX
  • Port 3 RX -> Port 2 TX

Xilinx Vivado Version

Design implemented and tested using Xilinx Vivado 2019.1.

Bitstream Generation

Step 1: Build IP Cores

make -C ip/

Step 2: Synthezise and Implement

Change into one of the project directory:

cd projects/cut_through/

or

cd projects/store_and_forward/

Then create the project, perform synthesis and finally implementation:

make project && make synth && make impl

Determine the location of the generated bitstream:

find . -name "*.bit"

About

NetFPGA-SUME network interface loopback design


Languages

Language:Tcl 61.7%Language:Verilog 19.4%Language:Makefile 9.8%Language:SystemVerilog 9.1%