FD_SET crash in multiplayer: 2.3 beta 3 (during checkSockets(), aka select(2))

[anonymous]

I'm not ot actually sure what was my last action before this crash. Gentoo GNU/Linux, the game was compiled from source in my home directory and wasn't "make install"'ed

[anonymous]

Another crash, and i don't know if it is connected to this. Here's the console outputs:

FIRST CRASH

error   |01:13:17: [recvDroidMove] Packet from 0 refers to non-existent droid 217, [Human : p0]
error   |01:14:22: [recvGroupOrder] Packet from 4 refers to non-existent droid 164876, [Human : p4]
error   |01:14:22: [recvDroidInfo] Packet from 4 refers to non-existent droid 164876, [Human : p4]
error   |01:14:50: [recvGroupOrder] Packet from 2 refers to non-existent droid 165554, [Human : p2]
error   |01:14:50: [recvDroidInfo] Packet from 2 refers to non-existent droid 165554, [Human : p2]
error   |01:14:51: [recvGroupOrder] Packet from 2 refers to non-existent droid 165642, [Human : p2]
error   |01:14:51: [recvDroidInfo] Packet from 2 refers to non-existent droid 165642, [Human : p2]
error   |01:17:04: [recvDroidMove] Packet from 4 refers to non-existent droid 163684, [Human : p4]
error   |01:23:28: [recvGroupOrder] Packet from 3 refers to non-existent droid 166907, [Human : p3]
error   |01:23:37: [recvGroupOrder] Packet from 3 refers to non-existent droid 166283, [Human : p3]
error   |01:23:48: [recvGroupOrder] Packet from 3 refers to non-existent droid 166651, [Human : p3]
error   |01:23:48: [recvGroupOrder] Packet from 3 refers to non-existent droid 165915, [Human : p3]
error   |01:23:48: [recvDroidMove] Packet from 3 refers to non-existent droid 165915, [Human : p3]
error   |01:23:55: [recvDroidInfo] Packet from 3 refers to non-existent droid 166251, [Human : p3]
error   |01:24:08: [recvDroidMove] Packet from 2 refers to non-existent droid 167346, [Human : p2]
error   |01:28:44: [recvDroidInfo] Packet from 4 refers to non-existent droid 167892, [Human : p4]
error   |01:29:51: [recvDroidInfo] Packet from 4 refers to non-existent droid 167004, [Human : p4]
error   |01:31:27: [recvDroidInfo] Packet from 3 refers to non-existent droid 168747, [Human : p3]
error   |01:31:27: [recvDroidMove] Packet from 3 refers to non-existent droid 168747, [Human : p3]
error   |01:31:27: [recvDroidInfo] Packet from 3 refers to non-existent droid 168747, [Human : p3]
error   |01:31:30: [recvDroidInfo] Packet from 3 refers to non-existent droid 169123, [Human : p3]
error   |01:31:30: [recvDroidMove] Packet from 3 refers to non-existent droid 169123, [Human : p3]
error   |01:31:30: [recvDroidInfo] Packet from 3 refers to non-existent droid 169123, [Human : p3]
last message repeated 2 times
error   |01:31:30: [recvDroidMove] Packet from 3 refers to non-existent droid 169123, [Human : p3]
error   |01:31:30: [recvDroidInfo] Packet from 3 refers to non-existent droid 169123, [Human : p3]
error   |01:31:33: [recvDroidInfo] Packet from 3 refers to non-existent droid 169587, [Human : p3]
error   |01:31:33: [recvDroidMove] Packet from 3 refers to non-existent droid 169587, [Human : p3]
error   |01:31:33: [recvDroidInfo] Packet from 3 refers to non-existent droid 169587, [Human : p3]
last message repeated 2 times
error   |01:31:33: [recvDroidMove] Packet from 3 refers to non-existent droid 169587, [Human : p3]
error   |01:31:33: [recvDroidInfo] Packet from 3 refers to non-existent droid 169587, [Human : p3]
last message repeated 2 times
last message repeated 1 times (total 3 repeats)
error   |01:31:39: [recvDroidMove] Packet from 3 refers to non-existent droid 168947, [Human : p3]
error   |01:31:39: [recvDroidInfo] Packet from 3 refers to non-existent droid 168947, [Human : p3]
error   |01:31:43: [recvDroidMove] Packet from 3 refers to non-existent droid 168835, [Human : p3]
error   |01:31:43: [recvDroidInfo] Packet from 3 refers to non-existent droid 168835, [Human : p3]
last message repeated 2 times
last message repeated 1 times (total 3 repeats)
error   |01:31:47: [recvDroidInfo] Packet from 3 refers to non-existent droid 168931, [Human : p3]
error   |01:31:53: [recvDroidInfo] Packet from 3 refers to non-existent droid 168859, [Human : p3]
error   |01:33:43: [recvGroupOrder] Packet from 4 refers to non-existent droid 169348, [Human : p4]
error   |01:33:43: [recvDroidInfo] Packet from 4 refers to non-existent droid 169348, [Human : p4]
error   |01:35:01: [recvDroidMove] Packet from 2 refers to non-existent droid 168762, [Human : p2]
error   |01:35:10: [NETbcast] Failed to send message: Bad file descriptor
error   |01:35:10: [socketClose] Failed to close socket: Bad file descriptor
Saved dump file to '/tmp/warzone2100.gdmp-jP0KV3'
If you create a bugreport regarding this crash, please include this file.
Segmentation fault.

