Scotfleiger

My error I was looking at the wrong offset (3C00 not 3E00). The .dll works on my system. I will pass it on to my user to test. FSUIPC6.log

Further information. With P3Dv4 0x3C00 also appears to point to the current AIR FILE for the loaded aircraft and not the Flt Sim path.

One of my LINDA users has installed FSUIPC 6.0.7 with P3Dv4.5 using the c:/users/{name}/prepar3d v4 add-ons/FSUIPC6 option. He does not have P3Dv5 installed. LINDA reads offsets 0x3E00 to read the path to the Flt Sim and 0x3124 to read the Flt Sim version number (eg 45). On his system the former lists the add-on path to FSUIPC6 and the latter is blank. On my system running P3Dv5 the Flt Sim path and version are read correctly from the offsets. With P3Dv4.5 the results are intermittent sometimes the offsets are populated correctly and are blank at other times. Can you check this from your end?

The hanging background task is what started this whole conversation. It is labelled 'Prepar3d...' of size 560KB. The process of elimination you suggested identified that it was the VRI device that was stopping FSUIPC6 terminating. I say it is stopping FSUIPC6 terminating as the fusipc6.log cannot have its name changed until this task ends. LINDA is long terminated by this time - at the 25-30s point with 6.0.3g. The only VRI driver is the COMPORT FDTI device driver. No other VRI software is running or used. I take back what said. It was not a failure. LINDA consists of 2 parts - GUI and LUA. The GUI LINDA.EXE is the on-screen interface for assigning functions to buttons and interfaces with FSUIPC. It can be started manually (before P3D) or via the FSUIPC [programs] block. The LUA code is triggered by FSUIPC running ipcReady.lua that calls linda.lua which in turn calls init.lua for loading all other LINDA code. This code can operate without the GUI. When I say 'manually close' LINDA it is me clicking on the red quit icon (top right) to close the application. I will need to double check but I believe the LUA code is left loaded and running when the GUI terminates. I will take a look at the GUI code to see what can be done. Andrew

Hi Pete I am sorry to report that 6.0.3h is a failure. On exiting P3D the auto start program LINDA displayed 'Not responding' and nothing closed even after over 2 minutes. This was both with and without logging. 6.0.3h is only killing one of the 2 LINDA LUA thread (init.lua). The other (linda.lua) is not reported. See previous logs. Further to my earlier statement, LINDA closes and ends in task in less than a second when manually closed. Andrew FSUIPC6-603h-NoLog.log FSUIPC6-603h-YesLog.log

Hi Pete I agree the delays times are a little excessive if you are closing and wanting to reopen P3D. Typically the shutdown took 30+ seconds to complete so I am not complaining as I and users need a full shutdown without manually checking the Task Manager. I started my stopwatch when I clicked on the Yes button on the exit P3D dialog. It was 25-30s to FSUIPC closing the external apps (eg. LINDA), sending the CMDRST and updating fsuipc6.ini. These events all happened within a second. When FSUIPC closes LINDA it displays ‘not responding’ for several seconds, waits for several more seconds and then closes. When I close LINDA manually it takes less than 5s. It was a further 25-30s before the ‘hanging’ P3D background task (or stub as I called it) closed. At this point the fsuipc6.log was released and could be renamed. The total time to close everything was under 60s (30+30). I only have 2 comms ports defined COM3 and COM4 for the 2 different VRi Combos I use for testing. I will run tests on the new dll later. Andrew

That sounds good. I suggest a comment be added to the release notes about leaving 60s to close down. Andrew

Morning John I ran the series of tests with 6.0.3g as requested with several runs of each. In all cases the shutdown was completed successfully. It around 30s to close LINDA, reset VRI (when VRIDisableCMDRST=no) and right to fsuipc6.ini and around 60s to close the background task stub. NoLog = LogExtras = 0 YesLog = LogExtras = x2005 NoVRI = VRIDisableCMDRST=no YesVRI = VRIDisableCMDRST=yes Good work, thank you Andrew FSUIPC6-603g-NoLog-NoVRI.log FSUIPC6-603g-NoLog-NoVRI-LINDA running outside.log FSUIPC6-603g-YesLog-NoVRI.log FSUIPC6-603g-YesLog-YesVRI.log

Hi John I ran 6.0.3e with VRIDisableCMDRST set to Yes and No and LogExtras=0. With CMDRST disable everything closed correctly after a much longer delay. However, with CMDRST enabled the hanging task remained and the FSUIPC6.log was not closed fully. FSUIPC6-603e-NoCMDRST.log FSUIPC6-603e-YesCMSRST.log

