Review (#600): with low_latency_input on, the re-drain after the pacer
wait sampled the pad without host_pause_service, so ui_capture was a frame
stale and the overlay's opening chord (Start+Select) could reach the guest
for one frame. The re-drain now services the overlay before the sample.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
psx_host_set_input_sink: an overlay provider sees host SDL events first,
in the event drain and the host pause loop, and may consume them.
psx_host_set_ui_capture: while on, local game input reads released and
centred (netplay included, so peers receive "nothing pressed") and the
external-input gate stays closed. Debug builds gain host_key, which pushes
a real SDL key event for scripted runs. Inert without a provider.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Review (#599): who keeps the callback context alive until render-thread
calls are quiescent. Draw calls now run inside an in-flight count, and
set_draw_cb (replace or NULL) swaps the record and waits until no call
that read the old one is running, then frees it. When it returns, the
registrant may free ctx and its GL objects; it must unregister before
destroying them or the context. A change from inside the callback does
not wait. Documented in host_overlay.h. Tests: a reader thread drawing
while the owner unregisters and frees (300 rounds; fails without the
wait), and self-unregister from the callback.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Review (#599): fn and ctx were two separate stores read on the render
thread, so an unregister could pair a live fn with a NULL ctx. They are now
one immutable record behind an acquire/release atomic pointer. The GL
fixture runners compile a fixed source list; gpu_gl_renderer.c now calls
psx_host_overlay_draw, so every run_gl_*.py links host_overlay.c.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
psx_host_overlay_set_draw_cb: the OpenGL backend calls it after the game
frame and host OSD layers are composed, before present_shot capture and
the swap, on the thread that owns the GL context. Without a registered
callback nothing changes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Review (#598): resuming left the frame pacer's deadline in the past, so it
paid a pause back as up to ~12 unpaced frames. Resume now runs
present_session_reset, an audio cycle resync and a dynamic-resolution hold,
as a savestate load does. host_pause on=1 headless was a no-op that answered
ok; the headless frame path now enters host_pause_loop too.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
psx_host_pause_push/pop: a nesting, host-owned pause. While it is held the
runtime freezes the guest at the vblank present (as the save-state menu
does), keeps presenting the last frame and pauses audio. Push refuses
during netplay. Nothing pushes unless an overlay provider (or the debug
server's new host_pause command) asks, so behaviour is unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Review (#654): forget the ring's names after gl_renderer_render_thread_stop
(the worker presents and captures on its own thread, and stop drains it
and hands the context back), before the context is deleted. The call is
unconditional (an empty stub without debug tools). The fixture adds a
render-thread pass: frames captured on the worker and drained, more left
queued at shutdown, then checks the worker stopped first and every name
and pending readback is gone. Fails if the reset runs before the stop.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The seed path collapsed ls.texture_filter to a boolean, so a launcher's
xBR (3) was saved and applied as bilinear, and the rematch paths passed
the raw value through unchecked. psx_texture_filter_from_ui is now the
one translation (in-range values pass, others are nearest). Test:
video_texture_filter_test.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
After a renderer restart (lobby rematch, in-game menu -> match) the
capture read back into a stale pack-buffer name, i.e. into client memory
at offset 0, and crashed in glReadPixels (debug-tools builds).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
savestate save/load with slot >= SAVESTATE_SLOTS (12) was refused with the
generic message that blames LLE runs, which sent a test chasing the wrong
cause. It now replies "slot N out of range (0..11)". Guarded in
test_savestate_status_protocol.py.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
SIGPROF can land on a thread that is still being created. On macOS arm64
pthread_self() there hits a pointer-authentication trap (EXC_BREAKPOINT in
on_prof), which killed debug-tools builds at boot when a second instance
started lobby threads. The handler now compares the interrupted stack
pointer with the emulation thread's stack range; Linux keeps pthread_equal.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The lobby server starts a 2-player room the host relay cannot carry on ICE
(recomp-net-server start_transport, launch transport "ice"). This client
knew only "host" and the legacy server relay. It inferred a relay from
equal host/guest endpoints -- which is exactly what two peers behind one NAT
advertise -- and each peer dialled its own public address: the match stayed
at "linking" forever.
- psx_lobby_launch_transport.h: one pure decision (host / ice / relay /
direct) with a unit test covering ICE, same-NAT endpoints, host relay and
the legacy relay cases.
- PsxLobbyJoinInfo.transport_ice; the launch handler records it and never
turns it into a relay; launch fill and resolve_use_ice start ICE p2p
(signalled over the lobby) for it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Keep both netplay_peer_frontier.h and netplay_fmv_confirm.h includes; advance lib/recomp-net to a48d355 (main, includes #38).
