2008-06-10 16:32:39 +00:00
|
|
|
/* -*- Mode: C; tab-width: 4; indent-tabs-mode: t; c-basic-offset: 4 -*- */
|
2004-06-24 14:18:37 +00:00
|
|
|
/* NetworkManager -- Network link manager
|
|
|
|
|
*
|
|
|
|
|
* This program is free software; you can redistribute it and/or modify
|
|
|
|
|
* it under the terms of the GNU General Public License as published by
|
|
|
|
|
* the Free Software Foundation; either version 2 of the License, or
|
|
|
|
|
* (at your option) any later version.
|
|
|
|
|
*
|
|
|
|
|
* This program is distributed in the hope that it will be useful,
|
|
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
|
|
* GNU General Public License for more details.
|
|
|
|
|
*
|
2008-06-26 18:31:52 +00:00
|
|
|
* You should have received a copy of the GNU General Public License along
|
|
|
|
|
* with this program; if not, write to the Free Software Foundation, Inc.,
|
|
|
|
|
* 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
|
2004-06-24 14:18:37 +00:00
|
|
|
*
|
2013-02-05 17:29:15 +01:00
|
|
|
* Copyright (C) 2004 - 2013 Red Hat, Inc.
|
2008-11-03 04:13:42 +00:00
|
|
|
* Copyright (C) 2007 - 2008 Novell, Inc.
|
2004-06-24 14:18:37 +00:00
|
|
|
*/
|
|
|
|
|
|
2016-02-19 14:57:48 +01:00
|
|
|
#include "nm-default.h"
|
2014-11-13 10:07:02 -05:00
|
|
|
|
2016-09-29 13:49:01 +02:00
|
|
|
#include "nm-policy.h"
|
|
|
|
|
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
#include <unistd.h>
|
|
|
|
|
#include <netdb.h>
|
2004-06-24 14:18:37 +00:00
|
|
|
|
|
|
|
|
#include "NetworkManagerUtils.h"
|
2016-09-28 19:37:10 +02:00
|
|
|
#include "nm-act-request.h"
|
2018-11-20 14:51:22 +01:00
|
|
|
#include "nm-keep-alive.h"
|
2016-11-21 00:43:52 +01:00
|
|
|
#include "devices/nm-device.h"
|
2008-03-13 20:30:39 +00:00
|
|
|
#include "nm-setting-ip4-config.h"
|
2007-11-07 16:06:43 +00:00
|
|
|
#include "nm-setting-connection.h"
|
2016-11-21 00:43:52 +01:00
|
|
|
#include "platform/nm-platform.h"
|
|
|
|
|
#include "dns/nm-dns-manager.h"
|
|
|
|
|
#include "vpn/nm-vpn-manager.h"
|
2014-08-16 01:33:46 +02:00
|
|
|
#include "nm-auth-utils.h"
|
2011-10-07 15:58:09 +02:00
|
|
|
#include "nm-firewall-manager.h"
|
2012-06-03 19:32:16 -05:00
|
|
|
#include "nm-dispatcher.h"
|
2012-08-21 17:49:41 +02:00
|
|
|
#include "nm-utils.h"
|
2014-08-26 15:49:18 +02:00
|
|
|
#include "nm-core-internal.h"
|
2014-07-17 17:06:44 -04:00
|
|
|
#include "nm-manager.h"
|
2016-11-21 00:43:52 +01:00
|
|
|
#include "settings/nm-settings.h"
|
|
|
|
|
#include "settings/nm-settings-connection.h"
|
2017-11-27 09:07:28 +01:00
|
|
|
#include "settings/nm-agent-manager.h"
|
2014-07-17 17:06:44 -04:00
|
|
|
#include "nm-dhcp4-config.h"
|
|
|
|
|
#include "nm-dhcp6-config.h"
|
2017-03-02 19:18:49 +01:00
|
|
|
#include "nm-config.h"
|
2017-04-17 20:17:45 +02:00
|
|
|
#include "nm-netns.h"
|
2017-04-23 00:40:46 +02:00
|
|
|
#include "nm-hostname-manager.h"
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
|
2016-09-29 13:49:01 +02:00
|
|
|
/*****************************************************************************/
|
2016-03-03 09:20:18 +01:00
|
|
|
|
2016-09-29 13:49:01 +02:00
|
|
|
NM_GOBJECT_PROPERTIES_DEFINE (NMPolicy,
|
|
|
|
|
PROP_MANAGER,
|
|
|
|
|
PROP_SETTINGS,
|
2018-06-28 18:05:05 +02:00
|
|
|
PROP_DEFAULT_IP4_AC,
|
|
|
|
|
PROP_DEFAULT_IP6_AC,
|
|
|
|
|
PROP_ACTIVATING_IP4_AC,
|
|
|
|
|
PROP_ACTIVATING_IP6_AC,
|
2016-09-29 13:49:01 +02:00
|
|
|
);
|
2016-04-04 13:58:44 +02:00
|
|
|
|
2016-09-29 13:49:01 +02:00
|
|
|
typedef struct {
|
2007-02-08 15:34:26 +00:00
|
|
|
NMManager *manager;
|
2017-04-17 20:17:45 +02:00
|
|
|
NMNetns *netns;
|
2015-08-10 17:07:40 +02:00
|
|
|
NMFirewallManager *firewall_manager;
|
2017-11-22 12:15:24 +01:00
|
|
|
CList pending_activation_checks;
|
2016-04-04 13:44:30 +02:00
|
|
|
|
2017-11-27 09:07:28 +01:00
|
|
|
NMAgentManager *agent_mgr;
|
|
|
|
|
|
2016-04-04 13:44:30 +02:00
|
|
|
GHashTable *devices;
|
2017-09-25 22:20:47 +02:00
|
|
|
GHashTable *pending_active_connections;
|
2008-03-07 Dan Williams <dcbw@redhat.com>
First pass of multiple active device support. Expect bugs.
* src/nm-ip4-config.c
src/nm-ip4-config.h
- (nm_ip4_config_get_secondary, nm_ip4_config_set_secondary): remove;
there are better ways to do this in the named manager
* src/nm-device.c
src/nm-device.h
- (nm_device_can_activate): return whether the device can activate a
connection right now; taking into account things like carrier state
and rfkill state
- (nm_device_get_best_auto_connection): renamed from
nm_device_get_best_connection
- (real_act_stage4_get_ip4_config): MTU stuff is now handled in the
device subclasses themselves, so that each device can override the
MTU from it's NMSetting subclass if needed
- (nm_device_set_ip4_config): set MTU when setting up routes and stuff
in NetworkManagerSystem.c, not here
* src/named-manager/nm-named-manager.c
src/named-manager/nm-named-manager.h
- (nm_named_manager_name_owner_changed,
nm_named_manager_dbus_connection_changed): fix for changes to
rewrite_resolv_conf()
- (compute_nameservers): don't need the NMNamedManager at all, remove
from parameter list
- (merge_one_ip4_config): new function; merge ip4 configs together
- (rewrite_resolv_conf): write out resolv.conf from all the stored
ip4 configs; the VPN config takes precedence, then the best
device config, then the rest of the configs
- (get_domain_for_config): take the NMNamedManager as an argument
to check whether the config is the VPN config
- (add_ip4_config_to_named): fixups for removal of the 'secondary'
attribute from ip4 configs
- (add_all_ip4_configs_to_named): add all the configs in priority order
- (remove_ip4_config_from_named): fix for changes to
get_domain_for_config()
- (nm_named_manager_add_ip4_config): assign the config to the right slot
based on its type; callers must pass in the type now
- (get_last_default_domain): remove, unused
- (nm_named_manager_remove_ip4_config): handle config slots correctly
* src/nm-device-802-11-wireless.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): handle MTU override
* src/nm-device-802-3-ethernet.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): new function; handle MTU override
* src/vpn-manager/nm-vpn-connection.c
- (nm_vpn_connection_ip4_config_get): don't need to set the 'secondary'
attribute on the ip4 config
* src/NetworkManagerPolicy.c
- (nm_policy_auto_get_best_device): remove
- (nm_policy_device_change_check): remove
- (update_default_route): new function; set the default route via
the specified device
- (get_device_priority): new function; return the priority number of
a device type WRT which one should have the default route. Order is
(highest to lowest) wired, wireless, GSM, CDMA.
- (update_routing_and_dns): new function; determine which device should
have the default route, then update the routing table and DNS
- (maybe_auto_activate_device): new function; if a device is now
available for activation, find out what connection it would like to
activate and do it
- (schedule_activate_check): new function; if a device can be activated
now, schedule the activation. Each device may have only one
pending activation at a given time.
- (device_state_changed): if activation was canceled, try again,
possibly with another connection; if the device was activated,
update routing and DNS; if the device was deactivated, try again
with another connection
- (device_carrier_changed): if there is no carrier, deactivate the
device; otherwise schedule an activation check for the device
- (wireless_networks_changed): schedule an activation check for the
device
- (device_added): keep track of the signal handler IDs so they can
be removed when the device goes away
- (device_removed): remove any signal handlers that might be attached
to the device; update routing and DNS
- (schedule_activate_all): new function
- (connections_added, connection_added, connection_updated): when
connections change, schedule all devices for an activation check
- (connection_removed): when a device is deactivated because its
connection was removed, schedule another activation check for it
- (nm_policy_destroy): destroy pending activations and disconnect
all device signal handlers
* src/nm-manager.c
- (nm_manager_activate_device): if the device was already actived,
deactivate it
- (deactivate_old_device): remove
- (connection_added_default_handler, impl_manager_activate_device):
don't deactivate other devices when activating this one
* src/backends/NetworkManagerGentoo.c
src/backends/NetworkManagerFrugalware.c
src/backends/NetworkManagerPaldo.c
src/backends/NetworkManagerRedHat.c
src/backends/NetworkManagerSlackware.c
src/backends/NetworkManagerArch.c
src/backends/NetworkManagerSuSE.c
src/backends/NetworkManagerDebian.c
- (nm_system_get_mtu): remove; MTU should be provided through the
distro's system settings service plugin instead
- (nm_system_device_add_default_route_via_device): remove
- (nm_system_device_add_default_route_via_device_with_iface): remove
- (nm_system_device_replace_default_route): new function; call
generic implementation
* src/backends/NetworkManagerGeneric.c
src/backends/NetworkManagerGeneric.h
- (nm_generic_device_add_default_route_via_device,
nm_generic_device_add_default_route_via_device_with_iface): remove
- (nm_generic_device_replace_default_route): replace the default route
with the given route via some gateway
* src/NetworkManagerSystem.c
src/NetworkManagerSystem.h
- (nm_system_device_set_from_ip4_config): let the policy handle updates
to routing and DNS; but set the MTU here
- (nm_system_vpn_device_set_from_ip4_config): set the route with the
ip_iface of the active device; use the standard MTU setting function
- (nm_system_set_mtu): remove
- (nm_system_device_set_mtu): consolidate MTU setting code in one place
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@3391 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-03-07 19:41:32 +00:00
|
|
|
|
2012-08-21 17:49:41 +02:00
|
|
|
GSList *pending_secondaries;
|
|
|
|
|
|
2010-10-27 20:05:23 -05:00
|
|
|
NMSettings *settings;
|
2010-10-27 15:47:10 -05:00
|
|
|
|
2017-04-23 00:40:46 +02:00
|
|
|
NMHostnameManager *hostname_manager;
|
|
|
|
|
|
2018-06-28 18:05:05 +02:00
|
|
|
NMActiveConnection *default_ac4, *activating_ac4;
|
|
|
|
|
NMActiveConnection *default_ac6, *activating_ac6;
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
|
2017-04-23 19:30:07 +02:00
|
|
|
struct {
|
|
|
|
|
GInetAddress *addr;
|
|
|
|
|
GResolver *resolver;
|
|
|
|
|
GCancellable *cancellable;
|
|
|
|
|
} lookup;
|
|
|
|
|
|
2013-02-05 17:29:15 +01:00
|
|
|
NMDnsManager *dns_manager;
|
|
|
|
|
gulong config_changed_id;
|
2010-01-27 17:13:35 -08:00
|
|
|
|
2016-01-04 09:38:46 +01:00
|
|
|
guint reset_retries_id; /* idle handler for resetting the retries count */
|
2011-06-17 12:43:28 +02:00
|
|
|
|
2016-04-13 14:14:36 +02:00
|
|
|
guint schedule_activate_all_id; /* idle handler for schedule_activate_all(). */
|
|
|
|
|
|
2017-03-02 19:18:49 +01:00
|
|
|
NMPolicyHostnameMode hostname_mode;
|
2010-01-27 17:13:35 -08:00
|
|
|
char *orig_hostname; /* hostname at NM start time */
|
2010-06-17 23:01:40 -07:00
|
|
|
char *cur_hostname; /* hostname we want to assign */
|
2017-02-17 16:22:30 +01:00
|
|
|
char *last_hostname; /* last hostname NM set (to detect if someone else changed it in the meanwhile) */
|
|
|
|
|
gboolean changing_hostname; /* hostname set operation still in progress */
|
2017-03-03 18:32:08 +01:00
|
|
|
gboolean dhcp_hostname; /* current hostname was set from dhcp */
|
2016-10-31 23:31:14 +01:00
|
|
|
|
|
|
|
|
GArray *ip6_prefix_delegations; /* pool of ip6 prefixes delegated to all devices */
|
2016-09-29 13:49:01 +02:00
|
|
|
} NMPolicyPrivate;
|
|
|
|
|
|
|
|
|
|
struct _NMPolicy {
|
|
|
|
|
GObject parent;
|
|
|
|
|
NMPolicyPrivate _priv;
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
struct _NMPolicyClass {
|
|
|
|
|
GObjectClass parent;
|
2016-04-04 13:58:44 +02:00
|
|
|
};
|
2013-08-22 10:10:17 -04:00
|
|
|
|
2016-09-29 13:49:01 +02:00
|
|
|
G_DEFINE_TYPE (NMPolicy, nm_policy, G_TYPE_OBJECT)
|
|
|
|
|
|
|
|
|
|
#define NM_POLICY_GET_PRIVATE(self) _NM_GET_PRIVATE (self, NMPolicy, NM_IS_POLICY)
|
|
|
|
|
|
|
|
|
|
static NMPolicy *
|
|
|
|
|
_PRIV_TO_SELF (NMPolicyPrivate *priv)
|
2016-04-04 13:58:44 +02:00
|
|
|
{
|
2016-09-29 13:49:01 +02:00
|
|
|
NMPolicy *self;
|
|
|
|
|
|
|
|
|
|
nm_assert (priv);
|
|
|
|
|
|
|
|
|
|
self = (NMPolicy *) (((char *) priv) - G_STRUCT_OFFSET (NMPolicy, _priv));
|
|
|
|
|
|
2016-04-04 13:58:44 +02:00
|
|
|
nm_assert (NM_IS_POLICY (self));
|
2016-09-29 13:49:01 +02:00
|
|
|
return self;
|
2016-04-04 13:58:44 +02:00
|
|
|
}
|
2013-08-22 10:10:17 -04:00
|
|
|
|
2016-09-29 13:49:01 +02:00
|
|
|
/*****************************************************************************/
|
2007-02-08 15:34:26 +00:00
|
|
|
|
2016-09-29 13:49:01 +02:00
|
|
|
#define _NMLOG_PREFIX_NAME "policy"
|
|
|
|
|
#define _NMLOG(level, domain, ...) \
|
|
|
|
|
G_STMT_START { \
|
2017-03-01 10:20:01 +00:00
|
|
|
nm_log ((level), (domain), NULL, NULL, \
|
2016-09-29 13:49:01 +02:00
|
|
|
"%s" _NM_UTILS_MACRO_FIRST (__VA_ARGS__), \
|
|
|
|
|
_NMLOG_PREFIX_NAME": " \
|
|
|
|
|
_NM_UTILS_MACRO_REST (__VA_ARGS__)); \
|
|
|
|
|
} G_STMT_END
|
|
|
|
|
|
|
|
|
|
/*****************************************************************************/
|
2013-08-22 13:06:51 -04:00
|
|
|
|
2016-03-29 10:46:07 +02:00
|
|
|
static void schedule_activate_all (NMPolicy *self);
|
2017-09-25 22:20:47 +02:00
|
|
|
static void schedule_activate_check (NMPolicy *self, NMDevice *device);
|
2018-06-28 18:05:05 +02:00
|
|
|
static NMDevice *get_default_device (NMPolicy *self, int addr_family);
|
2011-12-06 16:33:00 -06:00
|
|
|
|
2016-09-29 13:49:01 +02:00
|
|
|
/*****************************************************************************/
|
2011-12-06 16:33:00 -06:00
|
|
|
|
2019-02-09 11:46:47 +01:00
|
|
|
static void
|
|
|
|
|
_dns_manager_set_ip_config (NMDnsManager *dns_manager,
|
|
|
|
|
NMIPConfig *ip_config,
|
|
|
|
|
NMDnsIPConfigType ip_config_type,
|
|
|
|
|
NMDevice *device)
|
|
|
|
|
{
|
|
|
|
|
if ( NM_IN_SET (ip_config_type, NM_DNS_IP_CONFIG_TYPE_DEFAULT,
|
|
|
|
|
NM_DNS_IP_CONFIG_TYPE_BEST_DEVICE)
|
|
|
|
|
&& device
|
|
|
|
|
&& nm_device_get_route_metric_default (nm_device_get_device_type (device)) == NM_VPN_ROUTE_METRIC_DEFAULT) {
|
|
|
|
|
/* some device types are inherently VPN. */
|
|
|
|
|
ip_config_type = NM_DNS_IP_CONFIG_TYPE_VPN;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
nm_dns_manager_set_ip_config (dns_manager, ip_config, ip_config_type);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*****************************************************************************/
|
|
|
|
|
|
2016-10-31 23:31:14 +01:00
|
|
|
typedef struct {
|
|
|
|
|
NMPlatformIP6Address prefix;
|
|
|
|
|
NMDevice *device; /* The requesting ("uplink") device */
|
|
|
|
|
guint64 next_subnet; /* Cache of the next subnet number to be
|
|
|
|
|
* assigned from this prefix */
|
|
|
|
|
GHashTable *subnets; /* ifindex -> NMPlatformIP6Address */
|
|
|
|
|
} IP6PrefixDelegation;
|
|
|
|
|
|
|
|
|
|
static void
|
|
|
|
|
_clear_ip6_subnet (gpointer key, gpointer value, gpointer user_data)
|
|
|
|
|
{
|
|
|
|
|
NMPlatformIP6Address *subnet = value;
|
|
|
|
|
NMDevice *device = nm_manager_get_device_by_ifindex (nm_manager_get (),
|
|
|
|
|
GPOINTER_TO_INT (key));
|
|
|
|
|
|
2016-11-09 18:06:53 +01:00
|
|
|
if (device) {
|
|
|
|
|
/* We can not remove a subnet we already started announcing.
|
|
|
|
|
* Just un-prefer it. */
|
|
|
|
|
subnet->preferred = 0;
|
|
|
|
|
nm_device_use_ip6_subnet (device, subnet);
|
|
|
|
|
}
|
2016-10-31 23:31:14 +01:00
|
|
|
g_slice_free (NMPlatformIP6Address, subnet);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void
|
|
|
|
|
clear_ip6_prefix_delegation (gpointer data)
|
|
|
|
|
{
|
|
|
|
|
IP6PrefixDelegation *delegation = data;
|
2018-11-26 16:49:51 +01:00
|
|
|
char sbuf[NM_UTILS_INET_ADDRSTRLEN];
|
2016-10-31 23:31:14 +01:00
|
|
|
|
|
|
|
|
_LOGD (LOGD_IP6, "ipv6-pd: undelegating prefix %s/%d",
|
2018-11-26 16:49:51 +01:00
|
|
|
nm_utils_inet6_ntop (&delegation->prefix.address, sbuf),
|
2016-10-31 23:31:14 +01:00
|
|
|
delegation->prefix.plen);
|
|
|
|
|
|
|
|
|
|
g_hash_table_foreach (delegation->subnets, _clear_ip6_subnet, NULL);
|
|
|
|
|
g_hash_table_destroy (delegation->subnets);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void
|
|
|
|
|
expire_ip6_delegations (NMPolicy *self)
|
|
|
|
|
{
|
|
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
|
|
|
|
guint32 now = nm_utils_get_monotonic_timestamp_s ();
|
2016-11-09 17:58:11 +01:00
|
|
|
IP6PrefixDelegation *delegation = NULL;
|
|
|
|
|
guint i;
|
2016-10-31 23:31:14 +01:00
|
|
|
|
|
|
|
|
for (i = 0; i < priv->ip6_prefix_delegations->len; i++) {
|
|
|
|
|
delegation = &g_array_index (priv->ip6_prefix_delegations,
|
|
|
|
|
IP6PrefixDelegation, i);
|
|
|
|
|
if (delegation->prefix.timestamp + delegation->prefix.lifetime < now)
|
|
|
|
|
g_array_remove_index_fast (priv->ip6_prefix_delegations, i);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Try to obtain a new subnet for a particular active connection from given
|
|
|
|
|
* delegated prefix, possibly reusing the existing subnet.
|
|
|
|
|
* Return value of FALSE indicates no more subnets are available from
|
|
|
|
|
* this prefix (and other prefix should be used -- and requested if necessary).
|
|
|
|
|
*/
|
|
|
|
|
static gboolean
|
|
|
|
|
ip6_subnet_from_delegation (IP6PrefixDelegation *delegation, NMDevice *device)
|
|
|
|
|
{
|
|
|
|
|
NMPlatformIP6Address *subnet;
|
|
|
|
|
int ifindex = nm_device_get_ifindex (device);
|
2018-11-26 16:49:51 +01:00
|
|
|
char sbuf[NM_UTILS_INET_ADDRSTRLEN];
|
2016-10-31 23:31:14 +01:00
|
|
|
|
|
|
|
|
subnet = g_hash_table_lookup (delegation->subnets, GINT_TO_POINTER (ifindex));
|
|
|
|
|
if (!subnet) {
|
|
|
|
|
/* Check for out-of-prefixes condition. */
|
|
|
|
|
if (delegation->next_subnet >= (1 << (64 - delegation->prefix.plen))) {
|
|
|
|
|
_LOGD (LOGD_IP6, "ipv6-pd: no more prefixes in %s/%d",
|
2018-11-26 16:49:51 +01:00
|
|
|
nm_utils_inet6_ntop (&delegation->prefix.address, sbuf),
|
2016-10-31 23:31:14 +01:00
|
|
|
delegation->prefix.plen);
|
|
|
|
|
return FALSE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Allocate a new subnet. */
|
|
|
|
|
subnet = g_slice_new0 (NMPlatformIP6Address);
|
|
|
|
|
g_hash_table_insert (delegation->subnets, GINT_TO_POINTER (ifindex), subnet);
|
|
|
|
|
|
|
|
|
|
subnet->plen = 64;
|
|
|
|
|
subnet->address.s6_addr32[0] = delegation->prefix.address.s6_addr32[0]
|
|
|
|
|
| htonl (delegation->next_subnet >> 32);
|
|
|
|
|
subnet->address.s6_addr32[1] = delegation->prefix.address.s6_addr32[1]
|
|
|
|
|
| htonl (delegation->next_subnet);
|
|
|
|
|
|
|
|
|
|
/* Out subnet pool management is pretty unsophisticated. We only add
|
|
|
|
|
* the subnets and index them by ifindex. That keeps the implementation
|
|
|
|
|
* simple and the dead entries make it easy to reuse the same subnet on
|
|
|
|
|
* subsequent activations. On the other hand they may waste the subnet
|
|
|
|
|
* space. */
|
|
|
|
|
delegation->next_subnet++;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
subnet->timestamp = delegation->prefix.timestamp;
|
|
|
|
|
subnet->lifetime = delegation->prefix.lifetime;
|
|
|
|
|
subnet->preferred = delegation->prefix.preferred;
|
|
|
|
|
|
|
|
|
|
_LOGD (LOGD_IP6, "ipv6-pd: %s allocated from a /%d prefix on %s",
|
2018-11-26 16:49:51 +01:00
|
|
|
nm_utils_inet6_ntop (&subnet->address, sbuf),
|
2016-10-31 23:31:14 +01:00
|
|
|
delegation->prefix.plen,
|
|
|
|
|
nm_device_get_iface (device));
|
|
|
|
|
|
|
|
|
|
nm_device_use_ip6_subnet (device, subnet);
|
|
|
|
|
|
|
|
|
|
return TRUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Try to obtain a subnet from each prefix delegated to given requesting
|
|
|
|
|
* ("uplink") device and assign it to the downlink device.
|
|
|
|
|
* Requests a new prefix if no subnet could be found.
|
|
|
|
|
*/
|
|
|
|
|
static void
|
|
|
|
|
ip6_subnet_from_device (NMPolicy *self, NMDevice *from_device, NMDevice *device)
|
|
|
|
|
{
|
|
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
|
|
|
|
IP6PrefixDelegation *delegation = NULL;
|
|
|
|
|
gboolean got_subnet = FALSE;
|
2016-11-09 17:58:11 +01:00
|
|
|
guint have_prefixes = 0;
|
|
|
|
|
guint i;
|
2016-10-31 23:31:14 +01:00
|
|
|
|
|
|
|
|
expire_ip6_delegations (self);
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < priv->ip6_prefix_delegations->len; i++) {
|
|
|
|
|
delegation = &g_array_index (priv->ip6_prefix_delegations,
|
|
|
|
|
IP6PrefixDelegation, i);
|
|
|
|
|
|
|
|
|
|
if (delegation->device != from_device)
|
|
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
if (ip6_subnet_from_delegation (delegation, device))
|
|
|
|
|
got_subnet = TRUE;
|
|
|
|
|
have_prefixes++;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (!got_subnet) {
|
2016-11-09 17:58:11 +01:00
|
|
|
_LOGI (LOGD_IP6, "ipv6-pd: none of %u prefixes of %s can be shared on %s",
|
2016-10-31 23:31:14 +01:00
|
|
|
have_prefixes, nm_device_get_iface (from_device),
|
|
|
|
|
nm_device_get_iface (device));
|
|
|
|
|
nm_device_request_ip6_prefixes (from_device, have_prefixes + 1);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void
|
|
|
|
|
ip6_remove_device_prefix_delegations (NMPolicy *self, NMDevice *device)
|
|
|
|
|
{
|
|
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2016-11-09 17:58:11 +01:00
|
|
|
IP6PrefixDelegation *delegation = NULL;
|
|
|
|
|
guint i;
|
2016-10-31 23:31:14 +01:00
|
|
|
|
|
|
|
|
for (i = 0; i < priv->ip6_prefix_delegations->len; i++) {
|
2016-11-09 17:58:11 +01:00
|
|
|
delegation = &g_array_index (priv->ip6_prefix_delegations,
|
2016-10-31 23:31:14 +01:00
|
|
|
IP6PrefixDelegation, i);
|
|
|
|
|
if (delegation->device == device)
|
|
|
|
|
g_array_remove_index_fast (priv->ip6_prefix_delegations, i);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void
|
|
|
|
|
device_ip6_prefix_delegated (NMDevice *device,
|
|
|
|
|
NMPlatformIP6Address *prefix,
|
|
|
|
|
gpointer user_data)
|
|
|
|
|
{
|
|
|
|
|
NMPolicyPrivate *priv = user_data;
|
|
|
|
|
NMPolicy *self = _PRIV_TO_SELF (priv);
|
2016-11-09 17:58:11 +01:00
|
|
|
IP6PrefixDelegation *delegation = NULL;
|
|
|
|
|
guint i;
|
2017-11-23 21:30:09 +01:00
|
|
|
const CList *tmp_list;
|
|
|
|
|
NMActiveConnection *ac;
|
2018-11-26 16:49:51 +01:00
|
|
|
char sbuf[NM_UTILS_INET_ADDRSTRLEN];
|
2016-10-31 23:31:14 +01:00
|
|
|
|
|
|
|
|
_LOGI (LOGD_IP6, "ipv6-pd: received a prefix %s/%d from %s",
|
2018-11-26 16:49:51 +01:00
|
|
|
nm_utils_inet6_ntop (&prefix->address, sbuf),
|
2016-10-31 23:31:14 +01:00
|
|
|
prefix->plen,
|
|
|
|
|
nm_device_get_iface (device));
|
|
|
|
|
|
|
|
|
|
expire_ip6_delegations (self);
|
|
|
|
|
|
2016-11-09 17:58:11 +01:00
|
|
|
for (i = 0; i < priv->ip6_prefix_delegations->len; i++) {
|
|
|
|
|
/* Look for an already known prefix to update. */
|
|
|
|
|
delegation = &g_array_index (priv->ip6_prefix_delegations, IP6PrefixDelegation, i);
|
|
|
|
|
if (IN6_ARE_ADDR_EQUAL (&delegation->prefix.address, &prefix->address))
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (i == priv->ip6_prefix_delegations->len) {
|
|
|
|
|
/* Allocate a delegation delegation for new prefix. */
|
|
|
|
|
g_array_set_size (priv->ip6_prefix_delegations, i + 1);
|
|
|
|
|
delegation = &g_array_index (priv->ip6_prefix_delegations, IP6PrefixDelegation, i);
|
2017-11-15 16:06:43 +01:00
|
|
|
delegation->subnets = g_hash_table_new (nm_direct_hash, NULL);
|
2016-10-31 23:31:14 +01:00
|
|
|
delegation->next_subnet = 0;
|
2016-11-09 17:58:11 +01:00
|
|
|
}
|
2016-10-31 23:31:14 +01:00
|
|
|
|
|
|
|
|
delegation->device = device;
|
|
|
|
|
delegation->prefix = *prefix;
|
|
|
|
|
|
|
|
|
|
/* The newly activated connections are added to the list beginning,
|
|
|
|
|
* so traversing it from the beginning makes it likely for newly
|
|
|
|
|
* activated connections that have no subnet assigned to be served
|
|
|
|
|
* first. That is a simple yet fair policy, which is good. */
|
2017-11-23 21:30:09 +01:00
|
|
|
nm_manager_for_each_active_connection (priv->manager, ac, tmp_list) {
|
|
|
|
|
NMDevice *to_device;
|
2016-10-31 23:31:14 +01:00
|
|
|
|
2017-11-23 21:30:09 +01:00
|
|
|
to_device = nm_active_connection_get_device (ac);
|
2016-10-31 23:31:14 +01:00
|
|
|
if (nm_device_needs_ip6_subnet (to_device))
|
|
|
|
|
ip6_subnet_from_delegation (delegation, to_device);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void
|
|
|
|
|
device_ip6_subnet_needed (NMDevice *device,
|
|
|
|
|
gpointer user_data)
|
|
|
|
|
{
|
|
|
|
|
NMPolicyPrivate *priv = user_data;
|
|
|
|
|
NMPolicy *self = _PRIV_TO_SELF (priv);
|
|
|
|
|
|
|
|
|
|
_LOGD (LOGD_IP6, "ipv6-pd: %s needs a subnet",
|
|
|
|
|
nm_device_get_iface (device));
|
|
|
|
|
|
2018-06-28 18:05:05 +02:00
|
|
|
if (!priv->default_ac6) {
|
2016-10-31 23:31:14 +01:00
|
|
|
/* We request the prefixes when the default IPv6 device is set. */
|
|
|
|
|
_LOGI (LOGD_IP6, "ipv6-pd: no device to obtain a subnet to share on %s from",
|
|
|
|
|
nm_device_get_iface (device));
|
|
|
|
|
return;
|
|
|
|
|
}
|
2018-06-28 18:05:05 +02:00
|
|
|
ip6_subnet_from_device (self, get_default_device (self, AF_INET6), device);
|
|
|
|
|
nm_device_copy_ip6_dns_config (device, get_default_device (self, AF_INET6));
|
2016-10-31 23:31:14 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*****************************************************************************/
|
|
|
|
|
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
static NMDevice *
|
2018-06-28 18:05:05 +02:00
|
|
|
get_default_device (NMPolicy *self, int addr_family)
|
|
|
|
|
{
|
|
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
|
|
|
|
NMActiveConnection *ac;
|
|
|
|
|
|
|
|
|
|
nm_assert_addr_family (addr_family);
|
|
|
|
|
|
|
|
|
|
ac = (addr_family == AF_INET) ? priv->default_ac4 : priv->default_ac6;
|
|
|
|
|
|
|
|
|
|
return ac ? nm_active_connection_get_device (ac) : NULL;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static NMActiveConnection *
|
|
|
|
|
get_best_active_connection (NMPolicy *self,
|
|
|
|
|
int addr_family,
|
|
|
|
|
gboolean fully_activated)
|
2008-03-07 Dan Williams <dcbw@redhat.com>
First pass of multiple active device support. Expect bugs.
