Gregor Richards
55157e934d
input_latency_frames is now configurable and has a range
2017-02-01 22:54:03 -05:00
Gregor Richards
5f5a8dc6d7
CRC searching should really start from run_ptr, not input_ptr
2017-02-01 22:54:03 -05:00
Gregor Richards
561eb42c84
Don't enable input latency if we're not connected
2017-02-01 22:54:03 -05:00
Gregor Richards
c4cb94db19
New approach to input latency
2017-02-01 22:54:03 -05:00
Twinaphex
2596de53ec
Merge pull request #4523 from fr500/master
...
add support for no content cores
2017-01-29 04:51:50 +01:00
Gregor Richards
3d24e844a4
We're in client mode if either server OR direct_host is set
2017-01-28 22:39:17 -05:00
radius
cfb882f4fc
add support for no content cores
2017-01-28 22:12:34 -05:00
Gregor Richards
12e8deedb4
Only disable netplay in deinit_netplay if it was actually on
2017-01-25 22:08:05 -05:00
Gregor Richards
263470a66c
Set netplay_enabled to false during deinit_netplay
...
This simply prevents the odd behavior of netplay automatically
restarting, or trying to restart, when you load a new game.
2017-01-25 21:23:55 -05:00
fr500
29a1d504d7
allow different server ports again
2017-01-25 17:48:43 -05:00
fr500
af14da93d5
fix whitespace
2017-01-25 17:43:56 -05:00
fr500
c097c2062e
fix when port is 0
2017-01-25 17:43:17 -05:00
twinaphex
178d959cad
Fix C89_BUILD
2017-01-22 21:24:57 +01:00
twinaphex
36b2c48cb5
Fix CXX_BUILD
2017-01-22 21:19:39 +01:00
Twinaphex
3ff158b907
Merge pull request #4436 from fr500/lobby
...
Basic lobby system (don't merge, just for tracking)
2017-01-22 21:08:43 +01:00
radius
a0e6a24f38
use deferred netplay init
2017-01-22 13:07:30 -05:00
radius
33cc168eb9
allow connecting to lobby port instead of the port in settings
2017-01-22 12:29:29 -05:00
Gregor Richards
addcbb896a
Fixing an incompatibility between stateless mode and the password prompt
...
Previously, the host would time out waiting for the guest to enter a
password, as the timeout was not conditionalized on whether the guest
was actually playing. This fixes that.
2017-01-22 10:43:09 -05:00
twinaphex
96c8ca5a09
Header update #1
2017-01-22 13:40:32 +01:00
twinaphex
768ce0854c
Make driver_set_nonblock_state a public function
2017-01-22 12:47:17 +01:00
Gregor Richards
8c304dfda7
Make the client send an INFO even if it's wrong
...
This simply gives the server a reason why the client disconnected,
rather than a generic, unexplained disconnection.
2017-01-21 13:18:31 -05:00
radius
0de43b954e
can now connect when starting a game too
2017-01-20 16:04:10 -05:00
radius
dcca2f8131
only announce if host
2017-01-20 14:33:57 -05:00
radius
8a1fb24fe9
update url
2017-01-19 23:21:22 -05:00
radius
930cde21a2
crc is announced now (for cores that don't have needs_fullpath set)
2017-01-19 21:43:46 -05:00
radius
3dd6c91493
start populating the list
2017-01-19 08:56:56 -05:00
radius
95a2da81e4
start adding ui elements to insert room list into
2017-01-18 22:46:48 -05:00
radius
72cdb390cb
reannounce roughly every minute
2017-01-18 21:31:02 -05:00
radius
ca2c2a8ce7
separate this
2017-01-18 21:12:01 -05:00
twinaphex
9a87bada55
Style nit
2017-01-17 19:05:33 +01:00
radius
f2beb49fd8
Announce netplay games to lobby.libretro.com
2017-01-16 22:24:35 -05:00
Gregor Richards
a0f236a83f
Use autosave_lock and autosave_unlock during SRAM negotiation
...
