Lennart Poettering

This has been requrested before, and I understand why you want it, but I am still not convinced this is a good idea...

"systemctl poweroff" shuts down going through the usual clean service termination logic. "systemctl poweroff -f" skips the clean service termination logic, and goes directly into the final killing and umount...

OK, giving the hint to retry with "-ff" certainly is a good idea. Reopening this issue hence, and renaming.

I posted PR #3310 now that clarifies that "systemctl reboot -ff" doesn't contact PID1, and hence works also when PID 1 has crashed or hangs.

Hmm, I wonder what the best approach to detecting the thunderbolt docking state would be. Ideally they'd provide an input device exposing the SW_DOCK switch. @gicmo any idea? what's the...

(hmm, it might even suffice if boltd implements SW_DOCK in userspace via uinput. that would be pretty straightforward I guess, and not require kernel patching. Doing this in the kernel...

hmm, ok. so, what's the plan then? @gicmo, can i convince you to implement SW_DOCK via uinput in boltd and some heuristic? it's not trivial given this is all async......

(it might be slightly better to do the tag thing, i.e. solve this in logind and leave bolt as it is, sounds more future proof to me. opinions?) (if we...

> IIRC, the primary heuristic we have chosen so far in GNOME is whether an external monitor is available. People are generally unhappy with this though and a better (but...

> I think the trigger for me considering going down that route was the whole `SuspendThenHibernate` thing. We had patches to implement it in GNOME, but the patched code paths...