* src/nm-ip4-config.c
src/nm-ip4-config.h
- (nm_ip4_config_get_secondary, nm_ip4_config_set_secondary): remove;
there are better ways to do this in the named manager
* src/nm-device.c
src/nm-device.h
- (nm_device_can_activate): return whether the device can activate a
connection right now; taking into account things like carrier state
and rfkill state
- (nm_device_get_best_auto_connection): renamed from
nm_device_get_best_connection
- (real_act_stage4_get_ip4_config): MTU stuff is now handled in the
device subclasses themselves, so that each device can override the
MTU from it's NMSetting subclass if needed
- (nm_device_set_ip4_config): set MTU when setting up routes and stuff
in NetworkManagerSystem.c, not here
* src/named-manager/nm-named-manager.c
src/named-manager/nm-named-manager.h
- (nm_named_manager_name_owner_changed,
nm_named_manager_dbus_connection_changed): fix for changes to
rewrite_resolv_conf()
- (compute_nameservers): don't need the NMNamedManager at all, remove
from parameter list
- (merge_one_ip4_config): new function; merge ip4 configs together
- (rewrite_resolv_conf): write out resolv.conf from all the stored
ip4 configs; the VPN config takes precedence, then the best
device config, then the rest of the configs
- (get_domain_for_config): take the NMNamedManager as an argument
to check whether the config is the VPN config
- (add_ip4_config_to_named): fixups for removal of the 'secondary'
attribute from ip4 configs
- (add_all_ip4_configs_to_named): add all the configs in priority order
- (remove_ip4_config_from_named): fix for changes to
get_domain_for_config()
- (nm_named_manager_add_ip4_config): assign the config to the right slot
based on its type; callers must pass in the type now
- (get_last_default_domain): remove, unused
- (nm_named_manager_remove_ip4_config): handle config slots correctly
* src/nm-device-802-11-wireless.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): handle MTU override
* src/nm-device-802-3-ethernet.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): new function; handle MTU override
* src/vpn-manager/nm-vpn-connection.c
- (nm_vpn_connection_ip4_config_get): don't need to set the 'secondary'
attribute on the ip4 config
* src/NetworkManagerPolicy.c
- (nm_policy_auto_get_best_device): remove
- (nm_policy_device_change_check): remove
- (update_default_route): new function; set the default route via
the specified device
- (get_device_priority): new function; return the priority number of
a device type WRT which one should have the default route. Order is
(highest to lowest) wired, wireless, GSM, CDMA.
- (update_routing_and_dns): new function; determine which device should
have the default route, then update the routing table and DNS
- (maybe_auto_activate_device): new function; if a device is now
available for activation, find out what connection it would like to
activate and do it
- (schedule_activate_check): new function; if a device can be activated
now, schedule the activation. Each device may have only one
pending activation at a given time.
- (device_state_changed): if activation was canceled, try again,
possibly with another connection; if the device was activated,
update routing and DNS; if the device was deactivated, try again
with another connection
- (device_carrier_changed): if there is no carrier, deactivate the
device; otherwise schedule an activation check for the device
- (wireless_networks_changed): schedule an activation check for the
device
- (device_added): keep track of the signal handler IDs so they can
be removed when the device goes away
- (device_removed): remove any signal handlers that might be attached
to the device; update routing and DNS
- (schedule_activate_all): new function
- (connections_added, connection_added, connection_updated): when
connections change, schedule all devices for an activation check
- (connection_removed): when a device is deactivated because its
connection was removed, schedule another activation check for it
- (nm_policy_destroy): destroy pending activations and disconnect
all device signal handlers
* src/nm-manager.c
- (nm_manager_activate_device): if the device was already actived,
deactivate it
- (deactivate_old_device): remove
- (connection_added_default_handler, impl_manager_activate_device):
don't deactivate other devices when activating this one
* src/backends/NetworkManagerGentoo.c
src/backends/NetworkManagerFrugalware.c
src/backends/NetworkManagerPaldo.c
src/backends/NetworkManagerRedHat.c
src/backends/NetworkManagerSlackware.c
src/backends/NetworkManagerArch.c
src/backends/NetworkManagerSuSE.c
src/backends/NetworkManagerDebian.c
- (nm_system_get_mtu): remove; MTU should be provided through the
distro's system settings service plugin instead
- (nm_system_device_add_default_route_via_device): remove
- (nm_system_device_add_default_route_via_device_with_iface): remove
- (nm_system_device_replace_default_route): new function; call
generic implementation
* src/backends/NetworkManagerGeneric.c
src/backends/NetworkManagerGeneric.h
- (nm_generic_device_add_default_route_via_device,
nm_generic_device_add_default_route_via_device_with_iface): remove
- (nm_generic_device_replace_default_route): replace the default route
with the given route via some gateway
* src/NetworkManagerSystem.c
src/NetworkManagerSystem.h
- (nm_system_device_set_from_ip4_config): let the policy handle updates
to routing and DNS; but set the MTU here
- (nm_system_vpn_device_set_from_ip4_config): set the route with the
ip_iface of the active device; use the standard MTU setting function
- (nm_system_set_mtu): remove
- (nm_system_device_set_mtu): consolidate MTU setting code in one place
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@3391 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-03-07 19:41:32 +00:00
|
|
|
{
|
2014-02-27 16:27:57 +01:00
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2018-03-29 08:52:45 +02:00
|
|
|
const CList *tmp_lst;
|
core: track devices in manager via embedded CList
Instead of using a GSList for tracking the devices, use a CList.
I think a CList is in most cases the more suitable data structure
then GSList:
- you can find out in O(1) whether the object is linked. That
is nice, for example to assert in NMDevice's destructor that
the object was unlinked, and we will use that later in
nm_manager_get_device_by_path().
- you can unlink the element in O(1) and you can unlink the
element without having access to the link's head
- Contrary to GSList, this does not require an extra slice
allocation for the link node. It quite possibliy consumes
slightly less memory because the CList structure is embedded
in a struct that we already allocate. Even if slice allocation
would be perfect to only consume 2*sizeof(gpointer) for the link
note, it would at most be as-good as CList. Quite possibly,
there is an overhead though.
- CList possibly has better memory locality, because the link
structure and the data are close to each other.
Something which could be seen as disavantage, is that with CList
one device can only be tracked in one NMManager instance at a time.
But that is fine. There exists only one NMManager instance for now,
and even if we would ever introduce multiple managers, we probably
would not associate one NMDevice instance with multiple managers.
The advantages are arguably not huge, but CList is IMHO clearly the
more suited data structure. No need to stick to a suboptimal data
structure for the job. Refactor it.
2018-03-23 21:51:07 +01:00
|
|
|
NMDevice *device;
|
core: rework handling of default-routes and drop NMDefaultRouteManager
Remove NMDefaultRouteManager. Instead, add the default-route to the
NMIP4Config/NMIP6Config instance.
This basically reverts commit e8824f6a5205ffcf761abd3e0897a22b254c7797.
We added NMDefaultRouteManager because we used the corresponding to `ip
route replace` when configuring routes. That would replace default-routes
on other interfaces so we needed a central manager to coordinate routes.
Now, we use the corresponding of `ip route append` to configure routes,
and each interface can configure routes indepdentently.
In NMDevice, when creating the default-route, ignore @auto_method for
external devices. We shall not touch these devices.
Especially the code in NMPolicy regarding selection of the best-device
seems wrong. It probably needs further adjustments in the future.
Especially get_best_ip_config() should be replaced, because this
distinction VPN vs. devices seems wrong to me.
Thereby, remove the @ignore_never_default argument. It was added by
commit bb750260045239ab85574366bae8102eff8058cc, I don't think it's
needed anymore.
This brings another change. Now that we track default-routes in
NMIP4Config/NMIP6Config, they are also exposed on D-Bus like regular
routes. I think that makes sense, but it is a change in behavior, as
previously such routes were not exposed there.
2017-08-30 11:46:42 +02:00
|
|
|
guint32 best_metric = G_MAXUINT32;
|
|
|
|
|
gboolean best_is_fully_activated = FALSE;
|
2018-06-28 18:05:05 +02:00
|
|
|
NMActiveConnection *best_ac, *prev_ac;
|
core: rework handling of default-routes and drop NMDefaultRouteManager
Remove NMDefaultRouteManager. Instead, add the default-route to the
NMIP4Config/NMIP6Config instance.
This basically reverts commit e8824f6a5205ffcf761abd3e0897a22b254c7797.
We added NMDefaultRouteManager because we used the corresponding to `ip
route replace` when configuring routes. That would replace default-routes
on other interfaces so we needed a central manager to coordinate routes.
Now, we use the corresponding of `ip route append` to configure routes,
and each interface can configure routes indepdentently.
In NMDevice, when creating the default-route, ignore @auto_method for
external devices. We shall not touch these devices.
Especially the code in NMPolicy regarding selection of the best-device
seems wrong. It probably needs further adjustments in the future.
Especially get_best_ip_config() should be replaced, because this
distinction VPN vs. devices seems wrong to me.
Thereby, remove the @ignore_never_default argument. It was added by
commit bb750260045239ab85574366bae8102eff8058cc, I don't think it's
needed anymore.
This brings another change. Now that we track default-routes in
NMIP4Config/NMIP6Config, they are also exposed on D-Bus like regular
routes. I think that makes sense, but it is a change in behavior, as
previously such routes were not exposed there.
2017-08-30 11:46:42 +02:00
|
|
|
|
|
|
|
|
nm_assert (NM_IN_SET (addr_family, AF_INET, AF_INET6));
|
|
|
|
|
|
2018-06-28 18:05:05 +02:00
|
|
|
/* we prefer the current AC in case of identical metric.
|
|
|
|
|
* Hence, try that one first. */
|
|
|
|
|
prev_ac = addr_family == AF_INET
|
|
|
|
|
? (fully_activated ? priv->default_ac4 : priv->activating_ac4)
|
|
|
|
|
: (fully_activated ? priv->default_ac6 : priv->activating_ac6);
|
|
|
|
|
best_ac = NULL;
|
core: rework handling of default-routes and drop NMDefaultRouteManager
Remove NMDefaultRouteManager. Instead, add the default-route to the
NMIP4Config/NMIP6Config instance.
This basically reverts commit e8824f6a5205ffcf761abd3e0897a22b254c7797.
We added NMDefaultRouteManager because we used the corresponding to `ip
route replace` when configuring routes. That would replace default-routes
on other interfaces so we needed a central manager to coordinate routes.
Now, we use the corresponding of `ip route append` to configure routes,
and each interface can configure routes indepdentently.
In NMDevice, when creating the default-route, ignore @auto_method for
external devices. We shall not touch these devices.
Especially the code in NMPolicy regarding selection of the best-device
seems wrong. It probably needs further adjustments in the future.
Especially get_best_ip_config() should be replaced, because this
distinction VPN vs. devices seems wrong to me.
Thereby, remove the @ignore_never_default argument. It was added by
commit bb750260045239ab85574366bae8102eff8058cc, I don't think it's
needed anymore.
This brings another change. Now that we track default-routes in
NMIP4Config/NMIP6Config, they are also exposed on D-Bus like regular
routes. I think that makes sense, but it is a change in behavior, as
previously such routes were not exposed there.
2017-08-30 11:46:42 +02:00
|
|
|
|
2018-03-29 08:52:45 +02:00
|
|
|
nm_manager_for_each_device (priv->manager, device, tmp_lst) {
|
core: rework handling of default-routes and drop NMDefaultRouteManager
Remove NMDefaultRouteManager. Instead, add the default-route to the
NMIP4Config/NMIP6Config instance.
This basically reverts commit e8824f6a5205ffcf761abd3e0897a22b254c7797.
We added NMDefaultRouteManager because we used the corresponding to `ip
route replace` when configuring routes. That would replace default-routes
on other interfaces so we needed a central manager to coordinate routes.
Now, we use the corresponding of `ip route append` to configure routes,
and each interface can configure routes indepdentently.
In NMDevice, when creating the default-route, ignore @auto_method for
external devices. We shall not touch these devices.
Especially the code in NMPolicy regarding selection of the best-device
seems wrong. It probably needs further adjustments in the future.
Especially get_best_ip_config() should be replaced, because this
distinction VPN vs. devices seems wrong to me.
Thereby, remove the @ignore_never_default argument. It was added by
commit bb750260045239ab85574366bae8102eff8058cc, I don't think it's
needed anymore.
This brings another change. Now that we track default-routes in
NMIP4Config/NMIP6Config, they are also exposed on D-Bus like regular
routes. I think that makes sense, but it is a change in behavior, as
previously such routes were not exposed there.
2017-08-30 11:46:42 +02:00
|
|
|
NMDeviceState state;
|
|
|
|
|
const NMPObject *r;
|
2018-06-28 18:05:05 +02:00
|
|
|
NMActiveConnection *ac;
|
core: rework handling of default-routes and drop NMDefaultRouteManager
Remove NMDefaultRouteManager. Instead, add the default-route to the
NMIP4Config/NMIP6Config instance.
This basically reverts commit e8824f6a5205ffcf761abd3e0897a22b254c7797.
We added NMDefaultRouteManager because we used the corresponding to `ip
route replace` when configuring routes. That would replace default-routes
on other interfaces so we needed a central manager to coordinate routes.
Now, we use the corresponding of `ip route append` to configure routes,
and each interface can configure routes indepdentently.
In NMDevice, when creating the default-route, ignore @auto_method for
external devices. We shall not touch these devices.
Especially the code in NMPolicy regarding selection of the best-device
seems wrong. It probably needs further adjustments in the future.
Especially get_best_ip_config() should be replaced, because this
distinction VPN vs. devices seems wrong to me.
Thereby, remove the @ignore_never_default argument. It was added by
commit bb750260045239ab85574366bae8102eff8058cc, I don't think it's
needed anymore.
This brings another change. Now that we track default-routes in
NMIP4Config/NMIP6Config, they are also exposed on D-Bus like regular
routes. I think that makes sense, but it is a change in behavior, as
previously such routes were not exposed there.
2017-08-30 11:46:42 +02:00
|
|
|
NMConnection *connection;
|
|
|
|
|
guint32 metric;
|
|
|
|
|
gboolean is_fully_activated;
|
2014-10-31 23:26:48 +01:00
|
|
|
|
core: rework handling of default-routes and drop NMDefaultRouteManager
Remove NMDefaultRouteManager. Instead, add the default-route to the
NMIP4Config/NMIP6Config instance.
This basically reverts commit e8824f6a5205ffcf761abd3e0897a22b254c7797.
We added NMDefaultRouteManager because we used the corresponding to `ip
route replace` when configuring routes. That would replace default-routes
on other interfaces so we needed a central manager to coordinate routes.
Now, we use the corresponding of `ip route append` to configure routes,
and each interface can configure routes indepdentently.
In NMDevice, when creating the default-route, ignore @auto_method for
external devices. We shall not touch these devices.
Especially the code in NMPolicy regarding selection of the best-device
seems wrong. It probably needs further adjustments in the future.
Especially get_best_ip_config() should be replaced, because this
distinction VPN vs. devices seems wrong to me.
Thereby, remove the @ignore_never_default argument. It was added by
commit bb750260045239ab85574366bae8102eff8058cc, I don't think it's
needed anymore.
This brings another change. Now that we track default-routes in
NMIP4Config/NMIP6Config, they are also exposed on D-Bus like regular
routes. I think that makes sense, but it is a change in behavior, as
previously such routes were not exposed there.
2017-08-30 11:46:42 +02:00
|
|
|
state = nm_device_get_state (device);
|
|
|
|
|
if ( state <= NM_DEVICE_STATE_DISCONNECTED
|
|
|
|
|
|| state >= NM_DEVICE_STATE_DEACTIVATING)
|
|
|
|
|
continue;
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
|
core: rework handling of default-routes and drop NMDefaultRouteManager
Remove NMDefaultRouteManager. Instead, add the default-route to the
NMIP4Config/NMIP6Config instance.
This basically reverts commit e8824f6a5205ffcf761abd3e0897a22b254c7797.
We added NMDefaultRouteManager because we used the corresponding to `ip
route replace` when configuring routes. That would replace default-routes
on other interfaces so we needed a central manager to coordinate routes.
Now, we use the corresponding of `ip route append` to configure routes,
and each interface can configure routes indepdentently.
In NMDevice, when creating the default-route, ignore @auto_method for
external devices. We shall not touch these devices.
Especially the code in NMPolicy regarding selection of the best-device
seems wrong. It probably needs further adjustments in the future.
Especially get_best_ip_config() should be replaced, because this
distinction VPN vs. devices seems wrong to me.
Thereby, remove the @ignore_never_default argument. It was added by
commit bb750260045239ab85574366bae8102eff8058cc, I don't think it's
needed anymore.
This brings another change. Now that we track default-routes in
NMIP4Config/NMIP6Config, they are also exposed on D-Bus like regular
routes. I think that makes sense, but it is a change in behavior, as
previously such routes were not exposed there.
2017-08-30 11:46:42 +02:00
|
|
|
if (nm_device_sys_iface_state_is_external (device))
|
|
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
r = nm_device_get_best_default_route (device, addr_family);
|
|
|
|
|
if (r) {
|
2018-01-23 12:24:29 +01:00
|
|
|
/* NOTE: the best route might have rt_source NM_IP_CONFIG_SOURCE_VPN,
|
core: rework handling of default-routes and drop NMDefaultRouteManager
Remove NMDefaultRouteManager. Instead, add the default-route to the
NMIP4Config/NMIP6Config instance.
This basically reverts commit e8824f6a5205ffcf761abd3e0897a22b254c7797.
We added NMDefaultRouteManager because we used the corresponding to `ip
route replace` when configuring routes. That would replace default-routes
on other interfaces so we needed a central manager to coordinate routes.
Now, we use the corresponding of `ip route append` to configure routes,
and each interface can configure routes indepdentently.
In NMDevice, when creating the default-route, ignore @auto_method for
external devices. We shall not touch these devices.
Especially the code in NMPolicy regarding selection of the best-device
seems wrong. It probably needs further adjustments in the future.
Especially get_best_ip_config() should be replaced, because this
distinction VPN vs. devices seems wrong to me.
Thereby, remove the @ignore_never_default argument. It was added by
commit bb750260045239ab85574366bae8102eff8058cc, I don't think it's
needed anymore.
This brings another change. Now that we track default-routes in
NMIP4Config/NMIP6Config, they are also exposed on D-Bus like regular
routes. I think that makes sense, but it is a change in behavior, as
previously such routes were not exposed there.
2017-08-30 11:46:42 +02:00
|
|
|
* which means it was injected by a VPN, not added by device.
|
|
|
|
|
*
|
|
|
|
|
* In this case, is it really the best device? Why do we even need the best
|
|
|
|
|
* device?? */
|
|
|
|
|
metric = nm_utils_ip_route_metric_normalize (addr_family,
|
|
|
|
|
NMP_OBJECT_CAST_IP_ROUTE (r)->metric);
|
|
|
|
|
is_fully_activated = TRUE;
|
|
|
|
|
} else if ( !fully_activated
|
|
|
|
|
&& (connection = nm_device_get_applied_connection (device))
|
|
|
|
|
&& nm_utils_connection_has_default_route (connection, addr_family, NULL)) {
|
|
|
|
|
metric = nm_utils_ip_route_metric_normalize (addr_family,
|
2017-09-28 17:14:24 +02:00
|
|
|
nm_device_get_route_metric (device, addr_family));
|
core: rework handling of default-routes and drop NMDefaultRouteManager
Remove NMDefaultRouteManager. Instead, add the default-route to the
NMIP4Config/NMIP6Config instance.
This basically reverts commit e8824f6a5205ffcf761abd3e0897a22b254c7797.
We added NMDefaultRouteManager because we used the corresponding to `ip
route replace` when configuring routes. That would replace default-routes
on other interfaces so we needed a central manager to coordinate routes.
Now, we use the corresponding of `ip route append` to configure routes,
and each interface can configure routes indepdentently.
In NMDevice, when creating the default-route, ignore @auto_method for
external devices. We shall not touch these devices.
Especially the code in NMPolicy regarding selection of the best-device
seems wrong. It probably needs further adjustments in the future.
Especially get_best_ip_config() should be replaced, because this
distinction VPN vs. devices seems wrong to me.
Thereby, remove the @ignore_never_default argument. It was added by
commit bb750260045239ab85574366bae8102eff8058cc, I don't think it's
needed anymore.
This brings another change. Now that we track default-routes in
NMIP4Config/NMIP6Config, they are also exposed on D-Bus like regular
routes. I think that makes sense, but it is a change in behavior, as
previously such routes were not exposed there.
2017-08-30 11:46:42 +02:00
|
|
|
is_fully_activated = FALSE;
|
|
|
|
|
} else
|
|
|
|
|
continue;
|
2014-10-31 23:26:48 +01:00
|
|
|
|
2018-06-28 18:05:05 +02:00
|
|
|
ac = (NMActiveConnection *) nm_device_get_act_request (device);
|
|
|
|
|
nm_assert (ac);
|
|
|
|
|
|
|
|
|
|
if ( !best_ac
|
core: rework handling of default-routes and drop NMDefaultRouteManager
Remove NMDefaultRouteManager. Instead, add the default-route to the
NMIP4Config/NMIP6Config instance.
This basically reverts commit e8824f6a5205ffcf761abd3e0897a22b254c7797.
We added NMDefaultRouteManager because we used the corresponding to `ip
route replace` when configuring routes. That would replace default-routes
on other interfaces so we needed a central manager to coordinate routes.
Now, we use the corresponding of `ip route append` to configure routes,
and each interface can configure routes indepdentently.
In NMDevice, when creating the default-route, ignore @auto_method for
external devices. We shall not touch these devices.
Especially the code in NMPolicy regarding selection of the best-device
seems wrong. It probably needs further adjustments in the future.
Especially get_best_ip_config() should be replaced, because this
distinction VPN vs. devices seems wrong to me.
Thereby, remove the @ignore_never_default argument. It was added by
commit bb750260045239ab85574366bae8102eff8058cc, I don't think it's
needed anymore.
This brings another change. Now that we track default-routes in
NMIP4Config/NMIP6Config, they are also exposed on D-Bus like regular
routes. I think that makes sense, but it is a change in behavior, as
previously such routes were not exposed there.
2017-08-30 11:46:42 +02:00
|
|
|
|| (!best_is_fully_activated && is_fully_activated)
|
|
|
|
|
|| ( metric < best_metric
|
2018-06-28 18:05:05 +02:00
|
|
|
|| (metric == best_metric && ac == prev_ac))) {
|
|
|
|
|
best_ac = ac;
|
core: rework handling of default-routes and drop NMDefaultRouteManager
Remove NMDefaultRouteManager. Instead, add the default-route to the
NMIP4Config/NMIP6Config instance.
This basically reverts commit e8824f6a5205ffcf761abd3e0897a22b254c7797.
We added NMDefaultRouteManager because we used the corresponding to `ip
route replace` when configuring routes. That would replace default-routes
on other interfaces so we needed a central manager to coordinate routes.
Now, we use the corresponding of `ip route append` to configure routes,
and each interface can configure routes indepdentently.
In NMDevice, when creating the default-route, ignore @auto_method for
external devices. We shall not touch these devices.
Especially the code in NMPolicy regarding selection of the best-device
seems wrong. It probably needs further adjustments in the future.
Especially get_best_ip_config() should be replaced, because this
distinction VPN vs. devices seems wrong to me.
Thereby, remove the @ignore_never_default argument. It was added by
commit bb750260045239ab85574366bae8102eff8058cc, I don't think it's
needed anymore.
This brings another change. Now that we track default-routes in
NMIP4Config/NMIP6Config, they are also exposed on D-Bus like regular
routes. I think that makes sense, but it is a change in behavior, as
previously such routes were not exposed there.
2017-08-30 11:46:42 +02:00
|
|
|
best_metric = metric;
|
|
|
|
|
best_is_fully_activated = is_fully_activated;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if ( !fully_activated
|
2018-06-28 18:05:05 +02:00
|
|
|
&& best_ac
|
core: rework handling of default-routes and drop NMDefaultRouteManager
Remove NMDefaultRouteManager. Instead, add the default-route to the
NMIP4Config/NMIP6Config instance.
This basically reverts commit e8824f6a5205ffcf761abd3e0897a22b254c7797.
We added NMDefaultRouteManager because we used the corresponding to `ip
route replace` when configuring routes. That would replace default-routes
on other interfaces so we needed a central manager to coordinate routes.
Now, we use the corresponding of `ip route append` to configure routes,
and each interface can configure routes indepdentently.
In NMDevice, when creating the default-route, ignore @auto_method for
external devices. We shall not touch these devices.
Especially the code in NMPolicy regarding selection of the best-device
seems wrong. It probably needs further adjustments in the future.
Especially get_best_ip_config() should be replaced, because this
distinction VPN vs. devices seems wrong to me.
Thereby, remove the @ignore_never_default argument. It was added by
commit bb750260045239ab85574366bae8102eff8058cc, I don't think it's
needed anymore.
This brings another change. Now that we track default-routes in
NMIP4Config/NMIP6Config, they are also exposed on D-Bus like regular
routes. I think that makes sense, but it is a change in behavior, as
previously such routes were not exposed there.
2017-08-30 11:46:42 +02:00
|
|
|
&& best_is_fully_activated) {
|
2018-06-28 18:05:05 +02:00
|
|
|
/* There's a best activating AC only if the best device
|
core: rework handling of default-routes and drop NMDefaultRouteManager
Remove NMDefaultRouteManager. Instead, add the default-route to the
NMIP4Config/NMIP6Config instance.
This basically reverts commit e8824f6a5205ffcf761abd3e0897a22b254c7797.
We added NMDefaultRouteManager because we used the corresponding to `ip
route replace` when configuring routes. That would replace default-routes
on other interfaces so we needed a central manager to coordinate routes.
Now, we use the corresponding of `ip route append` to configure routes,
and each interface can configure routes indepdentently.
In NMDevice, when creating the default-route, ignore @auto_method for
external devices. We shall not touch these devices.
Especially the code in NMPolicy regarding selection of the best-device
seems wrong. It probably needs further adjustments in the future.
Especially get_best_ip_config() should be replaced, because this
distinction VPN vs. devices seems wrong to me.
Thereby, remove the @ignore_never_default argument. It was added by
commit bb750260045239ab85574366bae8102eff8058cc, I don't think it's
needed anymore.
This brings another change. Now that we track default-routes in
NMIP4Config/NMIP6Config, they are also exposed on D-Bus like regular
routes. I think that makes sense, but it is a change in behavior, as
previously such routes were not exposed there.
2017-08-30 11:46:42 +02:00
|
|
|
* among all activating and already-activated devices is a
|
|
|
|
|
* still-activating one. */
|
|
|
|
|
return NULL;
|
|
|
|
|
}
|
|
|
|
|
|
2018-06-28 18:05:05 +02:00
|
|
|
return best_ac;
|
2010-04-23 22:11:22 -07:00
|
|
|
}
|
|
|
|
|
|
2017-03-10 15:17:02 +01:00
|
|
|
static gboolean
|
|
|
|
|
all_devices_not_active (NMPolicy *self)
|
|
|
|
|
{
|
|
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2018-03-29 08:52:45 +02:00
|
|
|
const CList *tmp_lst;
|
core: track devices in manager via embedded CList
Instead of using a GSList for tracking the devices, use a CList.
I think a CList is in most cases the more suitable data structure
then GSList:
- you can find out in O(1) whether the object is linked. That
is nice, for example to assert in NMDevice's destructor that
the object was unlinked, and we will use that later in
nm_manager_get_device_by_path().