The SRAM transfer in netplay handshake now uses autosave_lock and
autosave_unlock. Will possibly fix a hang/crash bug on Android with
netplay and autosave conflicting.
2017-01-15 18:10:24 -05:00
Gregor Richards
e9e7878b5c
netplay_buf should use O_NOSIGNAL to avoid SIGPIPE
2017-01-15 10:59:25 -05:00
twinaphex
80bc91477a
Remove header include
2017-01-09 02:19:16 +01:00
Gregor Richards
20c0de352c
Fix to the netplay_is_alive function when server isn't playing
...
Fixes a bug by which very strange things would happen if the server went
to spectator mode with exactly one connected client.
2017-01-02 20:03:00 -05:00
Gregor Richards
ee55e2d783
LOAD_SAVESTATE can't use player frame timing from a server that's not playing
2017-01-02 18:56:08 -05:00
Gregor Richards
cd281d5757
Reverse catch-up, i.e., server-demanded stalling
...
Previously, if two clients were connected to the same server and one of
them was ahead of the server, the only way to rectify that situation was
for the client to get so far ahead that it stalled, as the server could
only catch up with an ahead client if all clients were ahead. That's
unrealistic. This gives the server the alternate option of demanding
that a client stall. This keeps things nicely in line even with >2
players.
2016-12-24 15:25:03 -05:00
twinaphex
70a1a5f38e
Buildfixes
2016-12-24 01:48:36 +01:00
twinaphex
64dc7daeca
(MSVC) MSVC build fixes
2016-12-24 01:45:37 +01:00
twinaphex
77e5cdbfde
Revert "Rename runloop_msg_queue_push to video_driver_msg_queue_push"
...
This reverts commit ad7386913c
.
2016-12-22 23:36:11 +01:00
twinaphex
ad7386913c
Rename runloop_msg_queue_push to video_driver_msg_queue_push
...
and move code to video_driver.c
2016-12-22 13:15:02 +01:00
Gregor Richards
3df38fa40e
Silence two compiler warnings
2016-12-21 10:13:31 -05:00
Gregor Richards
f0bb0a77f8
Make netplay no longer cache passwords
...
Netplay now always loads passwords from the configuration, so that
passwords can be changed mid-session.
2016-12-21 09:51:50 -05:00
Gregor Richards
c780e7db43
Accept the variable size quirk
...
Since the quirks protocol was that a core could report variable
savestate size, but the host then tells it "no", we should actually
accept the variable size quirk in netplay, since RetroArch refuses to
allow cores to actually produce variable-size states.
2016-12-20 22:17:15 -05:00
Gregor Richards
7bb26eb5c5
Fix some minor checked-return bugs
...
Should resolve Coverity reports 158862, 158861, 158860.
158863 is spurious but I have no idea how to convince Coverity of that.
2016-12-20 19:16:31 -05:00
Gregor Richards
8f35de9da0
Adding blocks in the empty-RECV case to (hopefully) suppress warnings
2016-12-19 17:31:24 -05:00
Twinaphex
1d36c5effa
Silence warnings - remove unused variables
2016-12-19 23:19:59 +01:00
Alcaro
fc1c83d8c0
Shut up false positive warning
2016-12-19 16:38:55 +01:00
Gregor Richards
11a3063fc6
Forgot the other half of that last bugfix
2016-12-18 20:08:21 -05:00
Gregor Richards
8f8e6dfc5f
Unset the netplay callbacks when we deinit netplay
...
This is a bugfix to master that hadn't migrated to my branch.
2016-12-18 19:54:01 -05:00
Gregor Richards
dcd4b3046b
Making negative check_frames be "check only" mode
2016-12-18 19:28:44 -05:00
Gregor Richards
677ffa9ebd
Support different forms of compression from different clients.
2016-12-18 19:28:44 -05:00
Gregor Richards
84c33634a6
Communicate paused-ness during initial connection SYNC.
2016-12-18 19:28:44 -05:00
Gregor Richards
60b81e3a51
Nominal support for server sending blank INFO
2016-12-18 19:28:44 -05:00
Gregor Richards
e79f30604f
Better stall timing
...
