Powering off modems

[Dan Williams]

On Thu, 2014-02-20 at 17:12 +0100, Aleksander Morgado wrote:
> On Thu, Feb 20, 2014 at 4:58 PM, Dan Williams <dcbw at redhat.com> wrote:
> >> I just merged some commits which enable ModemManager to tell the modem
> >> to power-off completely (no new API, just allowing OFF in
> >> SetPowerState()). The idea behind these changes is to let users of
> >> ModemManager tell the modem to power-off before really cutting the
> >> power source of the device. It is not a general use case (i.e. not
> >> everyone should do that), only do it if you are able to cut the power
> >> to the device and if you're about to do it after setting the modem in
> >> OFF mode.
> >>
> >> Currently it is implemented for QMI, Wavecom and Cinterion devices
> >> only. For QMI devices, the device goes in OFFLINE mode, which can only
> >> be recovered doing an actually physical power cycle (or with a RESET
> >> command, which MM won't do automatically). For Cinterion devices, the
> >> power off will make the ttys unusable (if RS232 device) or will really
> >> remove the ttys (if USB device). For Wavecom devices, it ends up
> >> removing the ttys and also rebooting the device after some seconds.
> >>
> >> I was going to include CFUN=0 as default implementation, but didn't do
> >> it. Maybe we should.
> >
> > The problem with this is when you tether a phone.  CFUN=0 tells the
> > phone to turn off, which almost always is not what you want when you're
> > just shutting down or suspending your laptop.  And it's typically hard
> > to identify when something is a phone and when it's not, either through
> > kernel drivers (most just show up as option or usbserial) or AT
> > responses.
> 
> So maybe it is better to just keep this OFF functionality per-plugin;
> and not provide the CFUN=0 default, right?

Yeah.

Dan