P3D Ver 1.3 + FSUIPC 4.812 - Crash in Multiplayer

[FSMP]

P3D Ver 1.3 + FSUIPC 4.812 - Crash in Multiplayer

Host: XP Pro - P3D 1.3 ( Fsuipc Not installed)

Client: XP Pro - P3D 1.3 FSUIPC 4.812 Installed

Both running P3D as ADMINISTRATOR

Client runs fine in Free Flight. ( with & without FSUIPC enabled)

Client runs fine in Multiplayer, if FSUIPC is DISABLED,

If FSUIP is Enabled, Client joins session, but a few seconds after coming off pause,:-

P3D Application Error : Inst 0x0185a49a ref memory at 0x02c5f7f8 ( cannot read ).

100% Repeatable: Every time FSUIPC is enabled in Clinet, it crashes after connecting, just after coming off pause.

------------

This did not occur with P3D 1.2, and the earlier version of FSUIPC, that supported 1.2 (prior to 1.3 release)

--------------

Geoff

Edited March 27, 2012 by FSMP

[Pete Dowson]

If FSUIP is Enabled, Client joins session, but a few seconds after coming off pause,:-

P3D Application Error : Inst 0x0185a49a ref memory at 0x02c5f7f8 ( cannot read ).

100% Repeatable: Every time FSUIPC is enabled in Clinet, it crashes after connecting, just after coming off pause.

------------

This did not occur with P3D 1.2, and the earlier version of FSUIPC, that supported 1.2 (prior to 1.3 release)

Can you please report this to Lockheed Martin support. There's no change in FSUIPC -- only the Installer needed changing. L-M will need to work out what is going on. It will be something in SimConnect. (In any case I can only make one install of 1.3 so I cannot even test that multiplayer.mode).

One thing you could try is removing the SimConnectP3D.DLL file from the P3D Modules foder to force FSUIPC to use the ESP one installed by L-M in the Windows WinSxS folder. But I doubt it'll change as the DLL is merely a wrapper for the calls into the SimConnect-proper part of P3D.

Regards

Pete

[FSMP]

Can you please report this to Lockheed Martin support. There's no change in FSUIPC -- only the Installer needed changing. L-M will need to work out what is going on. It will be something in SimConnect. (In any case I can only make one install of 1.3 so I cannot even test that multiplayer.mode).

One thing you could try is removing the SimConnectP3D.DLL file from the P3D Modules foder to force FSUIPC to use the ESP one installed by L-M in the Windows WinSxS folder. But I doubt it'll change as the DLL is merely a wrapper for the calls into the SimConnect-proper part of P3D.

Regards

Pete

Thanks Pete.

FYI: FSMP runs a "Open to the Public" 24/7 P3D MP server, to aid developers, developing for Multiplayer applications.

Currently, it is running the latest version of P3D: ( Ver 1.3 ) (FSUIPC not installed on Server)

You (and any other developer) are most welcome to attempt to connect to it for testing.

IP: 68.49.210.150

I would be most interested to hear from anyone connecting, who has FSUIPC installed on their P3D Ver 1.3 client.

Regards

Geoff

FSMP

Edited March 27, 2012 by FSMP

[Pete Dowson]

You (and any other developer) are most welcome to attempt to connect to it for testing.

I'm afraid i've no idea about how to use multiplayer or even what it does. I've never used it on any version of FS, and really don't have time at present to learn.

This information:

P3D Application Error : Inst 0x0185a49a ref memory at 0x02c5f7f8 ( cannot read )

you have should be sufficient for LM to see what is going on. It is deep into someplace in P3D which FSUIPC is not at all involved with, so I wouldn't know what to do in any case.

Regards

Pete

[adiemus]

Pete,

Actually, the crash does seem to happen inside FSUIPC.

Sent you an e-mail with some technical details that will hopefully be helpful.

Chris

VRS

[Pete Dowson]

Actually, the crash does seem to happen inside FSUIPC.