- you can unlink the element in O(1) and you can unlink the
element without having access to the link's head
- Contrary to GSList, this does not require an extra slice
allocation for the link node. It quite possibliy consumes
slightly less memory because the CList structure is embedded
in a struct that we already allocate. Even if slice allocation
would be perfect to only consume 2*sizeof(gpointer) for the link
note, it would at most be as-good as CList. Quite possibly,
there is an overhead though.
- CList possibly has better memory locality, because the link
structure and the data are close to each other.
Something which could be seen as disavantage, is that with CList
one device can only be tracked in one NMManager instance at a time.
But that is fine. There exists only one NMManager instance for now,
and even if we would ever introduce multiple managers, we probably
would not associate one NMDevice instance with multiple managers.
The advantages are arguably not huge, but CList is IMHO clearly the
more suited data structure. No need to stick to a suboptimal data
structure for the job. Refactor it.
2018-03-23 21:51:07 +01:00
|
|
|
NMDevice *device;
|
2017-03-10 15:17:02 +01:00
|
|
|
|
2018-03-29 08:52:45 +02:00
|
|
|
nm_manager_for_each_device (priv->manager, device, tmp_lst) {
|
2017-03-10 15:17:02 +01:00
|
|
|
NMDeviceState state;
|
|
|
|
|
|
core: track devices in manager via embedded CList
Instead of using a GSList for tracking the devices, use a CList.
I think a CList is in most cases the more suitable data structure
then GSList:
- you can find out in O(1) whether the object is linked. That
is nice, for example to assert in NMDevice's destructor that
the object was unlinked, and we will use that later in
nm_manager_get_device_by_path().
- you can unlink the element in O(1) and you can unlink the
element without having access to the link's head
- Contrary to GSList, this does not require an extra slice
allocation for the link node. It quite possibliy consumes
slightly less memory because the CList structure is embedded
in a struct that we already allocate. Even if slice allocation
would be perfect to only consume 2*sizeof(gpointer) for the link
note, it would at most be as-good as CList. Quite possibly,
there is an overhead though.
- CList possibly has better memory locality, because the link
structure and the data are close to each other.
Something which could be seen as disavantage, is that with CList
one device can only be tracked in one NMManager instance at a time.
But that is fine. There exists only one NMManager instance for now,
and even if we would ever introduce multiple managers, we probably
would not associate one NMDevice instance with multiple managers.
The advantages are arguably not huge, but CList is IMHO clearly the
more suited data structure. No need to stick to a suboptimal data
structure for the job. Refactor it.
2018-03-23 21:51:07 +01:00
|
|
|
state = nm_device_get_state (device);
|
2017-03-10 15:17:02 +01:00
|
|
|
if ( state <= NM_DEVICE_STATE_DISCONNECTED
|
|
|
|
|
|| state >= NM_DEVICE_STATE_DEACTIVATING) {
|
|
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
return FALSE;
|
|
|
|
|
}
|
|
|
|
|
return TRUE;
|
|
|
|
|
}
|
|
|
|
|
|
2013-02-24 13:51:56 +01:00
|
|
|
#define FALLBACK_HOSTNAME4 "localhost.localdomain"
|
|
|
|
|
|
2016-03-29 10:46:07 +02:00
|
|
|
static void
|
|
|
|
|
settings_set_hostname_cb (const char *hostname,
|
|
|
|
|
gboolean result,
|
|
|
|
|
gpointer user_data)
|
2013-02-24 13:51:56 +01:00
|
|
|
{
|
2017-02-17 16:22:30 +01:00
|
|
|
NMPolicy *self = NM_POLICY (user_data);
|
|
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2016-02-18 18:19:30 +01:00
|
|
|
int ret = 0;
|
2019-01-31 13:29:21 +01:00
|
|
|
int errsv;
|
2016-02-18 18:19:30 +01:00
|
|
|
|
|
|
|
|
if (!result) {
|
2017-03-03 12:11:57 +01:00
|
|
|
_LOGT (LOGD_DNS, "set-hostname: hostname set via dbus failed, fallback to \"sethostname\"");
|
2016-02-18 18:19:30 +01:00
|
|
|
ret = sethostname (hostname, strlen (hostname));
|
|
|
|
|
if (ret != 0) {
|
2019-01-31 13:29:21 +01:00
|
|
|
errsv = errno;
|
2017-03-03 12:11:57 +01:00
|
|
|
_LOGW (LOGD_DNS, "set-hostname: couldn't set the system hostname to '%s': (%d) %s",
|
2019-01-31 17:22:18 +01:00
|
|
|
hostname, errsv, nm_strerror_native (errsv));
|
2016-02-18 18:19:30 +01:00
|
|
|
if (errsv == EPERM)
|
2017-03-03 12:11:57 +01:00
|
|
|
_LOGW (LOGD_DNS, "set-hostname: you should use hostnamed when systemd hardening is in effect!");
|
2016-02-18 18:19:30 +01:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2017-02-17 16:22:30 +01:00
|
|
|
priv->changing_hostname = FALSE;
|
2016-02-18 18:19:30 +01:00
|
|
|
if (!ret)
|
2017-03-08 14:11:22 +01:00
|
|
|
nm_dispatcher_call_hostname (NULL, NULL, NULL);
|
2017-02-17 16:22:30 +01:00
|
|
|
g_object_unref (self);
|
2016-02-18 18:19:30 +01:00
|
|
|
}
|
|
|
|
|
|
2017-02-17 15:33:27 +01:00
|
|
|
#define HOST_NAME_BUFSIZE (HOST_NAME_MAX + 2)
|
|
|
|
|
|
|
|
|
|
static char *
|
2017-04-23 19:54:39 +02:00
|
|
|
_get_hostname (NMPolicy *self)
|
2017-02-17 15:33:27 +01:00
|
|
|
{
|
|
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2017-04-23 19:54:39 +02:00
|
|
|
char *hostname = NULL;
|
2019-01-31 13:29:21 +01:00
|
|
|
int errsv;
|
2017-02-17 15:33:27 +01:00
|
|
|
|
2017-02-17 16:22:30 +01:00
|
|
|
/* If there is an in-progress hostname change, return
|
|
|
|
|
* the last hostname set as would be set soon...
|
|
|
|
|
*/
|
|
|
|
|
if (priv->changing_hostname) {
|
2017-03-03 12:11:57 +01:00
|
|
|
_LOGT (LOGD_DNS, "get-hostname: \"%s\" (last on set)", priv->last_hostname);
|
2017-04-23 19:54:39 +02:00
|
|
|
return g_strdup (priv->last_hostname);
|
2017-02-17 16:22:30 +01:00
|
|
|
}
|
|
|
|
|
|
2017-02-17 15:33:27 +01:00
|
|
|
/* try to get the hostname via dbus... */
|
2017-04-23 19:54:39 +02:00
|
|
|
if (nm_hostname_manager_get_transient_hostname (priv->hostname_manager, &hostname)) {
|
|
|
|
|
_LOGT (LOGD_DNS, "get-hostname: \"%s\" (from dbus)", hostname);
|
|
|
|
|
return hostname;
|
2017-02-17 15:33:27 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* ...or retrieve it by yourself */
|
2017-04-23 19:54:39 +02:00
|
|
|
hostname = g_malloc (HOST_NAME_BUFSIZE);
|
|
|
|
|
if (gethostname (hostname, HOST_NAME_BUFSIZE -1) != 0) {
|
2019-01-31 13:29:21 +01:00
|
|
|
errsv = errno;
|
2017-03-03 12:11:57 +01:00
|
|
|
_LOGT (LOGD_DNS, "get-hostname: couldn't get the system hostname: (%d) %s",
|
2019-01-31 17:08:03 +01:00
|
|
|
errsv, nm_strerror_native (errsv));
|
2017-04-23 19:54:39 +02:00
|
|
|
g_free (hostname);
|
2017-02-17 15:33:27 +01:00
|
|
|
return NULL;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* the name may be truncated... */
|
2017-04-23 19:54:39 +02:00
|
|
|
hostname[HOST_NAME_BUFSIZE - 1] = '\0';
|
|
|
|
|
if (strlen (hostname) >= HOST_NAME_BUFSIZE -1) {
|
|
|
|
|
_LOGT (LOGD_DNS, "get-hostname: system hostname too long: \"%s\"", hostname);
|
|
|
|
|
g_free (hostname);
|
2017-02-17 15:33:27 +01:00
|
|
|
return NULL;
|
|
|
|
|
}
|
|
|
|
|
|
2017-04-23 19:54:39 +02:00
|
|
|
_LOGT (LOGD_DNS, "get-hostname: \"%s\"", hostname);
|
|
|
|
|
return hostname;
|
2017-02-17 15:33:27 +01:00
|
|
|
}
|
|
|
|
|
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
static void
|
2016-03-29 10:46:07 +02:00
|
|
|
_set_hostname (NMPolicy *self,
|
2010-06-17 23:01:40 -07:00
|
|
|
const char *new_hostname,
|
|
|
|
|
const char *msg)
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
{
|
2016-03-29 10:46:07 +02:00
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2017-02-17 15:33:27 +01:00
|
|
|
gs_free char *old_hostname = NULL;
|
2016-02-22 10:10:15 +01:00
|
|
|
const char *name;
|
2013-08-22 10:10:17 -04:00
|
|
|
|
2011-11-02 15:21:52 -05:00
|
|
|
/* The incoming hostname *can* be NULL, which will get translated to
|
|
|
|
|
* 'localhost.localdomain' or such in the hostname policy code, but we
|
|
|
|
|
* keep cur_hostname = NULL in the case because we need to know that
|
|
|
|
|
* there was no valid hostname to start with.
|
|
|
|
|
*/
|
|
|
|
|
|
2015-04-17 09:07:15 +02:00
|
|
|
/* Clear lookup addresses if we have a hostname, so that we don't
|
2013-02-24 13:51:56 +01:00
|
|
|
* restart the reverse lookup thread later.
|
2013-02-05 17:29:15 +01:00
|
|
|
*/
|
2013-02-24 13:51:56 +01:00
|
|
|
if (new_hostname)
|
2017-04-23 19:30:07 +02:00
|
|
|
g_clear_object (&priv->lookup.addr);
|
2013-02-05 17:29:15 +01:00
|
|
|
|
2017-02-27 15:47:43 +01:00
|
|
|
/* Update the DNS only if the hostname is actually
|
|
|
|
|
* going to change.
|
|
|
|
|
*/
|
|
|
|
|
if (!nm_streq0 (priv->cur_hostname, new_hostname)) {
|
2016-08-17 18:40:17 +02:00
|
|
|
g_free (priv->cur_hostname);
|
|
|
|
|
priv->cur_hostname = g_strdup (new_hostname);
|
|
|
|
|
|
|
|
|
|
/* Notify the DNS manager of the hostname change so that the domain part, if
|
|
|
|
|
* present, can be added to the search list.
|
|
|
|
|
*/
|
2017-03-10 15:17:02 +01:00
|
|
|
nm_dns_manager_set_hostname (priv->dns_manager, priv->cur_hostname,
|
|
|
|
|
all_devices_not_active (self));
|
2016-08-17 18:40:17 +02:00
|
|
|
}
|
2010-06-17 23:01:40 -07:00
|
|
|
|
2016-02-22 10:10:15 +01:00
|
|
|
/* Finally, set kernel hostname */
|
2016-08-17 18:40:17 +02:00
|
|
|
if (!new_hostname)
|
2016-02-22 10:10:15 +01:00
|
|
|
name = FALLBACK_HOSTNAME4;
|
2016-08-17 18:40:17 +02:00
|
|
|
else if (!new_hostname[0]) {
|
2016-02-22 10:10:15 +01:00
|
|
|
g_warn_if_reached ();
|
|
|
|
|
name = FALLBACK_HOSTNAME4;
|
|
|
|
|
} else
|
2016-08-17 18:40:17 +02:00
|
|
|
name = new_hostname;
|
2016-02-22 10:10:15 +01:00
|
|
|
|
2017-02-17 15:33:27 +01:00
|
|
|
/* Don't set the hostname if it isn't actually changing */
|
2017-04-23 19:54:39 +02:00
|
|
|
if ( (old_hostname = _get_hostname (self))
|
2017-02-17 15:33:27 +01:00
|
|
|
&& (nm_streq (name, old_hostname))) {
|
2017-03-03 12:11:57 +01:00
|
|
|
_LOGT (LOGD_DNS, "set-hostname: hostname already set to '%s' (%s)", name, msg);
|
2017-02-17 15:33:27 +01:00
|
|
|
return;
|
2016-02-22 10:10:15 +01:00
|
|
|
}
|
|
|
|
|
|
2017-02-17 16:22:30 +01:00
|
|
|
/* Keep track of the last set hostname */
|
|
|
|
|
g_free (priv->last_hostname);
|
|
|
|
|
priv->last_hostname = g_strdup (name);
|
|
|
|
|
priv->changing_hostname = TRUE;
|
|
|
|
|
|
2017-03-03 12:11:57 +01:00
|
|
|
_LOGI (LOGD_DNS, "set-hostname: set hostname to '%s' (%s)", name, msg);
|
2016-02-22 10:10:15 +01:00
|
|
|
|
|
|
|
|
/* Ask NMSettings to update the transient hostname using its
|
|
|
|
|
* systemd-hostnamed proxy */
|
2017-04-23 00:40:46 +02:00
|
|
|
nm_hostname_manager_set_transient_hostname (priv->hostname_manager,
|
|
|
|
|
name,
|
|
|
|
|
settings_set_hostname_cb,
|
|
|
|
|
g_object_ref (self));
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void
|
2013-02-24 13:51:56 +01:00
|
|
|
lookup_callback (GObject *source,
|
|
|
|
|
GAsyncResult *result,
|
2010-04-19 16:59:34 -07:00
|
|
|
gpointer user_data)
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
{
|
2017-04-23 19:14:19 +02:00
|
|
|
NMPolicy *self;
|
|
|
|
|
NMPolicyPrivate *priv;
|
|
|
|
|
gs_free char *hostname = NULL;
|
|
|
|
|
gs_free_error GError *error = NULL;
|
2013-02-24 13:51:56 +01:00
|
|
|
|
|
|
|
|
hostname = g_resolver_lookup_by_address_finish (G_RESOLVER (source), result, &error);
|
2017-04-23 19:14:19 +02:00
|
|
|
if (g_error_matches (error, G_IO_ERROR, G_IO_ERROR_CANCELLED))
|
2013-02-24 13:51:56 +01:00
|
|
|
return;
|
2017-04-23 19:14:19 +02:00
|
|
|
|
|
|
|
|
self = user_data;
|
|
|
|
|
priv = NM_POLICY_GET_PRIVATE (self);
|
|
|
|
|
|
2017-04-23 19:30:07 +02:00
|
|
|
g_clear_object (&priv->lookup.cancellable);
|
2013-02-24 13:51:56 +01:00
|
|
|
|
|
|
|
|
if (hostname)
|
2016-03-29 10:46:07 +02:00
|
|
|
_set_hostname (self, hostname, "from address lookup");
|
2017-04-23 19:14:19 +02:00
|
|
|
else
|
2016-03-29 10:46:07 +02:00
|
|
|
_set_hostname (self, NULL, error->message);
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
}
|
|
|
|
|
|
2017-04-23 19:24:21 +02:00
|
|
|
static void
|
|
|
|
|
lookup_by_address (NMPolicy *self)
|
|
|
|
|
{
|
|
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
|
|
|
|
|
2017-04-23 19:30:07 +02:00
|
|
|
nm_clear_g_cancellable (&priv->lookup.cancellable);
|
|
|
|
|
priv->lookup.cancellable = g_cancellable_new ();
|
|
|
|
|
g_resolver_lookup_by_address_async (priv->lookup.resolver,
|
|
|
|
|
priv->lookup.addr,
|
|
|
|
|
priv->lookup.cancellable,
|
2017-04-23 19:24:21 +02:00
|
|
|
lookup_callback, self);
|
|
|
|
|
}
|
|
|
|
|
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
static void
|
2017-09-06 13:01:55 +02:00
|
|
|
update_system_hostname (NMPolicy *self, const char *msg)
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
{
|
2016-03-29 10:46:07 +02:00
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2017-04-23 14:20:37 +02:00
|
|
|
const char *configured_hostname;
|
2017-02-17 16:22:30 +01:00
|
|
|
gs_free char *temp_hostname = NULL;
|
2010-06-08 10:40:08 -07:00
|
|
|
const char *dhcp_hostname, *p;
|
policy: ignore nameservers when starting lookup thread (rh #1031763)
When generating connections at startup for active interfaces, the
generation code may not always be able to read DNS information for
the connection. Thus, the device's IP4Config won't have any
nameservers and the device won't be considered for reverse-address
lookup. However, since any device that gets this far is already
the "best" device and has the default route, and thus should be the
one used for reverse-address lookup.
Second, reorganize the code better handle dual-stack in the
future by checking the IP configs directly, instead of the
devices. Since 'best4' and 'best6' may be different devices,
we want to operate on the IP configs, not devices, to handle
situations where the best IP4Config may not be suitable for
reverse lookup, but the best IP6Config is.
https://bugzilla.redhat.com/show_bug.cgi?id=1031763
2013-11-19 19:40:28 -06:00
|
|
|
NMIP4Config *ip4_config;
|
|
|
|
|
NMIP6Config *ip6_config;
|
2017-02-17 16:22:30 +01:00
|
|
|
gboolean external_hostname = FALSE;
|
2017-07-07 23:34:41 +02:00
|
|
|
const NMPlatformIP4Address *addr4;
|
|
|
|
|
const NMPlatformIP6Address *addr6;
|
2018-06-28 18:05:05 +02:00
|
|
|
NMDevice *device;
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
|
2016-03-29 10:46:07 +02:00
|
|
|
g_return_if_fail (self != NULL);
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
|
2017-03-03 12:11:57 +01:00
|
|
|
if (priv->hostname_mode == NM_POLICY_HOSTNAME_MODE_NONE) {
|
|
|
|
|
_LOGT (LOGD_DNS, "set-hostname: hostname is unmanaged");
|
2017-03-02 19:18:49 +01:00
|
|
|
return;
|
2017-03-03 12:11:57 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
_LOGT (LOGD_DNS, "set-hostname: updating hostname (%s)", msg);
|
2017-03-02 19:18:49 +01:00
|
|
|
|
2017-04-23 19:30:07 +02:00
|
|
|
nm_clear_g_cancellable (&priv->lookup.cancellable);
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
|
2017-02-17 16:22:30 +01:00
|
|
|
/* Check if the hostname was set externally to NM, so that in that case
|
|
|
|
|
* we can avoid to fallback to the one we got when we started.
|
|
|
|
|
* Consider "not specific" hostnames as equal. */
|
2017-04-23 19:54:39 +02:00
|
|
|
if ( (temp_hostname = _get_hostname (self))
|
2017-02-17 16:22:30 +01:00
|
|
|
&& !nm_streq0 (temp_hostname, priv->last_hostname)
|
|
|
|
|
&& ( nm_utils_is_specific_hostname (temp_hostname)
|
|
|
|
|
|| nm_utils_is_specific_hostname (priv->last_hostname))) {
|
|
|
|
|
external_hostname = TRUE;
|
2017-03-03 12:11:57 +01:00
|
|
|
_LOGI (LOGD_DNS, "set-hostname: current hostname was changed outside NetworkManager: '%s'",
|
2017-02-17 16:22:30 +01:00
|
|
|
temp_hostname);
|
2017-03-03 18:32:08 +01:00
|
|
|
priv->dhcp_hostname = FALSE;
|
2017-02-17 16:22:30 +01:00
|
|
|
|
2017-04-23 20:31:31 +02:00
|
|
|
if (!nm_utils_is_specific_hostname (temp_hostname))
|
|
|
|
|
nm_clear_g_free (&temp_hostname);
|
2017-02-17 16:22:30 +01:00
|
|
|
if (!nm_streq0 (temp_hostname, priv->orig_hostname)) {
|
|
|
|
|
/* Update original (fallback) hostname */
|
|
|
|
|
g_free (priv->orig_hostname);
|
2017-04-23 20:31:31 +02:00
|
|
|
priv->orig_hostname = g_steal_pointer (&temp_hostname);
|
|
|
|
|
_LOGT (LOGD_DNS, "hostname-original: update to %s%s%s",
|
|
|
|
|
NM_PRINT_FMT_QUOTE_STRING (priv->orig_hostname));
|
2017-02-17 16:22:30 +01:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2010-01-27 17:13:35 -08:00
|
|
|
/* Hostname precedence order:
|
|
|
|
|
*
|
2010-10-27 20:22:14 -05:00
|
|
|
* 1) a configured hostname (from settings)
|
2010-01-27 17:13:35 -08:00
|
|
|
* 2) automatic hostname from the default device's config (DHCP, VPN, etc)
|
2017-02-17 16:22:30 +01:00
|
|
|
* 3) the last hostname set outside NM
|
2010-01-27 17:13:35 -08:00
|
|
|
* 4) reverse-DNS of the best device's IPv4 address
|
|
|
|
|
*
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
*/
|
|
|
|
|
|
2010-01-27 17:13:35 -08:00
|
|
|
/* Try a persistent hostname first */
|
2017-04-23 14:20:37 +02:00
|
|
|
configured_hostname = nm_hostname_manager_get_hostname (priv->hostname_manager);
|
2014-06-30 13:59:49 +02:00
|
|
|
if (configured_hostname && nm_utils_is_specific_hostname (configured_hostname)) {
|
2016-03-29 10:46:07 +02:00
|
|
|
_set_hostname (self, configured_hostname, "from system configuration");
|
2017-03-03 18:32:08 +01:00
|
|
|
priv->dhcp_hostname = FALSE;
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
2018-06-28 18:05:05 +02:00
|
|
|
if (priv->default_ac4) {
|
libnm, core, cli, tui: fix the capitalization of various types
GLib/Gtk have mostly settled on the convention that two-letter
acronyms in type names remain all-caps (eg, "IO"), but longer acronyms
become initial-caps-only (eg, "Tcp").
NM was inconsistent, with most long acronyms using initial caps only
(Adsl, Cdma, Dcb, Gsm, Olpc, Vlan), but others using all caps (DHCP,
PPP, PPPOE, VPN). Fix libnm and src/ to use initial-caps only for all
three-or-more-letter-long acronyms (and update nmcli and nmtui for the
libnm changes).
2014-06-26 13:44:36 -04:00
|
|
|
NMDhcp4Config *dhcp4_config;
|
2010-06-08 10:40:08 -07:00
|
|
|
|
|
|
|
|
/* Grab a hostname out of the device's DHCP4 config */
|
2018-06-28 18:05:05 +02:00
|
|
|
dhcp4_config = nm_device_get_dhcp4_config (get_default_device (self, AF_INET));
|
2010-06-08 10:40:08 -07:00
|
|
|
if (dhcp4_config) {
|
2017-09-06 13:13:22 +02:00
|
|
|
dhcp_hostname = nm_dhcp4_config_get_option (dhcp4_config, "host_name");
|
|
|
|
|
if (dhcp_hostname && dhcp_hostname[0]) {
|
|
|
|
|
p = nm_str_skip_leading_spaces (dhcp_hostname);
|
|
|
|
|
if (p[0]) {
|
|
|
|
|
_set_hostname (self, p, "from DHCPv4");
|
|
|
|
|
priv->dhcp_hostname = TRUE;
|
|
|
|
|
return;
|
2010-01-27 17:13:35 -08:00
|
|
|
}
|
2017-03-03 12:11:57 +01:00
|
|
|
_LOGW (LOGD_DNS, "set-hostname: DHCPv4-provided hostname '%s' looks invalid; ignoring it",
|
2016-03-03 09:20:18 +01:00
|
|
|
dhcp_hostname);
|
2010-06-08 10:40:08 -07:00
|
|
|
}
|
|
|
|
|
}
|
2017-09-06 13:21:39 +02:00
|
|
|
}
|
|
|
|
|
|
2018-06-28 18:05:05 +02:00
|
|
|
if (priv->default_ac6) {
|
libnm, core, cli, tui: fix the capitalization of various types
GLib/Gtk have mostly settled on the convention that two-letter
acronyms in type names remain all-caps (eg, "IO"), but longer acronyms
become initial-caps-only (eg, "Tcp").
NM was inconsistent, with most long acronyms using initial caps only
(Adsl, Cdma, Dcb, Gsm, Olpc, Vlan), but others using all caps (DHCP,
PPP, PPPOE, VPN). Fix libnm and src/ to use initial-caps only for all
three-or-more-letter-long acronyms (and update nmcli and nmtui for the
libnm changes).
2014-06-26 13:44:36 -04:00
|
|
|
NMDhcp6Config *dhcp6_config;
|
2010-06-08 10:40:08 -07:00
|
|
|
|
2011-07-12 14:47:33 +02:00
|
|
|
/* Grab a hostname out of the device's DHCP6 config */
|
2018-06-28 18:05:05 +02:00
|
|
|
dhcp6_config = nm_device_get_dhcp6_config (get_default_device (self, AF_INET6));
|
2010-06-08 10:40:08 -07:00
|
|
|
if (dhcp6_config) {
|
2017-09-06 13:13:22 +02:00
|
|
|
dhcp_hostname = nm_dhcp6_config_get_option (dhcp6_config, "host_name");
|
|
|
|
|
if (dhcp_hostname && dhcp_hostname[0]) {
|
|
|
|
|
p = nm_str_skip_leading_spaces (dhcp_hostname);
|
|
|
|
|
if (p[0]) {
|
|
|
|
|
_set_hostname (self, p, "from DHCPv6");
|
|
|
|
|
priv->dhcp_hostname = TRUE;
|
|
|
|
|
return;
|
2010-06-08 10:40:08 -07:00
|
|
|
}
|
2017-03-03 12:11:57 +01:00
|
|
|
_LOGW (LOGD_DNS, "set-hostname: DHCPv6-provided hostname '%s' looks invalid; ignoring it",
|
2016-03-03 09:20:18 +01:00
|
|
|
dhcp_hostname);
|
2010-01-27 17:13:35 -08:00
|
|
|
}
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2017-02-17 16:22:30 +01:00
|
|
|
/* If an hostname was set outside NetworkManager keep it */
|
|
|
|
|
if (external_hostname)
|
|
|
|
|
return;
|
|
|
|
|
|
2017-03-03 18:32:08 +01:00
|
|
|
if (priv->hostname_mode == NM_POLICY_HOSTNAME_MODE_DHCP) {
|
|
|
|
|
/* In dhcp hostname-mode, the hostname is updated only if it comes from
|
|
|
|
|
* a DHCP host-name option: if last set was from a host-name option and
|
|
|
|
|
* we are here than that connection is gone (with its host-name option),
|
|
|
|
|
* so reset the hostname to the previous value
|
|
|
|
|
*/
|
|
|
|
|
if (priv->dhcp_hostname) {
|
|
|
|
|
_set_hostname (self, priv->orig_hostname, "reset dhcp hostname");
|
|
|
|
|
priv->dhcp_hostname = FALSE;
|
|
|
|
|
}
|
2017-03-02 19:18:49 +01:00
|
|
|
return;
|
2017-03-03 18:32:08 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
priv->dhcp_hostname = FALSE;
|
2017-03-02 19:18:49 +01:00
|
|
|
|
2018-06-28 18:05:05 +02:00
|
|
|
if (!priv->default_ac4 && !priv->default_ac6) {
|
2017-02-17 16:22:30 +01:00
|
|
|
/* No best device; fall back to the last hostname set externally
|
|
|
|
|
* to NM or if there wasn't one, 'localhost.localdomain'
|
|
|
|
|
*/
|
|
|
|
|
_set_hostname (self, priv->orig_hostname, "no default device");
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* If no automatically-configured hostname, try using the last hostname
|
|
|
|
|
* set externally to NM
|
2010-01-27 17:13:35 -08:00
|
|
|
*/
|
2013-08-22 10:10:17 -04:00
|
|
|
if (priv->orig_hostname) {
|
2016-03-29 10:46:07 +02:00
|
|
|
_set_hostname (self, priv->orig_hostname, "from system startup");
|
2010-01-27 17:13:35 -08:00
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
2010-06-08 10:40:08 -07:00
|
|
|
/* No configured hostname, no automatically determined hostname, and no
|
|
|
|
|
* bootup hostname. Start reverse DNS of the current IPv4 or IPv6 address.
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
*/
|
2018-06-28 18:05:05 +02:00
|
|
|
device = get_default_device (self, AF_INET);
|
|
|
|
|
ip4_config = device ? nm_device_get_ip4_config (device) : NULL;
|
|
|
|
|
|
|
|
|
|
device = get_default_device (self, AF_INET6);
|
|
|
|
|
ip6_config = device ? nm_device_get_ip6_config (device) : NULL;
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
|
2017-07-07 23:34:41 +02:00
|
|
|
if ( ip4_config
|
|
|
|
|
&& (addr4 = nm_ip4_config_get_first_address (ip4_config))) {
|
2017-04-23 19:30:07 +02:00
|
|
|
g_clear_object (&priv->lookup.addr);
|
|
|
|
|
priv->lookup.addr = g_inet_address_new_from_bytes ((guint8 *) &addr4->address,
|
2014-01-16 14:14:42 -05:00
|
|
|
G_SOCKET_FAMILY_IPV4);
|
2017-07-07 23:34:41 +02:00
|
|
|
} else if ( ip6_config
|
|
|
|
|
&& (addr6 = nm_ip6_config_get_first_address (ip6_config))) {
|
2017-04-23 19:30:07 +02:00
|
|
|
g_clear_object (&priv->lookup.addr);
|
|
|
|
|
priv->lookup.addr = g_inet_address_new_from_bytes ((guint8 *) &addr6->address,
|
2014-01-16 14:14:42 -05:00
|
|
|
G_SOCKET_FAMILY_IPV6);
|
policy: ignore nameservers when starting lookup thread (rh #1031763)
When generating connections at startup for active interfaces, the
generation code may not always be able to read DNS information for
the connection. Thus, the device's IP4Config won't have any
nameservers and the device won't be considered for reverse-address
lookup. However, since any device that gets this far is already
the "best" device and has the default route, and thus should be the
one used for reverse-address lookup.
