Master moved the textured-dot test stubs into gpu_exec_stubs.inc; the PGXP stubs move there too. pgxp_note_rect_bypass stub takes (int). Stamps recomputed.
#459's fixture predates #468; fixture_link stubs the undefined data symbol as a function, which reads non-zero and aborts.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Closes#474. The win64 binary zip carries no license file; stage the notice beside tcc.exe and in licenses/, and make the zip verifier require both.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Fixes the false FMV flag after every snapshot load made before the first video (#475 cause 1); v1 payloads still load. Drops the unread host-frame stamp.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Many PS1 renderers drop a primitive whose depth index falls past the end
of their ordering table. Raising that limit can write past the table, so
a title can now list such guards in game.toml with the register that
carries the index and the last safe value. While a trusted mod has called
psx_mod_set_draw_distance_clamp(1), generated main-EXE code clamps that
register before the guard runs, so far geometry is kept in the farthest
slot instead of popping in later.
- Off by default and reset at every session start, so netplay and every
title that lists no sites run unchanged; with no sites the generated
code and declarations are byte-identical.
- Main executable only: the full instruction word is the guard,
generation fails on a mismatch or a load-delay hazard, captured
overlay code keeps its own code (the sites are not part of the overlay
config identity), and the dirty-RAM interpreter applies the clamp
where the game's text image holds the listed word.
- Compatibility: code_generator.cpp is a codegen-hash input, so the
codegen hash and the overlay-cache tag change for every title, sites
or not. Savestates from earlier builds are refused, and overlay shards
rebuild once on first visit. Generated code and behaviour are
unchanged for a title without sites or with the switch off.
- Tests: draw_distance_codegen_test (config rules, emission, guards,
declarations, overlay identity) and draw_distance_interp_test (store,
switch, interpreted guards); the session-reset guard covers the switch.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
With every drawn vertex precise, a title can still show cracks where the game
culled a face on its integer winding. R4's road emitter runs NCLIP on each
far row and skips rows whose native height rounds to 0. Natively their
neighbours meet at the same integer y. With PGXP they sit at their exact y,
and the skipped row becomes a strip of sky across the road beyond the start
gantry, at 4K and in 2P. tri_mixed cannot see this: the missing face draws
no triangle.
Precise culling (DuckStation's "PGXP culling", minimal footprint):
- While geometry correction is armed, gte_nclip computes the exact
determinant of the three SXY register shadows (pgxp_gte_nclip_precise:
live, current words, the GPU's own truncation agreement and tolerance).
When its sign is nonzero and differs from the integer MAC0's, MAC0 takes
that sign, with the rounded exact doubled area (at least 1) as magnitude.
FLAG keeps the integer overflow bits.
- It changes guest-visible MAC0, so it is a `culling` option of
psx.enhancement.pgxp, false in the builtin, with no [video] key, so netplay
(which clears every mod) never runs it. It needs geometry correction.
It is held off in every pass compared against another execution: the
speculative and replay passes, and the whole overlay shadow diff
(psx_overlay_shadow_diff_active), interpreter pass included.
- Counters whenever geometry correction is armed: nclip_precise,
nclip_disagree (exact sign differs: the culled-face crack exposure),
nclip_corrected. TCP pgxp {"culling":0|1}; env PSX_PGXP_CULLING.
Session arming in one place (src/pgxp_session.cpp, psx_pgxp_session_arm):
main.cpp's renderer setup now calls the same code the test drives. The arming
takes the mod request (reads and clears it), so a session with an empty plan
cannot inherit an earlier session's request even if the reset were skipped.
pgxp_session_test now links the real mod runtime, the builtin plugin and the
builtin manifest, and covers:
- a default-on override;
- a netplay clear, with and without the reset;
- the off switch;
- the options;
- env overrides;
- pgxp_mod_only.
[video] pgxp_mod_only (game.toml, default false): for a title that ships
PGXP through the mod, the mod is the one switch. The [video]
geometry_correction / perspective_texturing / pgxp_cpu_mode values are not
applied, from game.toml or the player's settings.toml. The launcher also
hides its Perspective textures row (has_geometry_precision = 0). That row
showed "off" while a default-on mod corrected textures, and it could keep
texture correction on after the player switched the mod off.
geom_correction also reports texture_correction and the culling switch.