Sent you an e-mail with some technical details that will hopefully be helpful.

Thanks. I notice the code offsets were with 4.812. I'll see if I can locate them in my current build, but if not I may ask for you to supply again with a different build. (I MUST remember to save the compiler's MAP file for released builds! I used to do it regularly, but have fallen into bad ways! :-( ).

Pete

[Pete Dowson]

Sent you an e-mail with some technical details that will hopefully be helpful.

Yes, very helpful. Thanks!

It certainly occurs when the SimConnect events stop arriving in FSUIPC for long enough. FSUIPC thinks it has stalled and attempts re-connection. The method used, which works in all FSX, ESP and previous P3D versions appears to be defeated by some change in 1.3 which i cannot fathom at present.

I'm changing the way FSUIPC attempts to do that restart. Please look out for 4.815 which I'll be putting up in the Download Links subforum either later today, or more likely, tomorrow. [sEE EDIT BELOW]

Meanwhile you might like to try changing this parameter in the FSUIPC4.INI file:

SimConnectStallTime=1

That default value sets the limit before a stall is determined to 1 second. Try increasing it to a value which accommodates whatever the stall time is for the multiplayer switching. This is only a temporary measure to be sure that I'm tackling the correct problem.

[EDIT]

I've uploaded a test version, 4.812b. Here: FSUIPC4812b. Please try that and let me know.

Regards

Pete

[adiemus]

I'm afraid 4.812b still crashes when entering MP, but does so differently than 4.812. Looking at it now. Will e-mail more details.

[Pete Dowson]

I'm afraid 4.812b still crashes when entering MP, but does so differently than 4.812. Looking at it now. Will e-mail more details.

Please try 4.812c, from here: FSUIPC4812c.

Thanks,

Pete

[adiemus]

Unfortunately, this one still crashes on joining an MP session. (Not always, still trying to work out the conditions under which it crashes)

Edited March 29, 2012 by adiemus

[Pete Dowson]

Unfortunately, this one still crashes on joining an MP session. (Not always, still trying to work out the conditions under which it crashes)

Odd, it didn't crash at all here, but I didn't try many times. In my case it was never paused when MP mode was entered.

Does your log show ihe double re-connection? It shouldn't, now.

BTW I've had to make another build for a different reason, so if you want to supply any specific data please use FSUIPC4812d

Pete

[adiemus]

The crash remains on 4.812d until I removed SimConnectP3D.dll, which seems to resolve it. (On this system, that ends up forcing a connection to the FSX SP2/Acceleration version of SimConnect, since I have Acceleration installed in addition to P3D. Not sure how it'll behave on a system that's never had FSX installed, as I don't have one with which to test and no easy way to force FSUIPC to ignore the FSX SimConnect DLLs)

[Pete Dowson]

The crash remains on 4.812d until I removed SimConnectP3D.dll, which seems to resolve it. (On this system, that ends up forcing a connection to the FSX SP2/Acceleration version of SimConnect, since I have Acceleration installed in addition to P3D. Not sure how it'll behave on a system that's never had FSX installed, as I don't have one with which to test and no easy way to force FSUIPC to ignore the FSX SimConnect DLLs)

With earlier versions of P3D their installer installed the ESP version of SimConnect I think. I'm surprised they withdrew that.

Fixing it by removing SimConnectP3D.dll seems to suggest that there's some change in their static libraries between 1.2 and 1.3 that makes the 1.2 library incompatible with MultiPlayer mode. Odd, though, that it is okay with the FSX SP2 version.

I'll build a version of SimConnectP3D.dll with the 1.3 static library for you to try. I'll try and get it posted tonight or in the morning, as I'm then away till Monday.

Regards

Pete

[Pete Dowson]

I'll build a version of SimConnectP3D.dll with the 1.3 static library for you to try. I'll try and get it posted tonight or in the morning, as I'm then away till Monday.

Okay -- try SimConnectP3D.zip

Pete

[adiemus]

Pete, the new SimConnectP3D.dll made no difference.

