I believe the keyboard backlight is controlled by the EmbeddedController firmware.
I have a similar issue with 4th gen Intel i7 4500U (Lenovo Thinkpad Yoga S1).
I am unable to reproduce this bug on my own. It tends to happen when I put my laptop in my bag.
I believe I had this issue before bookworm.
Also, bookworm from what I know does not use pm-suspend (systemd is in charge of the suspend).
You can remove pm-utils if installed to check it does not play a role.
I know for one that this laptop's latest BOIS update is not ACPICA conformant which can trigger issues with Linux kernel (a few data fields are accessed as global instead of being prepended with their path: D1F0, D1F1, D1F2 instead of \_SB.PCI0.D1F0, etc. This triggers acpi errors when the kernel calls ACPI functions that access these).
But this could be another issue I am not aware of (I don't know if resume from suspend from the Linux kernel involves any of these broken ACPI calls).
The fact that it worked with slackware and opensuse is of interest. You should open a bug report with these details on the Debian bug report tracker to track down this issue and share the reports between the affected users.
Could be these distributions disable a feature or have a kernel patch not in Debian.
Also, I saw reports about such hang on CPU microcode load by the kernel that could be worked around by adding dis_ucode_ldr to the kernel prompt, but they seem to be about easily reproducible setup (this was about suspending the laptop when AC unplugged on HP laptop, later fixed by BIOS and microcode updates).
Mind in my case I did not test the keyboard backlight key working. I also had the Function key led on, the suspend led pulsating (it is solid when the laptop is on), and the microphone mute led on (while I did not press it to mute the microphone).
The laptop is hot as far as I recall when this happens.
So it might be overheating protection
https://www.youtube.com/watch?v=6BsD4til36Q
But why does it overheat is another story, as the time to suspend might not let it overheat even if I put it in the bag.
The above mic mute led could hint that the workaround for
https://wiki.archlinux.org/title/Len..._Carbon_(Gen_7)
to put "snd_hda_intel.dmic_detect=0" on the kernel command line might help.
Still, I need to be able to reproduce the issue somewhat to check this helps.