[Linux/Flatpak] Cemu crash

[trinukso]

Cemu crashes at BOTW game startup. Version 2.0-55 from flatpak

flatpak run info.cemu.Cemu

(Cemu_relwithdebinfo:23): Gtk-CRITICAL **: 09:48:00.937: gtk_box_gadget_distribute: assertion 'size >= 0' failed in GtkScrollbar
Violación de segmento!
Error: signal 11:
Cemu_relwithdebinfo(eglSwapInterval+0) [0x5570925d8b78]+0) [0x5570925d8b78]
/usr/lib/x86_64-linux-gnu/libc.so.6(+0x3ee80) [0x7f06dda51e80]
Cemu_relwithdebinfo(eglSwapInterval+0) [0x5570925d8b78]+0) [0x5570925d8b78]

Stacktrace and additional info written to:
/home/vicen/.var/app/info.cemu.Cemu/data/Cemu/log.txt

Thanks.

log.txt

[Squall-Leonhart]

Not nearly enough information provided system info wise.

[trinukso]

Not nearly enough information provided system info wise.

Apologies. ArchLinux. Flatpak 1.15.4

Intel(R) Core(TM) i5-10400F CPU @ 2.90GHz 6.5.7-zen1-1-zen / Wayland (Gnome) NVIDIA GeForce RTX 2060/PCIe/SSE2

[trinukso]

The commit 3bbae71c2949a615b2a402d1b2672115e47390eb964ca7cab05692447c7ba91d (flatpak) works.

[ColinKinloch]

Oh well, I guess dark mode will have to wait for wxWidgets 3.2.3.x. Rolling back, https://github.com/flathub/info.cemu.Cemu/pull/81

[ColinKinloch]

The rolled back flatpak is now published. You can close this or maybe rename it to something like "wxWidgets 3.2.3 crash"

[ColinKinloch]

I just noticed, NVIDIA + wayland. Are you are forcing Cemu to use x11 with --nosocket=wayland?

[trinukso]

I just noticed, NVIDIA + wayland. Are you are forcing Cemu to use x11 with --nosocket=wayland?

No. Cemu use XWayland.

[ColinKinloch]

That's what I meant. XWayland is an x11 implementation on top of wayland. Setting --nosocket=wayland will make it seem from Cemus point of view that it's using x11. Otherwise the Cemu flatpak uses wayland which on NVIDIA can be buggy.

[trinukso]

That's what I meant. XWayland is an x11 implementation on top of wayland. Setting --nosocket=wayland will make it seem from Cemus point of view that it's using x11. Otherwise the Cemu flatpak uses wayland which on NVIDIA can be buggy.

So I should use this option?

[ColinKinloch]

Probably. If you've not noticed an issue maybe the drivers have been fixed. Here's some relevant info: https://github.com/cemu-project/Cemu/issues/792

[Clueninja]

I have the same issue running BoTW I tried building from source and the AUR both crashed in the same way yours did. Tested on both Xorg and Wayland (with and without --nosocket=wayland)

System Details Report

---

Hardware Information:

• Memory: 16.0 GiB
• Processor: Intel® Core™ i7-9750H × 12
• Hybrid Graphics: Quadro P600 + Intel UHD Graphics 630

Software Information:

• Firmware Version: F.23
• OS Name: EndeavourOS
• OS Build: 2022.08.05
• OS Type: 64-bit
• GNOME Version: 46
• Windowing System: X11
• Kernel Version: Linux 6.9.1-arch1-1

[Clueninja]

https://pastebin.com/ZBFy4uCh Appears to be a possible duplicate of #398 #466, #513, and #781 though some have experienced success using the intel gpu unfortunately I have not

[ColinKinloch]

This issue is specifically about the flatpak version, and not the AUR. --nosocket=wayland only applies to flatpak. Also I think this bug was related to the specific version of wxWidgets a few releases ago.

Out of interest why is your distro from 2022?

[Clueninja]

Yeah, so I tried the flatpak one first on wayland, with and without --nosocket=wayland then switched to x11 and tested that, then installed aur package and tested that, and then built from git. All of them crashed in the same exact way, similar to this issue and similar to a lot of the issues I tagged. I don't know why its 2022, I guess being arch based that's when I installed the os, all my packages are bleeding edge, though for flatpak cemu that shouldn't matter. If you want me to test anything or link a full crash dump (apart from the log I couldn't find anything) please let me know. I've had this issue for quite a while since I switched from a pc to my laptop or from running through wine to native, (they happened around the same time) so I'm very eager to get one of, if not, my favorite game running again! Cheers!

[Clueninja]

@ColinKinloch I just turned off Hybrid graphics in the bios, now its running only on the nvidia gpu, I got the exact same error. Also I tried the AppImage build and that also crashed with the output: Segmentation fault! Error: signal 11: /tmp/.mount_Cemu-2HIIgML/AppRun.wrapped(PPCInterpreterContainer<PPCItpCafeOSUsermode>::PPCInterpreter_LWZ(PPCInterpreter_t*, unsigned int)+0x26) [0x5d66a83741e6] /usr/lib/libc.so.6(+0x3cae0) [0x7d81925eeae0] /tmp/.mount_Cemu-2HIIgML/AppRun.wrapped(PPCInterpreterContainer<PPCItpCafeOSUsermode>::PPCInterpreter_LWZ(PPCInterpreter_t*, unsigned int)+0x26) [0x5d66a83741e6] /tmp/.mount_Cemu-2HIIgML/AppRun.wrapped(coreinit::__OSFiberThreadEntry(void*)+0x368) [0x5d66a85119e8] /usr/lib/libc.so.6(+0x56be0) [0x7d8192608be0]

I hope this helps narrow it down!

[ColinKinloch]

My point was that your issue is unrelated to this issue report. You should open a new issue or find one that's more relevant.

In fact I think this issue should have been closed in October.

[Clueninja]

My point was that your issue is unrelated to this issue report. You should open a new issue or find one that's more relevant.

In fact I think this issue should have been closed in October.

Okay, just looked like the same error message I got when I tested with the flatpak version. I couldnt find any more recent issues with a similar error so I'll just make a new issue for it. Cheers!