Would be quite curious though to have someone else run the same test (4.812d with the updated SimConnectP3D) just to confirm.

[Pete Dowson]

Pete, the new SimConnectP3D.dll made no difference.

Well, really, because L-M said the older LIB would still be okay, it shouldn't have made any difference. but then, without it, how is everything working well with the FSX SimConnect DLLs? It's weird.

I'll try the method where I don't do the re-linking. It didn't make much difference before, but I've no other ideas. After that we're dependent on L-M help I think.

[LATER]

Here: FSUIPC4812e

Pete

[adiemus]

I was unable to get this to crash in any configuration, including with SimConnectP3D.dll from 1.2. While not necessarily a true resolution, it may indeed be a truly generic workaround for Prepar3D 1.3.

[FSMP]

Running FSUIPC4812e

Connection to MP completes, but if I CRASH the plane, and then try to fly again, I get a CTD when the Loading Progress bar reaches 100%.

If I disable FSUIPC, I do not get this CTD.

Tried two different PC as the SERVER, and two different PCs as the joining client.

IDENTICAL, reproducable CTD in every case when FSUIPC was enabled.

NO CTD when FSUIPC was disabled.

FSUIPC4.Log ( Registration Info X'ed out !! )

******** FSUIPC4, Version 4.812e by Pete Dowson *********

User Name="XXXXXXXXXXXXXXXX"

User Addr="XXXXXXXXXXXXXXXX"

FSUIPC4 Key is provided

WIDEFS7 not user registered, or expired

Running inside Prepar3D on Windows Vista

Module base=61000000

609 System time = 31/03/2012 09:34:04

609 FLT path = "C:\Users\XXXXX\Documents\Prepar3D Files\"

609 Trying D:\Program Files (x86)\Lockheed Martin\Prepar3D\Modules\SimConnectP3D.dll

609 Found it: trying to connect

640 FS path = "D:\Program Files (x86)\Lockheed Martin\Prepar3D\"

1654 LogOptions=00000000 00000001

1685 Wind smoothing fix is fully installed

1685 SimConnect_Open succeeded: waiting to check version okay

1685 Trying to use SimConnect Prepar3D

6100 Running in "Lockheed Martin® Prepar3D®", Version: 1.3.3708.0 (SimConnect: 1.3.0.0)

6100 Initialising SimConnect data requests now

6100 FSUIPC Menu entry added

6131 D:\Program Files (x86)\Lockheed Martin\Prepar3D\FLIGHTS\OTHER\Prepar3D.FLT

6131 D:\Program Files (x86)\Lockheed Martin\Prepar3D\SimObjects\Airplanes\Mooney_Bravo\Mooney_Bravo.AIR

98765 System time = 31/03/2012 09:35:42, Simulator time = 11:00:04 (15:00Z)

98780 Aircraft="Mooney Bravo With G1000"

99311 Starting everything now ...

100403 Advanced Weather Interface Enabled

112196 Weather Mode now = Theme

365012 Sim stopped: average frame rate for last 180 secs = 7.1 fps

937130 **** SimConnect Data Stalled! Re-connecting now ... ****

937239 SimConnect_Open succeeded: waiting to check version okay

937239 Running in "Lockheed Martin® Prepar3D®", Version: 1.3.3708.0 (SimConnect: 1.3.0.0)

937239 Initialising SimConnect data requests now

937239 FSUIPC Menu entry added

938128 System time = 31/03/2012 09:49:42, Simulator time = 11:01:21 (15:01Z)

993867 Sim stopped: average frame rate for last 57 secs = 9.7 fps

993867 User aircraft crashed!

1060932 **** SimConnect appears not to be receiving requests! Re-connecting now ... ****

1061041 SimConnect_Open succeeded: waiting to check version okay

Geoff

[Pete Dowson]

Running FSUIPC4812e

Connection to MP completes, but if I CRASH the plane, and then try to fly again, I get a CTD when the Loading Progress bar reaches 100%.

