Commit Graph

66 Commits

Author SHA1 Message Date
Gregor Richards
65e308c92d Reset player flip state when we lose Netplay connection
This fixes a bug whereby reconnections would be out of sync due to the
flip state of the server and client being different. The server now
resets its flip state on disconnect.
2016-09-22 21:56:19 -04:00
Gregor Richards
06550ec289 Check if setting CLOEXEC fails on the Netplay server port. 2016-09-22 21:02:26 -04:00
Gregor Richards
1503b3d42f Don't serialize frame 0 in netplay, as the core may not yet be ready 2016-09-22 17:36:46 -04:00
Twinaphex
34944cac0f remove unused variables 2016-09-22 22:01:35 +02:00
Twinaphex
b2240c5783 Merge pull request #3631 from GregorR/netplay-reconnect
Support for Netplay reconnection
2016-09-22 14:27:38 +02:00
Gregor Richards
4a83e6dee9 Both sides should send the same header so either can check the other 2016-09-22 07:06:26 -04:00
Gregor Richards
42f8ed4338 Support for Netplay reconnection
In net server mode, when the client disconnects, the server now returns
to listening mode, in anticipation of the client reconnecting.
2016-09-21 22:37:43 -04:00
twinaphex
017bf34d71 Prevent null pointer dereference defect 2016-09-22 03:59:02 +02:00
Gregor Richards
4b370cf55e Re-add TCP_NODELAY in host mode. 2016-09-21 17:31:19 -04:00
Gregor Richards
c2d9e7a538 Support for late connection to Netplay host mode
This changes netplay host mode's behavior in net (normal) mode from
immediately blocking to waiting for a connection while allowing the game
to run, like spectator mode.
2016-09-21 17:26:02 -04:00
twinaphex
b548362869 Use encoding_crc32 2016-09-21 12:31:40 +02:00
Gregor Richards
851cd0fe6e Removing now-unused BSV-related functions from Netplay 2016-09-20 22:32:23 -04:00
twinaphex
d7e9c198e6 Cleanups to netplay 2016-09-17 18:21:29 +02:00
Twinaphex
e21662b3e9 Merge pull request #3607 from GregorR/netplay-savestates
Netplay savestate loading frontend changes
2016-09-17 18:15:49 +02:00
Alcaro
ec2fd5ebd7 Fix warning 2016-09-17 17:55:42 +02:00
Gregor Richards
d53373a5cb Netplay savestate loading frontend changes
Support for the frontend to inform Netplay when a savestate has been
loaded, so Netplay can in turn inform the peer.
2016-09-17 11:24:23 -04:00
twinaphex
a55feae619 CXX_BUILD fixes 2016-09-16 13:38:06 +02:00
Gregor Richards
ad336df709 Reimplemented Netplay spectate mode
Spectate mode is now far more similar to net (normal) mode, and, more
importantly, it works. In addition, spectate mode will not fast-forward
to catch up with the server if it lags too far behind.
2016-09-15 23:06:32 -04:00
Gregor Richards
4076b14568 Make Netplay flipping deterministic
The receiving side of a player-flip request now does a forced rewind to
assure that any already-computed frames are computed with the players on
the right side.
2016-09-15 23:06:32 -04:00
Gregor Richards
827009d2d3 A few corrections/clarifications in the Netplay README 2016-09-15 23:06:32 -04:00
Gregor Richards
2a0202ccf5 Making Netplay check frequency configurable. 2016-09-15 23:06:32 -04:00
Gregor Richards
9a91570db8 Adding new network commands to the Netplay README. 2016-09-15 23:06:32 -04:00
Gregor Richards
7271d1c3fa Support for frame CRCing
Every frame (soon to be configurable), the server does a CRC-32 hash and
sends it to the client. If the client finds that its own hash is
different from the server's, it requests a fresh savestate.

This is a last-ditch effort to sync if all else fails, and it's a
best-effort situation. The size of the buffer should assure that we
always still have the frame around to CRC, but I imagine there are edge
cases where we don't. If you're in an edge case, the CRC is ignored.
2016-09-15 23:06:32 -04:00
Gregor Richards
79eba578ff Remote state load bugfix: Don't skip an input frame 2016-09-15 23:06:32 -04:00
Gregor Richards
5f90f072ba Documentation fixup
Fixing up some of the function documentation, and documenting the packet
formats.
2016-09-15 23:06:32 -04:00
Gregor Richards
27188e102d Support for savestate loading over netplay
Assuming the core supports saving/loading states, and (crucially)
assuming the states are portable across the architectures on both sides
of the connection, Netplay now supports the transmission of savestates.
Right now the frontend doesn't actually send any such requests, as it's
not clear exactly where the code for that should be.

