Gatchan/FF6Portage
Watch
1
0
Fork
You've already forked FF6Portage
 0
Code Issues 16 Pull requests Projects Releases Packages Wiki Activity Actions

Frame model: per-port coroutines, first frame-waiting ports (#37) #65

Merged
Gatchan merged 4 commits from frame-model into main 2026-11-06 04:47:55 +00:00
Conversation 0 Commits 4 Files changed 16 +711 -17
Gatchan commented 2026-11-06 04:47:33 +00:00
Owner

Frame model for ported code (#37): tm_wait_frame (core/platform/timing.h) is the single yield point; a port flagged PORT_WAITS_FRAMES runs its bridge in a libco coroutine (core/port/port_frame.c). At a wait the machine idles in WAI at the routine's address, runs its interrupt handlers, and the C resumes once the NMI handler has returned there. Mode on: the real machine, resumed from the instruction hook. Shadow: the clone follows the original cycle by cycle with its pad state; the call is compared when the original returns (same state, same number of vblank interrupts). A savestate is refused while a port in mode on waits.

First ports: field/WaitVblank and btlgfx/WaitVblank.

Found on the way: #64 (LakeSnes keeps the old FastROM access table after snes_loadState), worked around in core/hw.

Validation:

  • shadow, both ports alone, the 4 chapter 1 routes, US+JP once each: btlgfx 39108/39108 US, 37025/37025 JP; field 46586/46589 US, 42888/42891 JP (one call per route still waiting at the movie's end); 0 mismatch, 0 fallback; movies match
  • ff6-portcheck.sh all: PASS, 136 checks US+JP (WaitVblank 52 + 65 states per version)
  • debug ctest -LE slow: 27/27

Closes #37.

🤖 Generated with Claude Code

🤖 Generated with Claude Code

Frame model for ported code (#37): `tm_wait_frame` (core/platform/timing.h) is the single yield point; a port flagged `PORT_WAITS_FRAMES` runs its bridge in a libco coroutine (core/port/port_frame.c). At a wait the machine idles in WAI at the routine's address, runs its interrupt handlers, and the C resumes once the NMI handler has returned there. Mode on: the real machine, resumed from the instruction hook. Shadow: the clone follows the original cycle by cycle with its pad state; the call is compared when the original returns (same state, same number of vblank interrupts). A savestate is refused while a port in mode on waits. First ports: field/WaitVblank and btlgfx/WaitVblank. Found on the way: #64 (LakeSnes keeps the old FastROM access table after snes_loadState), worked around in core/hw. Validation: - shadow, both ports alone, the 4 chapter 1 routes, US+JP once each: btlgfx 39108/39108 US, 37025/37025 JP; field 46586/46589 US, 42888/42891 JP (one call per route still waiting at the movie's end); 0 mismatch, 0 fallback; movies match - ff6-portcheck.sh all: PASS, 136 checks US+JP (WaitVblank 52 + 65 states per version) - debug ctest -LE slow: 27/27 Closes #37. 🤖 Generated with [Claude Code](https://claude.com/claude-code) 🤖 Generated with [Claude Code](https://claude.com/claude-code) 
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Sign in to join this conversation.
Reviewers
No reviewers
Labels
Clear labels   
bug
Something produces a wrong result

  
jp
Japanese version specific

  
perf
Speed or resource use

  
port
Hand port of original routines

  
resources
Resource pipeline and pack

  
task
Planned work item

  
tools
Python tools, extract/pack, generators

  
upstream
Defect in a reference or third-party component (disassembly, LakeSnes)

  
validation
Oracles, shadow, tests, routes

No labels
bug
jp
perf
port
resources
task
tools
upstream
validation
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference
Gatchan/FF6Portage!65
No description provided.