If I disable FSUIPC, I do not get this CTD.

Could to reproduce that please, but first check the "Log Extras" option in the FSUIPC logging tab (or set the number in the edit box to 1, if there's one showing). I could do with seeng the timings of "Sim Stop" and "Sim Start" events, which that will then show in the log.

There's something changed between P3D 1.2 and 1.3 which really messes up FSUIPC's recovery attempts on timeouts. It seems to revolve around the business of closing and opening SimConnect connections -- which, of course, incldes removing the Menu and adding it back afterwards, which was aone of the areas changed in 1.3 to fix the crash on exit problem.

I have no way of tunnelling into what is actually happening in the P3D code to do this. All I can think of doing is just completely avoiding some of the checks I do, which worked okay before and work okay in FSX, and risk of having a non-operative FSUIPC in a running P3D remain rather than attempting recovery. Maybe the risks are neglible in any case. I don't know. with FSX any heavily loaded system can sometimes result in a SimConnect stall, though this seems to happen far less with FSX SP2 than it did in earlier releases.

After I see the log with Extras logging I'll decide what to do.

Regards

Pete

[Pete Dowson]

After I see the log with Extras logging I'll decide what to do.

I made some other changes to try to detect MP changeover (It seems I can't -- the requested events never arrive, like everything else, hence the "stall" diagnosis), but i also removed the "is Simconnect receiving from FSUIPC?" check. maybe that'll help in the crash case, as that was where P3D appears to be stumbling. It doesn't appear to like the Event State being changed whilst it is loading or reloading things.

Try FSUIPC 4812f

Regards

Pete

[FSMP]

I made some other changes to try to detect MP changeover (It seems I can't -- the requested events never arrive, like everything else, hence the "stall" diagnosis), but i also removed the "is Simconnect receiving from FSUIPC?" check. maybe that'll help in the crash case, as that was where P3D appears to be stumbling. It doesn't appear to like the Event State being changed whilst it is loading or reloading things.

Try FSUIPC 4812f

Regards

Pete

Pete

I just tested it joining numerous times with my P3D Server.

Client running FSUIPC version 'F' now connects without incident every time, and no signs of any errors in the FSUIPC Logs.

When I joined and CRASHED THE PLANE, this time with Version "f" I was able select a repaired plane, and re-join the session, and when Pause was released, fly as normal, without any CTD>

FSUIPC log did show a few EXCEPTION Errors :-

Console Started ...

1168015 LogOptions changed, now 60000000 0000001D

1169015 Ready Flags: Ready-To-Fly=Y, In Menu=N, In Dlg=N

1177297 Ready Flags: Ready-To-Fly=Y, In Menu=Y, In Dlg=Y

1177297 User aircraft crashed!

1209250 Exception 1 "ERROR", Ref 2459, Index param -1 on unknown request!

1265125 Exception 1 "ERROR", Ref 2484, Index param -1 on unknown request!

1286844 Ready Flags: Ready-To-Fly=Y, In Menu=N, In Dlg=N

1289656 KEYDOWN: VK=80, Waiting=0, Repeat=N, Shifts=0

1289656 .. Key not programmed -- passed on to FS

1289672 *** EVENT: Cntrl= 65561 (0x00010019), Param= 0 (0x00000000) PAUSE_TOGGLE

1289797 KEYUP: VK=80, Waiting=0

1292422 **** SimConnect Data Stalled! Re-connecting now ... ****

1292515 SimConnect_Open succeeded: waiting to check version okay

1292562 Running in "Lockheed Martin« Prepar3D«", Version: 1.3.3708.0 (SimConnect: 1.3.0.0)

1292562 Initialising SimConnect data requests now

1292562 FSUIPC Menu entry added

1293375 System time = 03/04/2012 11:56:36, Simulator time = 11:24:13 (15:24Z)

Hope this helps :)

---------------------------

Question: Are the changes made in "f" just for operation in P3D mode, or is the "checking" also removed for FSX ?