This works in either direction, although I'll admit I have no idea what
happens if they both load at the same time.
2016-09-15 23:06:32 -04:00
Gregor Richards
4d11b1d674 Don't misinterpret "Netplay isn't loaded" as "must pause" :) 2016-09-15 23:06:32 -04:00
Gregor Richards
c3186c57ab Don't stall if connection lost with remote paused 2016-09-15 23:06:32 -04:00
Gregor Richards
96186438ec Remote pausing
Support for remote pausing, and with it, support for Netplay pausing the
frontend correctly. With this patch alone this doesn't work, since
there's no clean way for the frontend to tell Netplay that it's paused.
2016-09-15 23:06:32 -04:00
twinaphex
a71cd89379 Cleanup unused variable 2016-09-15 21:26:10 +02:00
Gregor Richards
9bc78d25a0 Minor stylistic fixes for clarity and consistency with the rest of
RetroArch.
2016-09-14 10:06:57 -04:00
Gregor Richards
b1a2e096e5 Use consistent quotes in Netplay README. 2016-09-14 10:03:26 -04:00
Gregor Richards
12bf3e4824 Rather than repeating the same sample for every frame while stalled,
just pause the sound while stalled. Both options are annoying, but this
option is less annoying.
2016-09-13 21:45:06 -04:00
Gregor Richards
ea0bb6f812 Just in case some systems don't have TCP_NODElAY, put that in an ifdef. 2016-09-13 21:39:10 -04:00
Gregor Richards
d4e074dbed Moved the advance of self_ptr to the same place as the advance of
self_frame_count, which is much clearer.
2016-09-13 21:32:57 -04:00
Gregor Richards
99b5ed92ed other should always be <= both real AND self. Before this fix, it was
possible (albeit unlikely) for the remote to get so far ahead of us that
they actually overwrote our own current data :)
2016-09-13 21:32:57 -04:00
Gregor Richards
1267e5e867 A few clarifications regarding the buffer's have_remote and used_real
data. simulation shouldn't touch 'em.
2016-09-13 21:32:57 -04:00
Gregor Richards
a0cfdb8a9c naks now cause disconnection. 2016-09-13 21:32:57 -04:00
Gregor Richards
b140b16b5d Removed Netplay positive acknowledgement messages: They didn't document
their corresponding sent message, and so couldn't be used for
acknowledgement anyway. Negative acknowledgement is sufficient.
2016-09-13 21:32:57 -04:00
Gregor Richards
6829b80c6b Reimplemented disconnection based on stalls. If we stall for 10 seconds,
disconnect.
2016-09-13 21:32:57 -04:00
Gregor Richards
8aa48cd3f9 Reinstituted "standard" timeout-based disconnection, which only works
for stall_frames==0 due to the fixes to not stall the UI.
2016-09-13 21:32:57 -04:00
Gregor Richards
0ccc39769d Adding my copyright lines to files I've touched. 2016-09-13 21:32:57 -04:00
Gregor Richards
ae8e695644 Fixing indentation to align with the rest of RetroArch. 2016-09-13 21:32:57 -04:00
Gregor Richards
f9f4e15d33 Removing commented-out code (mostly old UDP stuff) 2016-09-13 21:32:57 -04:00
Gregor Richards
c5fe0ec6be Updating the Netplay README to be true of the current implementation. 2016-09-13 21:32:57 -04:00
Gregor Richards
147d739197 Fixed stall_frames=0 mode to only block if frames are actually needed.
Rather than counting on the complexicon of used_real calculations, set
used_real when... real is used. (Problem: If the core doesn't read input
at all, used_real won't be set; todo: test with handhelds.) Minor
resilience fixes.
2016-09-13 21:32:57 -04:00
Gregor Richards
c7d0bf90f6 Bugfixes to bring Netplay Nouveau from "kinda working" to "stably
working":

(1) Fixups to the stall logic to make sure it always receives frames
while stalling :)

(2) Disused the used_real field. It was misconfigured and would
frequently claim to be using real data when real data hadn't been
used... this means more replays for now, but used_real will be readded.
(TODO)

(3) Stall duration is now related to sync frames, and thus configurable.

(4) Delta frames were having the good ol' initialization problem, as
frame==0 was indistinguishable from unused. Quickfixed by adding a
"used" field, but maybe there's a better way.

(5) If serialization fails, switch immediately to blocking mode
(stall_frames = 0). Blocking mode barely works, but if serialization
fails, no mode will work!

(6) I'm not sure which bug my replaying-from-previous-frame was trying
to fix, but the correct behavior is to replay from the last frame we had
vital information, not the frame prior. Notionally this should just be
an efficiency thing, but unsigned arithmetic at 0 made this a "just
ignore all input from now on" thing.
2016-09-13 21:32:57 -04:00
Gregor Richards
07e869ccae is_simulated was confusing and poorly named. Using have_remote in its
place, which is simply true if we've received remote data. Could also
just use read_frame_crount instead, but this keeps the info frame-local.
2016-09-13 21:32:57 -04:00
Gregor Richards
4f16a19f5e Rather than stalling by blocking and becoming unresponsive, stall by
replaying the same frame. TODO: Maybe mute the audio?
2016-09-13 21:32:57 -04:00
Gregor Richards
5edfbeafb0 Switched Netplay over to TCP. A lot of the stalling logic had to change
for this, and in particular, it now sometimes stalls in a way that makes
it very difficult to actually input anything (whoops :) ). Simply
setting the sync frames higher avoids that. With supported cores, this
is incredibly risilient, but when it fails, it mostly fails to freezing,
which is less than ideal.

TODO: Stall frames should be configurable. All the UDP code is still
there but commented out, should be gutted. The original fast-forward
code is now commented out, but really both fast-forward and stalling
should be options; the only complication is that it needs to send
simulated self-input for fast-forward.
2016-09-13 21:32:57 -04:00