[widgetii]

Follow-up to #130, covering the two nits from its review.

Changes

• CV6xxBootParts.first_table_offset replaces offset_32. build_ddr_table() used to read the DDR header's structure_length from file_data again to rebuild where table 0 starts, even though parse_cv6xx_boot() had already validated that. The parser now keeps the validated offset, so the two can't drift. offset_32 had no other users; the browser parser already kept tableOffset.

• The parser's bounds are documented with the layouts they were set against, measured on the published images:

  Image	GSL header	GSL → DDR header	Board map
  boot-hi3516cv608-nor.bin, boot-hi3516cv610-*-nor.bin	0x800	0x400	8 bytes, 1 table
  boot-hi3519dv500-dmebpro-nor.bin	0x1200	0x100	8 bytes, 1 table

  MAX_GSL_HEADER_OFFSET (64 KiB) and MAX_SECTION_GAP (4 KiB) leave room past both layouts. MAX_DDR_TABLES = 8 because the board map has 8 entries, so a ninth table could never be selected. The browser constants point at the Python comment.

Verified

• No behaviour change: GSL, U-Boot and build_ddr_table() output are byte-identical to master for all six published images and board IDs 0–9.
• pytest tests/ --ignore=tests/fuzz: 1097 passed, 3 skipped. Fuzz: 16 passed.
• ruff and mypy are clean; node --test: 95 passed.

[qodo-free-for-open-source-projects]

PR Summary by Qodo

CV6xx: Reuse the validated DDR table offset and document parser bounds

✨ Enhancement 📝 Documentation 🕐 10-20 Minutes

AI Description

• Preserve the parser-validated first-table offset instead of rereading the DDR header during table
 construction.
• Document the published image layouts behind parser bounds and link the browser's matching
 constants.

Diagram

graph TD
  Image["CV6xx image"] --> Parser["Python parser"] --> Parts["Validated offsets"] --> Builder["DDR builder"] --> Table["Selected table"]
  Image --> Browser["Browser parser"]
  Parser -.->|bound rationale| Browser

Loading

High-Level Assessment

Keep the validated offset in the parsed parts: it removes a redundant header read without changing table selection. Recomputing or revalidating it in the builder would duplicate parser logic without a clear benefit.

Files changed (2) +26 / -11

Refactor (1) +24 / -11

hisilicon_cv6xx.pyCarry the validated first-table offset into DDR table construction +24/-11

Carry the validated first-table offset into DDR table construction

• Replaces the unused intermediate offset with the parser-validated first-table offset, which build_ddr_table() now uses without rereading the header. Comments explain the published layouts, field sizes, and board-map limits behind the existing parser bounds.

src/defib/protocol/hisilicon_cv6xx.py

Documentation (1) +2 / -0

protocol.jsLink browser parser bounds to their Python rationale +2/-0

Link browser parser bounds to their Python rationale

• Adds a comment identifying the matching Python constants as the source of the bounds rationale and asking that the two sets change together. Browser parsing behavior is unchanged.

web/protocol.js

[qodo-free-for-open-source-projects]

Code Review by Qodo

🐞 Bugs (0) 📘 Rule violations (0) 📎 Requirement gaps (0)

Great, no issues found!

Qodo reviewed your code and found no material issues that require review

Tip of the day

💡 Did you know, you can enable the Remediation agent and Qodo fixes findings in a dedicated fix PR

More tips ↗ | Customize Qodo ↗ | Qodo docs ↗