When the host is resuming from system suspend, QMI indications
sent by the modem at resume time can be lost. The exact reason why it
happens is still unknown. Until this is fixed, ModemManager currently
workarounds that in QMI mode by listening and reacting to AT URCs too,
which are being received reliably. In order to achieve that,
messaging_setup_unsolicited_events chains the parent's implementation
with its own, effectively setting up handlers for both AT and QMI
channels.
This worked fine on modems such as EG25 which enable SMS indications
by default. However, some modems, such as BM818, don't have these
indications enabled on boot and don't report incoming messages via AT
unless requested via AT+CNMI.
To make SMS handling on resume reliable on such modems, make sure
that MMBroadbandModemQmi also enables/disables unsolicited events
in the same way it already sets up handlers for them.
Signed-off-by: Sebastian Krzyszkowiak <dos@dosowisko.net>
register_in_network_cancelled() may be called early if the given
cancellable is already cancelled, and if so, we want it to remove the
timeout and signal handler, which should have been configured before
setting up the cancellation signal handler.
Returning an error in the registration set request completely breaks
the async operation requesting the explicit registration, even if the
modem may have already been registered via indications while waiting
for the set request to be completed.
We now ignore the registration set error if it returns a generic
failure with NwError=0, as in certain Qualcomm based devices. We
ignore unconditionally, without explicitly checking if we're
registered or not, because the upper logic will anyway do that.
Fixes https://gitlab.freedesktop.org/mobile-broadband/ModemManager/-/issues/723
So that it is reloaded fresh once re-enabled, otherwise we may be
using stale state values that are not in sync with the state reported
in the interface skeletons.
When a request to disable the modem arrives while the packet service
state wait is ongoing we were not correctly cancelling the operation.
The main reason for this is that this operation does not change the
modem state, and so the "wait for final state" logic in the disabling
sequence was not being considered.
We solve this by plugging in the Simple.Connect() operation
cancellable in the wait for packet service state operation. The
connection attempt will be cancelled during the disabling sequence as
well, and when that happens we will explicitly halt the packet service
state wait as well.
We now support modem-reported packet service state, instead of
guessing it in the 3GPP interface logic.
The DSD system status reporting a valid access technology is an
indication that the packet service is ready. Instead of holding back
the registration status until the DSD system status reports a valid
value, we will now report the modem as registered, and bind the DSD
system status exclusively to the packet service status.
In the simple connection attempt we're already waiting first for being
registered, and then for being attached in PS.
In certain protocols like QMI or MBIM we may be able to report an
exact packet service state, so there is no need to guess it, as the
guess may not always be right.
The logic will track automatically whether modem-reported packet
service states are available, and use them if so. Otherwise, it'll try
to guess as we were doing before (e.g. if registered in EPS, packet
service is considered attached).
When a request to set a new eps bearer settings is received, ignore the
profile-name when comparing the previous configuration with the new one,
since the modem's profile might already have a name that won't match the
settings provided by the user. Profile names are used in QMI modems.
If the profile name does not match the existing one, the modem will
detach and reattach to the network.
Users with QMI or MBIM capable modems may want to ensure that these
are never managed using plain AT commands, as that also involves using
PPP. This fallback to AT could happen if the QMI or MBIM port probing
fails for whatever reason.
The new `ID_MM_REQUIRED` udev tag allows specifying that a given port
MUST be successfully grabbed when creating a new modem object, or
otherwise the modem object will not be created at all (even if there
are other fallback control ports like AT that could have been used).
Use this tag with caution.
It is assumed that when this tag is used some other external process
may be monitoring the existence of the modem object in DBus as exposed
by ModemManager, and if it does not appear for any reason then the
modem would be reseted with some other mechanism (e.g. GPIOs, if
available). If no such mechanism to autorecover the modem is in place,
using this tag may leave the modem exposed in the kernel but ignored
by ModemManager.
This tag must be applied on the specific port for which the existence
and usability must be ensured.
E.g. flagging the MBIM port of the Fibocom L850 module as required:
$ vim /lib/udev/rules.d/78-mm-test.rules
ACTION!="add|change|move|bind", GOTO="mm_test_rules_end"
SUBSYSTEMS=="usb", ATTRS{bInterfaceNumber}=="?*", ENV{.MM_USBIFNUM}="$attr{bInterfaceNumber}"
ATTRS{idVendor}=="2cb7", ATTRS{idProduct}=="0007", ENV{.MM_USBIFNUM}=="00", ENV{ID_MM_REQUIRED}="1"
LABEL="mm_test_rules_end"
$ sudo udevadm control --reload
$ sudo udevadm trigger
$ sudo udevadm info -p /sys/class/usbmisc/cdc-wdm0
...
E: ID_MM_REQUIRED=1
E: ID_MM_CANDIDATE=1
To make it clearer that the initial list of flags must be the one
based on which ones are expected in the subsystem and which one the
plugin is requesting.
We can remove the port type hint udev tags for the wwan subsystem, as
that logic is now incorporated in the port type hint processing logic
in the daemon itself.
This makes it clearer to know what exact hints are being used, as the
logic is in a single place and it has proper logging of all possible
cases.
Do not do this in the plugin; instead, do it along with the logic that
looks for port type hints in udev, so that we can properly warn if
things don't match.
Now that we've dealt with everything that has been deprecated since
glib-2.44 until glib-2.56 (which we check), we can enable warnings that
guard against using the deprecated constructs.
We cancel AT commands (or sequences) on cancellation of the modem's
cancellable as well as of the optional user-provided one.
However, the GTask can be associated with only one. That's okay if the
user didn't supply one -- we just use the modem's cancellable only.
Otherwise we chain the cancellables together (also cancel the user one
with the modem one) at cost of some extra complexity.
This aims to make the above a little clearer by using hopefully
better names.
Suggested-by: Aleksander Morgado <aleksandermj@chromium.org>
No users left for the getter that takes a reference after a port of
MMBaseModemAT to GTask.
peek_cancellable() is used instead, because the GTask constructor
takes a reference itself.