We now stall not until we've reached parity (which makes no sense since
we expect latency), but only until we're not likely to stall again.
2016-12-18 19:28:44 -05:00
Gregor Richards
8195a5bcee
Try to catch up if we're falling behind our peer, not anyone
...
In the previous catch-up system, we would only try to catch up if we
were falling behind the farthest-behind peer. However, as they would
also only try to catch up to us, everyone basically agreed to the
worst-case latency. It makes more sense to try to be in parity with your
direct peer than with indirect connections.
2016-12-18 19:28:44 -05:00
Gregor Richards
b535412914
Adding nick-changing
...
This is just to disambiguate multiple connections with the same nick
(usually Anonymous), which will now become e.g. Anonymous (2).
2016-12-18 19:28:44 -05:00
Gregor Richards
7c2f12fbd9
Removign some leftover debug code
2016-12-18 19:28:44 -05:00
Gregor Richards
04266cf4f7
Run synchronization even when stalled
...
Previously, we could be stalled by one player but still reading data
from another, which would wedge the client because we would never act
upon the newly-read data. Now we act upon data even if we're stalled.
Fixes bugs in initial connection with high latency.
2016-12-18 19:28:44 -05:00
Gregor Richards
d6d96704fd
Don't panic about delta frames not being ready if we're not even fully
...
connected yet.
2016-12-18 19:28:43 -05:00
Gregor Richards
8b21014d07
Improvements to handshake protocol
...
Making the netplay handshake protocol send the core and content as an
explicit command, so that the other side can (notionally) choose to load
it. That isn't implemented, of course.
2016-12-18 19:28:43 -05:00
Gregor Richards
1fa60b396f
Improvements to the communication of netplay pausing.
2016-12-18 19:28:43 -05:00
Gregor Richards
db2c8de44c
Slightly changing how catch-up works to avoid spamming the console
2016-12-18 19:28:43 -05:00
Gregor Richards
1141825605
Documenting (somewhat) the new sync system.
2016-12-18 19:28:43 -05:00
Gregor Richards
a20e79bbb5
Server needs double the buffer frames since two connected clients could
...
be ahead and behind by the max frames
2016-12-18 19:28:43 -05:00
Gregor Richards
f7f6590156
Goodbye delay_frames! stateless_mode is the new delay_frames=0
2016-12-18 19:28:43 -05:00
Gregor Richards
45d732a014
New sync system
...
The idea:
* Use a fixed number of delay_frames (eventually to be fixed at 120,
currently still uses the config variable, 0 will still be an option)
* Determine how long it takes to simulate a frame.
* Stall only if resimulating the intervening frames would be
sufficiently annoying (currently fixed at three frames worth of
time)
Because clients always try to catch up, the actual frame delay works out
automatically to be minimally zero and maximally the latency. If one
client is underpowered but the other is fine, the powerful one will
automatically take up the slack. Seems like the most reasonable system.
2016-12-18 19:28:43 -05:00
Gregor Richards
6890456ac0
Updated error messages and fixed some stall-related disconnected-client
...
bugs.
2016-12-18 19:28:43 -05:00
Gregor Richards
6658826759
CRC validity checking. Ignore CRCs if they don't work.
2016-12-18 19:28:43 -05:00
Gregor Richards
a6cd8c3aa1
Error reporting when a password is refused, plus a bugfix so that
...
socket_receive_all_nonblocking actually reports disconnects.
2016-12-18 19:28:43 -05:00
Gregor Richards
df2600fbf4
Added error reporting when a player is not allowed to play.
2016-12-18 19:28:43 -05:00
Gregor Richards
bade067d9a
Support for catching up if the netplay peer is ahead of us.
2016-12-18 19:28:43 -05:00
Gregor Richards
4e01481b39
Only allow clients to load state if there's no further synchronization
2016-12-18 19:28:43 -05:00
Gregor Richards
70d04ec6a5
Apparently the frontend gets really pissy if it can't check the flip state
2016-12-18 19:28:43 -05:00
Gregor Richards
da7efcb939
Cleaning up netplay headers.
