docs: complete data-templates.adoc with CSR and instruction examples

[saurabh12nxf]

Problem

The doc/data-templates.adoc file is currently a TODO placeholder (only 4 lines), but it's linked from the README.adoc as a "How-to/recipe" guide (line 27).

New contributors who click this link expecting guidance on creating UDB data files find only "TODO", which creates a poor user experience and slows down the contribution process.

Impact

Current state:

• New contributors have no guidance on UDB data structure
• No examples of how to create CSR definitions
• No examples of how to create instruction definitions
• No examples of how to create extension definitions
• Slows down onboarding and contribution process

Affected users:

• New contributors trying to add CSRs
• Developers migrating from riscv-opcodes
• Anyone trying to understand UDB data format

Proposed Solution

Complete doc/data-templates.adoc with comprehensive documentation including:

1. CSR Definition Template

• Real example based on a simple CSR (using mstatus fields as reference)
• Explanation of each field (location, long_name, description, type(), reset_value())
• How to handle RV32 vs RV64 differences (location_rv32 vs location_rv64)
• How to specify extension dependencies using definedBy
• Examples of different field types (RO, RW, WARL, etc.)

2. Instruction Definition Template

• Basic instruction structure
• Encoding specification
• Assembly format
• Link to existing examples in spec/std/isa/inst/

3. Extension Definition Template

• How to define a new extension
• Required fields and their meanings
• Link to existing examples in spec/std/isa/ext/

4. Common Patterns

• Config-dependent fields
• Multi-XLEN support (RV32/RV64/RV128)
• Parameter definitions
• IDL function usage

5. Validation Steps

• How to validate new data files
• Running schema validation
• Running tests

Context

I am a prospective LFX Spring '26 mentee for the "AI-assisted extraction of architectural parameters from RISC-V specifications" project.

I've studied the UDB structure extensively while developing an LLM-based parameter extraction proof-of-concept:

• Analyzed mstatus.yaml (627 lines, 22 fields)
• Studied misa.yaml and parameter definitions
• Tested LLM extraction achieving 98% accuracy

I believe I can create comprehensive, beginner-friendly documentation based on my understanding of the UDB data structure.

Proposed Changes

• Complete doc/data-templates.adoc with ~200-300 lines of documentation
• Include real, working examples from existing UDB data
• Add clear explanations suitable for new contributors
• Link to relevant schema documentation
• Include validation workflow

Acceptance Criteria

• File contains real, working examples (not just placeholders)
• Examples are well-commented and explained
• Covers CSR, instruction, and extension templates at minimum
• Includes validation steps
• Links to relevant schema documentation (schemas.adoc, idl.adoc)
• Renders correctly as AsciiDoc
• Helps new contributors understand UDB data structure

Questions

1. Should I include examples for all data types (CSR, instruction, extension, parameter) or focus on CSRs first?
2. Should I include IDL function examples or keep it simple?
3. Any specific CSR you'd prefer as the main example besides mstatus?

I'm happy to adjust the scope based on maintainer feedback.