Hang on game exit on Linux

[malric]

I am occasionally (once each 2-3 games maybe) experiencing a hang when game should exit. Generally I kill the process manually (process is named "unknown" for some reason). The infolog does not contain a stack trace (example: http://pastebin.com/Und8yR57). Demo file has size 0. Operating system: Arch Linux.

Does not bother me much, but maybe should be fixed. Suggestions to debug this further?

Edit: more info. Seems a stack trace was output at stderr/stdout (http://pastebin.com/uhMhABiC). I have created a bug in mantis at: https://springrts.com/mantis/view.php?id=4952

[albator]

Happens to me since half a decade as well.
Usually when the game last a long time

[code_man]

Il join in on the bandwaggon here, its an old one since 96 for me.

It used to be unable to play in windowmode back then for me and when it hung at end it took the entire system with it, really annoying.
Nowdays this didnt happen to me and since i play in windowmode and can kill it if it hangs luckily.

I can also confirm its a rare bug and it corrupts the infolog backtrace.
Linux 64bit, portable btw.

[malric]

Had it today once more, added the back traces of all running threads at the moment of the hang in the bug report.

[Anarchid]

The hang happens somewhere in LuaUI shutdown. The hang happens to me as well. Until this thread, i assumed it only happened to me because i'm using primus.

How many people in this thread use primus?

[abma]

i don't have this issue, i'm using the OS ati-driver. any ATI users with this issue? Maybe its nvidia-only bug or primus related?!

[malric]

If by primus you mean you have Nvidia Optimus, then yes I have it as well. I use official nvidia drivers (as far as I remember) but don't remember how I configured it (was some time ago and it works so I did not touch it since).

[Anarchid]

If by primus you mean you have Nvidia Optimus

Yes; Nvidia Optimus via bumblebee via primusrun spring.

[malric]

With nvidia or nouveau drivers? (just for reference)

[Anarchid]

Nvidia.

Let's try to invalidate this suspicion. Does anyone have the bug but not nvidia optimus?

[albator]

I use/ed NVIDIA default prop driver shipped with Mint. So 343 now and 331 for a while. Both without optimus since I use a desktop.

[abma]

I use/ed NVIDIA default prop driver shipped with Mint. So 343 now and 331 for a while. Both without optimus since I use a desktop.

infolog.txt of a hang then please!

[Anarchid]

@abma are you ok with a hang infolog with optimus?

Maybe the issue should be mantized for easier trackings.

[abma]

there already exists a mantis report:
https://springrts.com/mantis/view.php?id=4952

imo it would be good when you create a new report, maybe its not the same, as its a hang midgame vs. hang at exit.

[Anarchid]

http://hastebin.com/daximowuki.xml

[code_man]

i don't have this issue, i'm using the OS ati-driver. any ATI users with this issue? Maybe its nvidia-only bug or primus related?!

Im using propeitary nvidia drivers, i have a 550gtx, i dont know primus is.

Spring is the only game with this problem.

[albator]

I use/ed NVIDIA default prop driver shipped with Mint. So 343 now and 331 for a while. Both without optimus since I use a desktop.

infolog.txt of a hang then please!

http://pastebin.com/w1YLwKxN

http://replays.springrts.com/replay/1b6 ... d2f65f8d2/

[albator]

Actually, it appears that the infolog is cropped....

So the origin of the problem might be related to the fact that spring is not writting anything to the infolog anymore.

Let me know if you need me to reproduce the replay with other parameters....

[albator]

Here is an example and the log is that time complete.

http://pastebin.com/r8ew13J0

http://replays.springrts.com/replay/51b ... 8516254ae/

[cleanrock]

Test if "LoadingMT = 0" removes the hang at exit.