Long ago there were problems where certain Huawei devices would
stop responding on various ports, and sometimes would crash
randomly. The theory at the time was that touching the secondary
ports made the device angry, thus the plugin simply opened the
ports and listened for unsolicited messages. But if the device
didn't send any during that 7 second period, MM would not detect
and secondary ports at all. Plus, it was always a hack.
Instead, the new theory is that the device crashes if unsolicited
messages are enabled (^CURC=1), the secondary port gets touched,
*and* then closed and left for a while. Fix that by turning
unsolicited messages off at probe time, on when the device is
enabled, and off again when the device is disabled like happens
for other modems. Thus when MM first detects the modem, it turns
off unsolicited messages and the serial buffer on the secondary
port doesn't fill up and crash the modem.
Second, this allows us to simplify the probing logic quite a bit
so that we can probe all ports we find, but we still wait to probe
the first port so we can turn off unsolicited messages and get
hints about what port is the secondary.
Instead of having two places that custom init stuff got processed
(a hook in the MMPluginBase class itself and a callback too) just
use a callback, and simplify it somewhat so that the plugin tracks
how many tries it cares about and what to do based on the response
or error.
Port the differences over to the common Icera code (there were only
two) and remove the duplicate code from the Samsung plugin. The
Icera NWSTATE regex had to be adjusted to capture "-1" in the first
element which wasn't handled before but which I've seen on the
Samsung Y3300 before the card has registered.
The GInterface structure for MMModemIcera isn't instance data, thus we
shouldn't be storing an instance pointer in it. Instead, make implemtors
store the intstance data in their private structure, and have them
implement an accessor for the Icera-private data. This makes everone
(especially GObject) happy. It's a bit of additional indirection, but
we still get to use the MM_MODEM_ICERA_GET_PRIVATE() and we still
get to cast the passed-in GInterface MMModemIcera into the various
GSM MMModem subclasses, which is all we ever wanted anyway.
Make sure all pending work, if any, is torn down when closing
the port, since it's closed, and is likely going to get unreffed
soon and we don't want anything running after the port is dead.
If the modem for some reason returns ERROR for a +CPIN when it's
unlocked, and subsequent PIN post-unlock pin checks return READY,
just treat the modem as unlocked don't return the +CPIN error.
Change interface to take the number of GSM characters
rather than the number of octets, so that it is possible to
distinguish the 7-character and 8-character cases.
It's only used during probing where some port types (Sierra CnS
and some Icera devices) may send streams of data that we can't
understand until we close the port. It interferes with some SMS
operations, so turn it off for ports opened after probing.
For 3G devices in Class A, AT+WWSM can be used to get or set the allowed mode:
* +WWSM: 0 (2G only)
* +WWSM: 1 (3G only)
* +WWSM: 2,0 (Any)
* +WWSM: 2,1 (2G preferred)
* +WWSM: 2,2 (3G preferred)
For 2G devices, there is no such command, so we will default to Any and allow
2G-only and 2G-preferred setups.
These modems can be configured to use different mobile classes. For each kind of
modem, the best mobile class is the highest one in the following order:
- Class A (3G only mode)
- Class B (PS or CS, GPRS/EDGE or GSM)
- Class CG (PS only, GPRS/EDGE)
- Class CC (CS only, GSM)
BUG=chromeos-partner:2999, chromeos-partner:3215
TEST=network_LockedSIM
Change-Id: I17c25c52fa5cf4cffa94e73bd827eaae9e687df0
Signed-off-by: Elly Jones <ellyjones@chromium.org>
Signed-off-by: Jason Glasgow <jglasgow@chromium.org>
Signed-off-by: Jun Woo Lee <jw86.lee@samsung.com>
If a port returns EAGAIN on write attempts, previously the code would
spin and attempt to resend the failed byte after send_delay
microseconds. This resulted in up to 3 second hard blocks in
the serial code when sending to ports that don't respond. While
in this blocking loop no other events or dbus commands could be
processed.
Instead, send each byte and reschedule sending the next byte in
send_delay microseconds, so that we can process other events in between
attempts to write to stupid ports.
This doesn't hugely decrease the amount of time that probing
requires, since we still need to probe all ports of the device
before exporting the modem to D-Bus, but it does let MM find
responsive ports much more quickly, and ensures that MM doesn't
block any D-Bus requests.
Two issues here, first we dont' need to chain up to the parent
because it's not doing anything we need (it's mainly for PPP-based
devices) and second we need to wait a bit for the disconnect command
to complete by specifying a callback, otherwise the command may get
discarded when the port is shut down afterward.
Don't need init retries since the modem seems pretty sane, and we
also don't need to call AT+CFUN=1 twice. Just once should be
fine. We also don't need any "flashing" since the modem doesn't
really do PPP and thus shouldn't need any of the serial port
carrier stuff to get its attention, since we'll never be using
PPP on any of its ports.
