Difference between revisions of "SystemVerilog formatter for our LowRISC-based guidelines (2-3G)"
From iis-projects
Line 6: | Line 6: | ||
[[Category:Tbenz]] | [[Category:Tbenz]] | ||
[[Category:Available]] | [[Category:Available]] | ||
+ | [[Category:Hot]] | ||
== Introduction == | == Introduction == |
Revision as of 19:35, 2 November 2020
Introduction
Coding conventions are an essential part of every large hardware (and software) project. They guide various non-functional aspects of the project and the source code like file organization, indentation, declarations, code alignment, naming conventions, and countless others.
Coding conventions serve very important purposes:
- Improving maintenance by making the code more familiar to the maintainers
- Improve readability and ease the process of code reviewing
- Improving the overall code quality by giving the project a more complete and clean feel
- Ensuring better compatibility to different tools by encouraging the use of simpler and better-supported constructs
At IIS we use the LowRISC guidelines [1] for systemverilog. Although the effort of following the guidelines continuously while writing new modules is very manageable, it becomes a major hassle when including 3rd-party and ill-formatted into our projects.
Having a tool that automatically formats any systemverilog source code to apply to the coding guidelines of LowRISC would massively reduce the manual refactoring effort while integrating 3rd-party code.
Project Content
The project can be divided in the following sub tasks:
- Abstract the coding guidelines to a set of machine-understandable rules
- Explore existing parsing frameworks for systemverilog to simplify the processing of source code (or choose to create the parser from scratch)
- Use python to check a given source files against each rule and report the line or construct in violation with the rule
- Create a python tool that applies the rules to the source code and produces nicely formatted code
Prerequisites
- Preferably: Experience with Python
- Preferably: Experience with a hardware description language (HDL)
Composition
- 20% rule abstraction
- 10% parser framework exploration
- 30% formatting violation reporting
- 40% implementing formatting script