Is it necessary to disable modem before powering it down?

[Dan Williams]

On Wed, 2014-02-19 at 14:21 -0800, Thieu Le wrote:
> Hello,
> 
> I am trying to solve a problem during our system suspend where the disable
> modem step takes a long time and it would be quicker to just go straight to
> modem power down.
> 
> mm-iface-modem.c:handle_set_power_state_auth_ready() checks to make sure
> the modem is disabled before allowing it to be powered down.  Is this done
> because some modems require it to be in a disabled state before allowing
> the power down?

Out of curiosity, which command takes too long?

I think this is done mostly to ensure that internal stuff is cleaned up
correctly; like timeouts, polling the modem for status, and other things
that may have been set up during the enable state.  I don't think most
modems would care about being disabled if you're just going to CFUN=0
them.

If we were able to ensure that these things were properly cleaned up,
then we could jump directly to the power down state.  The way to do that
would probably be to identify what gets torn down during 'disable',
split that out into separate functions that could be called from both
disable and power down, and then do that from both places.

Dan