Second, reorganize the code better handle dual-stack in the
future by checking the IP configs directly, instead of the
devices. Since 'best4' and 'best6' may be different devices,
we want to operate on the IP configs, not devices, to handle
situations where the best IP4Config may not be suitable for
reverse lookup, but the best IP6Config is.
https://bugzilla.redhat.com/show_bug.cgi?id=1031763
2013-11-19 19:40:28 -06:00
|
|
|
} else {
|
2014-01-16 14:14:42 -05:00
|
|
|
/* No valid IP config; fall back to localhost.localdomain */
|
2016-03-29 10:46:07 +02:00
|
|
|
_set_hostname (self, NULL, "no IP config");
|
policy: ignore nameservers when starting lookup thread (rh #1031763)
When generating connections at startup for active interfaces, the
generation code may not always be able to read DNS information for
the connection. Thus, the device's IP4Config won't have any
nameservers and the device won't be considered for reverse-address
lookup. However, since any device that gets this far is already
the "best" device and has the default route, and thus should be the
one used for reverse-address lookup.
Second, reorganize the code better handle dual-stack in the
future by checking the IP configs directly, instead of the
devices. Since 'best4' and 'best6' may be different devices,
we want to operate on the IP configs, not devices, to handle
situations where the best IP4Config may not be suitable for
reverse lookup, but the best IP6Config is.
https://bugzilla.redhat.com/show_bug.cgi?id=1031763
2013-11-19 19:40:28 -06:00
|
|
|
return;
|
2010-06-08 10:40:08 -07:00
|
|
|
}
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
|
2017-04-23 19:24:21 +02:00
|
|
|
lookup_by_address (self);
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void
|
2016-03-29 10:46:07 +02:00
|
|
|
update_default_ac (NMPolicy *self,
|
2017-11-23 21:53:04 +01:00
|
|
|
int addr_family,
|
|
|
|
|
NMActiveConnection *best)
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
{
|
2016-03-29 10:46:07 +02:00
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2017-11-23 21:30:09 +01:00
|
|
|
const CList *tmp_list;
|
|
|
|
|
NMActiveConnection *ac;
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
|
2012-05-29 10:01:07 -05:00
|
|
|
/* Clear the 'default[6]' flag on all active connections that aren't the new
|
|
|
|
|
* default active connection. We'll set the new default after; this ensures
|
|
|
|
|
* we don't ever have two marked 'default[6]' simultaneously.
|
|
|
|
|
*/
|
2017-11-23 21:30:09 +01:00
|
|
|
nm_manager_for_each_active_connection (priv->manager, ac, tmp_list) {
|
|
|
|
|
if (ac != best)
|
2017-11-23 21:53:04 +01:00
|
|
|
nm_active_connection_set_default (ac, addr_family, FALSE);
|
2012-05-29 10:01:07 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Mark new default active connection */
|
|
|
|
|
if (best)
|
2017-11-23 21:53:04 +01:00
|
|
|
nm_active_connection_set_default (best, addr_family, TRUE);
|
2012-05-29 10:01:07 -05:00
|
|
|
}
|
|
|
|
|
|
core: rework handling of default-routes and drop NMDefaultRouteManager
Remove NMDefaultRouteManager. Instead, add the default-route to the
NMIP4Config/NMIP6Config instance.
This basically reverts commit e8824f6a5205ffcf761abd3e0897a22b254c7797.
We added NMDefaultRouteManager because we used the corresponding to `ip
route replace` when configuring routes. That would replace default-routes
on other interfaces so we needed a central manager to coordinate routes.
Now, we use the corresponding of `ip route append` to configure routes,
and each interface can configure routes indepdentently.
In NMDevice, when creating the default-route, ignore @auto_method for
external devices. We shall not touch these devices.
Especially the code in NMPolicy regarding selection of the best-device
seems wrong. It probably needs further adjustments in the future.
Especially get_best_ip_config() should be replaced, because this
distinction VPN vs. devices seems wrong to me.
Thereby, remove the @ignore_never_default argument. It was added by
commit bb750260045239ab85574366bae8102eff8058cc, I don't think it's
needed anymore.
This brings another change. Now that we track default-routes in
NMIP4Config/NMIP6Config, they are also exposed on D-Bus like regular
routes. I think that makes sense, but it is a change in behavior, as
previously such routes were not exposed there.
2017-08-30 11:46:42 +02:00
|
|
|
static gpointer
|
|
|
|
|
get_best_ip_config (NMPolicy *self,
|
|
|
|
|
int addr_family,
|
|
|
|
|
const char **out_ip_iface,
|
|
|
|
|
NMActiveConnection **out_ac,
|
|
|
|
|
NMDevice **out_device,
|
|
|
|
|
NMVpnConnection **out_vpn)
|
2012-05-29 10:01:07 -05:00
|
|
|
{
|
core: rework handling of default-routes and drop NMDefaultRouteManager
Remove NMDefaultRouteManager. Instead, add the default-route to the
NMIP4Config/NMIP6Config instance.
This basically reverts commit e8824f6a5205ffcf761abd3e0897a22b254c7797.
We added NMDefaultRouteManager because we used the corresponding to `ip
route replace` when configuring routes. That would replace default-routes
on other interfaces so we needed a central manager to coordinate routes.
Now, we use the corresponding of `ip route append` to configure routes,
and each interface can configure routes indepdentently.
In NMDevice, when creating the default-route, ignore @auto_method for
external devices. We shall not touch these devices.
Especially the code in NMPolicy regarding selection of the best-device
seems wrong. It probably needs further adjustments in the future.
Especially get_best_ip_config() should be replaced, because this
distinction VPN vs. devices seems wrong to me.
Thereby, remove the @ignore_never_default argument. It was added by
commit bb750260045239ab85574366bae8102eff8058cc, I don't think it's
needed anymore.
This brings another change. Now that we track default-routes in
NMIP4Config/NMIP6Config, they are also exposed on D-Bus like regular
routes. I think that makes sense, but it is a change in behavior, as
previously such routes were not exposed there.
2017-08-30 11:46:42 +02:00
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2018-06-28 17:16:15 +02:00
|
|
|
gpointer conf, best_conf = NULL;
|
2017-11-23 21:30:09 +01:00
|
|
|
const CList *tmp_list;
|
|
|
|
|
NMActiveConnection *ac;
|
2018-06-28 17:16:15 +02:00
|
|
|
guint64 best_metric = G_MAXUINT64;
|
|
|
|
|
NMVpnConnection *best_vpn = NULL;
|
core: rework handling of default-routes and drop NMDefaultRouteManager
Remove NMDefaultRouteManager. Instead, add the default-route to the
NMIP4Config/NMIP6Config instance.
This basically reverts commit e8824f6a5205ffcf761abd3e0897a22b254c7797.
We added NMDefaultRouteManager because we used the corresponding to `ip
route replace` when configuring routes. That would replace default-routes
on other interfaces so we needed a central manager to coordinate routes.
Now, we use the corresponding of `ip route append` to configure routes,
and each interface can configure routes indepdentently.
In NMDevice, when creating the default-route, ignore @auto_method for
external devices. We shall not touch these devices.
Especially the code in NMPolicy regarding selection of the best-device
seems wrong. It probably needs further adjustments in the future.
Especially get_best_ip_config() should be replaced, because this
distinction VPN vs. devices seems wrong to me.
Thereby, remove the @ignore_never_default argument. It was added by
commit bb750260045239ab85574366bae8102eff8058cc, I don't think it's
needed anymore.
This brings another change. Now that we track default-routes in
NMIP4Config/NMIP6Config, they are also exposed on D-Bus like regular
routes. I think that makes sense, but it is a change in behavior, as
previously such routes were not exposed there.
2017-08-30 11:46:42 +02:00
|
|
|
|
|
|
|
|
nm_assert (NM_IN_SET (addr_family, AF_INET, AF_INET6));
|
|
|
|
|
|
2017-11-23 21:30:09 +01:00
|
|
|
nm_manager_for_each_active_connection (priv->manager, ac, tmp_list) {
|
core: rework handling of default-routes and drop NMDefaultRouteManager
Remove NMDefaultRouteManager. Instead, add the default-route to the
NMIP4Config/NMIP6Config instance.
This basically reverts commit e8824f6a5205ffcf761abd3e0897a22b254c7797.
We added NMDefaultRouteManager because we used the corresponding to `ip
route replace` when configuring routes. That would replace default-routes
on other interfaces so we needed a central manager to coordinate routes.
Now, we use the corresponding of `ip route append` to configure routes,
and each interface can configure routes indepdentently.
In NMDevice, when creating the default-route, ignore @auto_method for
external devices. We shall not touch these devices.
Especially the code in NMPolicy regarding selection of the best-device
seems wrong. It probably needs further adjustments in the future.
Especially get_best_ip_config() should be replaced, because this
distinction VPN vs. devices seems wrong to me.
Thereby, remove the @ignore_never_default argument. It was added by
commit bb750260045239ab85574366bae8102eff8058cc, I don't think it's
needed anymore.
This brings another change. Now that we track default-routes in
NMIP4Config/NMIP6Config, they are also exposed on D-Bus like regular
routes. I think that makes sense, but it is a change in behavior, as
previously such routes were not exposed there.
2017-08-30 11:46:42 +02:00
|
|
|
NMVpnConnection *candidate;
|
|
|
|
|
NMVpnConnectionState vpn_state;
|
2018-06-28 17:16:15 +02:00
|
|
|
const NMPObject *obj;
|
|
|
|
|
guint32 metric;
|
core: rework handling of default-routes and drop NMDefaultRouteManager
Remove NMDefaultRouteManager. Instead, add the default-route to the
NMIP4Config/NMIP6Config instance.
This basically reverts commit e8824f6a5205ffcf761abd3e0897a22b254c7797.
We added NMDefaultRouteManager because we used the corresponding to `ip
route replace` when configuring routes. That would replace default-routes
on other interfaces so we needed a central manager to coordinate routes.
Now, we use the corresponding of `ip route append` to configure routes,
and each interface can configure routes indepdentently.
In NMDevice, when creating the default-route, ignore @auto_method for
external devices. We shall not touch these devices.
Especially the code in NMPolicy regarding selection of the best-device
seems wrong. It probably needs further adjustments in the future.
Especially get_best_ip_config() should be replaced, because this
distinction VPN vs. devices seems wrong to me.
Thereby, remove the @ignore_never_default argument. It was added by
commit bb750260045239ab85574366bae8102eff8058cc, I don't think it's
needed anymore.
This brings another change. Now that we track default-routes in
NMIP4Config/NMIP6Config, they are also exposed on D-Bus like regular
routes. I think that makes sense, but it is a change in behavior, as
previously such routes were not exposed there.
2017-08-30 11:46:42 +02:00
|
|
|
|
2017-11-23 21:30:09 +01:00
|
|
|
if (!NM_IS_VPN_CONNECTION (ac))
|
core: rework handling of default-routes and drop NMDefaultRouteManager
Remove NMDefaultRouteManager. Instead, add the default-route to the
NMIP4Config/NMIP6Config instance.
This basically reverts commit e8824f6a5205ffcf761abd3e0897a22b254c7797.
We added NMDefaultRouteManager because we used the corresponding to `ip
route replace` when configuring routes. That would replace default-routes
on other interfaces so we needed a central manager to coordinate routes.
Now, we use the corresponding of `ip route append` to configure routes,
and each interface can configure routes indepdentently.
In NMDevice, when creating the default-route, ignore @auto_method for
external devices. We shall not touch these devices.
Especially the code in NMPolicy regarding selection of the best-device
seems wrong. It probably needs further adjustments in the future.
Especially get_best_ip_config() should be replaced, because this
distinction VPN vs. devices seems wrong to me.
Thereby, remove the @ignore_never_default argument. It was added by
commit bb750260045239ab85574366bae8102eff8058cc, I don't think it's
needed anymore.
This brings another change. Now that we track default-routes in
NMIP4Config/NMIP6Config, they are also exposed on D-Bus like regular
routes. I think that makes sense, but it is a change in behavior, as
previously such routes were not exposed there.
2017-08-30 11:46:42 +02:00
|
|
|
continue;
|
|
|
|
|
|
2017-11-23 21:30:09 +01:00
|
|
|
candidate = NM_VPN_CONNECTION (ac);
|
core: rework handling of default-routes and drop NMDefaultRouteManager
Remove NMDefaultRouteManager. Instead, add the default-route to the
NMIP4Config/NMIP6Config instance.
This basically reverts commit e8824f6a5205ffcf761abd3e0897a22b254c7797.
We added NMDefaultRouteManager because we used the corresponding to `ip
route replace` when configuring routes. That would replace default-routes
on other interfaces so we needed a central manager to coordinate routes.
Now, we use the corresponding of `ip route append` to configure routes,
and each interface can configure routes indepdentently.
In NMDevice, when creating the default-route, ignore @auto_method for
external devices. We shall not touch these devices.
Especially the code in NMPolicy regarding selection of the best-device
seems wrong. It probably needs further adjustments in the future.
Especially get_best_ip_config() should be replaced, because this
distinction VPN vs. devices seems wrong to me.
Thereby, remove the @ignore_never_default argument. It was added by
commit bb750260045239ab85574366bae8102eff8058cc, I don't think it's
needed anymore.
This brings another change. Now that we track default-routes in
NMIP4Config/NMIP6Config, they are also exposed on D-Bus like regular
routes. I think that makes sense, but it is a change in behavior, as
previously such routes were not exposed there.
2017-08-30 11:46:42 +02:00
|
|
|
|
|
|
|
|
vpn_state = nm_vpn_connection_get_vpn_state (candidate);
|
|
|
|
|
if (vpn_state != NM_VPN_CONNECTION_STATE_ACTIVATED)
|
|
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
if (addr_family == AF_INET)
|
|
|
|
|
conf = nm_vpn_connection_get_ip4_config (candidate);
|
|
|
|
|
else
|
|
|
|
|
conf = nm_vpn_connection_get_ip6_config (candidate);
|
|
|
|
|
if (!conf)
|
|
|
|
|
continue;
|
|
|
|
|
|
2018-06-28 17:16:15 +02:00
|
|
|
if (addr_family == AF_INET)
|
|
|
|
|
obj = nm_ip4_config_best_default_route_get (conf);
|
|
|
|
|
else
|
|
|
|
|
obj = nm_ip6_config_best_default_route_get (conf);
|
|
|
|
|
if (!obj)
|
|
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
metric = NMP_OBJECT_CAST_IPX_ROUTE (obj)->rx.metric;
|
|
|
|
|
if (metric <= best_metric) {
|
|
|
|
|
best_metric = metric;
|
|
|
|
|
best_conf = conf;
|
|
|
|
|
best_vpn = candidate;
|
core: rework handling of default-routes and drop NMDefaultRouteManager
Remove NMDefaultRouteManager. Instead, add the default-route to the
NMIP4Config/NMIP6Config instance.
This basically reverts commit e8824f6a5205ffcf761abd3e0897a22b254c7797.
We added NMDefaultRouteManager because we used the corresponding to `ip
route replace` when configuring routes. That would replace default-routes
on other interfaces so we needed a central manager to coordinate routes.
Now, we use the corresponding of `ip route append` to configure routes,
and each interface can configure routes indepdentently.
In NMDevice, when creating the default-route, ignore @auto_method for
external devices. We shall not touch these devices.
Especially the code in NMPolicy regarding selection of the best-device
seems wrong. It probably needs further adjustments in the future.
Especially get_best_ip_config() should be replaced, because this
distinction VPN vs. devices seems wrong to me.
Thereby, remove the @ignore_never_default argument. It was added by
commit bb750260045239ab85574366bae8102eff8058cc, I don't think it's
needed anymore.
This brings another change. Now that we track default-routes in
NMIP4Config/NMIP6Config, they are also exposed on D-Bus like regular
routes. I think that makes sense, but it is a change in behavior, as
previously such routes were not exposed there.
2017-08-30 11:46:42 +02:00
|
|
|
}
|
2018-06-28 17:16:15 +02:00
|
|
|
}
|
core: rework handling of default-routes and drop NMDefaultRouteManager
Remove NMDefaultRouteManager. Instead, add the default-route to the
NMIP4Config/NMIP6Config instance.
This basically reverts commit e8824f6a5205ffcf761abd3e0897a22b254c7797.
We added NMDefaultRouteManager because we used the corresponding to `ip
route replace` when configuring routes. That would replace default-routes
on other interfaces so we needed a central manager to coordinate routes.
Now, we use the corresponding of `ip route append` to configure routes,
and each interface can configure routes indepdentently.
In NMDevice, when creating the default-route, ignore @auto_method for
external devices. We shall not touch these devices.
Especially the code in NMPolicy regarding selection of the best-device
seems wrong. It probably needs further adjustments in the future.
Especially get_best_ip_config() should be replaced, because this
distinction VPN vs. devices seems wrong to me.
Thereby, remove the @ignore_never_default argument. It was added by
commit bb750260045239ab85574366bae8102eff8058cc, I don't think it's
needed anymore.
This brings another change. Now that we track default-routes in
NMIP4Config/NMIP6Config, they are also exposed on D-Bus like regular
routes. I think that makes sense, but it is a change in behavior, as
previously such routes were not exposed there.
2017-08-30 11:46:42 +02:00
|
|
|
|
2018-06-28 17:16:15 +02:00
|
|
|
if (best_metric != G_MAXUINT64) {
|
core: rework handling of default-routes and drop NMDefaultRouteManager
Remove NMDefaultRouteManager. Instead, add the default-route to the
NMIP4Config/NMIP6Config instance.
This basically reverts commit e8824f6a5205ffcf761abd3e0897a22b254c7797.
We added NMDefaultRouteManager because we used the corresponding to `ip
route replace` when configuring routes. That would replace default-routes
on other interfaces so we needed a central manager to coordinate routes.
Now, we use the corresponding of `ip route append` to configure routes,
and each interface can configure routes indepdentently.
In NMDevice, when creating the default-route, ignore @auto_method for
external devices. We shall not touch these devices.
Especially the code in NMPolicy regarding selection of the best-device
seems wrong. It probably needs further adjustments in the future.
Especially get_best_ip_config() should be replaced, because this
distinction VPN vs. devices seems wrong to me.
Thereby, remove the @ignore_never_default argument. It was added by
commit bb750260045239ab85574366bae8102eff8058cc, I don't think it's
needed anymore.
This brings another change. Now that we track default-routes in
NMIP4Config/NMIP6Config, they are also exposed on D-Bus like regular
routes. I think that makes sense, but it is a change in behavior, as
previously such routes were not exposed there.
2017-08-30 11:46:42 +02:00
|
|
|
NM_SET_OUT (out_device, NULL);
|
2018-06-28 17:16:15 +02:00
|
|
|
NM_SET_OUT (out_vpn, best_vpn);
|
|
|
|
|
NM_SET_OUT (out_ac, NM_ACTIVE_CONNECTION (best_vpn));
|
|
|
|
|
NM_SET_OUT (out_ip_iface, nm_vpn_connection_get_ip_iface (best_vpn, TRUE));
|
|
|
|
|
return best_conf;
|
core: rework handling of default-routes and drop NMDefaultRouteManager
Remove NMDefaultRouteManager. Instead, add the default-route to the
NMIP4Config/NMIP6Config instance.
This basically reverts commit e8824f6a5205ffcf761abd3e0897a22b254c7797.
We added NMDefaultRouteManager because we used the corresponding to `ip
route replace` when configuring routes. That would replace default-routes
on other interfaces so we needed a central manager to coordinate routes.
Now, we use the corresponding of `ip route append` to configure routes,
and each interface can configure routes indepdentently.
In NMDevice, when creating the default-route, ignore @auto_method for
external devices. We shall not touch these devices.
Especially the code in NMPolicy regarding selection of the best-device
seems wrong. It probably needs further adjustments in the future.
Especially get_best_ip_config() should be replaced, because this
distinction VPN vs. devices seems wrong to me.
Thereby, remove the @ignore_never_default argument. It was added by
commit bb750260045239ab85574366bae8102eff8058cc, I don't think it's
needed anymore.
This brings another change. Now that we track default-routes in
NMIP4Config/NMIP6Config, they are also exposed on D-Bus like regular
routes. I think that makes sense, but it is a change in behavior, as
previously such routes were not exposed there.
2017-08-30 11:46:42 +02:00
|
|
|
}
|
|
|
|
|
|
2018-06-28 18:05:05 +02:00
|
|
|
ac = get_best_active_connection (self, addr_family, TRUE);
|
|
|
|
|
if (ac) {
|
|
|
|
|
NMDevice *device = nm_active_connection_get_device (ac);
|
|
|
|
|
|
|
|
|
|
nm_assert (device);
|
core: rework handling of default-routes and drop NMDefaultRouteManager
Remove NMDefaultRouteManager. Instead, add the default-route to the
NMIP4Config/NMIP6Config instance.
This basically reverts commit e8824f6a5205ffcf761abd3e0897a22b254c7797.
We added NMDefaultRouteManager because we used the corresponding to `ip
route replace` when configuring routes. That would replace default-routes
on other interfaces so we needed a central manager to coordinate routes.
Now, we use the corresponding of `ip route append` to configure routes,
and each interface can configure routes indepdentently.
In NMDevice, when creating the default-route, ignore @auto_method for
external devices. We shall not touch these devices.
Especially the code in NMPolicy regarding selection of the best-device
seems wrong. It probably needs further adjustments in the future.
Especially get_best_ip_config() should be replaced, because this
distinction VPN vs. devices seems wrong to me.
Thereby, remove the @ignore_never_default argument. It was added by
commit bb750260045239ab85574366bae8102eff8058cc, I don't think it's
needed anymore.
This brings another change. Now that we track default-routes in
NMIP4Config/NMIP6Config, they are also exposed on D-Bus like regular
routes. I think that makes sense, but it is a change in behavior, as
previously such routes were not exposed there.
2017-08-30 11:46:42 +02:00
|
|
|
|
|
|
|
|
if (addr_family == AF_INET)
|
|
|
|
|
conf = nm_device_get_ip4_config (device);
|
|
|
|
|
else
|
|
|
|
|
conf = nm_device_get_ip6_config (device);
|
|
|
|
|
|
2018-06-28 18:05:05 +02:00
|
|
|
NM_SET_OUT (out_device, device);
|
|
|
|
|
NM_SET_OUT (out_vpn, NULL);
|
|
|
|
|
NM_SET_OUT (out_ac, ac);
|
|
|
|
|
NM_SET_OUT (out_ip_iface, nm_device_get_ip_iface (device));
|
|
|
|
|
return conf;
|
core: rework handling of default-routes and drop NMDefaultRouteManager
Remove NMDefaultRouteManager. Instead, add the default-route to the
NMIP4Config/NMIP6Config instance.
This basically reverts commit e8824f6a5205ffcf761abd3e0897a22b254c7797.
We added NMDefaultRouteManager because we used the corresponding to `ip
route replace` when configuring routes. That would replace default-routes
on other interfaces so we needed a central manager to coordinate routes.
Now, we use the corresponding of `ip route append` to configure routes,
and each interface can configure routes indepdentently.
In NMDevice, when creating the default-route, ignore @auto_method for
external devices. We shall not touch these devices.
Especially the code in NMPolicy regarding selection of the best-device
seems wrong. It probably needs further adjustments in the future.
Especially get_best_ip_config() should be replaced, because this
distinction VPN vs. devices seems wrong to me.
Thereby, remove the @ignore_never_default argument. It was added by
commit bb750260045239ab85574366bae8102eff8058cc, I don't think it's
needed anymore.
This brings another change. Now that we track default-routes in
NMIP4Config/NMIP6Config, they are also exposed on D-Bus like regular
routes. I think that makes sense, but it is a change in behavior, as
previously such routes were not exposed there.
2017-08-30 11:46:42 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
NM_SET_OUT (out_device, NULL);
|
|
|
|
|
NM_SET_OUT (out_vpn, NULL);
|
|
|
|
|
NM_SET_OUT (out_ac, NULL);
|
|
|
|
|
NM_SET_OUT (out_ip_iface, NULL);
|
|
|
|
|
return NULL;
|
2012-05-29 10:01:07 -05:00
|
|
|
}
|
2008-10-10 Dan Williams <dcbw@redhat.com>
Rework default route handling to consolidate decisions in the policy,
and to take active VPN connections into account when changing the default
route (bgo #545912)
* src/NetworkManager.c
- (main): pass the vpn_manager to the policy so it knows about active
VPN connections; clean up the named manager which wasn't done before
* src/NetworkManagerPolicy.c
src/NetworkManagerPolicy.h
- (nm_policy_new): get a clue about the vpn_manager
- (update_default_route): remove, fold into update_routing_and_dns()
- (update_routing_and_dns): handle active VPN connections too; an
active VPN connection becomes the default route if it does not have
server-specified or user-specified custom routes. Otherwise, the
best active device gets the default route
- (vpn_connection_activated, vpn_connection_deactivated, nm_policy_new,
nm_policy_destroy): track VPN connection activation and deactivation
and update the default route when appropriate
* src/NetworkManagerSystem.c
src/NetworkManagerSystem.h
- (nm_system_vpn_device_unset_from_ip4_config): remove, put functionality
in the VPN connection itself
- (nm_system_vpn_device_set_from_ip4_config,
nm_system_device_set_from_ip4_config): merge together to make
nm_system_apply_ip4_config()
- (add_vpn_gateway_route): add a route to the VPN's external gateway
via the parent device
- (nm_system_apply_ip4_config): simplify
- (add_ip4_route_to_gateway): new function; add a direct route to the
gateway if needed
- (nm_system_device_replace_default_ip4_route): simplify, break gateway
route stuff out into add_ip4_route_to_gateway() for clarity
* src/nm-device.c
- (nm_device_set_ip4_config): update for nm_system_apply_ip4_config()
* src/vpn-manager/nm-vpn-connection.c
src/vpn-manager/nm-vpn-connection.h
- (nm_vpn_connection_get_ip4_config, nm_vpn_connection_get_ip_iface,
nm_vpn_connection_get_parent_device): add
- (nm_vpn_connection_ip4_config_get): make the requirement of a tunnel
device explicit
- (connection_state_changed): update the named manager now that
nm_system_vpn_device_unset_from_ip4_config() is gone; do something
useful on errors
* src/vpn-manager/nm-vpn-manager.c
src/vpn-manager/nm-vpn-manager.h
- Add a 'connection-activated' signal
- (nm_vpn_manager_get_active_connections): new function; mainly for the
policy to find out about active VPN connections
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4167 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-10-10 23:05:45 +00:00
|
|
|
|
2012-05-29 10:01:07 -05:00
|
|
|
static void
|
2016-03-29 10:46:07 +02:00
|
|
|
update_ip4_routing (NMPolicy *self, gboolean force_update)
|
2012-05-29 10:01:07 -05:00
|
|
|
{
|
2016-03-29 10:46:07 +02:00
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2017-09-06 13:39:28 +02:00
|
|
|
NMDevice *best = NULL;
|
libnm, core, cli, tui: fix the capitalization of various types
GLib/Gtk have mostly settled on the convention that two-letter
acronyms in type names remain all-caps (eg, "IO"), but longer acronyms
become initial-caps-only (eg, "Tcp").
NM was inconsistent, with most long acronyms using initial caps only
(Adsl, Cdma, Dcb, Gsm, Olpc, Vlan), but others using all caps (DHCP,
PPP, PPPOE, VPN). Fix libnm and src/ to use initial-caps only for all
three-or-more-letter-long acronyms (and update nmcli and nmtui for the
libnm changes).
2014-06-26 13:44:36 -04:00
|
|
|
NMVpnConnection *vpn = NULL;
|
2012-05-29 10:01:07 -05:00
|
|
|
NMActiveConnection *best_ac = NULL;
|
|
|
|
|
const char *ip_iface = NULL;
|
2017-11-23 21:30:09 +01:00
|
|
|
const CList *tmp_list;
|
|
|
|
|
NMActiveConnection *ac;
|
2008-03-07 Dan Williams <dcbw@redhat.com>
First pass of multiple active device support. Expect bugs.
* src/nm-ip4-config.c
src/nm-ip4-config.h
- (nm_ip4_config_get_secondary, nm_ip4_config_set_secondary): remove;
there are better ways to do this in the named manager
* src/nm-device.c
src/nm-device.h
- (nm_device_can_activate): return whether the device can activate a
connection right now; taking into account things like carrier state
and rfkill state
- (nm_device_get_best_auto_connection): renamed from
nm_device_get_best_connection
- (real_act_stage4_get_ip4_config): MTU stuff is now handled in the
device subclasses themselves, so that each device can override the
MTU from it's NMSetting subclass if needed
- (nm_device_set_ip4_config): set MTU when setting up routes and stuff
in NetworkManagerSystem.c, not here
* src/named-manager/nm-named-manager.c
src/named-manager/nm-named-manager.h
- (nm_named_manager_name_owner_changed,
nm_named_manager_dbus_connection_changed): fix for changes to
rewrite_resolv_conf()
- (compute_nameservers): don't need the NMNamedManager at all, remove
from parameter list
- (merge_one_ip4_config): new function; merge ip4 configs together
- (rewrite_resolv_conf): write out resolv.conf from all the stored
ip4 configs; the VPN config takes precedence, then the best
device config, then the rest of the configs
- (get_domain_for_config): take the NMNamedManager as an argument
to check whether the config is the VPN config
- (add_ip4_config_to_named): fixups for removal of the 'secondary'
attribute from ip4 configs
- (add_all_ip4_configs_to_named): add all the configs in priority order
- (remove_ip4_config_from_named): fix for changes to
get_domain_for_config()
- (nm_named_manager_add_ip4_config): assign the config to the right slot
based on its type; callers must pass in the type now
- (get_last_default_domain): remove, unused
- (nm_named_manager_remove_ip4_config): handle config slots correctly
* src/nm-device-802-11-wireless.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): handle MTU override
* src/nm-device-802-3-ethernet.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): new function; handle MTU override
* src/vpn-manager/nm-vpn-connection.c
- (nm_vpn_connection_ip4_config_get): don't need to set the 'secondary'
attribute on the ip4 config
* src/NetworkManagerPolicy.c
- (nm_policy_auto_get_best_device): remove
- (nm_policy_device_change_check): remove
- (update_default_route): new function; set the default route via
the specified device
- (get_device_priority): new function; return the priority number of
a device type WRT which one should have the default route. Order is
(highest to lowest) wired, wireless, GSM, CDMA.
