Clone the connection upon activation. This makes it safe for the user
to modify the original connection while it is activated.
This involves several changes:
- NMActiveConnection gets @settings_connection and @applied_connection.
To support add-and-activate, we constructing a NMActiveConnection with
no connection set. Previously, we would set the "connection" field to
a temporary NMConnection. Now NMManager piggybacks this temporary
connection as object-data (TAG_ACTIVE_CONNETION_ADD_AND_ACTIVATE).
- get rid of the functions nm_active_connection_get_connection_type()
and nm_active_connection_get_connection_uuid(). From their names
it is unclear whether this returns the settings or applied connection.
The (few) callers should figure that out themselves.
- rename nm_active_connection_get_id() to
nm_active_connection_get_settings_connection_id(). This function
is only used internally for logging.
- dispatcher calls now get two connections as well. The
applied-connection is used for the connection data, while
the settings-connection is used for the connection path.
- needs special handling for properties that apply immediately
when changed (nm_device_reapply_settings_immediately()).
Co-Authored-By: Thomas Haller <thaller@redhat.com>
https://bugzilla.gnome.org/show_bug.cgi?id=724041
NMSecretAgent (and in turn NMAgentManager) used the @connection argument both
for the connection data, but also for the connection path. Detangle these, and
accept the path separate from the connection.
This makes NMSecretAgent and NMAgentManager truly operate on a plain
NMConnection, without the non-obvious requirement, that the path of the
connection must be set.
Refactor agent-manager to always invoke the complete function for
nm_agent_manager_get_secrets().
In general, the complete function is always invoked asnychronously
when starting the operation. On the other hand, when cancelling the
operation or disposing the manager with pending operations, we now
(always) synchronously invoke the callback.
This makes it simpler for the user to reliably cancel the request
and perform potential cleanup.
This behavior bubbles up through NMSettingsConnection and NMActRequest,
and other callers that make directly or indicrectly make use of
nm_agent_manager_get_secrets().
Instead of having the call_id of type guint32, make it an (opaque)
pointer type.
This has the advantage of strong typing and avoids the possiblity
of reusing an invalid integer (or overflow of the call-id counter).
OTOH, it has the disadvantage, that after a call_id is disposed,
it might be reused for future invocations (because malloc might
reuse the memory).
In fact, it is always an error to use a call_id that is already
completed. This commit also adds assertions to the cancel() calls
that the provided call_id is a pending call. Hence, such a bug
will be uncovered by assertions (that only might not tigger in
certain unlikely cases where a call-id got reused).
Note that for NMAgentManager, save_secrets() and delete_secrets()
both returned a call_id. But they didn't also provide a callback when
the operation completes. So the user trying to cancel such a call,
cannot know whether the operation is still in process and he cannot
avoid triggering an assertion.
Fix that by not returning a call-id for these operations. No caller
cared about it anyway.
For NMSettingsConnection, also track the internally scheduled requests
for so that we can cancel them on dispose.
The GATEWAY from /etc/sysconfig/network file is used as a default value when
no GATEWAY is in ifcfg file. However, we have to ignore that GATEWAY for
connections without static addresses. Otherwise such connections would be
invalid and would disappear after restart/reaload.
Some notes:
Putting GATEWAY into /etc/sysconfig/network is not recommended, because it
inherently belongs to the ifcfg file as it is a per-interface property.
The recommended practice is to specify GATEWAY in individual ifcfg files and
define DEFROUTE=no if the interface should not get the default route.
But we continue to read GATEWAY from /etc/sysconfig/network for compatibility
reasons.
See also
https://bugzilla.redhat.com/show_bug.cgi?id=896198#c25https://bugzilla.redhat.com/show_bug.cgi?id=896198#c27
Fixes: f17699f4e3https://bugzilla.redhat.com/show_bug.cgi?id=1262972
A GObject interface, like a class, has two different C types
associated with it; the type of the "class" struct (eg, GObjectClass,
GFileIface), and the type of instances of that class/interface (eg,
GObject, GFile).
NetworkManager was doing this wrong though, and using the same C type
to point to both the interface's class struct and to instances of the
interface. This ends up not actually breaking anything, since for
interface types, the instance type is a non-dereferenceable dummy type
anyway. But it's wrong, since if, eg, NMDeviceFactory is a struct type
containing members "start", "device_added", etc, then you should not
be using an NMDeviceFactory* to point to an object that does not
contain those members.
Fix this by splitting NMDeviceFactory into NMDeviceFactoryInterface
and NMDeviceFactory; by splitting NMConnectionProvider into
NMConnectionProviderInterface and NMConnectionProvider; and by
splitting NMSettingsPlugin into NMSettingsPluginInterface and
NMSettingsPlugin; and then use the right types in the right places.
As a bonus, this also lets us now use G_DEFINE_INTERFACE.
Since there have not been separate system and user settings services
since 0.8, the "system" in NMSystemConfigInterface is kind of
meaningless. Rename it to NMSettingsPlugin, which describes what it
does better.