R4, Helter Skelter: 28.7% of NCLIPs disagree in sign. With culling, the
ladder is gone in same-frame 4K captures (1P straight and crest, 2P VS), and
the game emits about 24% more triangles. Its packet buffers peak at 60%.
fp_identity over 12000 frames: stock, PGXP off, and PGXP on with culling off
are IDENTICAL; with culling on, the RAM writes differ from the attract race
on, while cyc/mmio/mc agree throughout.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Two cases that could still draw a triangle with precise and native corners:
- The rectangle shortcut bypass now needs all four vertices of an
axis-aligned textured quad to be dataflow-precise. A quad with only some
precise corners, such as a CPU-built sprite with one corner copied from
SXY, keeps the shortcut and counts as rect_partial; two triangles would
have mixed precise and native corners. R4's facade quads are fully
precise (rect_partial 0 in every R4 run).
- Under preserve-projection, RTPS now applies the GPU's truncation window
itself (pgxp_ppp_accept): an exact projection outside it is not
shadowed, the vertex keeps the IR path's shadow, which always agrees, and
it draws precise instead of being rejected to native at the GPU
(ppp_window_fallback counts these). The exact math is skipped while the
hooks record nothing (pgxp_active(): speculative passes).
Also:
- The prepare_precise_triangle comment no longer claims a native neighbour
can only disagree by the sub-pixel fraction. Under preserve-projection it
can disagree by up to the window.
- gpu_textured_dot_nw_shift_exec_test links with -dead_strip on Darwin,
as isc_store_test does. GNU --gc-sections kept it from linking on macOS,
so the rectangle cases never ran there. The test now covers no, one and
all four precise corners.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A textured quad whose four packet vertices form an axis-aligned rectangle in
integer screen space and in UV takes the 2D rectangle shortcut
(gr_draw_textured_rect), which draws the native integer position with no
perspective. That is right for a CPU-built sprite, but a GTE-projected world
quad can land on such a rectangle too: a facade seen straight on. Its
neighbours are drawn precise, so at a high internal resolution the shortcut
left a background-coloured seam along the shared edges. On R4's Helter
Skelter start straight it showed as an L-shaped sky line along a building
panel at 4K, with tolerance off, both with and without preserve projection.
When geometry or texture correction is on and any of the quad's vertices
resolves to a dataflow-precise position (pgxp_probe_precise_vertex: the same
decision as pgxp_get_precise_vertex, not counted), the quad is drawn as its
two triangles like any other world quad. A CPU-built sprite has no dataflow
shadow and keeps the shortcut. With PGXP off nothing changes. geom_correction
counts the bypassed quads (rect_bypass).
Tests: gpu_textured_dot_nw_shift_exec_test gains the quad case (PGXP off,
on without a precise vertex, on with one); pgxp_test checks that the probe
does not count; gpu_gp0_history_guard_test's gpu.c stubs gain the two new
pgxp entry points.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Two opt-in [video] keys for titles whose build carries the PGXP hooks, both
live-tunable through the TCP pgxp verb. Defaults keep today's behaviour.
pgxp_position_fallback (default true). false skips the position-cache tier
in pgxp_get_precise_vertex and stops gte.cpp filling it. At hook-build
coverage the cache only hands unrelated 3D fractions to CPU-built 2D
polygons (a HUD needle, menus). Beetle and DuckStation leave their vertex
cache off by default.
pgxp_preserve_projection (default false). true makes RTPS/RTPT shadow the
exact projection, from the unshifted MACs and a true divide
(pgxp_project_precise), instead of the GTE's integer-IR one. That removes
the residual wobble of near geometry. Guest SXY, MAC0 and FLAG are
untouched. Only sf=1 vertices with no IR, SZ3 or divide saturation qualify,
and the rest keep the IR path. The exact position can sit a little below the
guest integer or more than a pixel above it, so truncation agreement becomes
the window -4 < precise - native < 5 px, a sanity bound over the IR path's
own error (on R4 every on-screen vertex falls within -1.9 .. +2.6 px). A
packet half at or beyond the GTE saturation limit must still agree exactly.
The tolerance clamp now measures the offset both ways. [widescreen]
precise_nclip reads the same shadows, so in a title that sets both, its
branch sign follows the exact projection too.
Observability: geom_correction reports both switches, ppp_produced, and
geometry-corrected triangles split into all-precise, mixed and native. The
mixed share is the mesh-cracking exposure of G1.1.
