This patch makes declarations bind to definitions within the same module
to prevent the potential ambiguity if referenced directly.
AddressSanitizer think they violated one definition rule, although
those symbols are accessed by address through their modules and do
not depend on the order of the libararies loaded.
Otherwise we may leave a bearer connected when ModemManager doesn't
think it's connected. Prevents a CME ERROR 277 loop on connect when
the bearer hasn't been torn down correctly.
Wait for either an E2NAP unsolicited disconnect status or (for older
devices) an ENAP poll response before completing the disconnect.
Otherwise the client may start connecting again (such as
NetworkManager autoconnect retry) and the unsolicited E2NAP may
abort it, or the modem may return CME ERROR 277 ("not disconnected
yet") for the next connection attempt.
https://bugs.freedesktop.org/attachment.cgi?id=123525
There are a few key parts to this patch:
1) move the poll id into the Dial3gppContext structure as it is
conceptually part of the connection context data. This simplifies
context cleanup by keeping the poll id cleanup in one place.
2) move unsolicited connection status (E2NAP) handling into the
normal connection codepath, instead of completing the connection
context from the report_connection_status() E2NAP handler. This
simplifies connect code by not requiring checks for a NULL context
everywhere, and allows us to pass the Dial3gppContext structure
around instead of the Bearer object, since the Dial3gppContext
will never be comleted/freed outside the normal connection flow
codepaths like GLib idles and AT command requests.
3) use the connect context cancellable for all AT command requests
in the connect path. This lets us use the error return from
mm_base_modem_at_command_full_finish() to handle cancellation and
also not bother listening to the 'cancelled' signal of the
cancellable, since we can just check for cancellation the next time
the ENAP poll function runs.
https://bugs.freedesktop.org/show_bug.cgi?id=95304
If the modem thinks a PDP context is already active it'll return
583 errors from IPDPCFG and IPDPACT until the context is
deactivated. Deactivation was previously done after authentication,
but needs to be done before any part of the connect process to
ensure the PDP context is inactive.
The previous approach worked only if the context was being
deactivated already (which can take a bit of time) because it would
be deactivated after a few seconds and the connect could continue.
This approach works for more cases (like a MM crash and restart
while the modem is connected).
v4: modems/providers may not return DNS servers and not all modems
support DHCP, so lack of DNS servers should not indicate a bearer
IP method of "DHCP". IP config daemon/scripts already have to handle
missing DNS anyway.
v6: IPv6 requires SLAAC or DHCPv6 as part of the specification, so for
now we assume modems will support it. Provide all the info the modem
sent, but if there is any missing information use an IP method of
"DHCP" to indicate that info should be obtained via SLAAC/DHCPv6. Only
use an IP method of "STATIC" when all basic properties are given by
the modem.
The "0" case wasn't being handled properly:
$ sudo mmcli -m 0 --create-bearer="apn=,allow-roaming=0"
Error parsing properties string: 'Cannot get boolean from string '0''
When we were completing tasks in idle, the logic was like this:
* Schedule task completion in idle
* self->priv->task = NULL
* (idle) Task completion callback called
This meant that the self->priv->task was always set to NULL before the
completion callback was called, which is what we wanted as a new task may be
scheduled in the callback itself.
Now, without completing in idle, we were wrongly doing:
* Task completion callback called
* self->priv->task = NULL
This commit fixes the logic by making sure self->priv->task = NULL before any
task completion.
Initialized "cgmi_retries" variable from CustomInitContext with the
same value as the other retries, moreover the context is now allocated
with g_slice_new0.
Before this changes, when cgmi_retries assumed big values during the
probing of no AT-capable ports, the command AT+CGMI (mm-plugin-dell.c:custom_init_step)
was sent a semi-infinite number of times, blocking the plugin's initialization.
The options that require an argument should explicitly specify so.
Before:
--log-level=INFO Log level: one of [ERR, WARN, INFO, DEBUG]
--log-file Path to log file
After:
--log-level=[LEVEL] Log level: one of ERR, WARN, INFO, DEBUG
--log-file=[PATH] Path to log file
mm-modem-oma.c:400:51: error: implicit conversion from enumeration type 'MMOmaSessionType' to different enumeration type 'MMOmaFeature'
[-Werror,-Wenum-conversion]
g_return_val_if_fail (MM_IS_MODEM_OMA (self), MM_OMA_SESSION_TYPE_UNKNOWN);
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~
/usr/include/glib-2.0/glib/gmessages.h:382:11: note: expanded from macro 'g_return_val_if_fail'
return (val); \
~~~~~~ ^~~
This should be MMOmaFeature, not MMOmaSessionType. The end result is the same,
as both are 0.
