Debian 'nocloud' images hang during boot
I've been using the `nocloud-arm64` flavor of Debian cloud images for a while on top of UTM/QEMU on macOS. Recently, they started to hang during boot and never show the login prompt where I can log in as `root` with no password. The last image I found that worked fine is the one from January this year ([20260112-2355](https://cdimage.debian.org/cdimage/cloud/trixie/20260112-2355/)). The issue persists even with the images released today ([20260313-2415](https://cdimage.debian.org/cdimage/cloud/trixie/20260313-2415/)).
I've attached a terminal to it (editing the VM and going to `Devices -> Display -> New -> Serial`) and got the following output:
```
Welcome to your new installation of Debian GNU/Linux 13 (trixie)!
Please configure your system!
-- Press any key to proceed --
[ OK ] Finished ldconfig.service - Rebuild Dynamic Linker Cache.
Starting systemd-update-done.service - Update is Completed...
[ OK ] Finished systemd-update-done.service - Update is Completed.
[ OK ] Finished apparmor.service - Load AppArmor profiles.
```
After pressing `ENTER` a few times, I got:
```
‣ Please enter the new timezone name or number (empty to skip, "list" to list op
tions):
No data entered, skipping.
🔐 ‣ Please enter the new root password (empty to skip): (press TAB for no echo)
No password entered, skipping.
/etc/passwd written.
/etc/shadow written.
[ OK ] Finished systemd-firstboot.service - First Boot Wizard.
[ OK ] Reached target first-boot-complete.target - First Boot Complete.
```
At this stage, the boot proceeds and I can see the regular login prompt in the main VM display window (not the serial console). Still, I wasn't able to log in as `root` without a password as usual.
Given this situation, I'd like to ask: was this change intentional? Or is this a regression from an unrelated change?
issue