Files
psxrecomp/docs
Claude c0ea49832a cdrom: fix 0x1F801800 status bits per oracle (spurious ADPBUSY, missing BUSYSTS)
Bit 2 is ADPBUSY (XA-ADPCM playback in progress) and must idle at 0; we
raised it unconditionally under a mistaken 'ADPCM empty' reading. Bit 7
(BUSYSTS: command written, first response not yet ready) was never raised;
now mapped to the queued-command window, mirroring beetle's
PendingCommandCounter/phase busy window. Verified against beetle cdc.cpp
Read(A==0), which sets bit 2 never and bit 7 during command latency.

Boot-diff verification shows this does NOT change our boot command stream
(the early-boot branch divergence vs the oracle persists — documented in
docs/kula-world-findings.md round 4, with the next sharp step: add guest
PC to the oracle's cdcmd hook and diff the first divergent call site).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BdjxoRB1f9VQsNzZayCgGS
2026-07-02 20:18:24 +00:00
..
2026-05-20 12:31:11 -07:00