The BIOS exception handler ran interpreted for the life of every process,
1.22 billion instructions and 44.6% of all interpreted work in Breath of
Fire III, even though it is compiled and even though the stub that
unblessed it sat at exactly the address bios/SCPH1001.toml already
declared as an install slot. Two mechanisms existed and did not compose:
the emitter hooked the declared slot PC, while the runtime's bless
verifier memcmp'd the whole body including that slot, so the body was
MISMATCH forever and native code never reached the hook.
Three parts, none of which pays off alone:
1. Publish the declared ranges to the runtime. The emitter emits a
PsxKernelPatchRange table next to the kernel body table and routes it
through the backend descriptor; the bless verifier compares a body in
segments that skip them. CLEAN now means "every byte outside every
declared range still matches ROM", which is the claim the native code
actually depends on.
2. Generalise the slot model to what the patches are. The old model
assumed a 4-word lui/addiu/jalr/nop stub written over ROM zeros and
fired on `word != 0`. Measured against both images, only one of five
observed shapes fits: patchers also plant a `jr` over real ROM
instructions, rewrite an 11-word prologue and insert an `mfc0 Cause`,
and NOP out a driver routine. A slot is now a range with a length and
an explicit `resume` (jalr / fallthrough / none), detected by
comparing every word against its ROM-baked value.
3. Hand back from the interpreter at the range end. Kernel page 0 is
permanently dirty, because the handler saves registers there on every
exception, so straight-line interpretation would run the whole
function after the patch. The emitter registers the resume PC as a
continuation key and dirty_ram_interp surfaces there.
A PC inside a declared range must never be a native dispatch key: the
compiled bytes there are not what executes. Seven pre-existing
jal-return continuations violated that and wedged the boot at frame 0 in
an infinite dispatch loop (OpenBIOS RAM 0x357C). The emitter now drops
continuation keys inside a range so dispatch misses through to the
interpreter, and refuses to emit if a range covers a function entry,
which would silently remove a callable function instead.
Both profiles declare their measured ranges, and SCPH1001.toml finally
pins [program.image] sha256 to v2.2 (CRC32 37157331) so a v2.0 dump
refuses to generate rather than producing phantom discovery gaps.
Measured on Breath of Fire III (SLPS-00990, Psy-Q SDK 3.70), headless,
matched frames, PSX_KERNEL_PATCH_RANGES=0 for the OFF side:
OpenBIOS SCPH-1001 v2.2
bless mismatch 21 -> 0 74 -> 0
kernel bodies -87.7%/frame -97.8%/frame
all interp insns -42.6%/frame -94.2%/frame
The residual kernel-body work is exactly declared-words x entries on
every range, so nothing leaks past the hook. The A0/B0/C0 vector
trampolines are unchanged (+0.2%, noise): the profile excludes them by
design and they are a separate job. Card read traces are byte-identical
across the A/B on both images, all 32 entries, which exercises the two
ranges that replace the kernel's card handler.
Rule 18 holds throughout: the guest's patched instructions still execute
as written on the dirty-RAM interpreter. Only the code around them
changes backend. Nothing is synthesised.
Tests: runtime/tests/test_kernel_patch_ranges.c (new, the verifier's
accept/reject behaviour and the hand-back predicate) and the install-slot
range invariants in recompiler/tests/bios_address_model_test.cpp.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The launcher and studio ship as Retro Launcher and Retro Studio now,
from the RetroPortingToolKit org, so the old product name is gone from
display text, documentation and release asset filenames.
The split is by case, and it is not cosmetic. CamelCase RetComM was
always the product; lowercase retcomm and RETCOMM_ are identifiers --
the C++ namespace, the binaries, the data directories, the CMake and
environment variables, the Cloudflare service names, the readme markers.
Renaming those would orphan existing installs for no benefit, so they
are untouched: 1101 lowercase retcomm and 205 RETCOMM_ still stand in
the launcher alone.
Four things are deliberately still spelled RetComM, and three of them
would have been real breakage:
- RetComM-Data is CamelCase but it is a directory on disk --
<exe_dir>\RetComM-Data\{runtime,config,data} for a portable install.
Renaming it strands every portable user's config, toolchains and
catalog cache.
- TechnicallyComputers/RetComM-Launcher in readme_has_launcher() is the
historical half of a both-slugs test. Renaming it makes it a duplicate
of the line above and sends every port built before the transfer back
to reporting its launcher section missing.
- _LEGACY_LAUNCHER_RE matched "^## RetComM Launcher". A sweep renamed it,
which would have silently stopped it matching every README already out
there; it now accepts both headings.
- The portable-stub lookup gained the pre-rename filenames back as
fallbacks, because old zips stay on the releases page and a rollback
downloads one.