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
- The dirty-RAM interpreter's IRQ poll stride is host-only; a load resets
it to phase 0 while Live carried an arbitrary phase, so a replay could
take the VBlank in RAM code at another instruction. It is now
re-anchored at every tick boundary in Live and Replay alike.
- The av digest zeroes the three depth-24 present helpers at the end of
the GPU snapshot (upload span, present hold, previous height): host
presentation state a suppressed-present replay leaves different, not
guest-visible.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
PSX_RB_DET_SELFTEST on R4 4P showed a replayed save of tick T differing
from the live save of T. Each cause, and the fix:
- Replay's finish_frame poll passed no PC hint (Live used the IRQ resume
chain); it now uses psx_netplay_snap_resume_hint.
- Devices are serviced lazily up to the next event, so root counters and
cycles_since_vblank lagged the clock by a host-path amount at save; a
load re-anchors them at the clock and drops the lag. The save now
publishes pending cycle charges and services every device to the clock
first, and renders the SPU up to it.
- GTE / mul-div completion stamps and the load-delay give-back were
re-anchored at the restored clock on every load, so a stall in flight at
the snap point cost the live run cycles but the replay none. New
BS_SEC_CPUTIM carries them relative to the clock.
- The SPU CD-input FIFO was outside the snap and wiped on rollback loads;
new BS_SEC_SPUCD carries the queued frames and the load keeps them.
- LZCR is serialized as derived from LZCS (the raw backing stays 0 until
the first LZCS write; guests only ever read the derived value).
Older snaps without the new sections load as before.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Replay's finish_frame saved its tick snap with no PC hint while Live used
the IRQ resume / last check / cpu->pc chain; the same tick's snap could
carry a different resume PC (psx_netplay_snap_resume_hint now serves both).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Every K ticks the host rewinds 16 ticks with nothing mispredicted, so each
seat replays confirmed rows over ticks it ran live; with PSX_RB_SNAP_AUDIT
the diff tool then compares replayed saves with live ones field by field.
On R4 4P it shows replay saves taken at a different guest point than the
live save of the same tick (CPU pc, cycles_since_vblank, root counters).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The waiting-room pump sent set_ready every frame while the local seat was
not yet echoed as ready. Each set_ready is a lobby_update broadcast to the
whole room, so two waiting peers produced hundreds of updates; the host then
missed the seat update (seat 2 shown open, Start disabled) and, after Start,
the launch itself. Offer changes still go out immediately.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Per ring save, the section CRCs (and small sections in hex) of the
boot_state blob, marked live or replay. tools/rb_snap_audit_diff.py
checks a seat's replay against its own live save and seats against each
other, naming the first differing section and bytes. Off unless set.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A Live realign restores the snap at T, sets the session to T+1 and
longjmps back into the guest from the admit spin. The resume ran before
T+1 was admitted, so that frame latched no inputs and did not advance;
the next frame was then counted as T+1 and the realigned seat ran one
frame ahead of its tick label from there on (ring saves after a NACK
realign carried the next tick's clock, 2x the cycles in the first
tick). Only the realigned seat shifted, so it forked from every other
seat; with four seats this produced the "resim core diverge" splits
and the later hc-fork / hang cascade.
flush_resume now admits the tick first on the Live realign path and
keeps spinning until it can.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Verifies the four PS1 semi-transparency modes, set-mask, mask check and
texel STP at 4x internal resolution with PGXP depth on and off, directly
and on the render thread, against the PS1 formulas and the software
rasterizer (new gl_blend_accuracy_test).
Finding: with the historical surface bases (fills and texels stored as
k*255/31, everything else as k*8) single-layer blends are off by one 5-bit
step in places (1312 of the test's checked pixels, e.g. B+F/4 over a fill).
[video] accurate_blending = true stores every VRAM write as k*8 so GL's
8-bit blend truncated by the readback is exactly the PS1 arithmetic; whites
of fills/texels then present as 248 like uploads already do. Default off.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Filter mode 3 magnifies textures with an edge-directed rule after Hyllian's
xBR (written from the published description, no code reused): for the
texel corner nearest a sample, weighted YUV colour distances over a 5x5
neighbourhood decide whether an edge runs along the corner's diagonal; if
so the corner is cut with an anti-aliased boundary and takes the closer
side neighbour. Flat areas, straight edges, gradients, cutouts and 1:1 draws
stay nearest. Selectable from [video] texture_filtering, settings.toml,
PSX_TEXTURE_FILTER and the new texture_filter TCP command. Default unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Equal buttons hid sealed-row stick differences when chasing a 4-seat
resim divergence (36 guest cycles on the last replayed tick).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Shokunin