Hello. I can't reproduce, I tried running wireshark, I can capture packets, I can save to file, I can then load this file. No problem. I tried wireshark as root and non-root. I ran my test on a up-to-date Kali Rolling VM, and I can confirm that I have the same versions as you (wireshark 4.4.0-1, libqt6gui6:arm64 6.7.2+dfsg-4kali1).

Can you please make sure that your system is up-to-date (sudo apt update && sudo apt full-upgrade)? And please reboot your system after that, just to be 200% sure?

Note that I could only test on a amd64 VM, I don't have a arm64 machine around to test... It could be that you're hitting a arch-specific bug.

There are indeed arm64 qt6 bugs reported on the Debian bugtracker, similar to your bug:

• https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1085035
• https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1086032

Assuming you run Wayland, you could try to switch to X11 instead (the exact procedure to do that depends on whether you run GNOME, KDE, XFCE)

I tried both X11 session or Wayland session with GNOME, amd64 VM, and I still can't reproduce, so it really seems to be specific to arm64

Yes, I'm quite sure it is arm64 specific. The system is up to date and I have also rebooted.

Also, to cross check, I've reinstalled from a fresh image. This works fine. Then, I upgraded only wireshark and Qt and I got the same crash as on the fully up-to-date system.

The Debian bug reports look very much like the same problem.

I'm using XFCE, which only supports X11, I believe.

I'm using XFCE, which only supports X11, I believe.

You are correct.

If you're motivated, you could try another install with GNOME, check that you can reproduce the issue. If so, then enable Wayland (by default Kali's GNOME runs X11), and check that the bug can still be reproduced. That would help to know if it's a bug that is specific to X11 or Wayland, and maybe it would help us to find if a patch is available already. Or at least, it would help to refine the existing bug report.

To make sure you run wayland or x11, you can run:

$ printenv | grep SESSION_TYPE
XDG_SESSION_TYPE=wayland

To enable Wayland on Kali's GNOME, check the file /usr/lib/systemd/system/gdm.service.d/disable-wayland.conf, it gives the procedure to switch from X11 to Wayland.

I upgraded only wireshark and Qt and I got the same crash as on the fully up-to-date system.

The thing is, we can't roll back this change, now latest Qt6 is in kali-rolling, and it will be released in Kali 2024.4 in around 2 weeks or so.

OK, I tried the instructions, but I couldn't get a Wayland session to run. GDM still only offered to start X11 sessions. Again, the arm64 build for Raspberry may be a bit of a special case here ...

I know that the instructions work for amd64, I just tried it myself when I wrote the message, but that was for a amd64 VM, indeed, I didn't test arm64. However I don't think arm64 should be special in that regard... Do you use an image downloaded from https://www.kali.org/get-kali/#kali-arm?

GDM still only offered to start X11 sessions

After enabling Wayland, I just log in without touching any setting, I didn't even notice that GDM offers to start X11. By default it should log you in Wayland.

OK, I was missing the "dtoverlay=vc4-kms-v3d" line in /boot/config.txt

Now, I also tested with Wayland: Exactly the same crash happens there as well.

I checked the Qt6 bug tracker, I just can't find this issue reported. I also checked the Git tree, couldn't really find any commit that seems related.

There is a commit in the 6.8 branch that "prevent crash" in various Q*Dialog widgets. Despite the promising name, it doesn't look related at all to our issue: no mention of ARM, and seemingly fixing an issue reported more than 10 years ago... Cf.

• https://github.com/qt/qtbase/commit/81d55f772892445686862e49d0a6092668f8c725
• https://bugreports.qt.io/browse/QTBUG-54693

So we only know that this bug was reported twice on the Debian bugtracker (cf. https://bugs.kali.org/view.php?id=8997#c20040), however the Debian maintainers didn't respond. We can hope that the next Qt6 version in Debian will magically fix it. Or, reach out on the Qt6 upstream bugtracker to report the issue -- feel free to do so @birefringence if you have the time :)

OK, I tried to report it here: https://bugreports.qt.io/browse/QTBUG-131747

Hey, I think it should be fixed in kali-rolling, please sudo apt update && sudo apt full-upgrade, you should see new version 6.7.2+dfsg-4kali1+b1 of qt6 being installed. Can you please confirm? Thanks!

Yes, it's fixed! So that means I can close the Qt bug report again?

Yes, it's fixed!

Nice to hear.

We just rebuilt the qt6 package. So it means something changed in the build dependencies, there was something bogus with the earlier build (both in Kali and Debian). I will compare the build logs on Monday, out of curiosity.

So that means I can close the Qt bug report again?

Yep, sorry for wasting your time. If I find something interesting in the build logs, I'll mention it in the Qt bug report.

No problem, thanks for fixing this!

Could worth to add this info to the Debian bugs?

I checked the build logs, but it's hard to really see anything. The last 2 builds of qt6-base in Kali were on the 5th (producing package with the crash) and 29th (producing good package) of November. Of course there was quite some changes in the build-deps in between, so there's nothing obvious to see in the build logs.

I added some info to the Debian bugs, to mention that it should be fixed with a binNMU.

Closing the bug here. Thanks @birefringence for following up until the end!