Release assets become Retro-Launcher-* / Retro-Studio-*. That is safe
for installed clients: pick_launcher_asset() gates on *linux*/*windows*/
*macos* plus extension and arch, not on the product name, and the
Windows portable stub is identified by its RCM1 trailer rather than its
filename.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013DpaQqVguHFpBuCmdoFYWY
recomp-net, retcomm-rbengine (now rbengine), retcomm-toolchains (now
RetroPorting-Toolchains), the catalog and the launcher moved from
TechnicallyComputers to the RetroPortingToolKit org. GitHub still
redirects the old repo URLs, so nothing is broken today — but this repo
is what every port inherits its .gitmodules from, so the old owner
propagates into each new clone until it is fixed here.
.gitmodules now names the new URLs. The toolchain repo is updated in the
CI action default, the codegen host, fetch_toolchain.sh, toolchain_pack
and package_setup_host, and the scaffolder's recomp-net / rbengine
defaults follow.
Two things the mechanical sweep would have missed:
- render_launcher_block() held a second copy of the launcher slug split
across two string literals, so a search for the full slug did not see
it and every regenerated README kept the dead owner. It is now derived
from LAUNCHER_REPO, which is the only place the slug lives.
- readme_has_launcher() matched one exact slug. Pointed at the new one
alone it would report every port built before the transfer as missing
its launcher section, and append a second copy to any README carrying
neither the markers nor a "## RetComM Launcher" heading. It now
accepts both.
The scaffolder's default GitHub owner for new ports becomes
RetroPortingToolKit. Existing ports are unaffected — they carry their
own owner in the README and remote.
psxrecomp, snesrecomp and recomp-ui did not move and still point at
mstan. Submodule paths (lib/recomp-net, lib/retcomm-rbengine) are
unchanged: they are pinned by every port and are not repo slugs.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_018xfKkiEkh87zL1EpE5Ty55
Matched Ape Escape retail SCPH1001 LLE builds identify 0661ce77 as the visible boot regression: parent 155e269b reaches game code, while 0661ce77 stalls near BIOS frame 375. Revert its four dependent follow-ups too; retain PR #326 readback bounds and PR #328 Vulkan aspect changes.
The row-by-row raster dispatcher amplifies an existing malformed DMA list into massive GL work. Keep the broader CPU/DMA arbitration experiment separate for compatibility testing. Document attribution, validation, and the temporarily restored interlace limitation.
Validation: full Ape target reaches title/menu and frame 11669 on retail LLE; OpenBIOS control reaches frame 3701. Eight focused tests pass; retained GL readback fixture passes 71 checks at both 1x and 4x after linking local SDL with iconv. Beads: beads-iisy; DMA follow-up: beads-eio.3.122.
Smart App Control blocks unsigned executables outright. The packager now
runs tools/ci/sign_windows.sh over the staged tree (host exe, DLLs,
emitters) with SHA-256 + RFC 3161 timestamp; no certificate in the
environment means a notice and an unsigned package as before, while a
configured certificate that fails to sign is fatal. The workflow template
passes WINDOWS_SIGN_PFX_BASE64 / WINDOWS_SIGN_PFX_PASSWORD to the package
step. Docs cover the secrets and the limit: the game exe the player builds
locally after Generate is unsigned by nature.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01V1dT6h1ZshMVUUVKTPxsCD
- Host may sit in a spectator seat: session slot 0 stays the sim authority
with a muted pad and no port; slot->port map keeps players on the port of
their lobby seat (host first). Rematch path copies the same plan.
- Seat self-service on LAN (MOTK5 SEATMOVE / SWAPASK / SWAPANS / SWAPRES)
relayed by the host; online swap requests accept a gallery seat as target.
- Lobby chat ring + join/leave lines; country codes from the server.
- Keybinds: a key drives one input per player (binding S to Cross unbinds it
from Circle), and keybinds.ini is re-read after the rematch launcher too.
- LAN registry file anchored to the exe directory.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01V1dT6h1ZshMVUUVKTPxsCD
Seat 1 (P2) can bring its LOCAL slot-1 card to a match: it uploads the card
to the host at launch over the new recomp-net MEMCARD op, the host installs
it as the match's slot-2 card, and the usual card broadcast carries it to
everyone. The host's real slot-2 file is rebound to netplay/guest_card2.mcd
for the session and restored on shutdown; guests already sandbox both slots.
Lobby side: a per-seat memcard_offer on set_ready (has_card, share), a host
allow flag in match_caps, and guest_memcard_active settled ONCE by the host
at start and shipped with the launch, so a toggle racing the start cannot
split the room. LAN carries the same over MOTK5 UPDATE / MEMCARD / JOIN
tail / a trailing START line. PSX_NET_GUEST_MEMCARD=1 for CLI sessions;
a host waiting on a card that never comes stalls at a named phase
(mc_guest_wait), not a silent "ok".
