Regression in 7.1.6-1: Segfault when starting virtual machine

Description:

Since updating virtualbox to version 7.1.6-1, some of my VMs fail to start. The journal shows the following error:

Jan 24 21:29:58 zenpad systemd-coredump[36494]: [🡕] Process 36351 (VirtualBoxVM) of user 1000 dumped core.
                                                
                                                Stack trace of thread 36427:
                                                #0  0x00007831e14da00a n/a (VBoxDD.so + 0xda00a)
                                                #1  0xffffffffffffffff n/a (n/a + 0x0)
                                                ELF object binary architecture: AMD x86-64

Downgrading to 7.1.4-4 solves the issue. Note that I only downgraded the virtualbox package itself. virtualbox-host-dkms is still at 7.1.6-1.

Unfortunately, I have not yet figured out what distinguishes VMs that crash from VMs that run properly.

Additional info:

• package version(s):
  • virtualbox-7.1.6-1
  • virtualbox-host-dkms-7.1.6-1
• Possible upstream bug: https://www.virtualbox.org/ticket/22293

Steps to reproduce:

• Update to virtualbox-7.1.6-1
• Try to start all of your VMs
• Some may fail with above error. Some won't. Probably this depends on the configuration.