- (update_routing_and_dns): new function; determine which device should
have the default route, then update the routing table and DNS
- (maybe_auto_activate_device): new function; if a device is now
available for activation, find out what connection it would like to
activate and do it
- (schedule_activate_check): new function; if a device can be activated
now, schedule the activation. Each device may have only one
pending activation at a given time.
- (device_state_changed): if activation was canceled, try again,
possibly with another connection; if the device was activated,
update routing and DNS; if the device was deactivated, try again
with another connection
- (device_carrier_changed): if there is no carrier, deactivate the
device; otherwise schedule an activation check for the device
- (wireless_networks_changed): schedule an activation check for the
device
- (device_added): keep track of the signal handler IDs so they can
be removed when the device goes away
- (device_removed): remove any signal handlers that might be attached
to the device; update routing and DNS
- (schedule_activate_all): new function
- (connections_added, connection_added, connection_updated): when
connections change, schedule all devices for an activation check
- (connection_removed): when a device is deactivated because its
connection was removed, schedule another activation check for it
- (nm_policy_destroy): destroy pending activations and disconnect
all device signal handlers
* src/nm-manager.c
- (nm_manager_activate_device): if the device was already actived,
deactivate it
- (deactivate_old_device): remove
- (connection_added_default_handler, impl_manager_activate_device):
don't deactivate other devices when activating this one
* src/backends/NetworkManagerGentoo.c
src/backends/NetworkManagerFrugalware.c
src/backends/NetworkManagerPaldo.c
src/backends/NetworkManagerRedHat.c
src/backends/NetworkManagerSlackware.c
src/backends/NetworkManagerArch.c
src/backends/NetworkManagerSuSE.c
src/backends/NetworkManagerDebian.c
- (nm_system_get_mtu): remove; MTU should be provided through the
distro's system settings service plugin instead
- (nm_system_device_add_default_route_via_device): remove
- (nm_system_device_add_default_route_via_device_with_iface): remove
- (nm_system_device_replace_default_route): new function; call
generic implementation
* src/backends/NetworkManagerGeneric.c
src/backends/NetworkManagerGeneric.h
- (nm_generic_device_add_default_route_via_device,
nm_generic_device_add_default_route_via_device_with_iface): remove
- (nm_generic_device_replace_default_route): replace the default route
with the given route via some gateway
* src/NetworkManagerSystem.c
src/NetworkManagerSystem.h
- (nm_system_device_set_from_ip4_config): let the policy handle updates
to routing and DNS; but set the MTU here
- (nm_system_vpn_device_set_from_ip4_config): set the route with the
ip_iface of the active device; use the standard MTU setting function
- (nm_system_set_mtu): remove
- (nm_system_device_set_mtu): consolidate MTU setting code in one place
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@3391 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-03-07 19:41:32 +00:00
|
|
|
|
2012-05-29 10:01:07 -05:00
|
|
|
/* Note that we might have an IPv4 VPN tunneled over an IPv6-only device,
|
|
|
|
|
* so we can get (vpn != NULL && best == NULL).
|
2008-06-10 02:06:42 +00:00
|
|
|
*/
|
core: rework handling of default-routes and drop NMDefaultRouteManager
Remove NMDefaultRouteManager. Instead, add the default-route to the
NMIP4Config/NMIP6Config instance.
This basically reverts commit e8824f6a5205ffcf761abd3e0897a22b254c7797.
We added NMDefaultRouteManager because we used the corresponding to `ip
route replace` when configuring routes. That would replace default-routes
on other interfaces so we needed a central manager to coordinate routes.
Now, we use the corresponding of `ip route append` to configure routes,
and each interface can configure routes indepdentently.
In NMDevice, when creating the default-route, ignore @auto_method for
external devices. We shall not touch these devices.
Especially the code in NMPolicy regarding selection of the best-device
seems wrong. It probably needs further adjustments in the future.
Especially get_best_ip_config() should be replaced, because this
distinction VPN vs. devices seems wrong to me.
Thereby, remove the @ignore_never_default argument. It was added by
commit bb750260045239ab85574366bae8102eff8058cc, I don't think it's
needed anymore.
This brings another change. Now that we track default-routes in
NMIP4Config/NMIP6Config, they are also exposed on D-Bus like regular
routes. I think that makes sense, but it is a change in behavior, as
previously such routes were not exposed there.
2017-08-30 11:46:42 +02:00
|
|
|
if (!get_best_ip_config (self, AF_INET, &ip_iface, &best_ac, &best, &vpn)) {
|
2018-06-28 18:05:05 +02:00
|
|
|
if (nm_clear_g_object (&priv->default_ac4)) {
|
|
|
|
|
_LOGt (LOGD_DNS, "set-default-ac-4: %p", NULL);
|
|
|
|
|
_notify (self, PROP_DEFAULT_IP4_AC);
|
2017-09-06 13:39:28 +02:00
|
|
|
}
|
2012-05-29 10:01:07 -05:00
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
g_assert ((best || vpn) && best_ac);
|
2008-06-10 02:06:42 +00:00
|
|
|
|
2017-09-06 13:39:28 +02:00
|
|
|
if ( !force_update
|
2018-06-28 18:05:05 +02:00
|
|
|
&& best_ac
|
|
|
|
|
&& best_ac == priv->default_ac4)
|
2012-05-29 10:01:07 -05:00
|
|
|
return;
|
2008-10-10 Dan Williams <dcbw@redhat.com>
Rework default route handling to consolidate decisions in the policy,
and to take active VPN connections into account when changing the default
route (bgo #545912)
* src/NetworkManager.c
- (main): pass the vpn_manager to the policy so it knows about active
VPN connections; clean up the named manager which wasn't done before
* src/NetworkManagerPolicy.c
src/NetworkManagerPolicy.h
- (nm_policy_new): get a clue about the vpn_manager
- (update_default_route): remove, fold into update_routing_and_dns()
- (update_routing_and_dns): handle active VPN connections too; an
active VPN connection becomes the default route if it does not have
server-specified or user-specified custom routes. Otherwise, the
best active device gets the default route
- (vpn_connection_activated, vpn_connection_deactivated, nm_policy_new,
nm_policy_destroy): track VPN connection activation and deactivation
and update the default route when appropriate
* src/NetworkManagerSystem.c
src/NetworkManagerSystem.h
- (nm_system_vpn_device_unset_from_ip4_config): remove, put functionality
in the VPN connection itself
- (nm_system_vpn_device_set_from_ip4_config,
nm_system_device_set_from_ip4_config): merge together to make
nm_system_apply_ip4_config()
- (add_vpn_gateway_route): add a route to the VPN's external gateway
via the parent device
- (nm_system_apply_ip4_config): simplify
- (add_ip4_route_to_gateway): new function; add a direct route to the
gateway if needed
- (nm_system_device_replace_default_ip4_route): simplify, break gateway
route stuff out into add_ip4_route_to_gateway() for clarity
* src/nm-device.c
- (nm_device_set_ip4_config): update for nm_system_apply_ip4_config()
* src/vpn-manager/nm-vpn-connection.c
src/vpn-manager/nm-vpn-connection.h
- (nm_vpn_connection_get_ip4_config, nm_vpn_connection_get_ip_iface,
nm_vpn_connection_get_parent_device): add
- (nm_vpn_connection_ip4_config_get): make the requirement of a tunnel
device explicit
- (connection_state_changed): update the named manager now that
nm_system_vpn_device_unset_from_ip4_config() is gone; do something
useful on errors
* src/vpn-manager/nm-vpn-manager.c
src/vpn-manager/nm-vpn-manager.h
- Add a 'connection-activated' signal
- (nm_vpn_manager_get_active_connections): new function; mainly for the
policy to find out about active VPN connections
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4167 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-10-10 23:05:45 +00:00
|
|
|
|
2014-10-23 17:04:49 -05:00
|
|
|
if (best) {
|
2017-11-23 21:30:09 +01:00
|
|
|
nm_manager_for_each_active_connection (priv->manager, ac, tmp_list) {
|
|
|
|
|
if ( NM_IS_VPN_CONNECTION (ac)
|
|
|
|
|
&& nm_vpn_connection_get_ip4_config (NM_VPN_CONNECTION (ac))
|
|
|
|
|
&& !nm_active_connection_get_device (ac))
|
|
|
|
|
nm_active_connection_set_device (ac, best);
|
2014-10-23 17:04:49 -05:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2017-11-23 21:53:04 +01:00
|
|
|
update_default_ac (self, AF_INET, best_ac);
|
2012-05-29 10:01:07 -05:00
|
|
|
|
2018-06-28 18:05:05 +02:00
|
|
|
if (!nm_g_object_ref_set (&priv->default_ac4, best_ac))
|
2013-07-15 20:17:25 +02:00
|
|
|
return;
|
2018-06-28 18:05:05 +02:00
|
|
|
_LOGt (LOGD_DNS, "set-default-ac-4: %p", priv->default_ac4);
|
2013-07-15 20:17:25 +02:00
|
|
|
|
2016-03-03 09:20:18 +01:00
|
|
|
_LOGI (LOGD_CORE, "set '%s' (%s) as default for IPv4 routing and DNS",
|
2017-09-06 13:39:28 +02:00
|
|
|
nm_connection_get_id (nm_active_connection_get_applied_connection (best_ac)),
|
|
|
|
|
ip_iface);
|
2018-06-28 18:05:05 +02:00
|
|
|
_notify (self, PROP_DEFAULT_IP4_AC);
|
2010-04-23 22:11:22 -07:00
|
|
|
}
|
|
|
|
|
|
2016-10-31 23:31:14 +01:00
|
|
|
static void
|
|
|
|
|
update_ip6_dns_delegation (NMPolicy *self)
|
|
|
|
|
{
|
|
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2017-11-23 21:30:09 +01:00
|
|
|
NMDevice *device;
|
|
|
|
|
NMActiveConnection *ac;
|
|
|
|
|
const CList *tmp_list;
|
2016-10-31 23:31:14 +01:00
|
|
|
|
2017-11-23 21:30:09 +01:00
|
|
|
nm_manager_for_each_active_connection (priv->manager, ac, tmp_list) {
|
|
|
|
|
device = nm_active_connection_get_device (ac);
|
2016-10-31 23:31:14 +01:00
|
|
|
if (device && nm_device_needs_ip6_subnet (device))
|
2018-06-28 18:05:05 +02:00
|
|
|
nm_device_copy_ip6_dns_config (device, get_default_device (self, AF_INET6));
|
2016-10-31 23:31:14 +01:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void
|
|
|
|
|
update_ip6_prefix_delegation (NMPolicy *self)
|
|
|
|
|
{
|
|
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2017-11-23 21:30:09 +01:00
|
|
|
NMDevice *device;
|
|
|
|
|
NMActiveConnection *ac;
|
|
|
|
|
const CList *tmp_list;
|
2016-10-31 23:31:14 +01:00
|
|
|
|
|
|
|
|
/* There's new default IPv6 connection, try to get a prefix for everyone. */
|
2017-11-23 21:30:09 +01:00
|
|
|
nm_manager_for_each_active_connection (priv->manager, ac, tmp_list) {
|
|
|
|
|
device = nm_active_connection_get_device (ac);
|
2016-10-31 23:31:14 +01:00
|
|
|
if (device && nm_device_needs_ip6_subnet (device))
|
2018-06-28 18:05:05 +02:00
|
|
|
ip6_subnet_from_device (self, get_default_device (self, AF_INET6), device);
|
2016-10-31 23:31:14 +01:00
|
|
|
}
|
2012-05-29 10:01:07 -05:00
|
|
|
}
|
2010-04-23 22:11:22 -07:00
|
|
|
|
2012-05-29 10:01:07 -05:00
|
|
|
static void
|
2016-03-29 10:46:07 +02:00
|
|
|
update_ip6_routing (NMPolicy *self, gboolean force_update)
|
2012-05-29 10:01:07 -05:00
|
|
|
{
|
2016-03-29 10:46:07 +02:00
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2017-09-06 13:39:28 +02:00
|
|
|
NMDevice *best = NULL;
|
libnm, core, cli, tui: fix the capitalization of various types
GLib/Gtk have mostly settled on the convention that two-letter
acronyms in type names remain all-caps (eg, "IO"), but longer acronyms
become initial-caps-only (eg, "Tcp").
NM was inconsistent, with most long acronyms using initial caps only
(Adsl, Cdma, Dcb, Gsm, Olpc, Vlan), but others using all caps (DHCP,
PPP, PPPOE, VPN). Fix libnm and src/ to use initial-caps only for all
three-or-more-letter-long acronyms (and update nmcli and nmtui for the
libnm changes).
2014-06-26 13:44:36 -04:00
|
|
|
NMVpnConnection *vpn = NULL;
|
2012-05-29 10:01:07 -05:00
|
|
|
NMActiveConnection *best_ac = NULL;
|
|
|
|
|
const char *ip_iface = NULL;
|
2017-11-23 21:30:09 +01:00
|
|
|
NMActiveConnection *ac;
|
|
|
|
|
const CList *tmp_list;
|
2010-04-23 22:11:22 -07:00
|
|
|
|
2012-05-29 10:01:07 -05:00
|
|
|
/* Note that we might have an IPv6 VPN tunneled over an IPv4-only device,
|
|
|
|
|
* so we can get (vpn != NULL && best == NULL).
|
2010-04-23 22:11:22 -07:00
|
|
|
*/
|
core: rework handling of default-routes and drop NMDefaultRouteManager
Remove NMDefaultRouteManager. Instead, add the default-route to the
NMIP4Config/NMIP6Config instance.
This basically reverts commit e8824f6a5205ffcf761abd3e0897a22b254c7797.
We added NMDefaultRouteManager because we used the corresponding to `ip
route replace` when configuring routes. That would replace default-routes
on other interfaces so we needed a central manager to coordinate routes.
Now, we use the corresponding of `ip route append` to configure routes,
and each interface can configure routes indepdentently.
In NMDevice, when creating the default-route, ignore @auto_method for
external devices. We shall not touch these devices.
Especially the code in NMPolicy regarding selection of the best-device
seems wrong. It probably needs further adjustments in the future.
Especially get_best_ip_config() should be replaced, because this
distinction VPN vs. devices seems wrong to me.
Thereby, remove the @ignore_never_default argument. It was added by
commit bb750260045239ab85574366bae8102eff8058cc, I don't think it's
needed anymore.
This brings another change. Now that we track default-routes in
NMIP4Config/NMIP6Config, they are also exposed on D-Bus like regular
routes. I think that makes sense, but it is a change in behavior, as
previously such routes were not exposed there.
2017-08-30 11:46:42 +02:00
|
|
|
if (!get_best_ip_config (self, AF_INET6, &ip_iface, &best_ac, &best, &vpn)) {
|
2018-06-28 18:05:05 +02:00
|
|
|
if (nm_clear_g_object (&priv->default_ac6)) {
|
|
|
|
|
_LOGt (LOGD_DNS, "set-default-ac-6: %p", NULL);
|
|
|
|
|
_notify (self, PROP_DEFAULT_IP6_AC);
|
2017-09-06 13:39:28 +02:00
|
|
|
}
|
2012-05-29 10:01:07 -05:00
|
|
|
return;
|
2010-04-23 22:11:22 -07:00
|
|
|
}
|
2012-05-29 10:01:07 -05:00
|
|
|
g_assert ((best || vpn) && best_ac);
|
2010-04-23 22:11:22 -07:00
|
|
|
|
2017-09-06 13:39:28 +02:00
|
|
|
if ( !force_update
|
2018-06-28 18:05:05 +02:00
|
|
|
&& best_ac
|
|
|
|
|
&& best_ac == priv->default_ac6)
|
2012-05-29 10:01:07 -05:00
|
|
|
return;
|
2010-04-23 22:11:22 -07:00
|
|
|
|
2014-10-23 17:04:49 -05:00
|
|
|
if (best) {
|
2017-11-23 21:30:09 +01:00
|
|
|
nm_manager_for_each_active_connection (priv->manager, ac, tmp_list) {
|
|
|
|
|
if ( NM_IS_VPN_CONNECTION (ac)
|
|
|
|
|
&& nm_vpn_connection_get_ip6_config (NM_VPN_CONNECTION (ac))
|
|
|
|
|
&& !nm_active_connection_get_device (ac))
|
|
|
|
|
nm_active_connection_set_device (ac, best);
|
2014-10-23 17:04:49 -05:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2017-11-23 21:53:04 +01:00
|
|
|
update_default_ac (self, AF_INET6, best_ac);
|
2012-05-29 10:01:07 -05:00
|
|
|
|
2018-06-28 18:05:05 +02:00
|
|
|
if (!nm_g_object_ref_set (&priv->default_ac6, best_ac))
|
2013-07-15 20:17:25 +02:00
|
|
|
return;
|
2018-06-28 18:05:05 +02:00
|
|
|
_LOGt (LOGD_DNS, "set-default-ac-6: %p", priv->default_ac6);
|
2016-10-31 23:31:14 +01:00
|
|
|
|
|
|
|
|
update_ip6_prefix_delegation (self);
|
|
|
|
|
|
2016-03-03 09:20:18 +01:00
|
|
|
_LOGI (LOGD_CORE, "set '%s' (%s) as default for IPv6 routing and DNS",
|
2017-09-06 13:39:28 +02:00
|
|
|
nm_connection_get_id (nm_active_connection_get_applied_connection (best_ac)),
|
|
|
|
|
ip_iface);
|
2018-06-28 18:05:05 +02:00
|
|
|
_notify (self, PROP_DEFAULT_IP6_AC);
|
2010-04-23 16:37:45 -07:00
|
|
|
}
|
|
|
|
|
|
2017-11-08 13:44:28 +01:00
|
|
|
static void
|
|
|
|
|
update_ip_dns (NMPolicy *self, int addr_family)
|
|
|
|
|
{
|
|
|
|
|
gpointer ip_config;
|
|
|
|
|
const char *ip_iface = NULL;
|
|
|
|
|
NMVpnConnection *vpn = NULL;
|
2019-02-09 11:46:47 +01:00
|
|
|
NMDevice *device = NULL;
|
2017-11-08 13:44:28 +01:00
|
|
|
|
|
|
|
|
nm_assert_addr_family (addr_family);
|
|
|
|
|
|
2019-02-09 11:46:47 +01:00
|
|
|
ip_config = get_best_ip_config (self, addr_family, &ip_iface, NULL, &device, &vpn);
|
2017-11-08 13:44:28 +01:00
|
|
|
if (ip_config) {
|
|
|
|
|
/* Tell the DNS manager this config is preferred by re-adding it with
|
|
|
|
|
* a different IP config type.
|
|
|
|
|
*/
|
2019-02-09 11:46:47 +01:00
|
|
|
_dns_manager_set_ip_config (NM_POLICY_GET_PRIVATE (self)->dns_manager,
|
|
|
|
|
ip_config,
|
|
|
|
|
vpn
|
|
|
|
|
? NM_DNS_IP_CONFIG_TYPE_VPN
|
|
|
|
|
: NM_DNS_IP_CONFIG_TYPE_BEST_DEVICE,
|
|
|
|
|
device);
|
2017-11-08 13:44:28 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (addr_family == AF_INET6)
|
|
|
|
|
update_ip6_dns_delegation (self);
|
|
|
|
|
}
|
|
|
|
|
|
2010-04-23 16:37:45 -07:00
|
|
|
static void
|
2016-03-29 10:46:07 +02:00
|
|
|
update_routing_and_dns (NMPolicy *self, gboolean force_update)
|
2010-04-23 16:37:45 -07:00
|
|
|
{
|
2016-03-29 10:46:07 +02:00
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2012-05-29 10:01:07 -05:00
|
|
|
|
2013-12-19 11:05:43 -05:00
|
|
|
nm_dns_manager_begin_updates (priv->dns_manager, __func__);
|
2012-05-29 10:01:07 -05:00
|
|
|
|
2017-11-08 13:44:28 +01:00
|
|
|
update_ip_dns (self, AF_INET);
|
|
|
|
|
update_ip_dns (self, AF_INET6);
|
2012-05-29 10:01:07 -05:00
|
|
|
|
2016-03-29 10:46:07 +02:00
|
|
|
update_ip4_routing (self, force_update);
|
|
|
|
|
update_ip6_routing (self, force_update);
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
|
2010-04-23 16:37:45 -07:00
|
|
|
/* Update the system hostname */
|
2017-09-06 13:01:55 +02:00
|
|
|
update_system_hostname (self, "routing and dns");
|
2012-05-29 10:01:07 -05:00
|
|
|
|
2013-12-19 11:05:43 -05:00
|
|
|
nm_dns_manager_end_updates (priv->dns_manager, __func__);
|
2004-07-15 Dan Williams <dcbw@redhat.com>
* src/Makefile.am
- Turn on warnings
* src/NetworkManager.c
- nm_create_device_and_add_to_list(): call nm_device_deactivate() rather
that doing the deactivation ourselves
- Cancel an pending actions on a device if its being removed
- Break up link state checking a bit, make non-active wireless cards
deactivated to save power
- Remove unused variables
* src/NetworkManager.h
- Add support for "pending" device
* src/NetworkManagerAP.h
src/NetworkManagerAP.c
- Add support for determining whether and AP has encryption enabled or not
- AP address is now "struct ether_addr" rather than a string
* src/NetworkManagerDbus.h
src/NetworkManagerDbus.c
- Add signal NeedKeyForNetwork, method SetKeyForNetwork (testing only)
- Changes for AP address from struct ether_addr->string
* src/NetworkManagerDevice.h
src/NetworkManagerDevice.c
- Remove unused variables, fix warnings
- Add support for Pending Actions (things that block a device from being "active"
until they are completed).
- First pending action: Get a WEP key from the user
- Add nm_device_is_wire[d|less](), rename nm_device_is_wireless()
- Clean up explicit testing of dev->iface_type to use nm_device_is_wireless()
- Update wireless link checking to try to determine if the AP we are associated
with is correct, but the WEP key we are using is just wrong. If its wrong,
trigger the GetUserKey pending action on the device
- If dhclient can't get an IP address, it brings the device down. Bring it back
up in that case, otherwise we can't scan or link-check on it
- Add IP address change notifications at appropriate points (still needs some work)
- Add nm_device_need_ap_switch(), checks whether we need to switch access points or not
* src/NetworkManagerPolicy.h
src/NetworkManagerPolicy.c
- Split out "best" access point determiniation into separate function
- Make device activation 2-stage: first the device is pending, then
in the next iteration through it becomes "active" unless it has
pending actions
* src/NetworkManagerUtils.h
src/NetworkManagerUtils.c
- Clean up unused variables and warnings
- Wrap our debug macros in {} to prevent possible confusion
* src/NetworkManagerWireless.c
- Forgot to return current best priority, which lead to last available AP always
being chosen no matter what its priority was. Corrected.
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@15 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-07-15 16:51:48 +00:00
|
|
|
}
|
|
|
|
|
|
2013-08-22 13:06:51 -04:00
|
|
|
static void
|
2018-06-28 18:05:05 +02:00
|
|
|
check_activating_active_connections (NMPolicy *self)
|
2013-08-22 13:06:51 -04:00
|
|
|
{
|
2016-03-29 10:46:07 +02:00
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2018-06-28 18:05:05 +02:00
|
|
|
NMActiveConnection *best4, *best6 = NULL;
|
2013-08-22 13:06:51 -04:00
|
|
|
|
2018-06-28 18:05:05 +02:00
|
|
|
best4 = get_best_active_connection (self, AF_INET, FALSE);
|
|
|
|
|
best6 = get_best_active_connection (self, AF_INET6, FALSE);
|
2013-08-22 13:06:51 -04:00
|
|
|
|
2017-09-06 15:41:12 +02:00
|
|
|
g_object_freeze_notify (G_OBJECT (self));
|
2013-08-22 13:06:51 -04:00
|
|
|
|
2018-06-28 18:05:05 +02:00
|
|
|
if (nm_g_object_ref_set (&priv->activating_ac4, best4)) {
|
|
|
|
|
_LOGt (LOGD_DNS, "set-activating-ac-4: %p", priv->activating_ac4);
|
|
|
|
|
_notify (self, PROP_ACTIVATING_IP4_AC);
|
2013-08-22 13:06:51 -04:00
|
|
|
}
|
2018-06-28 18:05:05 +02:00
|
|
|
if (nm_g_object_ref_set (&priv->activating_ac6, best6)) {
|
|
|
|
|
_LOGt (LOGD_DNS, "set-activating-ac-6: %p", priv->activating_ac6);
|
|
|
|
|
_notify (self, PROP_ACTIVATING_IP6_AC);
|
2013-08-22 13:06:51 -04:00
|
|
|
}
|
|
|
|
|
|
2017-09-06 15:41:12 +02:00
|
|
|
g_object_thaw_notify (G_OBJECT (self));
|
2013-08-22 13:06:51 -04:00
|
|
|
}
|
|
|
|
|
|
2008-03-07 Dan Williams <dcbw@redhat.com>
First pass of multiple active device support. Expect bugs.
* src/nm-ip4-config.c
src/nm-ip4-config.h
- (nm_ip4_config_get_secondary, nm_ip4_config_set_secondary): remove;
there are better ways to do this in the named manager
* src/nm-device.c
src/nm-device.h
- (nm_device_can_activate): return whether the device can activate a
connection right now; taking into account things like carrier state
and rfkill state
- (nm_device_get_best_auto_connection): renamed from
nm_device_get_best_connection
- (real_act_stage4_get_ip4_config): MTU stuff is now handled in the
device subclasses themselves, so that each device can override the
MTU from it's NMSetting subclass if needed
- (nm_device_set_ip4_config): set MTU when setting up routes and stuff
in NetworkManagerSystem.c, not here
* src/named-manager/nm-named-manager.c
src/named-manager/nm-named-manager.h
- (nm_named_manager_name_owner_changed,
nm_named_manager_dbus_connection_changed): fix for changes to
rewrite_resolv_conf()
- (compute_nameservers): don't need the NMNamedManager at all, remove
from parameter list
- (merge_one_ip4_config): new function; merge ip4 configs together
- (rewrite_resolv_conf): write out resolv.conf from all the stored
ip4 configs; the VPN config takes precedence, then the best
device config, then the rest of the configs
- (get_domain_for_config): take the NMNamedManager as an argument
to check whether the config is the VPN config
- (add_ip4_config_to_named): fixups for removal of the 'secondary'
attribute from ip4 configs
- (add_all_ip4_configs_to_named): add all the configs in priority order
- (remove_ip4_config_from_named): fix for changes to
get_domain_for_config()
- (nm_named_manager_add_ip4_config): assign the config to the right slot
based on its type; callers must pass in the type now
- (get_last_default_domain): remove, unused
- (nm_named_manager_remove_ip4_config): handle config slots correctly
* src/nm-device-802-11-wireless.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): handle MTU override
* src/nm-device-802-3-ethernet.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): new function; handle MTU override
* src/vpn-manager/nm-vpn-connection.c
- (nm_vpn_connection_ip4_config_get): don't need to set the 'secondary'
attribute on the ip4 config
* src/NetworkManagerPolicy.c
- (nm_policy_auto_get_best_device): remove
- (nm_policy_device_change_check): remove
- (update_default_route): new function; set the default route via
the specified device
- (get_device_priority): new function; return the priority number of
a device type WRT which one should have the default route. Order is
(highest to lowest) wired, wireless, GSM, CDMA.
- (update_routing_and_dns): new function; determine which device should
have the default route, then update the routing table and DNS
- (maybe_auto_activate_device): new function; if a device is now
available for activation, find out what connection it would like to
activate and do it
- (schedule_activate_check): new function; if a device can be activated
now, schedule the activation. Each device may have only one
pending activation at a given time.
- (device_state_changed): if activation was canceled, try again,
possibly with another connection; if the device was activated,
update routing and DNS; if the device was deactivated, try again
with another connection
- (device_carrier_changed): if there is no carrier, deactivate the
device; otherwise schedule an activation check for the device
- (wireless_networks_changed): schedule an activation check for the
device
- (device_added): keep track of the signal handler IDs so they can
be removed when the device goes away
- (device_removed): remove any signal handlers that might be attached
to the device; update routing and DNS
- (schedule_activate_all): new function
- (connections_added, connection_added, connection_updated): when
connections change, schedule all devices for an activation check
- (connection_removed): when a device is deactivated because its
connection was removed, schedule another activation check for it
- (nm_policy_destroy): destroy pending activations and disconnect
all device signal handlers
* src/nm-manager.c
- (nm_manager_activate_device): if the device was already actived,
deactivate it
- (deactivate_old_device): remove
- (connection_added_default_handler, impl_manager_activate_device):
don't deactivate other devices when activating this one
* src/backends/NetworkManagerGentoo.c
src/backends/NetworkManagerFrugalware.c
src/backends/NetworkManagerPaldo.c
src/backends/NetworkManagerRedHat.c
src/backends/NetworkManagerSlackware.c
src/backends/NetworkManagerArch.c
src/backends/NetworkManagerSuSE.c
src/backends/NetworkManagerDebian.c
- (nm_system_get_mtu): remove; MTU should be provided through the
distro's system settings service plugin instead
- (nm_system_device_add_default_route_via_device): remove
- (nm_system_device_add_default_route_via_device_with_iface): remove
- (nm_system_device_replace_default_route): new function; call
generic implementation
* src/backends/NetworkManagerGeneric.c
src/backends/NetworkManagerGeneric.h
- (nm_generic_device_add_default_route_via_device,
nm_generic_device_add_default_route_via_device_with_iface): remove
- (nm_generic_device_replace_default_route): replace the default route
with the given route via some gateway
* src/NetworkManagerSystem.c
src/NetworkManagerSystem.h
- (nm_system_device_set_from_ip4_config): let the policy handle updates
to routing and DNS; but set the MTU here
- (nm_system_vpn_device_set_from_ip4_config): set the route with the
ip_iface of the active device; use the standard MTU setting function
- (nm_system_set_mtu): remove
- (nm_system_device_set_mtu): consolidate MTU setting code in one place
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@3391 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-03-07 19:41:32 +00:00
|
|
|
typedef struct {
|
2017-11-22 12:15:24 +01:00
|
|
|
CList pending_lst;
|
2008-03-07 Dan Williams <dcbw@redhat.com>
First pass of multiple active device support. Expect bugs.
