Comment 75 for bug 1252121

Daniel Hahler [2013-12-19 13:24 -0000]:
> With the normal suspend methd, PrepareForSleep=false showed up now (and
> the network came up again), (together with PrepareForSleep=true):

Interesting. To the other affected people: are you using hybrid
suspend or the "simple" (RAM only) suspend?

> Where is the PrepareForSleep=false signal emitted

That comes from logind, which is the component which provides the
"Suspend()" API for the desktop. It should send "PrepareForSleep
true", then it will wait until apps which registered a suspend
inhibitor are done with their pre-suspend actions, then do the actual
suspend, and after resume send the "PrepareForSleep=false" signal so
that clients can do their post-suspend actions.

> Is it something NetworkManager requires to wake up again?

Exactly. NM shuts down the network on PrepareForSleep=false, and
searches for networks/re-connects on PrepareForSleep=true.