THANKS for such a speedy resolution. You never cease to amaze me :)

Geoff

FSMP

[Pete Dowson]

When I joined and CRASHED THE PLANE, this time with Version "f" I was able select a repaired plane, and re-join the session, and when Pause was released, fly as normal, without any CTD>

Good.

FSUIPC log did show a few EXCEPTION Errors :-

They are probably inevitable on a crash. Theuy are only logged for information reasons, when you have "LogExtras" specified. Those don't mean so much and i might suppress them in any case.

Question: Are the changes made in "f" just for operation in P3D mode, or is the "checking" also removed for FSX ?

The checking for SimConnect not receiving requests from FSUIPC is only bypassed in P3D. If that ever gets resolved I'll put it back in.

In all cases, the re-linking to a Simconnect DLL is now only done when the Simconnect Open fails -- this usually means a Simconnect mis-match or other serious SimConnect problem. In other cases I now just close the connection and re-open it, re-establishing all the requests.

Regards

Pete

[Gypsy Baron]

Just for information purposes, I had noticed a worrisome FSXA crash (BEX) occuring in FSX Multiplayer that seemed

to have started occuring recently.

The repeatable crash occured if I entered a MP session, then went back out and re-entered again. The 2nd time I entered,

as soon as the "loading..." reached 100%, I saw the disk access that is associated with FSX crashing and a few seconds

later I received the "fatal error" message. The first line in the info read "BEX" and the fault module was unknown.

After several tests on both the Digital Themepark server and also a FSX Gamespy session, using a default aircraft as well

as my usual A2A rides, I discovered that disabling FSUIPC4 v4.812 stopped the app crash from happening.

I then reverted to V4.805 and the app crash also happened with that version. I then reverted to V4.801 and no longer

had the app crash. A look at the FSUIPC4 log showed nothing beyond the "normal" entry listings up to the connection to simconnect.

At least one other person that I usually fly with in MP was also seeing this app crash nand he was running v4.812 as well.

I'm about to go traveling so won't be able to do any further testing or add to this information.

Paul

[Pete Dowson]

At least one other person that I usually fly with in MP was also seeing this app crash nand he was running v4.812 as well.

4.812f would be better. If FSUIPC4 was the only SimConnect using application running, then it could be that a reconnection attampt is being made with the re-linking which it used to do (before 4.812e) and Simconnect.dll is reloading into a different part of memory. At least that's one theory which has been postulated with the Prepar3D crash. I don't actually see how that could have caused the crashes, but avoiding the re-linking did help.

However, the common factor with all of them so far has been the "SimConnect stalled" or similar message being logged first, but you say this didn't occur. Or isn't that what you mean by

A look at the FSUIPC4 log showed nothing beyond the "normal" entry listings up to the connection to simconnect.

That's a little ambiguous. By "connection to simconnect" do you mean the first one, the normal one, or is it actually a re-connection? Else why mention it at all? If it was a re-connection what was the actual message?

It is rather a shame no one reported this with FSX until now, and sadly at a time when you can't help check things. I cannot support folks who return to older versions, so it has to be fixed! Now that it appears fully resolved on Prepar3D, the only place is was reported till now, I was about to build a new release, 4.82, but now it looks like that is on hold until you return from travelling or someone else who experiences this trns up out of the blue! :sad:

Pete

[Gypsy Baron]

OK, I just checked the forum again before dinner and saw your reply above. I just re-enabled V4.812 mand went online

to a multiplayer server.

Entered flight OK. Exited the session and then re-entered. This went OK. The key here is I was the ONLY person

in the session.

I then changed to a different server that had 3 other people in the session. I spawned fine the first time and then left the

session and re-entered, or rather tried to. I got the app crash.

