Commit Graph
50029 Commits
Author SHA1 Message Date
Henrik RydgårdandClaude Opus 5.5 adaa14d0bb docs/adhoc-multiplayer-testing: Instances can share too little GPU
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-30 10:08:11 +01:00
Henrik RydgårdandClaude Opus 5.5 214b14fe63 Add --nickname, and give each adhoc_session.py instance its own
Games that show the system nickname in their lobbies can then tell the
instances apart. Also: press --times N --every S for menus that need
several presses, and start clears the last session's screenshots, which
could otherwise pass for this one's.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-30 10:08:10 +01:00
Henrik RydgårdandClaude Opus 5.5 f0c6a26087 adhoc_session.py: Wait for screens, and collect logs
until waits for a region of the screen to match a reference crop (or
a pixel its color), optionally pressing a button meanwhile, since
loading screens take a different time on every run and a press before
a menu is ready is lost. crop makes the references. log sets a log
channel's level and watch collects the log messages for a while.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-30 10:07:57 +01:00
Henrik RydgårdandClaude Opus 5.5 96e31f8c47 WebSocket debugger: Enabling a log channel turns logging on
Headless turns logging as a whole off without --log, so a channel
enabled with log.channel.set still logged nothing, and a client
watching the log events saw none.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-30 10:07:57 +01:00
Henrik RydgårdandClaude Opus 5.5 8f3c49fb9d Tools/adhoc_session.py: Play a game's multiplayer with several instances
Starts N headless instances of a game against each other, with or
without the relay, and drives them over the debugger: button presses,
side-by-side screenshots, frame rates, and scripts of those. Each
instance gets its own memory stick with a copy of the savedata, so
autosaves can't collide. docs/adhoc-multiplayer-testing.md describes the
setup, by hand too, and the network simulation options.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-30 10:07:57 +01:00
Henrik Rydgård d8ff2bd62e Merge pull request #22848 from hrydgard/adhoc-7-vanish
Ad hoc 7/9: Leaving the network, and timing out a peer that left, as on a PSP
2026-10-30 10:00:31 +01:00
Henrik Rydgård 3ca9da090a Merge pull request #22847 from hrydgard/adhoc-6-gamemode
Ad hoc 6/9: Fail entering game mode the way a PSP does
2026-10-30 09:22:49 +01:00
Henrik Rydgård f2f0f7d374 Merge pull request #22846 from hrydgard/adhoc-5-ctl-timing
Ad hoc 5/9: Accurate PSP timing for adhocctl events
2026-10-30 08:48:07 +01:00
Henrik Rydgård 86e4b2db02 Merge pull request #22845 from hrydgard/adhoc-4-ctl-thread
Ad hoc 4/9: Run the adhocctl client on the CPU thread, don't wait for the login
2026-10-30 07:33:44 +01:00
Henrik Rydgård 5d2f3a6bec Merge pull request #22844 from hrydgard/adhoc-3-matching
Ad hoc 3/9: sceNetAdhocMatching fixes from the pair tests
2026-10-30 00:32:30 +01:00
Henrik Rydgård 29eeb1f039 Merge pull request #22843 from hrydgard/adhoc-2-discover
Ad hoc 2/9: sceNetAdhocDiscover states, GetPeerInfo on the local MAC
2026-10-30 00:02:18 +01:00
Henrik Rydgård 7fab97c05f Merge pull request #22835 from hrydgard/adhoc-1-sockets
Ad hoc 1/9: Socket fixes from the single-console tests
2026-10-29 23:38:33 +01:00
Henrik RydgårdandClaude Opus 5.5 f449129dec sceNetAdhoc: Time a vanished peer's timeouts as a PSP does
A PSP checks a PTP connection's oldest unacknowledged data every
rexmt_int, counting from when the connection was established, and drops
the connection at the first check where that data is older than
(rexmt_cnt + 1) * rexmt_int. A matching peer times out keepalive_int *
init_count after its last keepalive, checked all the time instead of on
the console's own keepalive tick, which could add up to an interval.
