| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
This fixes segfaults resulting either from destroying a NULL ISI
client, or when multiple ISI modems become available, and get removed
in the wrong order. To support multiple ISI modems concurrently, a
modem object needs to be exposed via the gisi API, and tracked by the
ISI driver. This is a TODO item to add.
|
|
|
|
|
|
Refactor isimodem.c to add support for querying modem attributes over
the ISI protocol.
|
|
It should be included _first_ in all modules, and it is not proof
against multiple inclusions.
|
|
Adds support for using ttys through a serial device directly.
|
|
CLCC does not report clip_validity. So if clip_validity of witheld
has been reported in CLIP, but we're still somehow polling CLCC,
which happens if CRING is not supported, and RING is used instead,
this fact can be lost. Make sure we preserve clip_validity if
possible
|
|
According to 27.007, the number and number type are optional. We must
take care not to do anything unexpected in this case.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Add capability to read / write / stat files on the SIM. This now
supports cyclic, linear fixed and transparent SIM files.
Parse GET RESPONSE result to find structure of the file (cyclic, linear
fixed, or transparent), the file size and the record length.
Add both read and update capability for binary and record-based files.
Implement writing sim files through AT.
|
|
Also accept 0x9f as a success status word 1 for READ BINARY (GSM specific).
|
|
Use immediate mode reporting, which means that the individual CPBR
lines are reported up to the core immediately. This has a couple
advantages:
- We do not need to malloc/free a bunch of single variables
and copy them over. Helps performance.
- The lines are not buffered up and given to the plugin in one
shot, instead processing is performed piecemeal. This helps
with keeping memory consumption down to a minimum
|
|
Instead of using custom parsing, use the existing library
functionality in result iter open_list, next_range and close_list
|
|
|
|
|
|
|
|
|
|
Since the PDU line doesn't have a prefix, the parser can get confused
if other unsolicited notifications arrive. There is already explicit
support for unsolicited notifications with PDU, so we use that
instead
|
|
- Can receive 7Bit, 8Bit and UCS2 encoded messages
- 8Bit is assumed to be a datagram to a certain port
- Messages with Invalid format are ignored
- Concatenation not yet supported
|
|
|
|
- Breakup MCC/MNC LENGTH constant
- Don't hardcode numbers
- Fix >80 column length
- Fix test case to expect strings instead of shorts
|
|
This is to make sure both 2 and 3-digit MNC values are correctly
handled. Both the modem plugin API as well as the D-Bus API are
affected.
|
|
|
|
Rename encode_sms to sms_encode
Rename decode_sms to sms_decode
|
|
Handle Deliver / Status Report / Cell Broadcast messages and send the
appropriate acknowledgement PDUs (+CNMA) if required
|
|
|
|
|
|
|
|
|
|
Modem cleanup happens in two places: manager_free (ofono exit) and
manager_destroy (manually removing a specific modem). Both need
to unregister the entire set of registered interfaces.
manager_destroy function did not unregister all interfaces. Fixed
by adding a common function for both code paths.
|
|
|
|
|
|
|
|
|
|
|
|
The CallWaiting interface had one settable attribute after the latest
set of refactoring. Squash it onto the CallSettings interface where
it belonged in the first place
|
|
The use of a struct with status & cls seemed pointless when only
a single integer can work just as well.
Also the Fax & Data attributes have been squished. oFono does
not yet support those call types.
|
|
|
|
|
|
Future-Proof own numbers support. For now we only report voice numbers
Rename attributes to be a bit better
|
|
|