This is just:
git mv src/settings/nm-system-config-interface.h src/settings/nm-settings-plugin.h
git mv src/settings/nm-system-config-interface.c src/settings/nm-settings-plugin.c
perl -pi -e 's/SystemConfigInterface/SettingsPlugin/g;' \
-e 's/system_config_interface/settings_plugin/g;' \
-e 's/system-config-interface/settings-plugin/g;' \
-e 's/SYSTEM_CONFIG_INTERFACE/SETTINGS_PLUGIN/g;' \
-e 's/sc_plugin/settings_plugin/g;' \
-e 's/SC_PLUGIN/SETTINGS_PLUGIN/g;' \
-e 's/SC_IS_PLUGIN/SETTINGS_IS_PLUGIN/g;' \
-e 's/SC_TYPE_PLUGIN/SETTINGS_TYPE_PLUGIN/g;' \
-e 's/SCPlugin/SettingsPlugin/g;' \
-e 's/nm_system_config_factory/nm_settings_plugin_factory/g;' \
$(find src/settings -type f)
(followed by some whitespace fixups in nm-settings-plugin.c, and a
Makefile.am fix for the rename)
The signal "notify:g-name-owner" is only emitted for well-known
names, but not for unique connection names (":1.x") such as the secret
agent's connection. Also, it will not be emited for private connections.
That meant that when the client disconnected without explicitly
unregistering, the NMSecretAgent was not cleaned up and leaked
indefinitely.
As only one instance of secret agent with a certain 'identifier/uid'
pair can register, this bug also prevented the client from registering
until restart of NetworkManager.
Fixes: df6706813a
This code was unused, because we never enqueued any hashes
to the @asked list. Note that hashing also might give wrong
hash collisions, so this was buggy anyway.
Also, note that impl_agent_manager_register_with_capabilities()
already ensures that duplicate agents are not registered
in the first place (find_agent_by_identifier_and_uid()).
Refactor the handling of the asynchronous requests so that now
NMSecretAgent has the following properties:
- The callback will *always* be invoked exactly once (sans crashes).
Even if you cancel the call or if you dispose NMSecretAgent with
pending calls. That allows the caller to rely on being called back
and possibly cleanup the user-data.
- Callbacks are always invoked asynchronously with respect to their
start-call.
- You can cancel all 3 types of operations, not only the 'GetSecrets'
call. Note that this will still not cancel the calls 'DeleteSecrets'
and 'SaveSecrets' on a D-Bus level.
When cancelling, the callback will be invoked synchronously with
respect to the cancel call, with an GError indicating the cancellation
(G_IO_ERROR_CANCELLED).
- During dispose, the callback is also invoked synchronously, with
some other error reason.
This also fixes a crash where handling of the asynchronous data was
messed up and the priv->requests hash would end up to containing dangling
pointers.
https://bugzilla.redhat.com/show_bug.cgi?id=1253407
The @dbus_owner field was only cleaned up when the
proxy disconnected and leaked otherwise.
Also, don't clear @dbus_owner together with the proxy.
Otherwise, get_description() might yield different results
after the proxy got cleared. That can lead to problems because
NMAgentManager tracks the secrets agents by their "dbus-owner" --
IOW, NMAgentManager uses the "dbus-owner" as identifer for the
secret agent. Thus it must not change.
Fixes: 2a2fd1216b
The logging macros _LOGD(), etc. are specific to each
file as they format the message according to their context.
Still, they were cumbersome to define and their implementation
was repeated over and over (slightly different at times).
Move the declaration of these macros to "nm-logging.h".
The source file now only needs to define _NMLOG(), and either
_NMLOG_ENABLED() or _NMLOG_DOMAIN.
This reduces code duplication and encourages a common implementation
and usage of these macros.
Program received signal SIGSEGV, Segmentation fault.