Both measured with several settings in pspautotests
net/adhoc/pair/vanishtiming; DATA_TIMEOUT already matched.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 22:39:07 +01:00
Henrik RydgårdandClaude Opus 5.5 fa51700d58 sceNetAdhocctl: Fail entering game mode the way a PSP does
On a PSP (pspautotests net/adhoc/gamemode):
- a JoinEnterGameMode whose host never shows up gets ERROR with TIMEOUT
  when its timeout runs out, and is out of the group afterwards. The
  joiner's GAMEMODE, queued as soon as the server answered and held
  until the host appeared, came after 5 s anyway;
- a CreateEnterGameMode whose first member isn't this console returns
  0 and reports INVALID_ADDR;
- in a normal group, CreateEnterGameMode reports ALREADY_CONNECTED and
  stays in the group, and ExitGameMode returns INVALID_IBSS;
- ExitGameMode outside game mode returns 0 with no event.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 22:39:07 +01:00
Henrik RydgårdandClaude Opus 5.5 84c9f2d38d adhoc_multi: Show the output of tests that haven't been recorded yet
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 22:39:07 +01:00
Henrik RydgårdandClaude Opus 5.5 b58368405b proAdhoc: Keep pinging the server while emulation is paused
The adhocctl poll only runs while emulation does, so with the pause menu
open, or stopped in a debugger, for 15 s the server dropped the client,
and a game in a lobby lost its peers until it rejoined. A small thread
pings when the poll hasn't, as the friendFinder thread used to.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 22:39:07 +01:00
Henrik RydgårdandClaude Opus 5.5 ee86058ec7 sceNetAdhocctl: Leave the console out of its own peer lookups; give groups their own BSSID
From the basic and lifecycle pair tests, recorded on two PSPs:

- GetNameByAddr for the console's own MAC gives NO_ENTRY, and
  GetAddrByName doesn't find the console itself.
- GetPeerInfo with a size too small for the entry gives INVALID_BUFLEN.
- A PTP connection accepted over the relay reports the peer's port as
  the game sees it, not the relay's offset port.
- The built-in server makes up a locally administered BSSID for each
  group, which stays when the creator leaves. It was the MAC of the
  oldest member.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 22:39:07 +01:00
Henrik RydgårdandClaude Opus 5.5 1e3e9c7251 sceNetAdhocctl: Send Create, Connect and Scan without the 10 ms wait
On a PSP they return within about 30 µs (pspautotests net/adhoc/timing).
They waited 10 ms before trying to send the request; now the first try is
right away, and the call returns once it's sent.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 22:39:07 +01:00
Henrik RydgårdandClaude Opus 5.5 559b4e404b test.py: Add net/adhoc/ctlinit to tests_next
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 22:39:07 +01:00
Henrik RydgårdandClaude Opus 5.5 4197f2326a sceNetAdhoc: A console that leaves the network goes quiet, as on a PSP
The vanish pair test (recorded on two PSPs) now passes on both sides,
over the relay:

- From the Disconnect call, a console's PDP and PTP sends succeed and go
  nowhere, and nothing reaches it.
- A PTP connection to a peer that has left the group fails with
  DISCONNECTED after rexmt_int * rexmt_cnt. Relay sockets now get their
  retry settings (and are zeroed), and an accepted one counts as
  connected.
- Matching data to a peer that has left gets DATA_TIMEOUT after the
  retransmissions instead of a DATA_ACK.
- The matching timeout for a silent peer is keepalive_int * init_count,
  plus MATCHING_TIMEOUT_EXTRA_US (0) to tune for internet latency. It
  was (keepalive + rexmt) * count + 500ms, later than a PSP's.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 22:39:07 +01:00
Henrik RydgårdandClaude Opus 5.5 a2897c5b65 sceNetAdhocctl: Deliver CONNECT, SCAN and DISCONNECT at the PSP's times
A PSP delivers CONNECT 1.1 s after a Create, 0.53 s after a Join and
3.5-4 s after a Connect (which looks for the group first), SCAN 1.0 s
after a Scan and DISCONNECT 105 ms after a Disconnect, and the state
changes along with the event (pspautotests net/adhoc/timing and
net/adhoc/pair/timing). Deliver them then, counting from the request in
emulated time, or later if the server's answer takes longer. This
replaces the delays tuned for Dissidia 012, GTA: VCS and Tekken 5, and
Disconnect no longer blocks for 16.7 ms (Tekken 6): it returns at once,
as on a PSP.

