Salaga-RV

Salaga family of RISC-V hardware and some GUI software to run on it!

The purpose of this project is to improve my systems knowledge by implementing what I learnt in USC's EE356, EE457, and EE402 courses by Prof. Marco Paolieri, Prof. Gandhi Puvvada, and Prof. Bill Cheng respectivly.

I want to setup a hardware-software stack where I can run a bare-metal-C GUI program on the hardware I wrote!

Thanks to Bruno Levy's learn-fpga repository. It has been a huge help! It has most of the thing one needs to learn not just about logic-design for FPGAs, but writing bare-metal software as well.

Tools required

I use Ubuntu, and the following instructions should be similar for other UNIX-like systems.

For now, simulation is done using iverilog. However, I soon plan on adding support for verilator to speed up simulation time.

1) Installing iverilog

sudo apt install iverilog

2) Install RISC-V GNU Toolchain (Not required if you plan on just running code under the tests directory)

Follow instructions given at RISC-V GNU Toolchain GitHub repository. You will need the riscv64-unknown-elf-* binaries to be installed to compile software for Salaga-RV.

Try it out!

Clone the repository

git clone https://github.com/ndyashas/Salaga-RV.git

cd Salaga-RV

1) Running examples

Examples are split into two categories.

• gui
• no-gui

Go into the examples directory

cd examples

Examples with GUI output

cd no-gui

Code under gui are examples where the processor supports display functions such as setting color to a pixel, or drawing a rectangle!. Although the drivers are not very realistic, I felt the way it is implemented now is a good stage before I jump into a more realistic driver implementation.

Eg: To run the simple display-testing program, run the 001_display_test example as follows

cd 001_display_test

make Jala

The above command will compile the simulation model, and generate an executable. To run the simulation,

./obj_dir/Vtb

Examples with no GUI output

cd no-gui

Code under no-gui are examples where one can use functions such as putchar! (I'm yet to add support printf). The putchar function uses a memory-mapped UART port. The UART output is captured in the simulation and printed onto the console. This procedure is neatly explained in Bruno Levy's project here. Although the pinned_array is not required here, I have kept it nevertheless.

The procedure to run the examples are the same as in no-io

Eg: Let us simulate running the print_zig_zag example on the Jala core.

cd 002_print_zig_zag

make Jala

This should print out soomething as follows on your terminal

         
  *       
    *     
      *   
        * 
          
        * 
      *   
    *     
  *       
*         
  *       
    *     
      *   
        * 

The program completed in       41108 cycles

2) Running tests:

Go into the tests directory

cd tests

Run all tests at once

You can run all the tests at once on a given core by executing the run_unit_tests.sh script passing one of the processor cores as arguments.

Eg: To run all test cases against the Eka core, run the following command. You can test against the Jala core as well.

./run_unit_tests.sh Eka

*NOTE that Eka is passed as an argument to the run_unit tests.sh script to run the test against the Eka core.

Running individual tests

If a test fails, you may go into the directory of that specific test and run it.

Eg: If a test such as the tests/010_test_sw-junior_lw-senior2 fails for Jala, you can do the following to see why it failed

cd 010_test_sw-junior_lw-senior2

./run_test.sh Jala

*NOTE that Jala is passed as an argument to the run_test.sh script to run the test against the Jala core.