Skip to content

Code Generation

The messy.py script uses a JSON configuration file to generate the SystemC/SystemC-AMS models for the simulation. This document explains the structure of the JSON file and the code generation process.

JSON Configuration File

The JSON file defines the entire system to be simulated. Here is an example of a simple configuration:

{
    "resolution": "SC_MS",
    "tracing": {},
    "bus": {
        "vref": 3.3
    },
    "core": {
        "vref": 1.8,
        "iss": "gvsoc",
        "config": {
            "path": "./gvsoc_config.json",
            "axi_path": "/chip/axi_proxy"
        }
    },
    "peripherals": {
        "sensors": {
            "mic_click": {}
        },
        "harvesters": {
            "battery": {},
            "photovoltaic": {}
        }
    }
}

Top-Level Properties

  • resolution: The simulation time resolution. See Resolution for more details.
  • tracing: Tracing configuration. See Tracing for more details.
  • bus: The power bus configuration. See Power Bus for more details.
  • core: The RISC-V core configuration. See Core for more details.
  • peripherals: The peripherals connected to the core, such as sensors and harvesters.

Peripherals

The peripherals object contains two sub-objects:

  • sensors: A list of sensors to be included in the simulation. See Sensors for more details.
  • harvesters: A list of harvesters to be included in the simulation. See Harvesters for more details.

Code Generation Process

The messy.py script reads the JSON configuration file and generates the SystemC/SystemC-AMS models based on the templates located in the messy/codegen/templates directory. The generated code is placed in the messy/codegen/src and messy/codegen/include directories.

The generated code is then compiled and linked with the user application to create the final simulation executable.