Since the state now changes later, a Disconnect can come while a Create
is still pending. Keep track of whether the server has us in a group, so
the Disconnect still reaches it.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 22:39:07 +01:00
Henrik RydgårdandClaude Opus 5.5 5021e3ccc1 sceNetAdhocctlInit: Don't wait for the login to the server
On a PSP, Init returns at once and a Create, Connect or Scan right after
it works (pspautotests net/adhoc/ctlinit). The login now runs on in the
background, and those requests wait for it, as they already did when it
was lost. The first login no longer resets the busy flag, which would
have cleared it for a request already waiting on it.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 22:39:07 +01:00
Henrik RydgårdandClaude Opus 5.5 a276f00dd0 proAdhoc: Run the adhocctl client from a CoreTiming event
The friendFinder thread is replaced by friendFinderPoll(), which a
CoreTiming event calls every millisecond of emulated time on the CPU
thread. It handles everything the server has sent since the last poll,
instead of one event every 10 ms, and pings it at the same real-time
interval. Resolving the server and logging in, which can take seconds,
run on a thread of their own; the poll starts it and takes over the
session when it's done.

Everything else about the client is unchanged, including sceNetAdhocctlInit
waiting for the login. The savestate section goes to version 11 for the
poll event; older states get it scheduled on load.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 22:39:07 +01:00
Henrik RydgårdandClaude Opus 5.5 e5bafd423d sceNetAdhocMatching: Report every join request, after SetHelloOpt too
The rule that dropped the first join after the hello data changed came
from a race in pspautotests net/adhoc/pair/matching2, where the host
started looking for the REQUEST after it had come. A PSP reports the
join at once however soon after SetHelloOpt it arrives, also when the
host changes its hello data every second (net/adhoc/pair/matchingjoin).
Ridge Racer does that every 5 seconds, so its clients could never enter
a race.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 22:39:06 +01:00
Henrik RydgårdandClaude Opus 5.5 db7c9ede34 sceNetAdhocctlGetPeerInfo: NO_ENTRY for the local MAC
The console itself isn't in its peer list, and a PSP returns NO_ENTRY
for it (pspautotests net/adhoc/adhocctl).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 22:39:05 +01:00
Henrik RydgårdandClaude Opus 5.5 48c9bf9988 sceNetAdhocPdpRecv: Don't copy a datagram that doesn't fit
With a buffer smaller than the next datagram, a PSP returns
NOT_ENOUGH_SPACE with the datagram's size, copies nothing, and keeps
the datagram for the next call (pspautotests net/adhoc/pdp). The first
bytes were copied into the buffer.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 22:39:05 +01:00
Henrik RydgårdandClaude Opus 5.5 5cc3554667 sceNetAdhocMatching: Pass the pair matching tests
The matching, matching2 and matching3 pair tests (recorded on two PSPs)
now pass on both sides:

- Event and input messages are queued in order. They were stacks, so a
  burst came out reversed (PONG and PANG crossed), and two places spawned
  their messages backwards to compensate.
- A peer stays busy until its DATA_ACK is delivered (DATA_BUSY), and
  AbortSendData discards that DATA_ACK.
- A host's SelectTarget while its accept is in progress gives
  REQUEST_IN_PROGRESS.
- A child that leaves, a cancelled request and a denied request (client
  and P2P) go back to a known non-member peer: no TIMEOUT, hellos keep
  coming, and it can ask again; a cancelled request gives
  TARGET_NOT_READY. A client says BYE to a host it has left.
- After SetHelloOpt the next join from an unknown peer reaches the host
  without a REQUEST.
