https://bugs.winehq.org/show_bug.cgi?id=44480
Bug ID: 44480 Summary: Rabi-Ribi v1.88: Game crashes during startup Product: Wine Version: 3.1 Hardware: x86-64 OS: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: -unknown Assignee: wine-bugs@winehq.org Reporter: mf_messias@protonmail.com Distribution: ---
Created attachment 60430 --> https://bugs.winehq.org/attachment.cgi?id=60430 Latest backtrace
Description of problem: Game crashes during early loading, with or without DLCs.
How reproducible: Always reproducible.
Steps to Reproduce: 1. Load Rabi-Ribi from Steam (Windows) with default flags 2. Wait until it crashes.
Additional info: Changing startup flags don't see to affect it. A current "workaround" is downgrading the game back to version 1.80 in "Proprieties -> Betas" if one wishes to play it. Seems like startup code was changed and it might be related to the crash: https://steamcommunity.com/app/400910/discussions/0/1482109512320896025/
https://bugs.winehq.org/show_bug.cgi?id=44480
Fabian Maurer dark.shadow4@web.de changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |dark.shadow4@web.de
--- Comment #1 from Fabian Maurer dark.shadow4@web.de --- Does using wine-staging make any difference?
https://bugs.winehq.org/show_bug.cgi?id=44480
--- Comment #2 from Matheus F. mf_messias@protonmail.com --- Created attachment 60440 --> https://bugs.winehq.org/attachment.cgi?id=60440 Backtrace from wine-staging 2.20
https://bugs.winehq.org/show_bug.cgi?id=44480
--- Comment #3 from Matheus F. mf_messias@protonmail.com --- No, it doesn't, it pretty much gives the same page fault error. Although I am currently one version behind the latest wine-staging release (2.20 vs 2.21). I attached the backtrace just in case it is useful, but it seems the same.
https://bugs.winehq.org/show_bug.cgi?id=44480
Aru aru@rafaelfsilva.com.br changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |aru@rafaelfsilva.com.br
--- Comment #4 from Aru aru@rafaelfsilva.com.br --- Created attachment 60473 --> https://bugs.winehq.org/attachment.cgi?id=60473 playonlinux.log clean prefix Rabi-Ribi before downgrade
Currently running Rabi-Ribi through Steam via PlayOnLinux with Wine version 3.1 - 32 bits.
I can confirm the application crash happens and I get a similar backtrace with same addresses.
I found a workaround by downgrading the following packages in Arch Linux: Linux 4.14.15-1-ARCH #1 SMP PREEMPT Tue Jan 23 21:49:25 UTC 2018 x86_64 GNU/Linux
-> nvidia from: 2) nvidia-390.25-10-x86_64.pkg.tar.xz (local) to: 13) nvidia-387.34-21-x86_64.pkg.tar.xz (local)
-> nvidia-utils from: 2) nvidia-utils-390.25-2-x86_64.pkg.tar.xz (local) to: 5) nvidia-utils-387.34-5-x86_64.pkg.tar.xz (local)
-> linux from: 2) linux-4.15.3-1-x86_64.pkg.tar.xz (local) to: 11) linux-4.14.15-1-x86_64.pkg.tar.xz (local)
-> lib32-nvidia-utils from: 2) lib32-nvidia-utils-390.25-1-x86_64.pkg.tar.xz (local) to: 4) lib32-nvidia-utils-387.34-1-x86_64.pkg.tar.xz (local)
I'm still using Wine 3.1 now and the game works fine.
Right now I'm a little busy but I can gather more data after few hours.
Is there anything I can help for this issue? Would it help to try to downgrade all packages one by one until I find the culprit or that's not needed?
I'm not great at debugging these type of problems so I'm not sure how much I can help. I'd gladly follow instructions to gather more data to help to fix this issue.
I've also attached playonlinux.log from a clean Wine prefix on PlayOnLinux from before downgrading the packages.
I noticed these messages that seemed a little... off: 00a8:err:d3d:wined3d_debug_callback 0x86c8320: "GL_OUT_OF_MEMORY error generated. Failed to allocate CPU address space mapping for texture (consider building 64-bit app).". 00a8:err:d3d:wined3d_debug_callback 0x86c8320: "GL_INVALID_VALUE error generated. Size and/or offset out of range.".
I imagined that it was either a kernel or nvidia drivers regression which was what lead me to downgrading the packages.
I hope this information helps, even if a little. I'm open to testing more but for now I'd need someone to tell me what to test and what logs to attach.
Thanks! Have a great day!
https://bugs.winehq.org/show_bug.cgi?id=44480
--- Comment #5 from Fabian Maurer dark.shadow4@web.de --- Thanks for the detailed feedback, when I find the time I'll do some testing myself, as I too own the game - but have an AMD gpu. Maybe that makes a difference. When you make an apitrace, does replaying that reproduce the issue?
https://bugs.winehq.org/show_bug.cgi?id=44480
--- Comment #6 from Aru nieve.aru@gmail.com --- I'll test that tomorrow. I'm a little tired today.
I'm sorry if I should not be posting this but there's a possibility the bug is mostly related to NVIDIA drivers than to Wine. Especially since it works for you on the AMD card.
I found some topics about the 390.25 driver.
"Terrible performance regression with Nvidia 390.25 driver": https://bbs.archlinux.org/viewtopic.php?id=234241
and "Numerous Performance & Rendering Issues on 390.25": https://devtalk.nvidia.com/default/topic/1029484/linux/-linux-mint-18-3-nume...
Some of the regressions might affect Rabi-Ribi. As soon as I'm free tomorrow I'll post again about the apitrace.
https://bugs.winehq.org/show_bug.cgi?id=44480
--- Comment #7 from Matheus F. mf_messias@protonmail.com --- Well, I am on Intel/mesa, so I wonder if that is totally true. Still, I should check out terminal output and see if it is similar later (kinda skipped it due to Steam spamming like crazy).
https://bugs.winehq.org/show_bug.cgi?id=44480
--- Comment #8 from Aru nieve.aru@gmail.com --- It might be something different then.
I tested some more on each combination along with apitrace.
This is how I ran apitrace (32-bit version): apitrace32 trace wine Steam.exe
And then collected the wine-preloader.trace file on Rabi-Ribi directory.
To replay the apitrace I used (x86_64 version): apitrace replay wine-preloader.trace
Test cases: 1) Arch Linux, Kernel 4.14.15, NVIDIA 387.34-21: a) Game works fine. b) apitrace generated fine: apitrace_rabiribi_linux.4.14_nvidia_387.34.trace c) apitrace works fine on all test cases 1, 2 and 3.
2) Arch Linux, Kernel 4.15.3, NVIDIA 387.34-1: a) Game works fine. b) apitrace generated fine: apitrace_rabiribi_linux.4.15_nvidia_387.34.trace c) apitrace works fine on all test cases 1, 2 and 3.
3) Arch Linux, Kernel 4.15.3, NVIDIA 390.25-11: a) Game crashes. b) apitrace generated fine: apitrace_rabiribi_linux.4.15_nvidia_390.25.trace c) apitrace works fine on all test cases 1, 2 and 3, there are no crashes and apitrace exits nicely: output similar to the others and exit code 0.
So, I can still run the game if I use the latest kernel on repository if I patch NVIDIA 387.34-1 to be able to be compiled on 4.15 kernel. This is my current setup now.
Is there anything else I can test or any more information I can try to find?
I hope this is helpful. Sorry for taking long to report the tests.