2016-12-18 19:28:43 -05:00
Gregor Richards
9c6ac2b934
Only the server can flip users.
2016-12-18 19:28:43 -05:00
Gregor Richards
03415c261d
Added netplay spectator password separate from play password
2016-12-18 19:28:43 -05:00
Gregor Richards
694b7a9723
Don't allow more players to join than are actually being polled
2016-12-18 19:28:43 -05:00
Gregor Richards
7ad4e3f115
Per-connection stalling
2016-12-18 19:28:43 -05:00
Gregor Richards
1b22191869
Deduplicating some code.
2016-12-18 19:28:43 -05:00
Gregor Richards
f619789e48
Refactoring: netplay_common.c -> netplay_handshake.c/netplay_delta.c
...
Refactoring netplay_common into its two actual components, the handshake
and delta-frame related functions.
2016-12-18 19:28:43 -05:00
Gregor Richards
4e905bf524
Refactoring: netplay_sync.c
...
Renamed netplay_net.c to netplay_sync.c, as all that remains in that
file is synchronization-related functions.
2016-12-18 19:28:43 -05:00
Gregor Richards
257c5d3188
Refactoring netplay_init.c/netplay_net.c
...
Split the remainder of netplay.c into netplay_init.c and netplay_net.c
(which will soon be netplay_sync.c)
2016-12-18 19:28:43 -05:00
Gregor Richards
7e2465ef1f
Refactoring: Moving I/O functionality into netplay_io.c
2016-12-18 19:28:43 -05:00
Gregor Richards
8c59c7dd77
Starting to refactor: Separating frontend stuff into netplay_frontend.c
2016-12-18 19:28:43 -05:00
Gregor Richards
ab989d7f41
More careful disconnection.
2016-12-18 19:28:43 -05:00
Gregor Richards
c8bba9a981
Simplistic builtin randomizer.
2016-12-18 19:28:43 -05:00
Gregor Richards
28e331b5fd
Remove some magic numbers
2016-12-18 19:28:43 -05:00
Gregor Richards
6e6f2bfdbe
Use a proper password hash across the line.
2016-12-18 19:28:42 -05:00
Gregor Richards
763a657f82
Terrible first cut at password (sent in plain text D-8)
2016-12-18 19:28:42 -05:00
Gregor Richards
5a0328dc09
Transfer pad configuration over netplay.
2016-12-18 19:28:42 -05:00
Gregor Richards
888e45c88b
Updates to allow the server to toggle game/watch
2016-12-18 19:28:42 -05:00
Gregor Richards
3631ff74ff
Very, very partial support for the server spectating
2016-12-18 19:28:42 -05:00
Gregor Richards
3d7f1f6575
Updating Netplay README to be more true
2016-12-18 19:28:42 -05:00
Gregor Richards
f6f9905ae3
Made remote pausing connection-specific
2016-12-18 19:28:42 -05:00
Gregor Richards
6556af1100
force_send_savestate is global again
...
We cannot send a savestate to only one player, as sending a savestate is
a synchronization event invalidating all prior input.
2016-12-18 19:28:42 -05:00
Gregor Richards
3d34c0222f
Added simple (cheesy) announcements on join/leave
2016-12-18 19:28:42 -05:00
Gregor Richards
727743aa7b
Fixing a couple FIXMEs/TODOs
2016-12-18 19:28:42 -05:00
Gregor Richards
2130fd81a5
Fixed simulation for >2 players
2016-12-18 19:28:42 -05:00
Gregor Richards
262d77546b
Adding game/watch key
...
Adding a key to toggle between playing and spectating. This key takes
the place of the previous flip key, although player flipping does
continue to work (and must be rebound if you still want it)
2016-12-18 19:28:42 -05:00
Gregor Richards
90c15fc7a6
Still need to flush input data.
2016-12-18 19:28:42 -05:00
Gregor Richards
e7ce01ad3b
More renaming
...
Now that remote_input_state isn't always remote, it should just be
called real_input_state (and is).