Tests: pgxp_test (fallback off, window edges, saturation, two-sided
tolerance, the exact projection and its qualifiers, triangle census);
gte_register_access_test (RTPS/RTPT guest state equals the oracle with the
mode on, and the shadow is the exact projection); video settings test
(defaults, parsing, type check). gpu_gp0_history_guard_test's gpu.c stubs
gain pgxp_note_triangle.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The OpenGL textured program read the GP0(E2h) texture window from a
uniform, so every window change ended the open textured batch (batch
reason 5). Titles that tile far textures with per-primitive windows pay
for that in draws: Ridge Racer Type 4's split screen changes the window
about 515 times a game frame and drew about 180 batches, against 17 with
batching on. On the native-wide path every batch that reaches the margins
is drawn again into the wide surface; under heavy load on an M4, 2P with
widescreen at Match display ran at 39-43 guest VBlanks/s and 1P at 50,
against 50-58 and 57 with batching on.
The window now rides in each textured vertex (a_twin, the E2h low 20 bits
as one float) like texpage and CLUT already do, and the shader decodes it
there; the uniform and the window-replay queue's copy of it are gone.
With the default unchanged, batches still split at every window change,
so every vertex carries the batch's window and the image is the same as
before: the new test's oracle passes on the unmodified renderer too, with
the same digests at 1x, 4x, 9x and in the windowed 9x/18x mode.
[video] texture_window_batching = true (game.toml only, default off) lets
primitives with different windows share a batch. The image is the same:
an opaque batch draws its colour in one painter-ordered pass and
semi-transparent primitives are still drawn one per batch. While mask
checking is on the window stays a batch key, because an opaque batch
fixes the stencil for its STP=1 texels after its colour pass.
PSX_GL_TEXWIN_BATCH=0|1 overrides it for a run and the TCP command
gl_texwin_batch switches it live. Software and Vulkan are unchanged.
Tests: gl_scale_invariance_test gains mode "twin" (alternating windows
over opaque, semi-transparent, set-mask, mask-check and bilinear prims,
into the native-wide margins): off and on give the same native VRAM,
frame at S and wide surface, on draws 17 batches instead of 46 and flushes
only for the 5 mask-check window changes. Each run also checks an oracle
that does not depend on the renderer: raw rects through all four windows
into a 4-bit and a 15-bit page must give, pixel for pixel in the native
VRAM and at the centre of each pixel at S, the texel the PS1 window rule
picks from the VRAM as read back. A decode or packing error both settings
share fails it (checked with (v_twin >> 11) and with off_y left out of the
packed window). video_enhancement_settings_test checks the key's default
and opt-in.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
psxrecomp_add_game_runtime gains a PGXP option and forwards it to
psxrecomp_add_runtime_target, which already supports a PGXP primary
(-DPSX_PGXP=1, overlay flavor 2, published flavor file, no _pgxp suffix and no
clone). Until now a title could only get the hook flavor through
PSX_PGXP_VARIANT's side-by-side clone, or by leaving a bare PGXP in the
forwarded tail, where an earlier multi-value argument such as
CODEGEN_SETUP_SOURCES or GEN_FULL_GLOB swallowed it as a file name.
The generated C already carries the PGXP_*() hook sites, so a title's
committed generated/ does not change. Default unchanged: titles that do not
pass PGXP build the base flavor as before.
game_runtime_pgxp_test runs the real function body under cmake -P with the
runtime target stubbed.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
psx.enhancement.pgxp did nothing at boot. Its activation armed geometry and
texture correction, then main.cpp's renderer setup, which runs after
activation, applied the [video] baseline and switched both back off. Only the
[video] keys could turn PGXP on.
The activation now records a request (pgxp_mod_request), cleared at every
session start in reset_mod_owned_presentation(). The renderer setup arms PGXP
from the [video] baseline OR the request (pgxp_session.h); the validation env
overrides (PSX_GEOMETRY_CORRECTION, PSX_PERSPECTIVE_TEXTURING,
PSX_PGXP_CPU_MODE) still win. The g_video_* values stay the player's
settings, so the mod never leaks into a launcher seed, and a netplay session,
whose plan is cleared, keeps the [video] keys alone, as before.
No change for a title whose mod is off. pgxp_session_test drives the real
builtin activation and the resolver in main.cpp's order.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
test_window_fullscreen.cpp is a separate test program with its own main(). It includes SDL_main.h and is built with SDL_MAIN_HANDLED. The single-owner guard sdl3_main_single_include counts every file under runtime/tests, so it has failed since that test was added.
Name the file as a stand-alone program. Every other file under runtime/src, runtime/include and runtime/tests is still held to the one-owner rule.