- GetMembers lists the console itself first, also on a client.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 22:39:05 +01:00
Henrik RydgårdandClaude Opus 5.5 874080ced2 sceNetAdhocDiscover: Go through the states a PSP goes through
With nobody to find (pspautotests net/adhoc/discover): InitStart leaves
the result field alone and fails with 0x80411301 while running,
adhocctl enters the DISCOVER state after the first Update, and after a
Stop adhocctl goes back to DISCONNECTED with a DISCONNECT event, then
the status becomes COMPLETED with the result cancelled. RequestSuspend
while discovering fails with 0x80411303 instead of cancelling.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 22:39:05 +01:00
Henrik RydgårdandClaude Opus 5.5 d9ac972915 sceNetAdhocPtpSend, PtpRecv: DISCONNECTED once the peer has gone
After a Recv has found the peer gone, a PSP keeps returning
DISCONNECTED for Send and Recv on that socket, not NOT_CONNECTED
(pspautotests net/adhoc/ptp).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 22:39:05 +01:00
Henrik RydgårdandClaude Opus 5.5 e94ae7f190 sceNetAdhocMatchingStart2: Check the stack sizes and partitions
A stack size too small to be one (where a partition was passed) gives
STACKSIZE_TOO_SHORT, and a partition outside 1-6 then INVALID_ID
(pspautotests net/adhoc/matching). How small a stack still works isn't
known, so only sizes below 0x100 are refused.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 22:39:05 +01:00
Henrik RydgårdandClaude Opus 5.5 1649bd2072 sceNetAdhocPtpOpen: Accept the broadcast address as the destination
A PSP creates the socket (pspautotests net/adhoc/ptp); connecting it
just never succeeds.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 22:39:05 +01:00
Henrik RydgårdandClaude Opus 5.5 25ee41f135 sceNetAdhocMatching: Hand out context IDs in increasing order
A PSP numbers matching contexts 1 to 15 in increasing order and then
starts over, instead of reusing the lowest free ID (pspautotests
net/adhoc/matching).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 22:39:05 +01:00
Henrik RydgårdandClaude Opus 5.5 dc450f79b3 sceNetAdhoc: Hand out socket IDs in increasing order, like the PSP
A PSP numbers adhoc sockets in increasing order instead of reusing the
lowest free ID, and a PDP or PTP create that fails its address or port
check, or finds the PDP port in use, uses up an ID too (pspautotests
net/adhoc/pdp and net/adhoc/ptp). Find free slots starting after the last
one handed out.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 22:39:05 +01:00
Henrik RydgårdandClaude Opus 5.5 ce81c07a69 sceNetAdhocMatching: Start returns 0, not the matching thread's ID
The fake matching thread was started through sceKernelStartThread,
which switched threads inside the call, and the game got the new
thread's ID back. On a PSP, Start returns 0 (pspautotests
net/adhoc/matching). Create and start it like adhocctl's thread instead.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 22:39:05 +01:00
Henrik RydgårdandClaude Opus 5.5 97b18b95bb sceNetAdhocPollSocket: Report what a PSP reports
On a PSP (pspautotests net/adhoc/poll):
- a PTP listener reports SEND and FLUSH, and ACCEPT with a connection
  waiting;
- a connected PTP socket reports SEND, CONNECT and FLUSH, and RECV with
  data waiting;
- a connection whose peer has closed reports DISCONNECT alone, and the
  call returns EXCEPTION_EVENT;
- an unknown ID puts INVALID in the first entry;
- a blocking poll with nothing to report by its timeout returns TIMEOUT,
  and one that hits a disconnected socket or the poll alert returns at
  once.

A connection PtpOpen started counts as connected once it's up. Sockets
still connecting keep PPSSPP's own events (Fate Unlimited Codes,
Bleach), and a refused connect still isn't reported, since PPSSPP keeps
retrying it.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 22:39:05 +01:00
Henrik RydgårdandClaude Opus 5.5 ae906cb292 sceNetAdhocMatching: Don't list a member twice
A child the host accepts right after another one learns about that one
twice: from the sibling list in its accept packet, and from the BIRTH
packet the host sends about it. It added it both times, so with three or
four players the last one to join had an extra member (pspautotests
net/adhoc/group/matching, not recorded on PSPs yet).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 22:39:05 +01:00
Henrik RydgårdandClaude Opus 5.5 0d089972a3 sceNetAdhoc: Socket alerts don't stop calls that can complete at once
On a PSP (pspautotests net/adhoc/alerts), PtpConnect on a connected