2016-12-18 19:28:42 -05:00
Gregor Richards
f6630ad85a
More fixups for >2 player mode. Mostly was just sending packets that
...
shouldn't have been sent.
2016-12-18 19:28:42 -05:00
Gregor Richards
479955a085
More 3-player support. The initial connection is wonky with
...
delay_frames>0
2016-12-18 19:28:42 -05:00
Gregor Richards
d373684bd5
Off-by-one
2016-12-18 19:28:42 -05:00
Gregor Richards
b51cf8be2b
Real, actual (almost) >2-player support.
2016-12-18 19:28:42 -05:00
Gregor Richards
2cc8c5c467
Removing silly foo_ names used to help migration.
2016-12-18 19:28:42 -05:00
Gregor Richards
53c46530aa
Very first tidbits of true multiplayer support (minus actual multiple
...
players so far)
2016-12-18 19:28:42 -05:00
Gregor Richards
b5cd187077
Since there's now only one mode, removing netplay_callbacks entirely.
2016-12-18 19:28:42 -05:00
Gregor Richards
9b2270f5d4
Say goodbye to spectator mode (for now)
2016-12-18 19:28:42 -05:00
Gregor Richards
4768970d52
Moving force_send_savestate from netplay to connection.
2016-12-18 19:28:42 -05:00
Gregor Richards
ea722b49c8
Move other_addr from netplay to connection.
2016-12-18 19:28:42 -05:00
Gregor Richards
d1d29143b2
Move nickname field to connections. Spectate mode officially broken.
2016-12-18 19:28:42 -05:00
Gregor Richards
189cc6e5d6
Moving socket buffers to per-connection (currently breaks
...
delay_frames=0)
2016-12-18 19:28:42 -05:00
Gregor Richards
b334f04bd5
Removing RARCH_ from Netplay stall reasons.
2016-12-18 19:28:42 -05:00
Gregor Richards
1e1abf6951
First step of multiple connections. Still only one connection actually
...
works.
2016-12-18 19:28:42 -05:00
Gregor Richards
3908e25895
Separating local mode from remote mode.
2016-12-18 19:28:42 -05:00
Gregor Richards
775584cdac
Adding more README info.
2016-12-18 19:28:42 -05:00
Gregor Richards
170b590bdb
Client now starts one frame after finishing connection, for easier sync.
2016-12-18 19:28:42 -05:00
Gregor Richards
80be19a7da
Renames for readability, merging frame and sram into sync packet.
2016-12-18 19:28:42 -05:00
Gregor Richards
f89e54fcb7
Now that we don't have to start at frame 0, use the server frame count
2016-12-18 19:28:42 -05:00
Gregor Richards
b3092f6fde
Moving things that shouldn't have been public into netplay_private.h
2016-12-18 19:28:41 -05:00
Gregor Richards
9054b51f0e
Finally, bringing back the first frame (in nonblocking mode)
2016-12-18 19:28:41 -05:00
Gregor Richards
ce7686104d
Removing dead code
2016-12-18 19:28:41 -05:00
Gregor Richards
71e5cb756a
Support for delay_frames=0 in nonblocking mode.
2016-12-18 19:28:41 -05:00
Gregor Richards
1c8f056239
Nonblocking initial connection. Haven't tested with delay_frames=0 yet
2016-12-18 19:28:41 -05:00
Gregor Richards
3f8f9761f6
Fixes particularly affecting delay_frames=0 mode
2016-12-18 19:28:41 -05:00
Gregor Richards
91e7db3eca
This semicolon broke everything!
2016-12-18 19:28:40 -05:00
Gregor Richards
ba76528b8f
A more sophisticated status variable for later making the initial
...
connection nonblocking.
2016-12-18 19:28:39 -05:00
Gregor Richards
0f5eec9987
Use zbuffer_size instead of state_size for the packet buffer
2016-12-18 19:28:39 -05:00
Gregor Richards
1214541410
Added packet buffer resizing to cope with initialization quirks.
2016-12-18 19:28:38 -05:00
Gregor Richards
36a93e5697
Fix a few bits broken by merging.
