Comment 4 for bug 1229670

OK, continued to run powernapd in the debugger. My last comment does not turn out to by valuable, no polling on /dev/null. But some kind of polling is going on.

Line 92 in InputMonitor.py returns always directly after calling it (res = self._poll.poll(1000)). And always returns code 24. The InputMonitor is hooked to my USB keyboard. The keyboard is working properly while this. Killing the thread gets powernap into a sane state again.
Because I am seeing this from ubuntu 12.04 up to ubuntu 13.10 on two different i386 (32bit) boxes with different keyboards attached, I would scratch out some bad hardware behavior because of power save modes.

Let's see if I can bring some more info.