- Jan 24, 2025
-
-
Dan Williams authored
mbim-network: fix matching ActivationState values (issue #54) !222 Closes #54
-
"connected" is not a valid value for ActivationState; let's check "activated" instead. Found by Charlie King Fixes: #54 Signed-off-by:
Dan Williams <dan@ioncontrol.co>
-
Dan Williams authored
mbim-proxy: don't unlink abstract socket paths !221
-
- Jan 23, 2025
-
-
Dan Williams authored
Proxy sockets are abstract and thus don't exist on the filesystem. So there's nothing to clean up, and the g_unlink() just ends up deleting the qmi-proxy binary if you run it from the same directory where the binary lives. Found by Kasper Cheng <kaspercheng77@gmail.com> Signed-off-by:
Dan Williams <dan@ioncontrol.co>
-
- Nov 30, 2024
-
-
Aleksander Morgado authored
Just a rename, we keep '--query-registration-state' as a valid but hidden action. This is to be in line with the MBIM message name, which is "Register State".
-
Aleksander Morgado authored
The new '--set-register-state' action allows configuring automatic or manual registration, also specifying the MCCMNC to bind to and optionally the list of data modes to prefer. The '--register-automatic' action is considered deprecated, and can be replaced with '--set-register-state="action=automatic". Automatic registration setup: $ sudo mbimcli -d /dev/cdc-wdm1 -p --set-register-state=' action=automatic' Manual registration to a specific operator: $ sudo mbimcli -d /dev/cdc-wdm1 -p --set-register-state=' action=manual, provider-id=21407' Manual registration to a specific operator on limited access techs: $ sudo mbimcli -d /dev/cdc-wdm1 -p --set-register-state=' action=manual, provider-id=21407, data-class=lte|umts'
-
- Nov 29, 2024
-
-
Aleksander Morgado authored
-
Aleksander Morgado authored
Print them as empty fields ("") instead.
-
Aleksander Morgado authored
-
Aleksander Morgado authored
This allows modifying attach settings directly with mbimcli. $ sudo mbimcli -d /dev/cdc-wdm0 -p --ms-set-lte-attach-configuration=" default, ip-type=default,source=user,roaming-control=home,access-string=internet/ ip-type=default,source=user,roaming-control=partner,access-string=internet/ ip-type=default,source=user,roaming-control=non-partner,access-string=internet" [/dev/cdc-wdm0] LTE attach configuration available Configuration 0: IP type: default Roaming: home Source: user Access string: internet Username: n/a Password: n/a Compression: none Auth protocol: none Configuration 1: IP type: default Roaming: partner Source: user Access string: internet Username: n/a Password: n/a Compression: none Auth protocol: none Configuration 2: IP type: default Roaming: non-partner Source: user Access string: internet Username: n/a Password: n/a Compression: none Auth protocol: none
-
- Nov 27, 2024
-
-
Aleksander Morgado authored
-
Aleksander Morgado authored
Just to be in line with the C example.
-
Aleksander Morgado authored
-
- Aug 21, 2024
-
-
Aleksander Morgado authored
-
- Aug 06, 2024
-
-
Daniele Palmas authored
Errors definition available at: https://learn.microsoft.com/en-us/windows-hardware/drivers/network/mb-low-level-uicc-access
-
- Aug 04, 2024
-
-
Dan Williams authored
Some functions may want to use the enum helpers for input validation but handle printing the error themselves. Instead of calling g_printerr() directly from the enum helpers, pass the error back to the caller. Signed-off-by:
Dan Williams <dan@bigw.org>
-
- Jul 26, 2024
-
-
Aleksander Morgado authored
-
- Jun 19, 2024
-
-
Mank Wang authored
-
- Jun 17, 2024
-
- Jun 14, 2024
-
-
Aleksander Morgado authored
-
- Jun 11, 2024
-
-
- Jun 05, 2024
-
-
Mank Wang authored
Signed-off-by:
mank.wang <mank.wang@quectel.com>
-
- May 27, 2024
-
-
Lubomir Rintel authored
These were autofree'd on function return, but can actually be assigned new allocations on each inner loop interation. Move them inside the loop.
-
- May 23, 2024
-
-
Lubomir Rintel authored
Don't conditionalize setting trace_command and trace_value on split substrings being non-NULL. It makes a static analysis tool think they might be used uninitialized. That can-not happen, because the substrings can in fact never be NULL. Let's keep the check in form of an assert (perhaps to guard against a possible glib bug, etc.).
-
- May 13, 2024
-
-
Aleksander Morgado authored
-
Aleksander Morgado authored
-
- Apr 22, 2024
-
-
Aleksander Morgado authored
-
- Apr 08, 2024
-
-
At the MBIM service level we don't know if the embedded AT operation contains any personal info or not, so just assume it does.
-
At the MBIM service level we don't know if the embedded QMI message contains any personal info or not, so just assume it does.
-
- Apr 07, 2024
-
-
- Mar 28, 2024
-
-
Slark Xiao authored
When calling mbim-proxy with parameter '-v', it will report an error "cannot specify --verbose and --verbose-full at the same time". This issue was caused by the same shorName of --verbose and --verbose-full. Remove the shortName of '--verbose-full' align with mbimcli. Fixes #b1ca8388
-
- Mar 18, 2024
-
-
Yegor Yefremov authored
Enable insert_final_newline to comply with the current style.
-
- Feb 12, 2024
-
-
Yegor Yefremov authored
EditorConfig projects helps maintaining consistent coding styles across various editors and IDEs.
-
- Dec 13, 2023
-
-
Aleksander Morgado authored
-
- Dec 04, 2023
-
-
If the proxy is checking whether a given device is really open or not, it sends a caps query request, which may be replied with a FUNCTION_ERROR stating that the device is not open. If this happens, the proxy will explicitly force close the device, and try to reopen it again. The MBIM_DEVICE_SIGNAL_ERROR is always scheduled to be emitted after the response processor has been executed. In the case of the proxy, the response processor force-closes the MBIM device, which in turn would cancel any scheduled MBIM_DEVICE_SIGNAL_ERROR emission. There is really no need to block MBIM_DEVICE_SIGNAL_ERROR any more due to this.
-
The SIGNAL_ERROR signal is supposed to be emitted whenever we receive a FUNCTION_ERROR response. The signal is emitted *after* the message response processor has been executed, and this means that there is a chance that the response processor may have already closed the MBIM device. If this happens, we should not emit the SIGNAL_ERROR signal, as the user has already expressed their desire to fully close the ongoing MBIM session. This avoids the SIGNAL_ERROR signal being processed after the user has already requested a new MBIM device open following a forced close. Issue found by Maciej S. Szmigiero <mail@maciej.szmigiero.name>, and this patch implemented as an alternative to the one he wrote in !193 Fixes #43
-
Otherwise mbim_device_open -> device_open_context_step (DEVICE_OPEN_CONTEXT_STEP_FIRST) will refuse re-opening as it will think the device is already open. This refusal broke device re-opening on MBIM_PROTOCOL_ERROR_NOT_OPENED. Fixes: 0a119a7a ("libmbim-glib,device: ensure CLOSED state always in mbim_device_close()")
-
Aleksander Morgado authored
-
This adds a new group for sms functionality with two new functions: - --sms-delete - --sms-read
-