Will be used to report operations that are not run as they have been
attempted too many times already.
Modems will usually report this kind of error e.g. when attempting to
connect with the same invalid APN too many times. The modem at some
point decides that it won't attempt the APN connection with the
network any more, so it returns this error to the caller right away.
After the modem recovers, the ports are up but not yet responsive. When MM tries to probe the ports, they do not respond.
As a result, the probing tasks are left ongoing, waiting for a response from the modem.
If the modem drops off the bus for any reason, the cdc-wdm port probing task is not immediately canceled.
This patch fixes this issue by passing a cancellable object to mm_port_mbim_open().
In other words, the patch ensures that the cdc-wdm port probing task is canceled if the
modem drops off the bus before the task completes. This prevents the task from preventing
other port probing tasks from starting.
After LVFS upgrade, the module needs to be restarted. ModemManager
queries the module information, sometimes the module is not ready, and
the information queried by "mmcli -m any --firmware-status" is
incomplete,
It is now modified to re query once per second when obtaining
incomplete versions, up to a maximum of 16 times.
Signed-off-by: mank.wang <mank.wang@quectel.com>
When the subscriber ready status notification's ready state becomes "initialized"
from "nonInitialized", it indicates a new SIM is inserted, for which a modem
reprobe is needed.
Add PROCESS_NOTIFICATION_FLAG_SLOT_INFO_STATUS to enable and setup
flags, this should help SIM detection on inactive slots when the
modem is in a failed state.
This helps fix the following scenario, If a physical SIM card (pSIM) is
inserted into a device that is currently using its eSIM slot,
but does not have any eSIM profiles installed, the pSIM will not be detected.
The ID_MM_PROCESS tag should be set by the user when using the
allowlist filter, we must not ship any device flagged by default with
that tag.
Moreover, the two ENV actions in a single rule is something not
currently supported by the builtin udev rule parser that is used in
systems without udev (e.g. openwrt).
Fixes https://gitlab.freedesktop.org/mobile-broadband/ModemManager/-/issues/779
RM520N in USB mode exposes the following:
ttyUSB0 (if #0): QCDM/DIAG port
ttyUSB1 (if #1): GPS data port
ttyUSB2 (if #2): AT primary port
ttyUSB3 (if #3): AT secondary port
So, lets mark them so ModemManager can properly manage them.
Signed-off-by: Robert Marko <robert.marko@sartura.hr>
Adding a new force option to force send initial EPS bearer settings to
modem even if they match the existing ones. Currently settings are
not sent to the modem if they match old settings
This improves SIM hot swap check process, and it can help avoid unnecessary
modem reprobe during suspend-resume without SIM inserted.
The changes in this CL implemented the new logic ONLY for MBIM, and the
new logic has not been implemented for QMI and AT yet. Thus, we keep the
legacy `mm_iface_modem_check_for_sim_swap()' till the new logic for QMI
and AT has been implemented.
This fixes a bug where network-initiated USSD-Notify messages were not
processed (and forwarded to higher-level applications) at all, while
they should have been processed (and forwareded to higher-level
applications). Previously, when there was a USSD conversation between
the user and the network, the last message from the network (which, by
definition, is not a message with
user_action==QMI_VOICE_USER_ACTION_REQUIRED) would be silently
dropped.
The cause is a mismatch between the modem sending
user_action==QMI_VOICE_USER_ACTION_UNKNOWN while old ModemManager
expecting user_action==QMI_VOICE_USER_ACTION_NOT_REQUIRED.
iface-modem does not include config.h so it doesnt actually have the meson
generated build defines to check against.
I found this while trying to use AT commands over D-Bus by enabling
at_command_via_dbus, confirming that Meson sees it, defines it in config.h
but then I still got the:
"Cannot send AT command to modem: operation only allowed in debug mode".
Signed-off-by: Robert Marko <robert.marko@sartura.hr>
As we do in the similar modem state wait operations in the modem
interface, the caller of the wait operation may ignore the state
result as it's really not required in this case. Only looking at
whether an error is set or not is enough.
This by itself fixes a crash where
mm_iface_modem_3gpp_wait_for_packet_service_state_finish() may have
returned UNKNOWN without error set.
0x000079616c2564ed (libglib-2.0.so.0 - gtestutils.c: 3256) g_assertion_message
0x000079616c25654b (libglib-2.0.so.0 - gtestutils.c: 3282) g_assertion_message_expr
0x000057ba46231150 (ModemManager - mm-iface-modem-simple.c: 344) packet_service_attach_in_3gpp_network_step
0x000079616c0efb02 (libgio-2.0.so.0 - gtask.c: 1232) g_task_return_now
0x000079616c0eeb54 (libgio-2.0.so.0 - gtask.c: 1301) g_task_return
0x000057ba4620e5bd (ModemManager - mm-iface-modem-3gpp.c: 189) wait_for_packet_service_state_context_complete
0x000057ba4620b441 (ModemManager - mm-iface-modem-3gpp.c: 261) packet_service_state_changed
With GSM7 encoding, packedlen is the length of the unpacked string
after expanding septets to octets so it will be ~14% bigger than
the original string length. This means we have to be careful not
to copy too much data into the PDU buffer.
Similar issues exist in other branches of the same function.
Thanks rhezashan@gmail.com for the report.
Kernel driver mhi-net sets a custom MTU for the netdevice: use the
same value when resetting the netdevice to its initial state.
This is important especially when rmnet is used on top of mhi-net,
since rmnet netdevices get their initial MTU from the mhi-net one.
Aleksander Morgado
Madhav
mank.wang
Rukun Mao
Puliang Lu
Robert Marko
Ujjwal Pande
ming-pei.hsu
okaestne
Lukas Voegl
Daniel Hofmann
Dennis Gilmore
Nagi Marupaka
QuectelDuke
Eric Caruso
Daniele Palmas