L-VSX-500 (L07) - the carved SINTRAN data home¶

This folder IS the carved data for SINTRAN III VSX/500 L-version (L07). Everything cut out of the L07 SMD disk image - the paged system segments, the memory-resident images, the reverse-engineering tree, and the exact inputs used to produce them - lives here. It is the folder to READ from when you need L07 bytes, and the folder to EXTEND when you carve more.

Sibling versions: ../K-VSX-500/ (K05) and ../M-VSX-500/ (M06). See ../README.md for the version index.

The procedure that produced everything here is documented once, authoritatively, in: - ../../EXTRACTING-SEGMENTS.md - the paged system segments (the reconcile.py -> carve.py pipeline). - ../../EXTRACTING-RESIDENT-CODE.md - the memory-resident code/data the segment carver cannot produce. - ../../README.md - the carver overview and the one-line pipeline.

Do NOT re-derive the page/carve model - it is settled and proven against the raw disk in those docs.

What is in this folder¶

Path	What it is
`segments/`	The carved paged system segments: one big-endian `NNN-<NAME>.bin` + `NNN-<NAME>.meta.json` sidecar (carries the Ghidra base) per segment, plus `manifest.json` (all segments in one file) and `README.md` (the human-readable per-segment catalog: load address, pages, MADR, content type). 79 segment `.bin` files. START HERE for segment bytes.
`resident/`	The memory-resident images the segment carver skips (`madr==0`): `SINTRAN-DATA_commoncode.bin` (resident common code, Ghidra base `0x0`), `MACM-AREA-DATA_rtloader.bin` (RT-loader area, base `0x3000`), plus a disassembly. Pulled by raw NDFS block range - see `../../EXTRACTING-RESIDENT-CODE.md`.
`re/`	The L07 reverse-engineering tree: `mon-analysis/` (156 per-MON-call folders), `segments-ref/` (promoted, verified per-segment analyses), `MON-CALL-INDEX.md` (the authoritative MON index), `ND500-SYSTEM-MONITOR/`, `ND500-HANDLERS-OVERLAY.md`, disassemblies (`.dis`), Ghidra symbol dumps (`.ghidra-symbols.txt`), and handoffs. See `re/README.md`.
`inputs/`	The exact facts fed to the carver: `list-segment.txt` (live `@RT-LOADER LIST-SEGMENT`), `list-rt-programs.txt` (validation), `segment-table-bank3.bin` (the in-memory Segment Table dump giving MADR/SEGLE). Re-carving reads these.
`segment-facts.json`	The canonical reconciled facts (`reconcile.py` output) that `carve.py` consumes.
`section83-from-manual.json`	Release-manual section 8.3 ranges, parsed.
`json-discrepancies.txt`	Where the AI-derived segment JSON disagreed with the live/manual facts.
`mon-calls/`	Empty placeholder. Live MON-call analysis is under `re/mon-analysis/`.

How to EXTEND this (carve more L07 data)¶

A missing paged segment, or a re-carve: run the pipeline in ../../EXTRACTING-SEGMENTS.md (sections 4-6) using this folder's inputs/ and segment-facts.json; carve.py writes into segments/ and refreshes manifest.json + the per-segment meta.json.
Resident code/data (not in segments/): pull by raw NDFS block range per ../../EXTRACTING-RESIDENT-CODE.md into resident/. Resident DATA cells read zero on disk - capture them LIVE over DAP, do not disassemble them out of a carve.
A disassembly / analysis: write it into re/ (a promoted, verified segment analysis goes to re/segments-ref/NNN-<NAME>/; MON-call work to re/mon-analysis/<CODE>B-<NAME>/). Record the nd100-dis -b base and byte order used; tag every claim [V]/[I]/[OPEN]; name which symbol table an address came from.
Update the status-of-record doc in the SAME change - ../../../../SINTRAN/CARVING-HANDOFF.md for ND-100 work, ../../../../SINTRAN/ND500/ND500-STATUS-AND-INDEX.md for ND-500 work - and index the new file there. An un-indexed carve is invisible to the next session.

Byte-order reminder: every .bin here is big-endian (native ND-100), ready for a big-endian Ghidra processor as-is. Byte-swap ONLY for nd100-dis (it reads little-endian). ND-500 segments (030-S3SM5, 062-S3SSM5) are 32-bit byte-addressed - do NOT load them with the ND-100 processor.

Parent: ../README.md