When straight-line interpretation reaches a clean page, the interpreter
hands the PC back to static dispatch. If that PC is game text whose live
bytes differ from the boot EXE, dispatch refuses the compiled code and
enters the interpreter again, which hands back one instruction later. A
page of such text cost about a thousand dispatch round trips.
This happens during the retail BIOS intro on profiles whose shell runs
interpreted in RAM (the SCPH5552 profile has no shell seeds). In a game
whose boot EXE later owns that RAM, every shell instruction took the
round trip; on one PAL title the intro ran at 17 fields/s on the host
instead of 42.
Straight-line flow now keeps interpreting while the next PC is game text
that psx_game_text_native_ok() refuses, and re-tests every instruction,
so it hands back at the first PC dispatch accepts.
interp_refused_game_text links the production interpreter into a fixture:
from a dirty page into refused text, it hands back once and dispatch runs
the compiled code at the first accepted page. Before this change the
interpreter returned at the first clean page. A control checks that a
clean page outside game text is still handed back at once.
Preserve current segmented dispatch, active-plan mod hooks and software scale limits. Extend function filters through the active plan, retain the last-store pointer with overlay ABI 25, and keep default config fingerprints stable. Refresh BIOS emitter stamps after byte-identical regeneration and reconcile regression fixtures.
Bundled releases link OpenBIOS only; retail-BIOS-derived C never ships.
"Bring your own BIOS" now works without it: when a dump matching a shipped
profile (psx_bios_known_images.h) is selected, the runtime builds that
backend on the player's machine with the overlay toolchain the release
already carries, caches it, and loads it as one more registered backend.
- runtime/include/psx_bios_module.h: the module contract. An overlay-style
self-contained shared library (no imports from the executable) exporting
psx_bios_module_{abi,codegen_hash,init,backend}; reaches the runtime only
through OverlayCallbacks plus a new PsxBiosModuleCallbacks table (traps,
dirty-RAM/kernel-bless probes, fntrace, log_probe) and pointer aliases of
the globals the generated dispatcher touches. Its #defines rewrite the
emitter's own extern lines, so the generated C compiles unchanged.
- runtime/include/psx_bios_module_glue.c.inc: module-side forwarders and the
four exports, compiled after overlay_dispatch_preamble.c.inc.
- runtime/src/psx_bios_module.c: host side. Cache path
<exe>/cache/bios/<os-arch>/bm<abi>_<codegen hash>_f<flavor>/<STEM>_<crc>,
ABI + codegen-hash gated load, callback fill, and a blocking build via
overlay_toolchain/bios_module_build.py (gcc via compiler.txt or PATH,
else tcc), then psx_bios_register().
- psx_bios_backend.c: the registry is now a mutable table seeded from the
CMake-generated psx_bios_builtin_registry by a constructor, with
psx_bios_register(); every existing walker keeps working.
- main.cpp: validate_bios_for_launch acquires a module (announcing a first
build) before declaring a mismatch; discovery accepts a buildable dump;
the launcher verify reports it usable ("ready" / "compiled on first
launch") with no Generate & rebuild; the BIOS row shows whenever the
toolchain is present; bundled_only respects module support.
- emitter: the CPS-marker constructor is not emitted in a module build.
- overlay_api.h: OverlayCallbacks gains a struct tag;
overlay_loader_callbacks() exposes the host table before loader init.
- tools/bios_module_build.py (ships in the toolchain): identity + profile
SHA-256 gate, staged emitter tree, emit, three-TU compile, ctypes
self-check, atomic publish. release_stage.py stages psxrecomp-bios, every
bios/*.toml, the seeds and the builder into overlay_toolchain/.
- Tests: runtime/tests/test_bios_module_host.c against three fake modules
(good / wrong ABI / wrong hash) on real dlopen; recompiler/tests/
test_bios_module_build.py builds the bundled OpenBIOS end to end and reads
the descriptor back; guards extended.
Verified on Twisted Metal 4 (bundled build + staged toolchain, headless):
first launch with --bios SCPH1001.BIN builds and loads the module and boots
on image=SCPH-1001; the second launch loads it from cache.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
A shipped bundled build links OpenBIOS only and carries no CLI, sources
or toolchain, yet the codegen host forced the BIOS row on and a retail
pick answered with needs_regen ("Generate & rebuild to switch"), a step
that cannot run there.
- The BIOS row is shown only when a choice can mean something: a retail
backend is linked, nothing is linked yet (setup host), or the codegen
host actually wired Generate (prepare_with_progress). main.cpp decides
after host_apply instead of the host forcing has_bios.