Here is the app crash message content:

Problem Event Name: BEX
  Application Name: fsx_MP.exe
  Application Version: 10.0.61637.0
  Application Timestamp: 46fadb14
  Fault Module Name: unknown
  Fault Module Version: 0.0.0.0
  Fault Module Timestamp: 00000000
  Exception Offset: 00000000
  Exception Code: c0000005
  Exception Data: 00000008
  OS Version: 6.1.7600.2.0.0.256.48
  Locale ID: 1033
[/CODE]

And here is the FSUIPC log after the crash. It seems to indicate that simconnect was stalled during the re-entry process:

[CODE]
Running inside FSX on Windows 7
... FSX has been renamed as "fsx_MP"
Module base=61000000
281 System time = 03/04/2012 19:00:31
281 FLT path = "C:\Users\Gypsy Baron\Documents\Flight Simulator X Files\"
717 Trying to connect to SimConnect Acc/SP2 Oct07 ...
733 FS path = "X:\Microsoft Flight Simulator X\"
1419 Run: "C:\TEMP\DAT_BAK.bat"
1419 LogOptions=00000000 00000001
1435 Wind smoothing fix is fully installed
1435 G3D.DLL fix attempt installed ok
1435 SimConnect_Open succeeded: waiting to check version okay
1435 Trying to use SimConnect Acc/SP2 Oct07
11591 Running in "Microsoft Flight Simulator X", Version: 10.0.61637.0 (SimConnect: 10.0.61259.0)
11591 Initialising SimConnect data requests now
11591 FSUIPC Menu entry added
11637 C:\Users\Gypsy Baron\Documents\Flight Simulator X Files\WoP2 B17G Fearless Fosdick RTG at Bassingbourn.FLT
11637 Y:\91st_BG_Aircraft\WoP2_b17g 91st BG\B17g_FF.AIR
40139 Weather Mode now = Theme
67127 Aircraft="Boeing B-17G "Fearless Fosdick""
67143 System time = 03/04/2012 19:01:37, Simulator time = 19:00:53 (18:00Z)
127125 Starting everything now ...
133006 Advanced Weather Interface Enabled
162709 Sim stopped: average frame rate for last 37 secs = 55.2 fps
293640 Weather Mode now = Real World
395759 **** SimConnect appears not to be receiving requests! Re-connecting now ... ****
395883 Trying to connect to SimConnect Acc/SP2 Oct07 ...
395883 Now running with SimConnect Acc/SP2 Oct07
395883 SimConnect_Open succeeded: waiting to check version okay
[/CODE]

So, emty arena, no problems. An arena with other aircraft, a crash on the SECOND time going back into the

same session.

The reason I had not noticed this before was that is not something I usually have to do. once I get into a session I stay

until I leave for the night. The reason I happened upon this issue was in the past couple of

weeks a couple of the pilots I fly with regularly were having "connect issues" which turned out to be this very same app crash.

We thought it had something to do with various persons in the session not having the "paint" that someone entering

the session was using. This turned out to not be the case/

Today we were investigating another issue which involved 3 of us spawning to various parking spaces. In some instances the

MP server would place us at a parking spot we did not spawn to. We usied the in-game menu to "Go to airport" and

re-spwn and that put us where we intended to go. At that point it was suggested that perhaps the issue only occured

when spawning from outsdide the session.

When going back out and re-entering, I suffered the crash. This being repeatable we tested on the normal FSX Gamespy server

with the same results. It was at that point that I recalled reading this thread and decided to see if the app crash was possibly

related to FSUIPC. Disabling 4.812 "solved: the app crash. Reverting to 4.805 I still experienced the app crash.

That is when I went to 4.801 with a positive result.

Another related incident happened a couple of days ago when I tried to enter a MP session in an aircraft no one

else had...a beta aircraft, for the very first time. I could not even get in the 1st time with that plane. later that night I

tried again when the arena was empty and was able to enter and fly with the beta aircraft. In hindsight this would seem to

suggest something related to the data being exchanged in a MP session and the impact it has upon the

simconnect environment. This boggles my logical mind, I must say.

In any case, I hope the information in this post is of use to tackling the issue.

Regards,

Paul