* src/nm-ip4-config.c
src/nm-ip4-config.h
- (nm_ip4_config_get_secondary, nm_ip4_config_set_secondary): remove;
there are better ways to do this in the named manager
* src/nm-device.c
src/nm-device.h
- (nm_device_can_activate): return whether the device can activate a
connection right now; taking into account things like carrier state
and rfkill state
- (nm_device_get_best_auto_connection): renamed from
nm_device_get_best_connection
- (real_act_stage4_get_ip4_config): MTU stuff is now handled in the
device subclasses themselves, so that each device can override the
MTU from it's NMSetting subclass if needed
- (nm_device_set_ip4_config): set MTU when setting up routes and stuff
in NetworkManagerSystem.c, not here
* src/named-manager/nm-named-manager.c
src/named-manager/nm-named-manager.h
- (nm_named_manager_name_owner_changed,
nm_named_manager_dbus_connection_changed): fix for changes to
rewrite_resolv_conf()
- (compute_nameservers): don't need the NMNamedManager at all, remove
from parameter list
- (merge_one_ip4_config): new function; merge ip4 configs together
- (rewrite_resolv_conf): write out resolv.conf from all the stored
ip4 configs; the VPN config takes precedence, then the best
device config, then the rest of the configs
- (get_domain_for_config): take the NMNamedManager as an argument
to check whether the config is the VPN config
- (add_ip4_config_to_named): fixups for removal of the 'secondary'
attribute from ip4 configs
- (add_all_ip4_configs_to_named): add all the configs in priority order
- (remove_ip4_config_from_named): fix for changes to
get_domain_for_config()
- (nm_named_manager_add_ip4_config): assign the config to the right slot
based on its type; callers must pass in the type now
- (get_last_default_domain): remove, unused
- (nm_named_manager_remove_ip4_config): handle config slots correctly
* src/nm-device-802-11-wireless.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): handle MTU override
* src/nm-device-802-3-ethernet.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): new function; handle MTU override
* src/vpn-manager/nm-vpn-connection.c
- (nm_vpn_connection_ip4_config_get): don't need to set the 'secondary'
attribute on the ip4 config
* src/NetworkManagerPolicy.c
- (nm_policy_auto_get_best_device): remove
- (nm_policy_device_change_check): remove
- (update_default_route): new function; set the default route via
the specified device
- (get_device_priority): new function; return the priority number of
a device type WRT which one should have the default route. Order is
(highest to lowest) wired, wireless, GSM, CDMA.
- (update_routing_and_dns): new function; determine which device should
have the default route, then update the routing table and DNS
- (maybe_auto_activate_device): new function; if a device is now
available for activation, find out what connection it would like to
activate and do it
- (schedule_activate_check): new function; if a device can be activated
now, schedule the activation. Each device may have only one
pending activation at a given time.
- (device_state_changed): if activation was canceled, try again,
possibly with another connection; if the device was activated,
update routing and DNS; if the device was deactivated, try again
with another connection
- (device_carrier_changed): if there is no carrier, deactivate the
device; otherwise schedule an activation check for the device
- (wireless_networks_changed): schedule an activation check for the
device
- (device_added): keep track of the signal handler IDs so they can
be removed when the device goes away
- (device_removed): remove any signal handlers that might be attached
to the device; update routing and DNS
- (schedule_activate_all): new function
- (connections_added, connection_added, connection_updated): when
connections change, schedule all devices for an activation check
- (connection_removed): when a device is deactivated because its
connection was removed, schedule another activation check for it
- (nm_policy_destroy): destroy pending activations and disconnect
all device signal handlers
* src/nm-manager.c
- (nm_manager_activate_device): if the device was already actived,
deactivate it
- (deactivate_old_device): remove
- (connection_added_default_handler, impl_manager_activate_device):
don't deactivate other devices when activating this one
* src/backends/NetworkManagerGentoo.c
src/backends/NetworkManagerFrugalware.c
src/backends/NetworkManagerPaldo.c
src/backends/NetworkManagerRedHat.c
src/backends/NetworkManagerSlackware.c
src/backends/NetworkManagerArch.c
src/backends/NetworkManagerSuSE.c
src/backends/NetworkManagerDebian.c
- (nm_system_get_mtu): remove; MTU should be provided through the
distro's system settings service plugin instead
- (nm_system_device_add_default_route_via_device): remove
- (nm_system_device_add_default_route_via_device_with_iface): remove
- (nm_system_device_replace_default_route): new function; call
generic implementation
* src/backends/NetworkManagerGeneric.c
src/backends/NetworkManagerGeneric.h
- (nm_generic_device_add_default_route_via_device,
nm_generic_device_add_default_route_via_device_with_iface): remove
- (nm_generic_device_replace_default_route): replace the default route
with the given route via some gateway
* src/NetworkManagerSystem.c
src/NetworkManagerSystem.h
- (nm_system_device_set_from_ip4_config): let the policy handle updates
to routing and DNS; but set the MTU here
- (nm_system_vpn_device_set_from_ip4_config): set the route with the
ip_iface of the active device; use the standard MTU setting function
- (nm_system_set_mtu): remove
- (nm_system_device_set_mtu): consolidate MTU setting code in one place
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@3391 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-03-07 19:41:32 +00:00
|
|
|
NMPolicy *policy;
|
|
|
|
|
NMDevice *device;
|
2014-01-20 21:51:04 +01:00
|
|
|
guint autoactivate_id;
|
2008-03-07 Dan Williams <dcbw@redhat.com>
First pass of multiple active device support. Expect bugs.
* src/nm-ip4-config.c
src/nm-ip4-config.h
- (nm_ip4_config_get_secondary, nm_ip4_config_set_secondary): remove;
there are better ways to do this in the named manager
* src/nm-device.c
src/nm-device.h
- (nm_device_can_activate): return whether the device can activate a
connection right now; taking into account things like carrier state
and rfkill state
- (nm_device_get_best_auto_connection): renamed from
nm_device_get_best_connection
- (real_act_stage4_get_ip4_config): MTU stuff is now handled in the
device subclasses themselves, so that each device can override the
MTU from it's NMSetting subclass if needed
- (nm_device_set_ip4_config): set MTU when setting up routes and stuff
in NetworkManagerSystem.c, not here
* src/named-manager/nm-named-manager.c
src/named-manager/nm-named-manager.h
- (nm_named_manager_name_owner_changed,
nm_named_manager_dbus_connection_changed): fix for changes to
rewrite_resolv_conf()
- (compute_nameservers): don't need the NMNamedManager at all, remove
from parameter list
- (merge_one_ip4_config): new function; merge ip4 configs together
- (rewrite_resolv_conf): write out resolv.conf from all the stored
ip4 configs; the VPN config takes precedence, then the best
device config, then the rest of the configs
- (get_domain_for_config): take the NMNamedManager as an argument
to check whether the config is the VPN config
- (add_ip4_config_to_named): fixups for removal of the 'secondary'
attribute from ip4 configs
- (add_all_ip4_configs_to_named): add all the configs in priority order
- (remove_ip4_config_from_named): fix for changes to
get_domain_for_config()
- (nm_named_manager_add_ip4_config): assign the config to the right slot
based on its type; callers must pass in the type now
- (get_last_default_domain): remove, unused
- (nm_named_manager_remove_ip4_config): handle config slots correctly
* src/nm-device-802-11-wireless.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): handle MTU override
* src/nm-device-802-3-ethernet.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): new function; handle MTU override
* src/vpn-manager/nm-vpn-connection.c
- (nm_vpn_connection_ip4_config_get): don't need to set the 'secondary'
attribute on the ip4 config
* src/NetworkManagerPolicy.c
- (nm_policy_auto_get_best_device): remove
- (nm_policy_device_change_check): remove
- (update_default_route): new function; set the default route via
the specified device
- (get_device_priority): new function; return the priority number of
a device type WRT which one should have the default route. Order is
(highest to lowest) wired, wireless, GSM, CDMA.
- (update_routing_and_dns): new function; determine which device should
have the default route, then update the routing table and DNS
- (maybe_auto_activate_device): new function; if a device is now
available for activation, find out what connection it would like to
activate and do it
- (schedule_activate_check): new function; if a device can be activated
now, schedule the activation. Each device may have only one
pending activation at a given time.
- (device_state_changed): if activation was canceled, try again,
possibly with another connection; if the device was activated,
update routing and DNS; if the device was deactivated, try again
with another connection
- (device_carrier_changed): if there is no carrier, deactivate the
device; otherwise schedule an activation check for the device
- (wireless_networks_changed): schedule an activation check for the
device
- (device_added): keep track of the signal handler IDs so they can
be removed when the device goes away
- (device_removed): remove any signal handlers that might be attached
to the device; update routing and DNS
- (schedule_activate_all): new function
- (connections_added, connection_added, connection_updated): when
connections change, schedule all devices for an activation check
- (connection_removed): when a device is deactivated because its
connection was removed, schedule another activation check for it
- (nm_policy_destroy): destroy pending activations and disconnect
all device signal handlers
* src/nm-manager.c
- (nm_manager_activate_device): if the device was already actived,
deactivate it
- (deactivate_old_device): remove
- (connection_added_default_handler, impl_manager_activate_device):
don't deactivate other devices when activating this one
* src/backends/NetworkManagerGentoo.c
src/backends/NetworkManagerFrugalware.c
src/backends/NetworkManagerPaldo.c
src/backends/NetworkManagerRedHat.c
src/backends/NetworkManagerSlackware.c
src/backends/NetworkManagerArch.c
src/backends/NetworkManagerSuSE.c
src/backends/NetworkManagerDebian.c
- (nm_system_get_mtu): remove; MTU should be provided through the
distro's system settings service plugin instead
- (nm_system_device_add_default_route_via_device): remove
- (nm_system_device_add_default_route_via_device_with_iface): remove
- (nm_system_device_replace_default_route): new function; call
generic implementation
* src/backends/NetworkManagerGeneric.c
src/backends/NetworkManagerGeneric.h
- (nm_generic_device_add_default_route_via_device,
nm_generic_device_add_default_route_via_device_with_iface): remove
- (nm_generic_device_replace_default_route): replace the default route
with the given route via some gateway
* src/NetworkManagerSystem.c
src/NetworkManagerSystem.h
- (nm_system_device_set_from_ip4_config): let the policy handle updates
to routing and DNS; but set the MTU here
- (nm_system_vpn_device_set_from_ip4_config): set the route with the
ip_iface of the active device; use the standard MTU setting function
- (nm_system_set_mtu): remove
- (nm_system_device_set_mtu): consolidate MTU setting code in one place
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@3391 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-03-07 19:41:32 +00:00
|
|
|
} ActivateData;
|
|
|
|
|
|
2011-04-19 00:26:37 -05:00
|
|
|
static void
|
|
|
|
|
activate_data_free (ActivateData *data)
|
|
|
|
|
{
|
2017-01-27 13:57:15 +01:00
|
|
|
nm_device_remove_pending_action (data->device, NM_PENDING_ACTION_AUTOACTIVATE, TRUE);
|
2017-11-28 11:22:01 +01:00
|
|
|
c_list_unlink_stale (&data->pending_lst);
|
2017-11-02 16:39:42 +01:00
|
|
|
nm_clear_g_source (&data->autoactivate_id);
|
2011-04-19 00:26:37 -05:00
|
|
|
g_object_unref (data->device);
|
2016-03-29 10:48:01 +02:00
|
|
|
g_slice_free (ActivateData, data);
|
2011-04-19 00:26:37 -05:00
|
|
|
}
|
|
|
|
|
|
2017-09-25 22:20:47 +02:00
|
|
|
static void
|
|
|
|
|
pending_ac_gone (gpointer data, GObject *where_the_object_was)
|
|
|
|
|
{
|
|
|
|
|
NMPolicy *self = NM_POLICY (data);
|
|
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
|
|
|
|
|
|
|
|
|
/* Active connections should reach the DEACTIVATED state
|
|
|
|
|
* before disappearing. */
|
|
|
|
|
nm_assert_not_reached();
|
|
|
|
|
|
|
|
|
|
if (g_hash_table_remove (priv->pending_active_connections, where_the_object_was))
|
|
|
|
|
g_object_unref (self);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void
|
|
|
|
|
pending_ac_state_changed (NMActiveConnection *ac, guint state, guint reason, NMPolicy *self)
|
|
|
|
|
{
|
|
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
|
|
|
|
NMSettingsConnection *con;
|
|
|
|
|
|
|
|
|
|
if (state >= NM_ACTIVE_CONNECTION_STATE_DEACTIVATING) {
|
|
|
|
|
/* The AC is being deactivated before the device had a chance
|
|
|
|
|
* to move to PREPARE. Schedule a new auto-activation on the
|
|
|
|
|
* device, but block the current connection to avoid an activation
|
|
|
|
|
* loop.
|
|
|
|
|
*/
|
2018-05-25 02:31:59 +02:00
|
|
|
if (reason != NM_ACTIVE_CONNECTION_STATE_REASON_DEVICE_DISCONNECTED) {
|
|
|
|
|
con = nm_active_connection_get_settings_connection (ac);
|
|
|
|
|
nm_settings_connection_autoconnect_blocked_reason_set (con, NM_SETTINGS_AUTO_CONNECT_BLOCKED_REASON_FAILED, TRUE);
|
|
|
|
|
schedule_activate_check (self, nm_active_connection_get_device (ac));
|
|
|
|
|
}
|
2017-09-25 22:20:47 +02:00
|
|
|
|
|
|
|
|
/* Cleanup */
|
|
|
|
|
g_signal_handlers_disconnect_by_func (ac, pending_ac_state_changed, self);
|
|
|
|
|
if (!g_hash_table_remove (priv->pending_active_connections, ac))
|
|
|
|
|
nm_assert_not_reached ();
|
|
|
|
|
g_object_weak_unref (G_OBJECT (ac), pending_ac_gone, self);
|
|
|
|
|
g_object_unref (self);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2016-09-26 16:40:15 +02:00
|
|
|
static void
|
|
|
|
|
auto_activate_device (NMPolicy *self,
|
|
|
|
|
NMDevice *device)
|
2004-08-26 20:05:24 +00:00
|
|
|
{
|
2013-08-22 10:10:17 -04:00
|
|
|
NMPolicyPrivate *priv;
|
2015-07-14 16:53:24 +02:00
|
|
|
NMSettingsConnection *best_connection;
|
2016-09-26 15:34:07 +02:00
|
|
|
gs_free char *specific_object = NULL;
|
2017-02-03 15:13:03 +01:00
|
|
|
gs_free NMSettingsConnection **connections = NULL;
|
|
|
|
|
guint i, len;
|
2017-11-22 20:09:47 +01:00
|
|
|
gs_free_error GError *error = NULL;
|
2018-04-13 11:21:59 +02:00
|
|
|
gs_unref_object NMAuthSubject *subject = NULL;
|
2017-11-02 18:50:19 +01:00
|
|
|
NMActiveConnection *ac;
|
2008-01-18 02:56:20 +00:00
|
|
|
|
2016-09-26 16:40:15 +02:00
|
|
|
nm_assert (NM_IS_POLICY (self));
|
|
|
|
|
nm_assert (NM_IS_DEVICE (device));
|
2008-03-11 16:37:41 +00:00
|
|
|
|
2016-09-26 16:40:15 +02:00
|
|
|
priv = NM_POLICY_GET_PRIVATE (self);
|
2011-04-19 00:26:37 -05:00
|
|
|
|
2008-03-11 16:37:41 +00:00
|
|
|
// FIXME: if a device is already activating (or activated) with a connection
|
|
|
|
|
// but another connection now overrides the current one for that device,
|
|
|
|
|
// deactivate the device and activate the new connection instead of just
|
|
|
|
|
// bailing if the device is already active
|
2016-09-26 16:40:15 +02:00
|
|
|
if (nm_device_get_act_request (device))
|
|
|
|
|
return;
|
2008-03-11 16:37:41 +00:00
|
|
|
|
2017-11-02 18:48:02 +01:00
|
|
|
if (!nm_device_autoconnect_allowed (device))
|
|
|
|
|
return;
|
|
|
|
|
|
2018-04-18 11:08:05 +02:00
|
|
|
connections = nm_manager_get_activatable_connections (priv->manager, TRUE, TRUE, &len);
|
2017-02-03 15:13:03 +01:00
|
|
|
if (!connections[0])
|
2016-09-26 16:40:15 +02:00
|
|
|
return;
|
2014-08-26 15:49:18 +02:00
|
|
|
|
2014-08-26 15:18:47 +02:00
|
|
|
/* Find the first connection that should be auto-activated */
|
|
|
|
|
best_connection = NULL;
|
2017-02-03 15:13:03 +01:00
|
|
|
for (i = 0; i < len; i++) {
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
NMSettingsConnection *candidate = connections[i];
|
|
|
|
|
NMConnection *cand_conn;
|
2017-10-25 22:17:47 +02:00
|
|
|
NMSettingConnection *s_con;
|
|
|
|
|
const char *permission;
|
2010-12-23 07:52:41 -06:00
|
|
|
|
2017-12-05 12:20:41 +01:00
|
|
|
if (nm_settings_connection_autoconnect_is_blocked (candidate))
|
2014-08-26 15:18:47 +02:00
|
|
|
continue;
|
2017-10-25 22:17:47 +02:00
|
|
|
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
cand_conn = nm_settings_connection_get_connection (candidate);
|
|
|
|
|
|
|
|
|
|
s_con = nm_connection_get_setting_connection (cand_conn);
|
2017-10-25 22:17:47 +02:00
|
|
|
if (!nm_setting_connection_get_autoconnect (s_con))
|
|
|
|
|
continue;
|
|
|
|
|
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
permission = nm_utils_get_shared_wifi_permission (cand_conn);
|
2017-10-25 22:17:47 +02:00
|
|
|
if ( permission
|
|
|
|
|
&& !nm_settings_connection_check_permission (candidate, permission))
|
|
|
|
|
continue;
|
|
|
|
|
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
if (nm_device_can_auto_connect (device, candidate, &specific_object)) {
|
2015-07-14 16:53:24 +02:00
|
|
|
best_connection = candidate;
|
2014-08-26 15:18:47 +02:00
|
|
|
break;
|
|
|
|
|
}
|
2005-05-03 Dan Williams <dcbw@redhat.com>
* Kill dhcpcd. We now use "dhcdbd", a dbus daemon that controls dhclient.
This means that NetworkManager shouldn't have DHCP issues anymore. It also
means you need dhcdbd, which you can get here (get the latest one):
http://people.redhat.com/jvdias/dhcdbd/
Technically NetworkManager can use any DHCP daemon that uses the same DBUS
interface as dhcdbd.
* Rewrite device activation to facilitate the new DHCP infrastructure and
future improvements. Its now "activation request" based, ie there is a single
activation request composed of the device, access point, and other info which
follows the entire activation process. There are 5 stages of the activation
process which correspond to:
1) Device preparation
2) Device configuration (bring it up, set ESSID/Key/etc)
3) IP Config Start (fire off DHCP if we're using it)
4) IP Config Get (grab config from DHCP or static config files)
5) IP Config Commit (set device's IP address, DNS, etc)
Note that there is no longer a "scanning" step, since the access point must
be known _before_ activation starts. If the access point drops out or does
not exist for some reason, the entire activation process fails and must be
restarted for a different access point or device.
Patch from Bill Moss:
* gnome/applet/applet.c
- Fix type of vpn_failure dialog -> vpn_banner dialog
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@597 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2005-05-03 20:41:36 +00:00
|
|
|
}
|
2005-01-21 19:32:08 +00:00
|
|
|
|
2017-11-02 18:50:19 +01:00
|
|
|
if (!best_connection)
|
|
|
|
|
return;
|
2017-09-25 22:20:47 +02:00
|
|
|
|
2018-06-07 11:57:17 +02:00
|
|
|
_LOGI (LOGD_DEVICE, "auto-activating connection '%s' (%s)",
|
|
|
|
|
nm_settings_connection_get_id (best_connection),
|
|
|
|
|
nm_settings_connection_get_uuid (best_connection));
|
|
|
|
|
|
2017-11-02 18:50:19 +01:00
|
|
|
subject = nm_auth_subject_new_internal ();
|
|
|
|
|
ac = nm_manager_activate_connection (priv->manager,
|
|
|
|
|
best_connection,
|
|
|
|
|
NULL,
|
|
|
|
|
specific_object,
|
|
|
|
|
device,
|
|
|
|
|
subject,
|
|
|
|
|
NM_ACTIVATION_TYPE_MANAGED,
|
2018-03-28 17:18:04 +02:00
|
|
|
NM_ACTIVATION_REASON_AUTOCONNECT,
|
core: improve and fix keeping connection active based on "connection.permissions"
By setting "connection.permissions", a profile is restricted to a
particular user.
That means for example, that another user cannot see, modify, delete,
activate or deactivate the profile. It also means, that the profile
will only autoconnect when the user is logged in (has a session).
Note that root is always able to activate the profile. Likewise, the
user is also allowed to manually activate the own profile, even if no
session currently exists (which can easily happen with `sudo`).
When the user logs out (the session goes away), we want do disconnect
the profile, however there are conflicting goals here:
1) if the profile was activate by root user, then logging out the user
should not disconnect the profile. The patch fixes that by not
binding the activation to the connection, if the activation is done
by the root user.
2) if the profile was activated by the owner when it had no session,
then it should stay alive until the user logs in (once) and logs
out again. This is already handled by the previous commit.
Yes, this point is odd. If you first do
$ sudo -u $OTHER_USER nmcli connection up $PROFILE
the profile activates despite not having a session. If you then
$ ssh guest@localhost nmcli device
you'll still see the profile active. However, the moment the SSH session
ends, a session closes and the profile disconnects. It's unclear, how to
solve that any better. I think, a user who cares about this, should not
activate the profile without having a session in the first place.
There are quite some special cases, in particular with internal
activations. In those cases we need to decide whether to bind the
activation to the profile's visibility.
Also, expose the "bind" setting in the D-Bus API. Note, that in the future
this flag may be modified via D-Bus API. Like we may also add related API
that allows to tweak the lifetime of the activation.
Also, I think we broke handling of connection visiblity with 37e8c53eeed
"core: Introduce helper class to track connection keep alive". This
should be fixed now too, with improved behavior.
Fixes: 37e8c53eeed579fe34a68819cd12f3295d581394
https://bugzilla.redhat.com/show_bug.cgi?id=1530977
2018-11-21 13:30:16 +01:00
|
|
|
NM_ACTIVATION_STATE_FLAG_LIFETIME_BOUND_TO_PROFILE_VISIBILITY,
|
2017-11-02 18:50:19 +01:00
|
|
|
&error);
|
|
|
|
|
if (!ac) {
|
2017-11-22 20:09:47 +01:00
|
|
|
_LOGI (LOGD_DEVICE, "connection '%s' auto-activation failed: %s",
|
2017-11-02 18:50:19 +01:00
|
|
|
nm_settings_connection_get_id (best_connection),
|
|
|
|
|
error->message);
|
|
|
|
|
nm_settings_connection_autoconnect_blocked_reason_set (best_connection,
|
2017-11-24 10:53:39 +01:00
|
|
|
NM_SETTINGS_AUTO_CONNECT_BLOCKED_REASON_FAILED,
|
|
|
|
|
TRUE);
|
2017-11-02 18:50:19 +01:00
|
|
|
schedule_activate_check (self, device);
|
|
|
|
|
return;
|
|
|
|
|
}
|
2017-09-25 22:20:47 +02:00
|
|
|
|
2017-11-02 18:50:19 +01:00
|
|
|
/* Subscribe to AC state-changed signal to detect when the
|
|
|
|
|
* activation fails in early stages without changing device
|
|
|
|
|
* state.
|
|
|
|
|
*/
|
2018-01-02 15:47:37 +01:00
|
|
|
if (g_hash_table_add (priv->pending_active_connections, ac)) {
|
2017-11-02 18:50:19 +01:00
|
|
|
g_signal_connect (ac, NM_ACTIVE_CONNECTION_STATE_CHANGED,
|
|
|
|
|
G_CALLBACK (pending_ac_state_changed), g_object_ref (self));
|
|
|
|
|
g_object_weak_ref (G_OBJECT (ac), (GWeakNotify) pending_ac_gone, self);
|
2005-01-07 18:07:06 +00:00
|
|
|
}
|
2016-09-26 16:40:15 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static gboolean
|
|
|
|
|
auto_activate_device_cb (gpointer user_data)
|
|
|
|
|
{
|
|
|
|
|
ActivateData *data = user_data;
|
|
|
|
|
|
|
|
|
|
g_assert (data);
|
|
|
|
|
g_assert (NM_IS_POLICY (data->policy));
|
|
|
|
|
g_assert (NM_IS_DEVICE (data->device));
|
2005-01-07 18:07:06 +00:00
|
|
|
|
2016-09-26 16:40:15 +02:00
|
|
|
data->autoactivate_id = 0;
|
|
|
|
|
auto_activate_device (data->policy, data->device);
|
2011-04-19 00:26:37 -05:00
|
|
|
activate_data_free (data);
|
2014-01-20 21:54:05 +01:00
|
|
|
return G_SOURCE_REMOVE;
|
2005-01-07 18:07:06 +00:00
|
|
|
}
|
|
|
|
|
|
2011-04-19 00:26:37 -05:00
|
|
|
static ActivateData *
|
2017-11-22 12:15:24 +01:00
|
|
|
find_pending_activation (NMPolicy *self, NMDevice *device)
|
2011-04-19 00:26:37 -05:00
|
|
|
{
|
2017-11-22 12:15:24 +01:00
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
|
|
|
|
ActivateData *data;
|
2011-04-19 00:26:37 -05:00
|
|
|
|
2017-11-22 12:15:24 +01:00
|
|
|
c_list_for_each_entry (data, &priv->pending_activation_checks, pending_lst) {
|
|
|
|
|
if (data->device == device)
|
|
|
|
|
return data;
|
2011-04-19 00:26:37 -05:00
|
|
|
}
|
|
|
|
|
return NULL;
|
|
|
|
|
}
|
|
|
|
|
|
2007-02-08 15:34:26 +00:00
|
|
|
/*****************************************************************************/
|
|
|
|
|
|
2012-08-21 17:49:41 +02:00
|
|
|
typedef struct {
|
|
|
|
|
NMDevice *device;
|
|
|
|
|
GSList *secondaries;
|
|
|
|
|
} PendingSecondaryData;
|
|
|
|
|
|
|
|
|
|
static PendingSecondaryData *
|
|
|
|
|
pending_secondary_data_new (NMDevice *device, GSList *secondaries)
|
|
|
|
|
{
|
|
|
|
|
PendingSecondaryData *data;
|
|
|
|
|
|
2016-03-29 10:48:01 +02:00
|
|
|
data = g_slice_new (PendingSecondaryData);
|
2012-08-21 17:49:41 +02:00
|
|
|
data->device = g_object_ref (device);
|
|
|
|
|
data->secondaries = secondaries;
|
|
|
|
|
return data;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void
|
|
|
|
|
pending_secondary_data_free (PendingSecondaryData *data)
|
|
|
|
|
{
|
|
|
|
|
g_object_unref (data->device);
|
2013-10-07 17:35:03 -05:00
|
|
|
g_slist_free_full (data->secondaries, g_object_unref);
|
2016-03-29 10:48:01 +02:00
|
|
|
g_slice_free (PendingSecondaryData, data);
|
2012-08-21 17:49:41 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void
|
2016-03-29 10:46:07 +02:00
|
|
|
process_secondaries (NMPolicy *self,
|
2012-08-21 17:49:41 +02:00
|
|
|
NMActiveConnection *active,
|
|
|
|
|
gboolean connected)
|
|
|
|
|
{
|
2016-03-29 10:46:07 +02:00
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2015-11-04 15:48:28 +01:00
|
|
|
GSList *iter, *iter2, *next, *next2;
|
2012-08-21 17:49:41 +02:00
|
|
|
|
2013-10-07 17:35:03 -05:00
|
|
|
/* Loop through devices waiting for secondary connections to activate */
|
2015-11-04 15:48:28 +01:00
|
|
|
for (iter = priv->pending_secondaries; iter; iter = next) {
|
2012-08-21 17:49:41 +02:00
|
|
|
PendingSecondaryData *secondary_data = (PendingSecondaryData *) iter->data;
|
|
|
|
|
NMDevice *item_device = secondary_data->device;
|
|
|
|
|
|
2015-11-04 15:48:28 +01:00
|
|
|
next = g_slist_next (iter);
|
|
|
|
|
|
2013-10-07 17:35:03 -05:00
|
|
|
/* Look for 'active' in each device's secondary connections list */
|
2015-11-04 15:48:28 +01:00
|
|
|
for (iter2 = secondary_data->secondaries; iter2; iter2 = next2) {
|
2013-10-07 17:35:03 -05:00
|
|
|
NMActiveConnection *secondary_active = NM_ACTIVE_CONNECTION (iter2->data);
|
|
|
|
|
|
2015-11-04 15:48:28 +01:00
|
|
|
next2 = g_slist_next (iter2);
|
|
|
|
|
|
2013-10-07 17:35:03 -05:00
|
|
|
if (active != secondary_active)
|
|
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
if (connected) {
|
2016-03-03 09:20:18 +01:00
|
|
|
_LOGD (LOGD_DEVICE, "secondary connection '%s' succeeded; active path '%s'",
|
|
|
|
|
nm_active_connection_get_settings_connection_id (active),
|
core/dbus: rework D-Bus implementation to use lower layer GDBusConnection API
Previously, we used the generated GDBusInterfaceSkeleton types and glued
them via the NMExportedObject base class to our NM types. We also used
GDBusObjectManagerServer.
Don't do that anymore. The resulting code was more complicated despite (or
because?) using generated classes. It was hard to understand, complex, had
ordering-issues, and had a runtime and memory overhead.