- ae_bios_verify short-circuits a bundled-only build with a plain
"runs its bundled OpenBIOS only" message before the identity checks,
whose "this build expects SCPH-1001" wording names the pinned stem,
not a linked backend, and sets needs_regen only when Generate exists.
- test_bios_selection_guards pins both.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Retire the setup-host release model (CI wiped generated/ and shipped
emitters + sources + a wizard so every player regenerated and rebuilt the
title). Titles now commit generated/ and CI ships the compiled game.
- docs/ci/templates/game-release.yml replaces setup-release.yml: gate on
committed game C (tools/ci/check_generated.sh), emit the OpenBIOS backend
in CI (tools/ci/generate_openbios.sh), configure with
PSXRECOMP_REQUIRE_GAME_C=ON and assert the full path was taken, package
with tools/package_game_release.sh, verify the zip is a game not a kit.
- tools/package_game_release.sh: exe, assets, bundled OpenBIOS, configs, mod
catalog + overlay toolchain via release_stage.py, notices; refuses sources,
emitters, generated C, BIOS dumps. clear_generated.sh removed;
package_setup_host.sh marked retired.
- Scaffold: gitignore templates track generated/, package_release.sh.in
wrapper, setup_project.{sh,ps1} commit generated/ after Generate.
- Project Studio: audit flags ignored/untracked generated/, retail BIOS C
and setup-host workflows; merge_gitignore un-ignores generated/; ops emit
the bundled packager + workflow; generate_ci uses the new template.
- Docs: docs/ci/BUNDLED_RELEASES.md replaces HOST_ONLY_RELEASES.md.
- Tests: tools/tests/test_bundled_release_layout.py; fix pre-existing
false positives in the packager-surface and overlay-static tests.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
psx_ram_runtime_map_test links dma.c without dma_gpu_ll.c. The linked-list
kick now asks for the walk's next event while it holds the CPU, so the
fixture needs that symbol too. The stubbed start leaves the walk inactive,
so the hold never loops here.
The GPU linked-list walk reads its headers and payload on the guest clock
while the CPU keeps running. PSX-SPX ("CPU Operation during DMA") says the
opposite: DMA outruns CPU memory access, and a CPU access to RAM or I/O
waits while DMA moves data.
The difference is visible in the retail SCPH1001 intro. Right after
DrawOTag it re-copies its shared text packets from a template, which
clears their link words. With the CPU running alongside the walk, the
copy reached a packet before the walk did, and the walk followed link 0
into low RAM (docs/internal/RETAIL_BIOS_INTERLACE_ROLLBACK.md describes
the same race at BFC188D8 and defers the CPU/DMA arbitration fix).
The store that kicks a linked-list transfer now returns only when the walk
has ended: the kick advances the guest clock event by event and services
the devices until the list is done. A cyclic list still ends at the
walker's existing node limit, and the hold also stops if an advance moves
no list word. Kicks made inside the device service, during lockstep replay
or during a render pass are not held.
dma_gpu_list_cpu_hold links the production dma.c and dma_gpu_ll.c into a
fixture. It kicks A -> B -> end, clears B's header right after the kicking
store (the intro's packet copy), and checks that the GPU still received A's
and B's payload. Before this change the GPU received only A's. A control
checks that a kick from inside the device service is not held.
settings.toml ([disc] path, [bios] path, [memcard] dir/card1/card2) and
the disc.cfg / bios.cfg sidecars can hold relative paths, for example when
a player edits them by hand or a game folder is moved as a whole. The
runtime resolved them against the working directory. A game started from
a shortcut whose "Start in" differs, or from a frontend, then looked for
its disc, BIOS and memory cards in that directory: saves appeared missing
and new cards were created there.
Resolve every persisted path against the executable directory through two
adapters on host_resolve, the same anchor the bezel artwork path uses:
- resolve_persisted_disc_path() for [disc] path and disc.cfg;
- anchor_on_exe_dir() for bios.cfg, [bios] path, [memcard] dir, card1 and
card2, and the card paths the launcher hands back.
Absolute and UNC paths are unchanged, empty values stay empty, and a
relative --disc on the command line still resolves from the caller's
working directory.
persisted_settings_paths_test compiles the real adapters from main.cpp
into a probe, runs it from an unrelated working directory, and checks that
every persisted-path consumer in main.cpp uses them.
Matthew Stanley