Flatpak version fails to launch

[Treeways]

Hello!

The Flatpak version of FireDragon gives me this error, then crashes a few seconds later:

> flatpak run org.garudalinux.firedragon
[2] Sandbox: CanCreateUserNamespace() clone() failure: EPERM

This bug may be related to this bug in upstream Firefox, which is a duplicate of this: https://bugzilla.mozilla.org/show_bug.cgi?id=1756236

EDIT: The error thrown seems unrelated to the crash.

Here is a brief core dump generated by systemd.

        Signal: 11 (SEGV)
  Command Line: /app/lib/firedragon/firedragon --name org.garudalinux.firedragon
    Executable: /app/lib/firedragon/firedragon
 Control Group: /user.slice/user-1000.slice/user@1000.service/app.slice/app-flatpak-org.garudalinux.firedragon-1398913528.scope
          Unit: user@1000.service
     User Unit: app-flatpak-org.garudalinux.firedragon-1398913528.scope
         Slice: user-1000.slice
     Owner UID: 1000 (treeways)
       Storage: /var/lib/systemd/coredump/core.firedragon.1000.c181c52c904f487b802a71a667904614.16548.1756837828000000.zst (present)
  Size on Disk: 12.8M
       Message: Process 16548 (firedragon) of user 1000 dumped core.
                
                Stack trace of thread 77:
                #0  0x00007fe5ef9b33b6 n/a (/usr/lib/x86_64-linux-gnu/libvulkan.so.1.3.290 + 0x443b6)
                #1  0x00007fe5ef9a3e17 n/a (/usr/lib/x86_64-linux-gnu/libvulkan.so.1.3.290 + 0x34e17)
                #2  0x00007fe5ef9a405d n/a (/usr/lib/x86_64-linux-gnu/libvulkan.so.1.3.290 + 0x3505d)
                #3  0x00007fe5fbb15e13 n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libVkLayer_MESA_device_select.so + 0x5e13)
                #4  0x00007fe5ef9aa929 n/a (/usr/lib/x86_64-linux-gnu/libvulkan.so.1.3.290 + 0x3b929)
                #5  0x00007fe5f848daf7 n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libgallium-25.2.1.so + 0xc8daf7)
                #6  0x00007fe5f84914e9 n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libgallium-25.2.1.so + 0xc914e9)
                #7  0x00007fe5f78fd352 n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libgallium-25.2.1.so + 0xfd352)
                #8  0x00007fe5f8077f9e n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libgallium-25.2.1.so + 0x877f9e)
                #9  0x00007fe5f7901d50 n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libgallium-25.2.1.so + 0x101d50)
                #10 0x00007fe5f7788f2b n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libEGL_mesa.so.0.0.0 + 0x1ff2b)
                #11 0x00007fe5f777e948 n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libEGL_mesa.so.0.0.0 + 0x15948)
                #12 0x00007fe60f7c9b96 n/a (/app/lib/firedragon/libxul.so + 0x49c9b96)
                ELF object binary architecture: AMD x86-64

Notes

• The browser opens in safe mode (when eventually prompted).
• Moving ~/.var/app/org.garudalinux.firedragon/ to a different directory does not fix the crash.
• The AppImage and .tar.xz releases run fine.
• I'm running Fedora Linux 42 with nouveau drivers.

[Treeways]

I was able to reproduce the error in Firefox - however, even with the error, the browser successfully launches.
Interestingly, Flathub's Firefox immediately crashes, while the Fedora Flatpak repository's Firefox does not crash. On the contrary, Librewolf's Flathub Flatpak crashes after a few seconds in the same way.

This is leading me on to think something is up with the underlying FreeDesktop runtime that Flathub utilizes, although I'm not sure what to try from here to verify that.

[stefanwimmer128]

Interesting. But unfortunately I'm unable to reproduce this.

Additionally I can confirm that [2] Sandbox: CanCreateUserNamespace() clone() failure: EPERM is an expected error when running the flatpak.

You could also try the Floorp Flatpak. And since the core dump points to somewhere in the graphics stack, you could try switching out the nouveau driver with the proprietary one to test. You could also try booting a live environment of another distribution and test the Flatpaks there.

[Treeways]

Thanks for the feedback - I bet you're on the right track here. I distro-hopped over to Arch so I could use Chaotic AUR's FireDragon instead, and (as with running the .tar.gz release from Fedora) that works perfectly fine with a stock Arch installation. It runs even smoother on this OS, actually, and distro-hopping also solved some other problems I had.

Interestingly, however, the Flatpak still does not launch (also tested with Firefox, Librewolf and Floorp with the same stack trace). This time, the core dump points to somewhere in libc:

Arch Linux with FireDragon Flatpak

coredumpctl info 2703
           PID: 2703 (firedragon)
           UID: 1000 (treeways)
           GID: 1000 (treeways)
        Signal: 11 (SEGV)
     Timestamp: Thu 2025-09-11
     ...
  Command Line: /app/lib/firedragon/firedragon
    Executable: /app/lib/firedragon/firedragon
 Control Group: /user.slice/user-1000.slice/user@1000.service/app.slice/app-flatpak-org.garudalinux.firedragon-3989238049.scope
          Unit: user@1000.service
     User Unit: app-flatpak-org.garudalinux.firedragon-3989238049.scope
         Slice: user-1000.slice
     Owner UID: 1000 (treeways)
     ...
       Storage: /var/lib/systemd/coredump/core.firedragon.1000.3e286bee5cff431a999212bb1124a32b.2703.1757626562000000.zst (present)
  Size on Disk: 13M
       Message: Process 2703 (firedragon) of user 1000 dumped core.

                Stack trace of thread 77:
                #0  0x00007f6bff49b344 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x9a344)
                #1  0x00007f6bff44250e n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x4150e)
                #2  0x00007f6bf90bde26 n/a (/app/lib/firedragon/libxul.so + 0x7abde26)
                ELF object binary architecture: AMD x86-64

I've got the following graphics cards on this laptop:

• Integrated: Intel UHD Graphics 630
• Discrete: NVIDIA GeForce GTX 1650 Max Q GPU

This stack trace was generated with the stock firmware that comes out of the box in an Arch installation (no nouveau or nvidia packages are present):

> pacman -Q | grep -e nvidia -e intel
intel-ucode 20250812-1
linux-firmware-intel 20250808-1
linux-firmware-nvidia 20250808-1

I'll test with proprietary NVIDIA drivers and will let you know whether or not I get the same crash.

[Treeways]

This bug might be better reported on Bugzilla - should I ask there instead?