This patch refactors this entirely and uses the lower layer API GDBusConnection
directly. It replaces the generated code, GDBusInterfaceSkeleton, and
GDBusObjectManagerServer. All this is now done by NMDbusObject and NMDBusManager
and static descriptor instances of type GDBusInterfaceInfo.
This adds a net plus of more then 1300 lines of hand written code. I claim
that this implementation is easier to understand. Note that previously we
also required extensive and complex glue code to bind our objects to the
generated skeleton objects. Instead, now glue our objects directly to
GDBusConnection. The result is more immediate and gets rid of layers of
code in between.
Now that the D-Bus glue us more under our control, we can address issus and
bottlenecks better, instead of adding code to bend the generated skeletons
to our needs.
Note that the current implementation now only supports one D-Bus connection.
That was effectively the case already, although there were places (and still are)
where the code pretends it could also support connections from a private socket.
We dropped private socket support mainly because it was unused, untested and
buggy, but also because GDBusObjectManagerServer could not export the same
objects on multiple connections. Now, it would be rather straight forward to
fix that and re-introduce ObjectManager on each private connection. But this
commit doesn't do that yet, and the new code intentionally supports only one
D-Bus connection.
Also, the D-Bus startup was simplified. There is no retry, either nm_dbus_manager_start()
succeeds, or it detects the initrd case. In the initrd case, bus manager never tries to
connect to D-Bus. Since the initrd scenario is not yet used/tested, this is good enough
for the moment. It could be easily extended later, for example with polling whether the
system bus appears (like was done previously). Also, restart of D-Bus daemon isn't
supported either -- just like before.
Note how NMDBusManager now implements the ObjectManager D-Bus interface
directly.
Also, this fixes race issues in the server, by no longer delaying
PropertiesChanged signals. NMExportedObject would collect changed
properties and send the signal out in idle_emit_properties_changed()
on idle. This messes up the ordering of change events w.r.t. other
signals and events on the bus. Note that not only NMExportedObject
messed up the ordering. Also the generated code would hook into
notify() and process change events in and idle handle, exhibiting the
same ordering issue too.
No longer do that. PropertiesChanged signals will be sent right away
by hooking into dispatch_properties_changed(). This means, changing
a property in quick succession will no longer be combined and is
guaranteed to emit signals for each individual state. Quite possibly
we emit now more PropertiesChanged signals then before.
However, we are now able to group a set of changes by using standard
g_object_freeze_notify()/g_object_thaw_notify(). We probably should
make more use of that.
Also, now that our signals are all handled in the right order, we
might find places where we still emit them in the wrong order. But that
is then due to the order in which our GObjects emit signals, not due
to an ill behavior of the D-Bus glue. Possibly we need to identify
such ordering issues and fix them.
Numbers (for contrib/rpm --without debug on x86_64):
- the patch changes the code size of NetworkManager by
- 2809360 bytes
+ 2537528 bytes (-9.7%)
- Runtime measurements are harder because there is a large variance
during testing. In other words, the numbers are not reproducible.
Currently, the implementation performs no caching of GVariants at all,
but it would be rather simple to add it, if that turns out to be
useful.
Anyway, without strong claim, it seems that the new form tends to
perform slightly better. That would be no surprise.
$ time (for i in {1..1000}; do nmcli >/dev/null || break; echo -n .; done)
- real 1m39.355s
+ real 1m37.432s
$ time (for i in {1..2000}; do busctl call org.freedesktop.NetworkManager /org/freedesktop org.freedesktop.DBus.ObjectManager GetManagedObjects > /dev/null || break; echo -n .; done)
- real 0m26.843s
+ real 0m25.281s
- Regarding RSS size, just looking at the processes in similar
conditions, doesn't give a large difference. On my system they
consume about 19MB RSS. It seems that the new version has a
slightly smaller RSS size.
- 19356 RSS
+ 18660 RSS
2018-02-26 13:51:52 +01:00
|
|
|
nm_dbus_object_get_path (NM_DBUS_OBJECT (active)));
|
2013-10-07 17:35:03 -05:00
|
|
|
|
|
|
|
|
/* Secondary connection activated */
|
|
|
|
|
secondary_data->secondaries = g_slist_remove (secondary_data->secondaries, secondary_active);
|
|
|
|
|
g_object_unref (secondary_active);
|
|
|
|
|
if (!secondary_data->secondaries) {
|
|
|
|
|
/* No secondary UUID remained -> remove the secondary data item */
|
|
|
|
|
priv->pending_secondaries = g_slist_remove (priv->pending_secondaries, secondary_data);
|
|
|
|
|
pending_secondary_data_free (secondary_data);
|
2014-04-15 11:51:56 +02:00
|
|
|
if (nm_device_get_state (item_device) == NM_DEVICE_STATE_SECONDARIES)
|
|
|
|
|
nm_device_state_changed (item_device, NM_DEVICE_STATE_ACTIVATED, NM_DEVICE_STATE_REASON_NONE);
|
2013-10-07 17:35:03 -05:00
|
|
|
break;
|
2012-08-21 17:49:41 +02:00
|
|
|
}
|
2013-10-07 17:35:03 -05:00
|
|
|
} else {
|
2016-03-03 09:20:18 +01:00
|
|
|
_LOGD (LOGD_DEVICE, "secondary connection '%s' failed; active path '%s'",
|
|
|
|
|
nm_active_connection_get_settings_connection_id (active),
|
core/dbus: rework D-Bus implementation to use lower layer GDBusConnection API
Previously, we used the generated GDBusInterfaceSkeleton types and glued
them via the NMExportedObject base class to our NM types. We also used
GDBusObjectManagerServer.
Don't do that anymore. The resulting code was more complicated despite (or
because?) using generated classes. It was hard to understand, complex, had
ordering-issues, and had a runtime and memory overhead.
This patch refactors this entirely and uses the lower layer API GDBusConnection
directly. It replaces the generated code, GDBusInterfaceSkeleton, and
GDBusObjectManagerServer. All this is now done by NMDbusObject and NMDBusManager
and static descriptor instances of type GDBusInterfaceInfo.
This adds a net plus of more then 1300 lines of hand written code. I claim
that this implementation is easier to understand. Note that previously we
also required extensive and complex glue code to bind our objects to the
generated skeleton objects. Instead, now glue our objects directly to
GDBusConnection. The result is more immediate and gets rid of layers of
code in between.
Now that the D-Bus glue us more under our control, we can address issus and
bottlenecks better, instead of adding code to bend the generated skeletons
to our needs.
Note that the current implementation now only supports one D-Bus connection.
That was effectively the case already, although there were places (and still are)
where the code pretends it could also support connections from a private socket.
We dropped private socket support mainly because it was unused, untested and
buggy, but also because GDBusObjectManagerServer could not export the same
objects on multiple connections. Now, it would be rather straight forward to
fix that and re-introduce ObjectManager on each private connection. But this
commit doesn't do that yet, and the new code intentionally supports only one
D-Bus connection.
Also, the D-Bus startup was simplified. There is no retry, either nm_dbus_manager_start()
succeeds, or it detects the initrd case. In the initrd case, bus manager never tries to
connect to D-Bus. Since the initrd scenario is not yet used/tested, this is good enough
for the moment. It could be easily extended later, for example with polling whether the
system bus appears (like was done previously). Also, restart of D-Bus daemon isn't
supported either -- just like before.
Note how NMDBusManager now implements the ObjectManager D-Bus interface
directly.
Also, this fixes race issues in the server, by no longer delaying
PropertiesChanged signals. NMExportedObject would collect changed
properties and send the signal out in idle_emit_properties_changed()
on idle. This messes up the ordering of change events w.r.t. other
signals and events on the bus. Note that not only NMExportedObject
messed up the ordering. Also the generated code would hook into
notify() and process change events in and idle handle, exhibiting the
same ordering issue too.
No longer do that. PropertiesChanged signals will be sent right away
by hooking into dispatch_properties_changed(). This means, changing
a property in quick succession will no longer be combined and is
guaranteed to emit signals for each individual state. Quite possibly
we emit now more PropertiesChanged signals then before.
However, we are now able to group a set of changes by using standard
g_object_freeze_notify()/g_object_thaw_notify(). We probably should
make more use of that.
Also, now that our signals are all handled in the right order, we
might find places where we still emit them in the wrong order. But that
is then due to the order in which our GObjects emit signals, not due
to an ill behavior of the D-Bus glue. Possibly we need to identify
such ordering issues and fix them.
Numbers (for contrib/rpm --without debug on x86_64):
- the patch changes the code size of NetworkManager by
- 2809360 bytes
+ 2537528 bytes (-9.7%)
- Runtime measurements are harder because there is a large variance
during testing. In other words, the numbers are not reproducible.
Currently, the implementation performs no caching of GVariants at all,
but it would be rather simple to add it, if that turns out to be
useful.
Anyway, without strong claim, it seems that the new form tends to
perform slightly better. That would be no surprise.
$ time (for i in {1..1000}; do nmcli >/dev/null || break; echo -n .; done)
- real 1m39.355s
+ real 1m37.432s
$ time (for i in {1..2000}; do busctl call org.freedesktop.NetworkManager /org/freedesktop org.freedesktop.DBus.ObjectManager GetManagedObjects > /dev/null || break; echo -n .; done)
- real 0m26.843s
+ real 0m25.281s
- Regarding RSS size, just looking at the processes in similar
conditions, doesn't give a large difference. On my system they
consume about 19MB RSS. It seems that the new version has a
slightly smaller RSS size.
- 19356 RSS
+ 18660 RSS
2018-02-26 13:51:52 +01:00
|
|
|
nm_dbus_object_get_path (NM_DBUS_OBJECT (active)));
|
2013-10-07 17:35:03 -05:00
|
|
|
|
|
|
|
|
/* Secondary connection failed -> do not watch other connections */
|
|
|
|
|
priv->pending_secondaries = g_slist_remove (priv->pending_secondaries, secondary_data);
|
|
|
|
|
pending_secondary_data_free (secondary_data);
|
2014-04-15 11:51:56 +02:00
|
|
|
if ( nm_device_get_state (item_device) == NM_DEVICE_STATE_SECONDARIES
|
|
|
|
|
|| nm_device_get_state (item_device) == NM_DEVICE_STATE_ACTIVATED)
|
|
|
|
|
nm_device_state_changed (item_device, NM_DEVICE_STATE_FAILED,
|
|
|
|
|
NM_DEVICE_STATE_REASON_SECONDARY_CONNECTION_FAILED);
|
2013-10-07 17:35:03 -05:00
|
|
|
break;
|
2012-08-21 17:49:41 +02:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
static void
|
2017-04-23 14:20:37 +02:00
|
|
|
hostname_changed (NMHostnameManager *hostname_manager, GParamSpec *pspec, gpointer user_data)
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
{
|
2016-04-04 14:36:34 +02:00
|
|
|
NMPolicyPrivate *priv = user_data;
|
2016-09-29 13:49:01 +02:00
|
|
|
NMPolicy *self = _PRIV_TO_SELF (priv);
|
2016-04-04 14:36:34 +02:00
|
|
|
|
2017-09-06 13:01:55 +02:00
|
|
|
update_system_hostname (self, "hostname changed");
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
}
|
|
|
|
|
|
2017-11-22 14:52:54 +01:00
|
|
|
static gboolean
|
|
|
|
|
reset_autoconnect_all (NMPolicy *self,
|
|
|
|
|
NMDevice *device, /* if present, only reset connections compatible with @device */
|
2017-11-24 12:14:42 +01:00
|
|
|
gboolean only_no_secrets)
|
2011-05-20 13:01:04 +02:00
|
|
|
{
|
2016-03-29 10:46:07 +02:00
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2017-11-22 11:02:06 +01:00
|
|
|
NMSettingsConnection *const*connections = NULL;
|
2017-02-03 14:15:16 +01:00
|
|
|
guint i;
|
2017-12-12 10:22:59 +01:00
|
|
|
gboolean changed = FALSE;
|
2011-05-20 13:01:04 +02:00
|
|
|
|
2017-11-22 14:52:54 +01:00
|
|
|
_LOGD (LOGD_DEVICE, "re-enabling autoconnect for all connections%s%s%s",
|
|
|
|
|
device ? " on " : "",
|
|
|
|
|
device ? nm_device_get_iface (device) : "",
|
2017-11-24 12:14:42 +01:00
|
|
|
only_no_secrets ? " (only clear no-secrets flag)" : "");
|
2013-12-17 10:30:04 -05:00
|
|
|
|
2017-11-22 11:02:06 +01:00
|
|
|
connections = nm_settings_get_connections (priv->settings, NULL);
|
2017-02-03 14:15:16 +01:00
|
|
|
for (i = 0; connections[i]; i++) {
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
NMSettingsConnection *sett_conn = connections[i];
|
2017-02-03 14:15:16 +01:00
|
|
|
|
2017-11-22 14:52:54 +01:00
|
|
|
if ( device
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
&& !nm_device_check_connection_compatible (device,
|
|
|
|
|
nm_settings_connection_get_connection (sett_conn),
|
|
|
|
|
NULL))
|
2017-11-22 14:52:54 +01:00
|
|
|
continue;
|
2013-12-17 10:30:04 -05:00
|
|
|
|
2017-11-24 12:14:42 +01:00
|
|
|
if (only_no_secrets) {
|
|
|
|
|
/* we only reset the no-secrets blocked flag. */
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
if (nm_settings_connection_autoconnect_blocked_reason_set (sett_conn,
|
2017-11-24 12:14:42 +01:00
|
|
|
NM_SETTINGS_AUTO_CONNECT_BLOCKED_REASON_NO_SECRETS,
|
|
|
|
|
FALSE)) {
|
|
|
|
|
/* maybe the connection is still blocked afterwards for other reasons
|
|
|
|
|
* and in the larger picture nothing changed. But it's too complicated
|
|
|
|
|
* to find out exactly. Just assume, something changed to be sure. */
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
if (!nm_settings_connection_autoconnect_is_blocked (sett_conn))
|
2017-11-24 12:59:15 +01:00
|
|
|
changed = TRUE;
|
2017-11-24 12:14:42 +01:00
|
|
|
}
|
|
|
|
|
} else {
|
|
|
|
|
/* we reset the tries-count and any blocked-reason */
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
if (nm_settings_connection_autoconnect_retries_get (sett_conn) == 0)
|
2017-11-24 12:14:42 +01:00
|
|
|
changed = TRUE;
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
nm_settings_connection_autoconnect_retries_reset (sett_conn);
|
2017-11-24 12:14:42 +01:00
|
|
|
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
if (nm_settings_connection_autoconnect_blocked_reason_set (sett_conn,
|
2017-11-24 12:54:31 +01:00
|
|
|
NM_SETTINGS_AUTO_CONNECT_BLOCKED_REASON_ALL
|
|
|
|
|
& ~NM_SETTINGS_AUTO_CONNECT_BLOCKED_REASON_USER_REQUEST,
|
2017-11-24 12:59:15 +01:00
|
|
|
FALSE)) {
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
if (!nm_settings_connection_autoconnect_is_blocked (sett_conn))
|
2017-11-24 12:59:15 +01:00
|
|
|
changed = TRUE;
|
|
|
|
|
}
|
2017-11-24 12:14:42 +01:00
|
|
|
}
|
2011-06-28 15:48:12 +02:00
|
|
|
}
|
2017-11-22 14:52:54 +01:00
|
|
|
return changed;
|
2011-06-28 15:48:12 +02:00
|
|
|
}
|
|
|
|
|
|
2009-10-20 15:25:04 -07:00
|
|
|
static void
|
|
|
|
|
sleeping_changed (NMManager *manager, GParamSpec *pspec, gpointer user_data)
|
|
|
|
|
{
|
2016-04-04 14:36:34 +02:00
|
|
|
NMPolicyPrivate *priv = user_data;
|
2016-09-29 13:49:01 +02:00
|
|
|
NMPolicy *self = _PRIV_TO_SELF (priv);
|
2010-05-22 08:55:30 -07:00
|
|
|
gboolean sleeping = FALSE, enabled = FALSE;
|
2009-10-20 15:25:04 -07:00
|
|
|
|
|
|
|
|
g_object_get (G_OBJECT (manager), NM_MANAGER_SLEEPING, &sleeping, NULL);
|
2010-05-22 08:55:30 -07:00
|
|
|
g_object_get (G_OBJECT (manager), NM_MANAGER_NETWORKING_ENABLED, &enabled, NULL);
|
2009-10-20 15:25:04 -07:00
|
|
|
|
2011-04-01 18:50:58 -05:00
|
|
|
/* Reset retries on all connections so they'll checked on wakeup */
|
2017-11-29 09:37:05 +01:00
|
|
|
if (sleeping || !enabled)
|
|
|
|
|
reset_autoconnect_all (self, NULL, FALSE);
|
2009-10-20 15:25:04 -07:00
|
|
|
}
|
|
|
|
|
|
2007-02-08 15:34:26 +00:00
|
|
|
static void
|
2016-03-29 10:46:07 +02:00
|
|
|
schedule_activate_check (NMPolicy *self, NMDevice *device)
|
2007-02-08 15:34:26 +00:00
|
|
|
{
|
2016-03-29 10:46:07 +02:00
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2008-03-07 Dan Williams <dcbw@redhat.com>
First pass of multiple active device support. Expect bugs.
* src/nm-ip4-config.c
src/nm-ip4-config.h
- (nm_ip4_config_get_secondary, nm_ip4_config_set_secondary): remove;
there are better ways to do this in the named manager
* src/nm-device.c
src/nm-device.h
- (nm_device_can_activate): return whether the device can activate a
connection right now; taking into account things like carrier state
and rfkill state
- (nm_device_get_best_auto_connection): renamed from
nm_device_get_best_connection
- (real_act_stage4_get_ip4_config): MTU stuff is now handled in the
device subclasses themselves, so that each device can override the
MTU from it's NMSetting subclass if needed
- (nm_device_set_ip4_config): set MTU when setting up routes and stuff
in NetworkManagerSystem.c, not here
* src/named-manager/nm-named-manager.c
src/named-manager/nm-named-manager.h
- (nm_named_manager_name_owner_changed,
nm_named_manager_dbus_connection_changed): fix for changes to
rewrite_resolv_conf()
- (compute_nameservers): don't need the NMNamedManager at all, remove
from parameter list
- (merge_one_ip4_config): new function; merge ip4 configs together
- (rewrite_resolv_conf): write out resolv.conf from all the stored
ip4 configs; the VPN config takes precedence, then the best
device config, then the rest of the configs
- (get_domain_for_config): take the NMNamedManager as an argument
to check whether the config is the VPN config
- (add_ip4_config_to_named): fixups for removal of the 'secondary'
attribute from ip4 configs
- (add_all_ip4_configs_to_named): add all the configs in priority order
- (remove_ip4_config_from_named): fix for changes to
get_domain_for_config()
- (nm_named_manager_add_ip4_config): assign the config to the right slot
based on its type; callers must pass in the type now
- (get_last_default_domain): remove, unused
- (nm_named_manager_remove_ip4_config): handle config slots correctly
* src/nm-device-802-11-wireless.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): handle MTU override
* src/nm-device-802-3-ethernet.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): new function; handle MTU override
* src/vpn-manager/nm-vpn-connection.c
- (nm_vpn_connection_ip4_config_get): don't need to set the 'secondary'
attribute on the ip4 config
* src/NetworkManagerPolicy.c
- (nm_policy_auto_get_best_device): remove
- (nm_policy_device_change_check): remove
- (update_default_route): new function; set the default route via
the specified device
- (get_device_priority): new function; return the priority number of
a device type WRT which one should have the default route. Order is
(highest to lowest) wired, wireless, GSM, CDMA.
- (update_routing_and_dns): new function; determine which device should
have the default route, then update the routing table and DNS
- (maybe_auto_activate_device): new function; if a device is now
available for activation, find out what connection it would like to
activate and do it
- (schedule_activate_check): new function; if a device can be activated
now, schedule the activation. Each device may have only one
pending activation at a given time.
- (device_state_changed): if activation was canceled, try again,
possibly with another connection; if the device was activated,
update routing and DNS; if the device was deactivated, try again
with another connection
- (device_carrier_changed): if there is no carrier, deactivate the
device; otherwise schedule an activation check for the device
- (wireless_networks_changed): schedule an activation check for the
device
- (device_added): keep track of the signal handler IDs so they can
be removed when the device goes away
- (device_removed): remove any signal handlers that might be attached
to the device; update routing and DNS
- (schedule_activate_all): new function
- (connections_added, connection_added, connection_updated): when
connections change, schedule all devices for an activation check
- (connection_removed): when a device is deactivated because its
connection was removed, schedule another activation check for it
- (nm_policy_destroy): destroy pending activations and disconnect
all device signal handlers
* src/nm-manager.c
- (nm_manager_activate_device): if the device was already actived,
deactivate it
- (deactivate_old_device): remove
- (connection_added_default_handler, impl_manager_activate_device):
don't deactivate other devices when activating this one
* src/backends/NetworkManagerGentoo.c
src/backends/NetworkManagerFrugalware.c
src/backends/NetworkManagerPaldo.c
src/backends/NetworkManagerRedHat.c
src/backends/NetworkManagerSlackware.c
src/backends/NetworkManagerArch.c
src/backends/NetworkManagerSuSE.c
src/backends/NetworkManagerDebian.c
- (nm_system_get_mtu): remove; MTU should be provided through the
distro's system settings service plugin instead
- (nm_system_device_add_default_route_via_device): remove
- (nm_system_device_add_default_route_via_device_with_iface): remove
- (nm_system_device_replace_default_route): new function; call
generic implementation
* src/backends/NetworkManagerGeneric.c
src/backends/NetworkManagerGeneric.h
- (nm_generic_device_add_default_route_via_device,
nm_generic_device_add_default_route_via_device_with_iface): remove
- (nm_generic_device_replace_default_route): replace the default route
with the given route via some gateway
* src/NetworkManagerSystem.c
src/NetworkManagerSystem.h
- (nm_system_device_set_from_ip4_config): let the policy handle updates
to routing and DNS; but set the MTU here
- (nm_system_vpn_device_set_from_ip4_config): set the route with the
ip_iface of the active device; use the standard MTU setting function
- (nm_system_set_mtu): remove
- (nm_system_device_set_mtu): consolidate MTU setting code in one place
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@3391 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-03-07 19:41:32 +00:00
|
|
|
ActivateData *data;
|
2017-11-23 21:30:09 +01:00
|
|
|
NMActiveConnection *ac;
|
|
|
|
|
const CList *tmp_list;
|
2008-03-07 Dan Williams <dcbw@redhat.com>
First pass of multiple active device support. Expect bugs.
* src/nm-ip4-config.c
src/nm-ip4-config.h
- (nm_ip4_config_get_secondary, nm_ip4_config_set_secondary): remove;
there are better ways to do this in the named manager
* src/nm-device.c
src/nm-device.h
- (nm_device_can_activate): return whether the device can activate a
connection right now; taking into account things like carrier state
and rfkill state
- (nm_device_get_best_auto_connection): renamed from
nm_device_get_best_connection
- (real_act_stage4_get_ip4_config): MTU stuff is now handled in the
device subclasses themselves, so that each device can override the
MTU from it's NMSetting subclass if needed
- (nm_device_set_ip4_config): set MTU when setting up routes and stuff
in NetworkManagerSystem.c, not here
* src/named-manager/nm-named-manager.c
src/named-manager/nm-named-manager.h
- (nm_named_manager_name_owner_changed,
nm_named_manager_dbus_connection_changed): fix for changes to
rewrite_resolv_conf()
- (compute_nameservers): don't need the NMNamedManager at all, remove
from parameter list
- (merge_one_ip4_config): new function; merge ip4 configs together
- (rewrite_resolv_conf): write out resolv.conf from all the stored
ip4 configs; the VPN config takes precedence, then the best
device config, then the rest of the configs
- (get_domain_for_config): take the NMNamedManager as an argument
to check whether the config is the VPN config
- (add_ip4_config_to_named): fixups for removal of the 'secondary'
attribute from ip4 configs
- (add_all_ip4_configs_to_named): add all the configs in priority order
- (remove_ip4_config_from_named): fix for changes to
get_domain_for_config()
- (nm_named_manager_add_ip4_config): assign the config to the right slot
based on its type; callers must pass in the type now
- (get_last_default_domain): remove, unused
- (nm_named_manager_remove_ip4_config): handle config slots correctly
* src/nm-device-802-11-wireless.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): handle MTU override
* src/nm-device-802-3-ethernet.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): new function; handle MTU override
* src/vpn-manager/nm-vpn-connection.c
- (nm_vpn_connection_ip4_config_get): don't need to set the 'secondary'
attribute on the ip4 config
* src/NetworkManagerPolicy.c
- (nm_policy_auto_get_best_device): remove
- (nm_policy_device_change_check): remove
- (update_default_route): new function; set the default route via
the specified device
- (get_device_priority): new function; return the priority number of
a device type WRT which one should have the default route. Order is
(highest to lowest) wired, wireless, GSM, CDMA.
- (update_routing_and_dns): new function; determine which device should
have the default route, then update the routing table and DNS
- (maybe_auto_activate_device): new function; if a device is now
available for activation, find out what connection it would like to
activate and do it
- (schedule_activate_check): new function; if a device can be activated
now, schedule the activation. Each device may have only one
pending activation at a given time.
- (device_state_changed): if activation was canceled, try again,
possibly with another connection; if the device was activated,
update routing and DNS; if the device was deactivated, try again
with another connection
- (device_carrier_changed): if there is no carrier, deactivate the
device; otherwise schedule an activation check for the device
- (wireless_networks_changed): schedule an activation check for the
device
- (device_added): keep track of the signal handler IDs so they can
be removed when the device goes away
- (device_removed): remove any signal handlers that might be attached
to the device; update routing and DNS
- (schedule_activate_all): new function
- (connections_added, connection_added, connection_updated): when
connections change, schedule all devices for an activation check
- (connection_removed): when a device is deactivated because its
connection was removed, schedule another activation check for it
- (nm_policy_destroy): destroy pending activations and disconnect
all device signal handlers
* src/nm-manager.c
- (nm_manager_activate_device): if the device was already actived,
deactivate it
- (deactivate_old_device): remove
- (connection_added_default_handler, impl_manager_activate_device):
don't deactivate other devices when activating this one
* src/backends/NetworkManagerGentoo.c
src/backends/NetworkManagerFrugalware.c
src/backends/NetworkManagerPaldo.c
src/backends/NetworkManagerRedHat.c
src/backends/NetworkManagerSlackware.c
src/backends/NetworkManagerArch.c
src/backends/NetworkManagerSuSE.c
src/backends/NetworkManagerDebian.c
- (nm_system_get_mtu): remove; MTU should be provided through the
distro's system settings service plugin instead
- (nm_system_device_add_default_route_via_device): remove
- (nm_system_device_add_default_route_via_device_with_iface): remove
- (nm_system_device_replace_default_route): new function; call
generic implementation
* src/backends/NetworkManagerGeneric.c
src/backends/NetworkManagerGeneric.h
- (nm_generic_device_add_default_route_via_device,
nm_generic_device_add_default_route_via_device_with_iface): remove
- (nm_generic_device_replace_default_route): replace the default route
with the given route via some gateway
* src/NetworkManagerSystem.c
src/NetworkManagerSystem.h
- (nm_system_device_set_from_ip4_config): let the policy handle updates
to routing and DNS; but set the MTU here
- (nm_system_vpn_device_set_from_ip4_config): set the route with the
ip_iface of the active device; use the standard MTU setting function
- (nm_system_set_mtu): remove
- (nm_system_device_set_mtu): consolidate MTU setting code in one place
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@3391 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-03-07 19:41:32 +00:00
|
|
|
|
2013-08-22 10:10:17 -04:00
|
|
|
if (nm_manager_get_state (priv->manager) == NM_STATE_ASLEEP)
|
2008-03-07 Dan Williams <dcbw@redhat.com>
First pass of multiple active device support. Expect bugs.
* src/nm-ip4-config.c
src/nm-ip4-config.h
- (nm_ip4_config_get_secondary, nm_ip4_config_set_secondary): remove;
there are better ways to do this in the named manager
* src/nm-device.c
src/nm-device.h
- (nm_device_can_activate): return whether the device can activate a
connection right now; taking into account things like carrier state
and rfkill state
- (nm_device_get_best_auto_connection): renamed from
nm_device_get_best_connection
- (real_act_stage4_get_ip4_config): MTU stuff is now handled in the
device subclasses themselves, so that each device can override the
MTU from it's NMSetting subclass if needed
- (nm_device_set_ip4_config): set MTU when setting up routes and stuff
in NetworkManagerSystem.c, not here
* src/named-manager/nm-named-manager.c
src/named-manager/nm-named-manager.h
- (nm_named_manager_name_owner_changed,
nm_named_manager_dbus_connection_changed): fix for changes to
rewrite_resolv_conf()
- (compute_nameservers): don't need the NMNamedManager at all, remove
from parameter list
- (merge_one_ip4_config): new function; merge ip4 configs together
- (rewrite_resolv_conf): write out resolv.conf from all the stored
ip4 configs; the VPN config takes precedence, then the best
device config, then the rest of the configs
- (get_domain_for_config): take the NMNamedManager as an argument
to check whether the config is the VPN config
- (add_ip4_config_to_named): fixups for removal of the 'secondary'
attribute from ip4 configs
- (add_all_ip4_configs_to_named): add all the configs in priority order
- (remove_ip4_config_from_named): fix for changes to
get_domain_for_config()
- (nm_named_manager_add_ip4_config): assign the config to the right slot
based on its type; callers must pass in the type now
- (get_last_default_domain): remove, unused
- (nm_named_manager_remove_ip4_config): handle config slots correctly
* src/nm-device-802-11-wireless.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): handle MTU override
* src/nm-device-802-3-ethernet.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): new function; handle MTU override
* src/vpn-manager/nm-vpn-connection.c
- (nm_vpn_connection_ip4_config_get): don't need to set the 'secondary'
attribute on the ip4 config
* src/NetworkManagerPolicy.c
- (nm_policy_auto_get_best_device): remove
- (nm_policy_device_change_check): remove
- (update_default_route): new function; set the default route via
the specified device
- (get_device_priority): new function; return the priority number of
a device type WRT which one should have the default route. Order is
(highest to lowest) wired, wireless, GSM, CDMA.
