[PATCH 0/5] Power supply and resume ordering meddling

[Werner Almesberger]

Andy Green wrote:
> Well as you realized while writing that, "handled completely in
> userspace" is not the right answer here, the event-driven tracking of
> charger state from the kernel is needed to drive any userspace policy.

Heh, you caught me there ;-) But in fact, I added that cautious
"efficiently" because I didn't want to state that just polling
would already be good enough.

Of course, an event-driven interface is architecturally nicer,
given that the source of all this are interrupts.

- Werner