2016-12-18 19:28:38 -05:00
Gregor Richards
900e5a79ec
Use nonblocking sockets for Netplay to avoid some stalls.
2016-12-18 19:28:38 -05:00
Twinaphex
d604441cea
Merge pull request #4266 from GregorR/netplay-better-simulation
...
Better netplay input simulation
2016-12-18 07:13:52 +01:00
Gregor Richards
1245d1b5dd
Better netplay input simulation
...
In resimulation mode, we only copy the buttons. The reason for this
is nonobvious:
If we resimulated nothing, then the /duration/ with which any input
was pressed would be approximately correct, since the original
simulation came in as the input came in, but the /number of times/
the input was pressed would be wrong, as there would be an
advancing wavefront of real data overtaking the simulated data
(which is really just real data offset by some frames).
That's acceptable for arrows in most situations, since the amount
you move is tied to the duration, but unacceptable for buttons,
which will seem to jerkily be pressed numerous times with those
wavefronts.
2016-12-18 00:46:02 -05:00
Gregor Richards
17e08c7a28
When we disconnect netplay, fully deinit it
...
In particular, we need to disentangle the interceding netplay callbacks.
In previous versions, if you disconnected netplay but were using a core
that made netplay stall for connections (i.e., one that has no
savestates), netplay would continue to stall because it still
interceded, even though it was supposed to be off. This fixes that.
2016-12-17 23:08:59 -05:00
Gregor Richards
8c08a5399d
Fix a possible NULL dereference in load_savestate
...
In the unlikely situation that serial_info wasn't provided and the delta
frame wasn't ready (possibly an impossible situation) it previously
would have segfaulted. This fixes that.
2016-12-15 16:19:54 -05:00
twinaphex
b4c1927e27
(MSVC 2003) Buildfixes
2006-05-18 13:31:43 +02:00
Gregor Richards
9a4b28735b
HAVE_SOCKET_LEGACY -> no IPv6
...
I had previously assumed that if AF_INET6 is defined, IPv6 support was
present. Some psychopathic console SDKs which shall not be named
actually define AF_INET6 but none of the IPv6 structures. As there is
therefore no way to determine whether IPv6 support is present at
preprocessor time, I'm ust assuming that HAVE_SOCKET_LEGACY implies no
IPv6. This means in effect that no consoles get IPv6 support.
2016-12-10 05:27:25 -05:00
twinaphex
4bf806ef00
(Netplay) Some build fixes for PS3 and others
2016-12-10 09:41:24 +01:00
Brad Parker
d558862e21
make "Got connection from" translatable. Update JP translation.
2016-12-09 23:38:00 -05:00
Gregor Richards
a70e21f328
Suppress a warning due to weird declarations on Windows.
2016-12-05 20:50:05 -05:00
Gregor Richards
e7f20a3673
Bugfixes
...
Silence some Coverity warnings (including a real memory leak) and be
more careful about checking IPv6 mode (for backwards compatibility with
systems that don't support IPv6 but may run code compiled to expect IPv6)
2016-12-05 00:45:40 -05:00
Gregor Richards
2b437ccb8c
Don't need to explicitly demand IPPROTO_TCP (and doing so breaks some
...
SOCKET_LEGACY builds)
2016-12-03 16:33:48 -05:00
Spec-Chum
7a6a0ee7af
Fix MSVC compilation
2016-12-03 18:24:49 +00:00
Gregor Richards
f725f38061
Use strlcpy
2016-12-03 09:58:16 -05:00
Gregor Richards
98f0d4d45b
Better error messages when the Netplay handshake fails
2016-12-03 09:46:20 -05:00
Gregor Richards
3c7fc73735
sendto fix for MSVC
2016-12-02 23:15:53 -05:00
Twinaphex
d71df5c06f
Merge pull request #4138 from GregorR/netplay-port-unsigned-short
...
Be explicit about types
2016-12-03 05:11:02 +01:00
Gregor Richards
cc664db231
Be explicit about types.
2016-12-02 23:08:34 -05:00