socket and PtpFlush with nothing left to send return 0 with their alerts
set, and the poll alert reports ALERT alone in revents. A connection
PtpOpen started counts as connected once it's up, even if no call has
looked at it since.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 22:39:05 +01:00
Henrik RydgårdandClaude Opus 5.5 7e44dc7e8d sceNetAdhocPdpCreate: Give port 0 a port in 0x1000-0x7FFF
On a PSP, a PDP socket created on port 0 gets a free port between
0x1000 and 0x7FFF (pspautotests net/adhoc/limits). The host picked
one instead, usually 0x8000 or above once the port offset was taken
off. Pick it the way PtpOpen already does for PTP, and try another if
the host port is taken. Port 0 still doesn't become the fixed port the
offset alone would give (Phantasy Star Portable 2).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 22:39:05 +01:00
Henrik Rydgård cf17b0ac79 Merge pull request #22841 from hrydgard/more-ui-tweaks
More UI tweaks, allow buffered frames = 2 again
2026-10-29 15:48:24 +01:00
Henrik Rydgård b319b1783a Load round rect radius from theme. Set it to 0 in 1995. 2026-10-29 15:28:37 +01:00
Henrik Rydgård 9c5f87f33d Allow picking 2 frames to buffer again. 2026-10-29 15:25:34 +01:00
Henrik Rydgård d09a9ffa11 Correct corner radius of popups to match inner rounded things 2026-10-29 15:12:55 +01:00
Henrik Rydgård 37fa5896c1 Merge pull request #22839 from hrydgard/play-crash-fixes-4
Play crash fixes: vertex range, interrupt and vtimer edge cases, GL push buffer
2026-10-29 09:17:42 +01:00
Henrik Rydgård 1ed60bb835 Merge pull request #22836 from hrydgard/headless-ppge-slots
Headless/debugger fixes, and a crash loading states after a system dialog
2026-10-29 09:17:17 +01:00
Henrik Rydgård 5ceca071fe Merge pull request #22838 from hrydgard/lag-sync-master
Lag sync: Fix drift and pacing. Set built-in relay to pump every millisecond
2026-10-29 09:16:22 +01:00
Henrik RydgårdandClaude Opus 5.5 d5503978e7 Interrupts: Check whether a sub-interrupt is enabled when its turn comes
All the enabled sub-interrupt handlers were queued when the interrupt fired,
so a handler disabled by an earlier one in the same VBLANK still ran, and one
an earlier handler enabled waited for the next VBLANK. The PSP checks each
handler as it gets to it (intr/disablepending): skip a disabled one, and
queue a newly enabled one if it comes after the handler that's running.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 08:58:33 +01:00
Henrik RydgårdandClaude Opus 5.5 0987fdb169 VTimers: Don't run a pending timer that was cancelled, stopped or given a new handler
The interrupt ran it anyway, and after sceKernelCancelVTimerHandler or
sceKernelSetVTimerHandler with no handler that meant jumping to address 0.
This happens when one timer's handler cancels another timer that came due
together with it, or when a timer comes due while interrupts are suspended.
On the PSP none of these run (threads/vtimers/pending/cancelpending). The
pending entry becomes an invalid id, which the interrupt pops and skips, as
with deleted timers.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 08:58:32 +01:00
Henrik RydgårdandClaude Opus 5.5 fe6cfb6c77 Test: intr/releaseself, a subinterrupt handler releasing itself
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 08:22:19 +01:00
Henrik RydgårdandClaude Opus 5.5 b566634318 Add --lag-sync, and let it pace headless --realtime
The JIT cores only return from the run loop at the end of a frame, so
--realtime paces once per frame, like the app. With --lag-sync, lag
sync paces every emulated millisecond instead and --realtime's own
sleep stays out of its way. Two instances measuring round trips over
the relay (pspautotests net/adhoc/pair/timing) go from 15-31 ms to
3 ms.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 07:33:14 +01:00
Henrik RydgårdandClaude Opus 5.5 d40290b30d sceDisplay: Keep lag sync from drifting, and let it pace alone
Lag sync ("Force real clock sync") waited for 1 ms after the end of the
previous sync, so every sync's oversleep added up: GTA VCS lost 0.4 s
per minute against the wall clock. Wait instead for a fixed base plus
the emulated time since, which can't drift, and start over from now
only when more than 100 ms behind (a pause, or loading).

While it runs, the frame throttle's own waits are skipped. They were out
of phase with lag sync, which then ran flat out to catch up, so frames
were bursts again. The wait itself is the precise one the frame
throttle uses, instead of a yield() spin on Windows.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 07:33:14 +01:00