Hi John Thanks for the updated .dll. With LogExtras=1 and x2005/8197, 6.03.e closed currently with the expected delay. I run this test 3 times without any problems. See 1st log. I then set the VRIDisableCMDRST=No. This test failed (2nd log). The VRI Combo reset and LINDA.exe closed but the Prepar3d background task failed to terminate. I again repeated this test 3 times. It left a hanging task on each occasion. I then tried the ForceTerminateTime. At 5000ms and 1000ms the hanging task remained open on both tests (3rd log). I allowed at least 20s for it to work. I set LogExtras=0, VRIDisableCMDRST=yes and ForceTerminateTime=1000. This test failed (4th). I then used 10000ms without success. The log files shows logging ended but the hanging task remained. Finally I removed the ForceTerminateTime entry. This time FSUIPC6 terminated correctly with no hanging tasks after a 5-10s delay. This was repeatable. These results are a little strange as the 4th log shows that the CMDRST was sent and success and the log closed but something stopped the tasks ending. AndrewFSUIPC6-603e-4th.log FSUIPC6_603e-1st.log FSUIPC6-603e-2nd.log FSUIPC6-603e-3rd.log FSUIPC6-603e-5th.log

I agree that the P3Dv5 crashes is nothing to do with you. When the crash occurs all processes are forced to close anything and the issue I have raised does not occur on next startup. If FSUIPC6 shutdown hangs I would go with the TerminateProcess option after a set delay. We just need to see the VRI comms closed down. Andrew

These are the fsuipc6.log when P3Dv5 crashes on exit. All logging stops at closing DLLs. FSUIPC6-crash 2.log FSUIPC6-crash.log Report.wer

Hi Pete My apologies for having to leave earlier. Thank you for reminding me about the event viewer. As a Mac user I don't normally see crashes! When P3Dv5 crashes on exit it is a .NET runtime error (1926): Application: Prepar3D.exe Framework Version: v4.0.30319 Description: The process was terminated due to an unhandled exception. Exception Info: exception code c0000005, exception address 00007FFD306C7056 Followed by an application error (1000): Faulting application name: Prepar3D.exe, version: 5.0.21.34709, time stamp: 0x5e95060d Faulting module name: ntdll.dll, version: 10.0.18362.778, time stamp: 0x0c1bb301 Exception code: 0xc0000005 Fault offset: 0x0000000000007056 Faulting process ID: 0x43f0 Faulting application start time: 0x01d6174387489c1c Faulting application path: C:\P3Dv5\Prepar3D.exe Faulting module path: C:\WINDOWS\SYSTEM32\ntdll.dll Report ID: 401f7199-5d35-46a5-a14f-ddfe2e319586 Faulting package full name: Faulting package-relative application ID: with the resulting Windows Error (1001). See attached ,wer file: Fault bucket 1707778556859359639, type 4 Event Name: APPCRASH Response: Not available Cab Id: 0 Problem signature: P1: Prepar3D.exe P2: 5.0.21.34709 P3: 5e95060d P4: ntdll.dll P5: 10.0.18362.778 P6: 0c1bb301 P7: c0000005 P8: 0000000000007056 P9: P10: The full debug log is attached. Somehow the .ini file edits were not saved. I ran several further tests with full debug on and there was no further hung tasks or crashes. Andrew Report.wer Approach crash.evtx FSUIPC6.log

I re-ran the tests with FSUIPC Logging Extras checkbox ticked. I achieved no hanging tasks. 603d closed fully. To answer Pete's question about P3Dv5 crashing. I have no idea why this is happening. It is appears random with the dialog box reporting the crash. This appears before the P3D main window closes. Clicking on Debug offers no information. Must go. Will come back later to finish my reply. FSUIPC6_603d-Extras_1.log FSUIPC6-603d-Extras-2.log FSUIPC6-603d-extras-3.log

Sorry I didn’t set the Extras value correctly. I will report again later.

Hi John Test ran with 6.0.3d. Closed correctly. In a number of follow on tests, P3Dv5 crashed on shutdown. This occurred before [programs] started LINDA was terminated. I did a further test with LINDA running independently of FSUIPC6 and the result was successful (2nd). I see that FSUIPC6 did not fully close the log. One final test (3rd) in the same conditions terminated fully and the log is complete. Andrew FSUIPC6_603d-1st.log FSUIPC6_603d-2nd.log FSUIPC6_603d-3rd.log