g_type_check_instance_cast (type_instance=type_instance@entry=0x89f180, iface_type=9004512) at gtype.c:4060
4060 node = lookup_type_node_I (type_instance->g_class->g_type);
(gdb) bt
#0 0x00007ffff4b44e80 in g_type_check_instance_cast (type_instance=type_instance@entry=0x89f180, iface_type=9004512) at gtype.c:4060
#1 0x000000000056a460 in connection_visibility_changed (connection=0x89f680 [NMKeyfileConnection], pspec=<optimized out>, user_data=0x89f180) at settings/nm-settings.c:870
#5 0x00007ffff4b3b54f in <emit signal notify:visible on instance 0x89f680 [NMKeyfileConnection]> (instance=instance@entry=0x89f680, signal_id=<optimized out>, detail=<optimized out>) at gsignal.c:3393
#2 0x00007ffff4b200b5 in g_closure_invoke (closure=0x9131a0, return_value=return_value@entry=0x0, n_param_values=2, param_values=param_values@entry=0x7fffffffd540, invocation_hint=invocation_hint@entry=0x7fffffffd4c0) at gclosure.c:801
#3 0x00007ffff4b32499 in signal_emit_unlocked_R (node=node@entry=0x8696b0, detail=detail@entry=641, instance=instance@entry=0x89f680, emission_return=emission_return@entry=0x0, instance_and_params=instance_and_params@entry=0x7fffffffd540) at gsignal.c:3581
#4 0x00007ffff4b3b1a0 in g_signal_emit_valist (instance=<optimized out>, signal_id=<optimized out>, detail=<optimized out>, var_args=var_args@entry=0x7fffffffd710) at gsignal.c:3337
#6 0x00007ffff4b24665 in g_object_dispatch_properties_changed (object=0x89f680 [NMKeyfileConnection], n_pspecs=<optimized out>, pspecs=<optimized out>) at gobject.c:1056
#7 0x00007ffff4b26d11 in g_object_notify (pspec=0x8ce660 [GParamBoolean], object=0x89f680 [NMKeyfileConnection]) at gobject.c:1149
#8 0x00007ffff4b26d11 in g_object_notify (object=0x89f680 [NMKeyfileConnection], property_name=property_name@entry=0x5d2eb9 "visible") at gobject.c:1197
#9 0x0000000000497f85 in set_visible (self=self@entry=0x89f680 [NMKeyfileConnection], new_visible=new_visible@entry=0) at settings/nm-settings-connection.c:296
#10 0x0000000000498165 in dispose (object=0x89f680 [NMKeyfileConnection]) at settings/nm-settings-connection.c:2390
#11 0x00007ffff4b24fec in g_object_unref (_object=0x89f680) at gobject.c:3137
#12 0x00000000004a4a4f in dispose (object=0xa24260 [NMVpnConnection]) at nm-active-connection.c:904
#13 0x00007ffff4b24fec in g_object_unref (_object=0xa24260) at gobject.c:3137
#14 0x0000000000577636 in nm_vpn_service_stop_connections (service=0x8ff610 [NMVpnService], quitting=1, reason=NM_VPN_CONNECTION_STATE_REASON_SERVICE_STOPPED) at vpn-manager/nm-vpn-service.c:150
#15 0x0000000000576ea2 in dispose (object=0x921060 [NMVpnManager]) at vpn-manager/nm-vpn-manager.c:284
#16 0x00007ffff4b24fec in g_object_unref (_object=0x921060) at gobject.c:3137
#17 0x00000000004d0f05 in dispose (object=0x88a2b0 [NMManager]) at nm-manager.c:5061
#18 0x00007ffff4b24fec in g_object_unref (_object=0x88a2b0) at gobject.c:3137
#19 0x0000000000444e08 in _nm_singleton_instance_destroy () at NetworkManagerUtils.c:138
#20 0x00007ffff7de97b7 in _dl_fini () at dl-fini.c:252
#21 0x00007ffff4444778 in __run_exit_handlers (status=status@entry=0, listp=0x7ffff47d0618 <__exit_funcs>, run_list_atexit=run_list_atexit@entry=true) at exit.c:82
#22 0x00007ffff44447c5 in __GI_exit (status=status@entry=0) at exit.c:104
#23 0x0000000000445b80 in main (argc=1, argv=0x7fffffffdf08) at main.c:458
(gdb)
Advantages:
- use current best-pratice
- registers a weak-ref to clear @singleton_instance when the
instance gets destroyed
- logs creation and destruction of singleton
- on shutdown, destroy the singleton instance via
_nm_singleton_instance_register_destruction(). Note, that
we now have yet another reference to the singleton that is
owned by register-destruction.
We already avoid loading duplicate plugins by checking find_plugin().
That iterates the plugins @list and checks for duplicate names.
Additionally, also reject duplicates based on the @plugins list.
Also, move the check for "keyfile" before, so that all explicit
checks for (statically) known names are early and together.
There is no need to have a static @singleton variable.
The only caller of nm_settings_keyfile_plugin_new() is
NMSettings which owns the singleton instance.
A *_new() function should just create a new instance and
that's it. It's unexpected to reuse the same instance.
Some functions from nm-session-monitor.c have an implicit access to
nm_session_monitor_get(). This is non-obvious behavior.
Instead require the explicit session-monitor instance to be
provided -- where needed.
Move the check for a platform link before devtimeout_from_file(). The
check in the platform cache should be more performant and yield success
in most cases.
This can save reading and parsing the ifcfg-rh file.
If NIC related initialization takes a long time in udev processing, but we have
an ifindex from kernel, we still want to wait until udev is finished and the
device is really usable.
Check that by calling nm_platform_link_get_by_ifname() and checking @initialized,
which means udev is finished.
Based on a patch by t-nishimura@hf.jp.nec.com
Improved by thaller@redhat.comhttps://bugzilla.redhat.com/show_bug.cgi?id=1192633
Port remaining bits to gdbus and remove stray dbus-glib references
Drop the dbus-glib version check from configure, since nothing depends
on new dbus-glib any more.
Move nm-dbus-glib-types.h and nm-gvaluearray-compat.h from include/ to
libnm-util/ since they are now only used by libnm-util and libnm-glib.
