The new `mm_plugin_manager_find_device_support()' will be an async method which
waits for ports to be added in the given `MMDevice'. As soon as all added ports
are probed, the device support check is considered finished.
Before this, we only exported the modem to DBus when all ports were organized,
in order to make sure that we select as primary port the one we really want and
not the first AT port grabbed. Given that to get all the ports organized we also
needed to wait to get all the ports grabbed, we can now also defer the creation
of the modem object until all the ports get grabbed. This allows us to create
different types of objects based on the ports available (e.g. we can now create
QMI-supported modem objects if we see a QMI port around).
When the interface is shut down the skeleton object no longer exists. Given that
we're allowing this actions from outside the interface, we should be safe and
don't assume that the caller knows exactly the state of the interface.
Just when we request to cancel the connection we need to set the bearer as
being disconnected, so that we do not end up overwriting connection cancellation
requests.
Reported by: Thieu Le <thieule@chromium.org>
This patch fixes mmcli to set the default timeout on the manager proxy
interface (MmGdbusOrgFreedesktopModemManager1) instead of MMManager
itself. The later is not a GDBusProxy object.
We remove the `--command-timeout' option, as the AT command timeout is now
directly computed from our default DBus timeout (modifiable with `--timeout').
ITU rec v.250 (6.1.1) says:
The DTE should not include additional commands on the same command line after
the Z command because such commands may be ignored.
load_supported_storages may fail, which results in unallocated storage
memory. This patch modifies load_initial_sms_parts_from_storages and
is_storage_supported to handle that gracefully.
For 3GPP modems, we'll try with AT+WS46=?, which is the standard command to
query supported networks. If it returns error, we'll fallback to the defaults.
For CDMA modems, we'll try to gather CDMA1x and/or EV-DO support directly from
the AT+GCAP response.
Following the same logic as in commit 9ddcf92a, those contexts in the interface
which contain the setup of a timeout are cleared before removing the DBus
skeleton objects, so that we make sure they are not fired out once these have
been disposed.
expire_signal_quality() / mm_iface_modem_update_access_technologies()
may be called after the DBus object skeleton is removed in
mm_iface_modem_shutdown() and before their associated timeout source is
removed (in *_context_free) due to destruction of the modem object.
This patch fixes the issue by removing context objects before removing
the DBus skeleton in mm_iface_modem_shutdown().