SECOND CRASH

error   |02:01:06: [recvDroidMove] Packet from 3 refers to non-existent droid 164483, [Human : p3]
error   |02:02:20: [recvGroupOrder] Packet from 1 refers to non-existent droid 171001, [Human : p1]
error   |02:02:25: [recvDroidInfo] Packet from 1 refers to non-existent droid 171497, [Human : p1]
error   |02:02:25: [recvDroidInfo] Packet from 1 refers to non-existent droid 171033, [Human : p1]
error   |02:02:25: [recvDroidInfo] Packet from 1 refers to non-existent droid 171497, [Human : p1]
error   |02:02:25: [recvDroidInfo] Packet from 1 refers to non-existent droid 171033, [Human : p1]
error   |02:04:01: [recvGroupOrder] Packet from 1 refers to non-existent droid 172665, [Human : p1]
error   |02:04:01: [recvDroidInfo] Packet from 1 refers to non-existent droid 172665, [Human : p1]
error   |02:04:34: [NETbcast] Failed to send message: Connection reset by peer
Saved dump file to '/tmp/warzone2100.gdmp-glK2dX'
If you create a bugreport regarding this crash, please include this file.
Segmentation fault.

[anonymous]

Both times it happened when i was attacking and therefore moving units frequently.

[anonymous]

And this time it was too quickly, in the very beginning of the game.

error   |02:18:02: [NETbcast] Failed to send message: Bad file descriptor
error   |02:18:02: [socketClose] Failed to close socket: Bad file descriptor
Saved dump file to '/tmp/warzone2100.gdmp-ZfsHtR'
If you create a bugreport regarding this crash, please include this file.
Ошибка сегментирования

[anonymous]

Ошибка сегментирования = Segmentation fault (russian locale here;)

[hao]

The host says he hang/crashed on the 2nd and 2rd times. But there was another player who said he just saw the host leaving but the game was still responding.

[hao]

More crashes. It was a 8player map but only 5 humans and 1 ai were present.

error   |02:37:53: [NETbcast] Failed to send message: Bad file descriptor
error   |02:37:53: [socketClose] Failed to close socket: Bad file descriptor
Saved dump file to '/tmp/warzone2100.gdmp-X16BJB'
If you create a bugreport regarding this crash, please include this file.
Segmentation fault.

[hao]

LC_ALL="en_US.UTF-8" ~/Downloads/warzone2100-2.3_beta3/src/warzone2100
error   |02:55:44: [NETbcast] Failed to send message: Socket operation on non-socket
error   |02:55:44: [socketClose] Failed to close socket: Bad file descriptor
Saved dump file to '/tmp/warzone2100.gdmp-Q3JYzW'
If you create a bugreport regarding this crash, please include this file.

This has happened just when all my opponents have left, and i saw the last one leaving, and then it crashed.

[hao]

that anonymous is me actually.

[Cyp]

Don't know if it's the same crash exactly, but mine keeps crashing at netplay.c:642.

Program terminated with signal 11, Segmentation fault.
[New process 5186]
[New process 5191]
[New process 5193]
[New process 5192]
#0  0x000000000059cf9b in checkSockets (set=0x2a0d820, timeout=0) at netplay.c:642
642     netplay.c: No such file or directory.
        in netplay.c
(gdb) bt full
#0  0x000000000059cf9b in checkSockets (set=0x2a0d820, timeout=0) at netplay.c:642
        tv = {tv_sec = 0, tv_usec = 0}
        ret = <value optimized out>
        fds = {fds_bits = {0 <repeats 16 times>}}
        count = 1
        i = 0
        maxfd = 43880960
        __FUNCTION__ = "checkSockets"
#1  0x000000000059fe20 in NETrecv (type=0x7fff01f32f3f "") at netplay.c:2592
        received = 685640
        size = 1
        current = 0
        __FUNCTION__ = "NETrecv"
#2  0x000000000051a8c0 in recvMessage () at multiplay.c:610
        type = 0 '\0'
        __FUNCTION__ = "recvMessage"
#3  0x000000000051b6c1 in multiPlayerLoop () at multiplay.c:297
        joinCount = 0 '\0'
        __FUNCTION__ = "multiPlayerLoop"
#4  0x00000000004f2db5 in gameLoop () at loop.c:266
        psCurr = <value optimized out>
        psNext = <value optimized out>
        psCBuilding = <value optimized out>
        psNBuilding = <value optimized out>
        psCFeat = <value optimized out>
        psNFeat = <value optimized out>
        i = <value optimized out>
        widgval = <value optimized out>
        quitting = <value optimized out>
        intRetVal = INT_INTERCEPT
        clearMode = <value optimized out>
        __FUNCTION__ = "gameLoop"
#5  0x00000000004f4e7d in main (argc=<value optimized out>, argv=<value optimized out>) at main.c:659
        __FUNCTION__ = "main"