Two latent guest-sandbox bugs fixed on the way: a disabled slot 2 had no
bound path, so a host blob carrying a card for it failed to import and the
guest stalled forever; and leaving the sandbox re-bound a disabled slot to
the sandbox file. Both slots are now bound in the sandbox and restored per
slot (memcard_rebind_path, which can also unbind).
Lobby chat: the client keeps a ring fed by the server's echo (op:chat);
LAN rooms relay through the host (MOTK5 CHATREQ guest->host, MOTK5 CHAT
host->everyone). Wired to the launcher's chat_send/chat_count/chat_get.
Also carries the spectator-seat work that was in the tree alongside this
(spectator seats in the lobby client, spectator_wire_slot in the launch).
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01V1dT6h1ZshMVUUVKTPxsCD
sh psxrecomp/tools/generate_ci.sh # in the project
powershell -File psxrecomp\tools\generate_ci.ps1
... [root] [--check] [--force] [--dry-run] [--zip-prefix X] [--title X]
Project Studio already knows how to emit the setup-host release workflow
(emit_ci_workflow, behind `apply` and `git install-ci`), but reaching it
meant a migration plan or a commit-and-push. This is the single-purpose
face of that operation: it resolves the template from the project's own
psxrecomp submodule, takes the zip prefix from
scripts/package_setup_release.sh so CI uploads the zips the packager
built, writes .github/workflows/release.yml, and stops.
--check reports missing / placeholder / stale (step names the template
has and the file lacks) with exit 1, so a fleet can be audited without
writing anything; only --force overwrites. The .sh and .ps1 are thin
launchers over the one Python implementation.
Verified read-only against two ports: ApeEscapeRecomp reports stale
(four emitter/ccache steps missing), TombaRecomp reports current.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0173EhspvAdRQc9DZdDr97Hg
The static isolated-fragment pass resolved demands by bare entry address across
every capture: once any part had a variant at an address, that address counted
as served for every other capture of the band, and static_entry_sources kept
only the last capture's bytes. The dispatcher validates each variant against
the exact bytes it was compiled from, so in a band with N occupants exactly one
occupant ever received fragments (BoF3: BOSS055 35/35, PLP678 19/19, SCENA19
20/20; the 128-occupant and 181-occupant bands got none). 23,728 (entry, image)
demands were unserved; at the title screen the resident SCENA16 ran every
observed entry on the interpreter while every spanning piece came from
SCENA19/04/06.
- Demands are keyed (entry, image crc); parts carry image_crc;
select_static_fragment_demands / served_static_variant_keys are the pure
helpers (unit-tested).
- Fragment compiles run on the --jobs pool; the image table is sent once per
worker (initializer) instead of once per demand; results merge in demand
order and duplicates already served by an earlier-merged part are dropped,
so the output equals the sequential loop's.
- generate_interior_fragment_static returns (part, reason). Deterministic
verdicts on the bytes (generated-C audit, requested-entry audit, and a walk
that runs off the image, which the recompiler reports by throwing) are
memoized in <out-dir>/interior_fail_memo.txt and counted as skipped, never
as shard failures; --force-interior retries an entry. Only a
non-deterministic recompiler failure is a shard failure.
- Fragment parts are grouped one translation unit per image (BoF3: 779 units
instead of 20,635 files).
BoF3 full run: 20,250 fragments built, 3,570 memoized rejections, 238 s cold,
nothing lost from the previous dispatcher (99,778 -> 138,148 served keys).
Title screen: 12 interpreted SCENARIO-band entries -> 2, both unharvested PCs.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Make a bare interpreted PC explainable from the session-long per_pc snapshot
alone. DirtyRamPcEntry gains occ_crc / occ_ok (the static overlay variant whose
code ranges span the PC, via a per-PC scan of the static-match cache, and
whether it validated at that moment) and last_ext_ra (gpr[31] at the most
recent external entry, i.e. the caller that reached this interior). Both are
stamped only on external entries -- arrivals from native dispatch, not interp
block chaining -- and emitted per row in dirty_ram_stats.per_pc as occ_crc,
occ_ok and ext_ra.
Reading the pair: occ_ok=1 is an interior gap inside live native code (an
alias seed fixes it); occ_ok=0 with a CRC means the spanning piece is resident
but CRC-missing, so it was compiled from another section's bytes or rewritten
at run time and no seed helps; 0 means nothing compiled spans the PC.
psx_overlay_resident_crc_at() records each static-match cache entry's code
span (lo_min/hi_max) so the scan is a range test; DLL-path titles fall back to
the loader's last residency hash. The debug-server reply buffer grows 32 -> 64
KiB for the wider rows; the tail reserve still cuts between rows.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Matthew Stanley