Hi John As my VRI Combo panel did not have the previous COM connection issue (reported by some MCP owners), I had not included VRIDisableCMDRST in fsuipc6.ini. However, your question was very apt. Using 6.0.3a with VRIDisableCMDRST=Yes, P3Dv5 correctly closed with no hanging tasks. No CMDRST instruction was sent on startup or close down. Setting VRIDisableCMDRST=No caused the task to hang! Just to test this result, I changed the option back to Yes. On 3 occasions P3Dv5 crashed on exit. On the 4th attempt, the task hanged up again. On the 5th attempt, it worked correctly. On 2 further tests, P3Dv5 crashed and then hanged on the background task. I am not sure this helps. NOTE: With 6.0.3 the offset with the FSUIPC6 version number (that LINDA reads) is now correctly populated. Andrew FSUIPC6_603a_CMDRSTNo.log FSUIPC6_603a-CMDRSTYes.log FSUIPC6_603_CMDRSTYes_5th attempt.log FSUIPC6_603a_CMDRSTYes_hanging_.log

Hi John I ran both 6.0.3 and 603b. In both cases I was left with the hanging Prepar3D Scenario Startup task (size 640MB). This task had to be manually stopped to enable P3Dv5 to start with VRI Comb panel working. The FSUIPC5.log was locked by this task preventing the log being renamed. LINDA was used to allow the MCP to be tested. Exiting P3Dv5 with 6.0.3b is causing P3Dv5 to crash on exit. There is no hanging task in this case. I also ran P3Dv4.5 with 6.0.3 as a comparison. This left a hanging task Lockhead Martin Prepar3D (size 182MB). Also the MCP did not reset on FSUIPC6 shutdown (CMDRST instruction not sent). I will later setup FSUIPC5 with P3Dv4.5 again to see if I can reproduce the problem. I attach a series of logs renamed to indicate the different tests. Thank you again for your attention to this issue. Andrew FSUIPC6_603_1stRun.log FSUIPC6_603_2ndRun.log FSUIPC6_603b_1stRun.log FSUIPC6_603_P3Dv45.log

The problem is affecting FSUIPC6 with both P3D v5 and v4.5. It was not apparent in FSUIPC5 with P3D v4.5.

It was not a problem that affected P3Dv4.5 with FSUIPC5. The only change was one you put in to stop the CMDRST command being sent to the VRI Combo (fsuipc5.ini parameter VRIDisableCMDRST) for some users with particular hardware issues. Is it possible to force the VRI processes to close on shutdown?

Hi Pete P3Dv5 runs correctly regardless of whether the hanging task is there or not. LINDA button inputs to FSUIPC6 and FSUIPC6 axis inputs both work as normal. The main issue and how I spotted the problem is the display of FSUIPC6 log output (including LINDA lines) on the LINDA debug tracer. The logging is actually going to the fsuipc6.log as you can see. Also there is no communication with the VRI Combo Panel is the fault condition. There is something blocking both the VRI output and FSUIPC/LINDA communication via IPC. I don't think looking at changing the VRI USB output is practical for me or our users. FSUIPC6.log

Hi Pete Thank you for your valuable insight. I ran a series of tests*** with all devices removed (except CH Flightstick), with VRI Combo added, and with all devices except the VRI Combo. You are correct. It would appear to be the VRI Combo connected to COM3 that is the cause of the hanging background task. There are no software or drivers running with the VRI except the FDTI USB to Serial device drivers necessary to communicate to FSUIPC. This has not been an issue that I have seen previously. NOTE: *** the tests involved disconnecting/connecting devices in turn, starting P3Dv5, operating the Pan View on the CH Flightstick, closing P3Dv5 and observing the Task Manager. FSUIPC6 No Devices.log FSUIPC6 with VRI.log FSUIPC6with All Devices.log

Hi Thomas The logs provided were with LINDA added back but the hanging task was there is all cases. The attached log is with FSUIPC6 as the only Add-on in P3Dv5. It causes the same issue. What I did observed in Task Manager as P3D/FSUIPC6 shuts down: Background Task Prepar3D.exe sized at ~= 750KB. The VRI panel flashes as it resets. Size is 500KB. Background Task Prepar3D.exe shown briefly as 'suspended' Background Task Prepar3D.exe reduces to size 300KB and remains running. FSUIPC6.log

Hi John Sorry to bring you more problems. I stripped out all other add-ons including all LINDA code and still got the hanging background task. FSUIPC.log is also locked from being renamed until End Task is used. I installed FSUIPC6 6.0.2 into the user/add-ons/fsuipc6 folder and I am still seeing the problem (see screenshot below). I also this time got the duplicate FSUIPC copies running error dialog on one occasion. On exiting P3Dv5 FSUIPC6 is correctly closing the running programs (via [programs]) but the process is still hanging and has to be stopped manually. http://prntscr.com/s1rywg FSUIPC6.log

I spent over 2 hours fault finding this problem. I switched off all Add-ons and that cleared the problem. Turning FSUIPC6 back on produced the hanging background task on exiting P3D.