[hao]

And this time I was hosting, and it crashed again. I was moving the map with right mouse button on the minimap.

error   |04:02:36: [socketListen] Failed to set up IPv6 socket for li                                                           stening on port 2100: Address already in use
error   |04:02:36: [NEThostGame] Cannot connect to master self: Addre                                                           ss already in use
error   |04:02:43: [socketListen] Failed to set up IPv6 socket for li                                                           stening on port 2100: Address already in use
error   |04:02:43: [NEThostGame] Cannot connect to master self: Addre                                                           ss already in use
error   |04:02:55: [socketListen] Failed to set up IPv6 socket for li                                                           stening on port 2100: Address already in use
error   |04:02:55: [NEThostGame] Cannot connect to master self: Addre                                                           ss already in use
error   |04:03:00: [socketListen] Failed to set up IPv6 socket for li                                                           stening on port 2100: Address already in use
error   |04:03:00: [NEThostGame] Cannot connect to master self: Addre                                                           ss already in use
error   |04:03:07: [readLobbyResponse] Lobby error (406): Game unreac                                                           hable, failed to open a connection to: [87.255.17.177]:2100
error   |04:03:07: [socketAccept] accept failed: Invalid argument
last message repeated 2 times
last message repeated 2 times (total 4 repeats)
last message repeated 4 times (total 8 repeats)
last message repeated 8 times (total 16 repeats)
last message repeated 16 times (total 32 repeats)
last message repeated 32 times (total 64 repeats)
last message repeated 22 times (total 86 repeats)
error   |04:24:37: [recvDroidMove] Packet from 2 refers to non-existe                                                           nt droid 162002, [Human : p2]
info    |04:25:21: [seq_Play] unable to open 'sequences/victory.ogg' for playback
error   |04:25:36: [readLobbyResponse] Lobby error (406): Game unreachable, failed to open a connection to: [87.255.17.177]:2100
error   |04:25:36: [socketAccept] accept failed: Invalid argument
last message repeated 2 times
last message repeated 2 times (total 4 repeats)
last message repeated 4 times (total 8 repeats)
last message repeated 8 times (total 16 repeats)
last message repeated 16 times (total 32 repeats)
last message repeated 30 times (total 62 repeats)
error   |04:48:16: [NETbcast] Failed to send message: Connection reset by peer
>-(/-/Saved dump file to '/tmp/warzone2100.gdmp-GF1Qdm'
If you create a bugreport regarding this crash, please include this file.
Segmentation fault

[hao]

I have a reason to believe that these crashes were caused by someone's leaving the game.

[hao]

At least it almost always crashes when someone is leaving. I don't remember many "[...] has left" messages in 2.3 beta 3 AND 4. <--- the problem still exists!

[Giel]

I'm hoping this patch will aid in nailing down the cause. At the very least it should make the backtraces more useful.

lib/netplay/netplay.c

@@ -739 +739 @@
                for (i = 0; i < set->len; ++i)
                {
                        if (set->fds[i])
-                               FD_SET(set->fds[i]->fd[SOCK_CONNECTION], &fds);
+                       {
+                               const SOCKET fd = set->fds[i]->fd[SOCK_CONNECTION];
+
+                               FD_SET(fd, &fds);
+                       }
                }
 
                ret = select(maxfd + 1, &fds, NULL, NULL, &tv);

[Giel]

(In [9950]) Make backtraces regarding FD_SET related crashes (see #1136) more useful

[Giel]

(In [9951]) Backport r9950 from trunk into the 2.3 branch:

• Make backtraces regarding FD_SET related crashes (see #1136) more useful

[Git SVN Gateway <gateway@…>]

(In [401723958cb61f2110e28ecf20b9622b2c84d3c9]) Make backtraces regarding FD_SET related crashes (see #1136) more useful

git-svn-id: ​https://warzone2100.svn.sourceforge.net/svnroot/warzone2100/trunk@9950 4a71c877-e1ca-e34f-864e-861f7616d084

[Per]

(In [9954]) 2.3: Fix SIGBUS errors when linux box is hosting and players disconnect. Patch reviewed by Buginator and Giel. Closes ticket:1136

[Git SVN Gateway <gateway@…>]

In [401723958cb61f2110e28ecf20b9622b2c84d3c9]:

Make backtraces regarding FD_SET related crashes (see #1136) more useful

git-svn-id: ​https://warzone2100.svn.sourceforge.net/svnroot/warzone2100/trunk@9950 4a71c877-e1ca-e34f-864e-861f7616d084