- (update_routing_and_dns): new function; determine which device should
have the default route, then update the routing table and DNS
- (maybe_auto_activate_device): new function; if a device is now
available for activation, find out what connection it would like to
activate and do it
- (schedule_activate_check): new function; if a device can be activated
now, schedule the activation. Each device may have only one
pending activation at a given time.
- (device_state_changed): if activation was canceled, try again,
possibly with another connection; if the device was activated,
update routing and DNS; if the device was deactivated, try again
with another connection
- (device_carrier_changed): if there is no carrier, deactivate the
device; otherwise schedule an activation check for the device
- (wireless_networks_changed): schedule an activation check for the
device
- (device_added): keep track of the signal handler IDs so they can
be removed when the device goes away
- (device_removed): remove any signal handlers that might be attached
to the device; update routing and DNS
- (schedule_activate_all): new function
- (connections_added, connection_added, connection_updated): when
connections change, schedule all devices for an activation check
- (connection_removed): when a device is deactivated because its
connection was removed, schedule another activation check for it
- (nm_policy_destroy): destroy pending activations and disconnect
all device signal handlers
* src/nm-manager.c
- (nm_manager_activate_device): if the device was already actived,
deactivate it
- (deactivate_old_device): remove
- (connection_added_default_handler, impl_manager_activate_device):
don't deactivate other devices when activating this one
* src/backends/NetworkManagerGentoo.c
src/backends/NetworkManagerFrugalware.c
src/backends/NetworkManagerPaldo.c
src/backends/NetworkManagerRedHat.c
src/backends/NetworkManagerSlackware.c
src/backends/NetworkManagerArch.c
src/backends/NetworkManagerSuSE.c
src/backends/NetworkManagerDebian.c
- (nm_system_get_mtu): remove; MTU should be provided through the
distro's system settings service plugin instead
- (nm_system_device_add_default_route_via_device): remove
- (nm_system_device_add_default_route_via_device_with_iface): remove
- (nm_system_device_replace_default_route): new function; call
generic implementation
* src/backends/NetworkManagerGeneric.c
src/backends/NetworkManagerGeneric.h
- (nm_generic_device_add_default_route_via_device,
nm_generic_device_add_default_route_via_device_with_iface): remove
- (nm_generic_device_replace_default_route): replace the default route
with the given route via some gateway
* src/NetworkManagerSystem.c
src/NetworkManagerSystem.h
- (nm_system_device_set_from_ip4_config): let the policy handle updates
to routing and DNS; but set the MTU here
- (nm_system_vpn_device_set_from_ip4_config): set the route with the
ip_iface of the active device; use the standard MTU setting function
- (nm_system_set_mtu): remove
- (nm_system_device_set_mtu): consolidate MTU setting code in one place
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@3391 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-03-07 19:41:32 +00:00
|
|
|
return;
|
|
|
|
|
|
2009-09-14 13:33:06 -07:00
|
|
|
if (!nm_device_autoconnect_allowed (device))
|
2008-03-07 Dan Williams <dcbw@redhat.com>
First pass of multiple active device support. Expect bugs.
* src/nm-ip4-config.c
src/nm-ip4-config.h
- (nm_ip4_config_get_secondary, nm_ip4_config_set_secondary): remove;
there are better ways to do this in the named manager
* src/nm-device.c
src/nm-device.h
- (nm_device_can_activate): return whether the device can activate a
connection right now; taking into account things like carrier state
and rfkill state
- (nm_device_get_best_auto_connection): renamed from
nm_device_get_best_connection
- (real_act_stage4_get_ip4_config): MTU stuff is now handled in the
device subclasses themselves, so that each device can override the
MTU from it's NMSetting subclass if needed
- (nm_device_set_ip4_config): set MTU when setting up routes and stuff
in NetworkManagerSystem.c, not here
* src/named-manager/nm-named-manager.c
src/named-manager/nm-named-manager.h
- (nm_named_manager_name_owner_changed,
nm_named_manager_dbus_connection_changed): fix for changes to
rewrite_resolv_conf()
- (compute_nameservers): don't need the NMNamedManager at all, remove
from parameter list
- (merge_one_ip4_config): new function; merge ip4 configs together
- (rewrite_resolv_conf): write out resolv.conf from all the stored
ip4 configs; the VPN config takes precedence, then the best
device config, then the rest of the configs
- (get_domain_for_config): take the NMNamedManager as an argument
to check whether the config is the VPN config
- (add_ip4_config_to_named): fixups for removal of the 'secondary'
attribute from ip4 configs
- (add_all_ip4_configs_to_named): add all the configs in priority order
- (remove_ip4_config_from_named): fix for changes to
get_domain_for_config()
- (nm_named_manager_add_ip4_config): assign the config to the right slot
based on its type; callers must pass in the type now
- (get_last_default_domain): remove, unused
- (nm_named_manager_remove_ip4_config): handle config slots correctly
* src/nm-device-802-11-wireless.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): handle MTU override
* src/nm-device-802-3-ethernet.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): new function; handle MTU override
* src/vpn-manager/nm-vpn-connection.c
- (nm_vpn_connection_ip4_config_get): don't need to set the 'secondary'
attribute on the ip4 config
* src/NetworkManagerPolicy.c
- (nm_policy_auto_get_best_device): remove
- (nm_policy_device_change_check): remove
- (update_default_route): new function; set the default route via
the specified device
- (get_device_priority): new function; return the priority number of
a device type WRT which one should have the default route. Order is
(highest to lowest) wired, wireless, GSM, CDMA.
- (update_routing_and_dns): new function; determine which device should
have the default route, then update the routing table and DNS
- (maybe_auto_activate_device): new function; if a device is now
available for activation, find out what connection it would like to
activate and do it
- (schedule_activate_check): new function; if a device can be activated
now, schedule the activation. Each device may have only one
pending activation at a given time.
- (device_state_changed): if activation was canceled, try again,
possibly with another connection; if the device was activated,
update routing and DNS; if the device was deactivated, try again
with another connection
- (device_carrier_changed): if there is no carrier, deactivate the
device; otherwise schedule an activation check for the device
- (wireless_networks_changed): schedule an activation check for the
device
- (device_added): keep track of the signal handler IDs so they can
be removed when the device goes away
- (device_removed): remove any signal handlers that might be attached
to the device; update routing and DNS
- (schedule_activate_all): new function
- (connections_added, connection_added, connection_updated): when
connections change, schedule all devices for an activation check
- (connection_removed): when a device is deactivated because its
connection was removed, schedule another activation check for it
- (nm_policy_destroy): destroy pending activations and disconnect
all device signal handlers
* src/nm-manager.c
- (nm_manager_activate_device): if the device was already actived,
deactivate it
- (deactivate_old_device): remove
- (connection_added_default_handler, impl_manager_activate_device):
don't deactivate other devices when activating this one
* src/backends/NetworkManagerGentoo.c
src/backends/NetworkManagerFrugalware.c
src/backends/NetworkManagerPaldo.c
src/backends/NetworkManagerRedHat.c
src/backends/NetworkManagerSlackware.c
src/backends/NetworkManagerArch.c
src/backends/NetworkManagerSuSE.c
src/backends/NetworkManagerDebian.c
- (nm_system_get_mtu): remove; MTU should be provided through the
distro's system settings service plugin instead
- (nm_system_device_add_default_route_via_device): remove
- (nm_system_device_add_default_route_via_device_with_iface): remove
- (nm_system_device_replace_default_route): new function; call
generic implementation
* src/backends/NetworkManagerGeneric.c
src/backends/NetworkManagerGeneric.h
- (nm_generic_device_add_default_route_via_device,
nm_generic_device_add_default_route_via_device_with_iface): remove
- (nm_generic_device_replace_default_route): replace the default route
with the given route via some gateway
* src/NetworkManagerSystem.c
src/NetworkManagerSystem.h
- (nm_system_device_set_from_ip4_config): let the policy handle updates
to routing and DNS; but set the MTU here
- (nm_system_vpn_device_set_from_ip4_config): set the route with the
ip_iface of the active device; use the standard MTU setting function
- (nm_system_set_mtu): remove
- (nm_system_device_set_mtu): consolidate MTU setting code in one place
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@3391 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-03-07 19:41:32 +00:00
|
|
|
return;
|
|
|
|
|
|
2017-11-22 12:15:24 +01:00
|
|
|
if (find_pending_activation (self, device))
|
2014-02-13 14:45:25 -05:00
|
|
|
return;
|
|
|
|
|
|
2017-11-23 21:30:09 +01:00
|
|
|
nm_manager_for_each_active_connection (priv->manager, ac, tmp_list) {
|
|
|
|
|
if (nm_active_connection_get_device (ac) == device)
|
2013-08-28 16:19:20 -05:00
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
2017-01-27 13:57:15 +01:00
|
|
|
nm_device_add_pending_action (device, NM_PENDING_ACTION_AUTOACTIVATE, TRUE);
|
2014-02-13 14:45:25 -05:00
|
|
|
|
2016-03-29 10:48:01 +02:00
|
|
|
data = g_slice_new0 (ActivateData);
|
2016-03-29 10:46:07 +02:00
|
|
|
data->policy = self;
|
2014-02-13 14:45:25 -05:00
|
|
|
data->device = g_object_ref (device);
|
2016-09-26 16:40:15 +02:00
|
|
|
data->autoactivate_id = g_idle_add (auto_activate_device_cb, data);
|
2017-11-22 12:15:24 +01:00
|
|
|
c_list_link_tail (&priv->pending_activation_checks, &data->pending_lst);
|
2007-02-08 15:34:26 +00:00
|
|
|
}
|
|
|
|
|
|
2011-06-17 12:43:28 +02:00
|
|
|
static gboolean
|
|
|
|
|
reset_connections_retries (gpointer user_data)
|
|
|
|
|
{
|
2016-03-29 10:46:07 +02:00
|
|
|
NMPolicy *self = (NMPolicy *) user_data;
|
|
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2017-11-22 11:02:06 +01:00
|
|
|
NMSettingsConnection *const*connections = NULL;
|
2017-02-03 14:15:16 +01:00
|
|
|
guint i;
|
2013-12-10 20:39:48 +01:00
|
|
|
gint32 con_stamp, min_stamp, now;
|
2011-12-06 16:33:00 -06:00
|
|
|
gboolean changed = FALSE;
|
2011-06-17 12:43:28 +02:00
|
|
|
|
2013-08-22 10:10:17 -04:00
|
|
|
priv->reset_retries_id = 0;
|
2011-06-17 12:43:28 +02:00
|
|
|
|
2013-12-05 04:57:01 -05:00
|
|
|
min_stamp = 0;
|
2013-12-10 20:39:48 +01:00
|
|
|
now = nm_utils_get_monotonic_timestamp_s ();
|
2017-11-22 11:02:06 +01:00
|
|
|
connections = nm_settings_get_connections (priv->settings, NULL);
|
2017-02-03 14:15:16 +01:00
|
|
|
for (i = 0; connections[i]; i++) {
|
|
|
|
|
NMSettingsConnection *connection = connections[i];
|
2013-12-05 04:57:01 -05:00
|
|
|
|
2017-11-22 14:58:09 +01:00
|
|
|
con_stamp = nm_settings_connection_autoconnect_retries_blocked_until (connection);
|
2011-06-17 12:43:28 +02:00
|
|
|
if (con_stamp == 0)
|
|
|
|
|
continue;
|
2013-12-05 04:57:01 -05:00
|
|
|
|
2014-01-30 15:59:40 +01:00
|
|
|
if (con_stamp <= now) {
|
2017-10-24 08:26:25 +02:00
|
|
|
nm_settings_connection_autoconnect_retries_reset (connection);
|
2011-12-06 16:33:00 -06:00
|
|
|
changed = TRUE;
|
2013-12-05 04:57:01 -05:00
|
|
|
} else if (min_stamp == 0 || min_stamp > con_stamp)
|
2011-06-17 12:43:28 +02:00
|
|
|
min_stamp = con_stamp;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Schedule the handler again if there are some stamps left */
|
2013-12-05 04:57:01 -05:00
|
|
|
if (min_stamp != 0)
|
2016-03-29 10:46:07 +02:00
|
|
|
priv->reset_retries_id = g_timeout_add_seconds (min_stamp - now, reset_connections_retries, self);
|
2011-12-06 16:33:00 -06:00
|
|
|
|
|
|
|
|
/* If anything changed, try to activate the newly re-enabled connections */
|
|
|
|
|
if (changed)
|
2016-03-29 10:46:07 +02:00
|
|
|
schedule_activate_all (self);
|
2011-12-06 16:33:00 -06:00
|
|
|
|
2011-06-17 12:43:28 +02:00
|
|
|
return FALSE;
|
|
|
|
|
}
|
|
|
|
|
|
2017-11-22 17:03:22 +01:00
|
|
|
static void
|
|
|
|
|
_connection_autoconnect_retries_set (NMPolicy *self,
|
|
|
|
|
NMSettingsConnection *connection,
|
|
|
|
|
int tries)
|
|
|
|
|
{
|
|
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
|
|
|
|
|
|
|
|
|
nm_assert (NM_IS_SETTINGS_CONNECTION (connection));
|
|
|
|
|
nm_assert (tries >= 0);
|
|
|
|
|
|
|
|
|
|
nm_settings_connection_autoconnect_retries_set (connection, tries);
|
|
|
|
|
|
|
|
|
|
if (tries == 0) {
|
|
|
|
|
/* Schedule a handler to reset retries count */
|
|
|
|
|
if (!priv->reset_retries_id) {
|
|
|
|
|
gint32 retry_time = nm_settings_connection_autoconnect_retries_blocked_until (connection);
|
|
|
|
|
|
|
|
|
|
g_warn_if_fail (retry_time != 0);
|
|
|
|
|
priv->reset_retries_id = g_timeout_add_seconds (MAX (0, retry_time - nm_utils_get_monotonic_timestamp_s ()), reset_connections_retries, self);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2011-12-07 13:15:24 +01:00
|
|
|
static void
|
2016-03-29 10:46:07 +02:00
|
|
|
activate_slave_connections (NMPolicy *self, NMDevice *device)
|
2011-12-07 13:15:24 +01:00
|
|
|
{
|
2016-03-29 10:46:07 +02:00
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
const char *master_device;
|
|
|
|
|
const char *master_uuid_settings = NULL;
|
|
|
|
|
const char *master_uuid_applied = NULL;
|
2017-02-03 14:15:16 +01:00
|
|
|
guint i;
|
2014-03-05 20:56:15 +01:00
|
|
|
NMActRequest *req;
|
2016-06-30 14:13:19 +02:00
|
|
|
gboolean internal_activation = FALSE;
|
2017-11-22 11:02:06 +01:00
|
|
|
NMSettingsConnection *const*connections;
|
2017-11-24 12:47:59 +01:00
|
|
|
gboolean changed;
|
2011-12-07 13:15:24 +01:00
|
|
|
|
|
|
|
|
master_device = nm_device_get_iface (device);
|
|
|
|
|
g_assert (master_device);
|
|
|
|
|
|
2014-03-05 20:56:15 +01:00
|
|
|
req = nm_device_get_act_request (device);
|
2015-07-14 16:53:24 +02:00
|
|
|
if (req) {
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
NMConnection *connection;
|
|
|
|
|
NMSettingsConnection *sett_conn;
|
2016-06-30 14:13:19 +02:00
|
|
|
NMAuthSubject *subject;
|
2015-07-14 16:53:24 +02:00
|
|
|
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
connection = nm_active_connection_get_applied_connection (NM_ACTIVE_CONNECTION (req));
|
|
|
|
|
if (connection)
|
|
|
|
|
master_uuid_applied = nm_connection_get_uuid (connection);
|
|
|
|
|
|
|
|
|
|
sett_conn = nm_active_connection_get_settings_connection (NM_ACTIVE_CONNECTION (req));
|
|
|
|
|
if (sett_conn) {
|
|
|
|
|
master_uuid_settings = nm_settings_connection_get_uuid (sett_conn);
|
|
|
|
|
if (nm_streq0 (master_uuid_settings, master_uuid_applied))
|
2015-07-14 16:53:24 +02:00
|
|
|
master_uuid_settings = NULL;
|
|
|
|
|
}
|
2016-06-30 14:13:19 +02:00
|
|
|
|
|
|
|
|
subject = nm_active_connection_get_subject (NM_ACTIVE_CONNECTION (req));
|
|
|
|
|
internal_activation = subject && nm_auth_subject_is_internal (subject);
|
2015-07-14 16:53:24 +02:00
|
|
|
}
|
2014-03-05 20:56:15 +01:00
|
|
|
|
2017-11-24 12:47:59 +01:00
|
|
|
changed = FALSE;
|
2017-11-22 11:02:06 +01:00
|
|
|
connections = nm_settings_get_connections (priv->settings, NULL);
|
2017-09-28 16:43:06 +02:00
|
|
|
for (i = 0; connections[i]; i++) {
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
NMSettingsConnection *sett_conn = connections[i];
|
2017-09-28 16:43:06 +02:00
|
|
|
NMSettingConnection *s_slave_con;
|
|
|
|
|
const char *slave_master;
|
2016-06-30 14:13:19 +02:00
|
|
|
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
s_slave_con = nm_connection_get_setting_connection (nm_settings_connection_get_connection (sett_conn));
|
2017-09-28 16:43:06 +02:00
|
|
|
slave_master = nm_setting_connection_get_master (s_slave_con);
|
|
|
|
|
if (!slave_master)
|
|
|
|
|
continue;
|
2017-11-22 22:23:13 +01:00
|
|
|
if (!NM_IN_STRSET (slave_master, master_device,
|
|
|
|
|
master_uuid_applied,
|
|
|
|
|
master_uuid_settings))
|
|
|
|
|
continue;
|
2011-12-07 13:15:24 +01:00
|
|
|
|
2017-11-24 12:47:59 +01:00
|
|
|
if (!internal_activation) {
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
if (nm_settings_connection_autoconnect_retries_get (sett_conn) == 0)
|
2017-11-24 12:47:59 +01:00
|
|
|
changed = TRUE;
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
nm_settings_connection_autoconnect_retries_reset (sett_conn);
|
2017-11-24 12:47:59 +01:00
|
|
|
}
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
if (nm_settings_connection_autoconnect_blocked_reason_set (sett_conn,
|
2017-11-24 12:47:59 +01:00
|
|
|
NM_SETTINGS_AUTO_CONNECT_BLOCKED_REASON_FAILED,
|
2017-11-24 12:59:15 +01:00
|
|
|
FALSE)) {
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
if (!nm_settings_connection_autoconnect_is_blocked (sett_conn))
|
2017-11-24 12:59:15 +01:00
|
|
|
changed = TRUE;
|
|
|
|
|
}
|
2011-12-07 13:15:24 +01:00
|
|
|
}
|
|
|
|
|
|
2017-11-24 12:47:59 +01:00
|
|
|
if (changed)
|
|
|
|
|
schedule_activate_all (self);
|
2011-12-07 13:15:24 +01:00
|
|
|
}
|
|
|
|
|
|
2012-08-21 17:49:41 +02:00
|
|
|
static gboolean
|
2016-03-29 10:46:07 +02:00
|
|
|
activate_secondary_connections (NMPolicy *self,
|
2012-08-21 17:49:41 +02:00
|
|
|
NMConnection *connection,
|
|
|
|
|
NMDevice *device)
|
|
|
|
|
{
|
2016-03-29 10:46:07 +02:00
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2012-08-21 17:49:41 +02:00
|
|
|
NMSettingConnection *s_con;
|
|
|
|
|
NMActiveConnection *ac;
|
|
|
|
|
PendingSecondaryData *secondary_data;
|
|
|
|
|
GSList *secondary_ac_list = NULL;
|
|
|
|
|
GError *error = NULL;
|
|
|
|
|
guint32 i;
|
|
|
|
|
gboolean success = TRUE;
|
core: improve and fix keeping connection active based on "connection.permissions"
By setting "connection.permissions", a profile is restricted to a
particular user.
That means for example, that another user cannot see, modify, delete,
activate or deactivate the profile. It also means, that the profile
will only autoconnect when the user is logged in (has a session).
Note that root is always able to activate the profile. Likewise, the
user is also allowed to manually activate the own profile, even if no
session currently exists (which can easily happen with `sudo`).
When the user logs out (the session goes away), we want do disconnect
the profile, however there are conflicting goals here:
1) if the profile was activate by root user, then logging out the user
should not disconnect the profile. The patch fixes that by not
binding the activation to the connection, if the activation is done
by the root user.
2) if the profile was activated by the owner when it had no session,
then it should stay alive until the user logs in (once) and logs
out again. This is already handled by the previous commit.
Yes, this point is odd. If you first do
$ sudo -u $OTHER_USER nmcli connection up $PROFILE
the profile activates despite not having a session. If you then
$ ssh guest@localhost nmcli device
you'll still see the profile active. However, the moment the SSH session
ends, a session closes and the profile disconnects. It's unclear, how to
solve that any better. I think, a user who cares about this, should not
activate the profile without having a session in the first place.
There are quite some special cases, in particular with internal
activations. In those cases we need to decide whether to bind the
activation to the profile's visibility.
Also, expose the "bind" setting in the D-Bus API. Note, that in the future
this flag may be modified via D-Bus API. Like we may also add related API
that allows to tweak the lifetime of the activation.
Also, I think we broke handling of connection visiblity with 37e8c53eeed
"core: Introduce helper class to track connection keep alive". This
should be fixed now too, with improved behavior.
Fixes: 37e8c53eeed579fe34a68819cd12f3295d581394
https://bugzilla.redhat.com/show_bug.cgi?id=1530977
2018-11-21 13:30:16 +01:00
|
|
|
NMActivationStateFlags initial_state_flags;
|
2012-08-21 17:49:41 +02:00
|
|
|
|
|
|
|
|
s_con = nm_connection_get_setting_connection (connection);
|
core: improve and fix keeping connection active based on "connection.permissions"
By setting "connection.permissions", a profile is restricted to a
particular user.
That means for example, that another user cannot see, modify, delete,
activate or deactivate the profile. It also means, that the profile
will only autoconnect when the user is logged in (has a session).
Note that root is always able to activate the profile. Likewise, the
user is also allowed to manually activate the own profile, even if no
session currently exists (which can easily happen with `sudo`).
When the user logs out (the session goes away), we want do disconnect
the profile, however there are conflicting goals here:
1) if the profile was activate by root user, then logging out the user
should not disconnect the profile. The patch fixes that by not
binding the activation to the connection, if the activation is done
by the root user.
2) if the profile was activated by the owner when it had no session,
then it should stay alive until the user logs in (once) and logs
out again. This is already handled by the previous commit.
Yes, this point is odd. If you first do
$ sudo -u $OTHER_USER nmcli connection up $PROFILE
the profile activates despite not having a session. If you then
$ ssh guest@localhost nmcli device
you'll still see the profile active. However, the moment the SSH session
ends, a session closes and the profile disconnects. It's unclear, how to
solve that any better. I think, a user who cares about this, should not
activate the profile without having a session in the first place.
There are quite some special cases, in particular with internal
activations. In those cases we need to decide whether to bind the
activation to the profile's visibility.
Also, expose the "bind" setting in the D-Bus API. Note, that in the future
this flag may be modified via D-Bus API. Like we may also add related API
that allows to tweak the lifetime of the activation.
Also, I think we broke handling of connection visiblity with 37e8c53eeed
"core: Introduce helper class to track connection keep alive". This
should be fixed now too, with improved behavior.
Fixes: 37e8c53eeed579fe34a68819cd12f3295d581394
https://bugzilla.redhat.com/show_bug.cgi?id=1530977
2018-11-21 13:30:16 +01:00
|
|
|
nm_assert (NM_IS_SETTING_CONNECTION (s_con));
|
|
|
|
|
|
|
|
|
|
/* we propagate the activation's state flags. */
|
|
|
|
|
initial_state_flags = nm_device_get_activation_state_flags (device)
|
|
|
|
|
& NM_ACTIVATION_STATE_FLAG_LIFETIME_BOUND_TO_PROFILE_VISIBILITY;
|
2012-08-21 17:49:41 +02:00
|
|
|
|
|
|
|
|
for (i = 0; i < nm_setting_connection_get_num_secondaries (s_con); i++) {
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
NMSettingsConnection *sett_conn;
|
2012-08-21 17:49:41 +02:00
|
|
|
const char *sec_uuid = nm_setting_connection_get_secondary (s_con, i);
|
2013-11-06 10:35:59 -05:00
|
|
|
NMActRequest *req;
|
2012-08-21 17:49:41 +02:00
|
|
|
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
sett_conn = nm_settings_get_connection_by_uuid (priv->settings, sec_uuid);
|
|
|
|
|
if (!sett_conn) {
|
2016-03-03 09:20:18 +01:00
|
|
|
_LOGW (LOGD_DEVICE, "secondary connection '%s' auto-activation failed: The connection doesn't exist.",
|
|
|
|
|
sec_uuid);
|
2012-08-21 17:49:41 +02:00
|
|
|
success = FALSE;
|
|
|
|
|
break;
|
|
|
|
|
}
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
|
|
|
|
|
if (!nm_connection_is_type (nm_settings_connection_get_connection (sett_conn),
|
|
|
|
|
NM_SETTING_VPN_SETTING_NAME)) {
|
2016-03-03 09:20:18 +01:00
|
|
|
_LOGW (LOGD_DEVICE, "secondary connection '%s (%s)' auto-activation failed: The connection is not a VPN.",
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
nm_settings_connection_get_id (sett_conn), sec_uuid);
|
2013-11-06 10:35:59 -05:00
|
|
|
success = FALSE;
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
req = nm_device_get_act_request (device);
|
|
|
|
|
|
2016-03-03 09:20:18 +01:00
|
|
|
_LOGD (LOGD_DEVICE, "activating secondary connection '%s (%s)' for base connection '%s (%s)'",
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
nm_settings_connection_get_id (sett_conn), sec_uuid,
|
2016-03-03 09:20:18 +01:00
|
|
|
nm_connection_get_id (connection), nm_connection_get_uuid (connection));
|
2013-11-06 10:35:59 -05:00
|
|
|
ac = nm_manager_activate_connection (priv->manager,
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
sett_conn,
|
2016-09-07 17:47:26 +02:00
|
|
|
NULL,
|
core/dbus: rework D-Bus implementation to use lower layer GDBusConnection API
Previously, we used the generated GDBusInterfaceSkeleton types and glued
them via the NMExportedObject base class to our NM types. We also used
GDBusObjectManagerServer.
Don't do that anymore. The resulting code was more complicated despite (or
because?) using generated classes. It was hard to understand, complex, had
ordering-issues, and had a runtime and memory overhead.
This patch refactors this entirely and uses the lower layer API GDBusConnection
directly. It replaces the generated code, GDBusInterfaceSkeleton, and
GDBusObjectManagerServer. All this is now done by NMDbusObject and NMDBusManager
and static descriptor instances of type GDBusInterfaceInfo.
This adds a net plus of more then 1300 lines of hand written code. I claim
that this implementation is easier to understand. Note that previously we
also required extensive and complex glue code to bind our objects to the
generated skeleton objects. Instead, now glue our objects directly to
GDBusConnection. The result is more immediate and gets rid of layers of
code in between.
Now that the D-Bus glue us more under our control, we can address issus and
bottlenecks better, instead of adding code to bend the generated skeletons
to our needs.
Note that the current implementation now only supports one D-Bus connection.
That was effectively the case already, although there were places (and still are)
where the code pretends it could also support connections from a private socket.
We dropped private socket support mainly because it was unused, untested and
buggy, but also because GDBusObjectManagerServer could not export the same
objects on multiple connections. Now, it would be rather straight forward to
fix that and re-introduce ObjectManager on each private connection. But this
commit doesn't do that yet, and the new code intentionally supports only one
D-Bus connection.
Also, the D-Bus startup was simplified. There is no retry, either nm_dbus_manager_start()
succeeds, or it detects the initrd case. In the initrd case, bus manager never tries to
connect to D-Bus. Since the initrd scenario is not yet used/tested, this is good enough
for the moment. It could be easily extended later, for example with polling whether the
system bus appears (like was done previously). Also, restart of D-Bus daemon isn't
supported either -- just like before.
Note how NMDBusManager now implements the ObjectManager D-Bus interface
directly.
Also, this fixes race issues in the server, by no longer delaying
PropertiesChanged signals. NMExportedObject would collect changed
properties and send the signal out in idle_emit_properties_changed()
on idle. This messes up the ordering of change events w.r.t. other
signals and events on the bus. Note that not only NMExportedObject
messed up the ordering. Also the generated code would hook into
notify() and process change events in and idle handle, exhibiting the
same ordering issue too.
No longer do that. PropertiesChanged signals will be sent right away
by hooking into dispatch_properties_changed(). This means, changing
a property in quick succession will no longer be combined and is
guaranteed to emit signals for each individual state. Quite possibly
we emit now more PropertiesChanged signals then before.
However, we are now able to group a set of changes by using standard
g_object_freeze_notify()/g_object_thaw_notify(). We probably should
make more use of that.
Also, now that our signals are all handled in the right order, we
might find places where we still emit them in the wrong order. But that
is then due to the order in which our GObjects emit signals, not due
to an ill behavior of the D-Bus glue. Possibly we need to identify
such ordering issues and fix them.
Numbers (for contrib/rpm --without debug on x86_64):
- the patch changes the code size of NetworkManager by
- 2809360 bytes
+ 2537528 bytes (-9.7%)
- Runtime measurements are harder because there is a large variance
during testing. In other words, the numbers are not reproducible.
Currently, the implementation performs no caching of GVariants at all,
but it would be rather simple to add it, if that turns out to be
useful.
Anyway, without strong claim, it seems that the new form tends to
perform slightly better. That would be no surprise.
$ time (for i in {1..1000}; do nmcli >/dev/null || break; echo -n .; done)
- real 1m39.355s
+ real 1m37.432s
$ time (for i in {1..2000}; do busctl call org.freedesktop.NetworkManager /org/freedesktop org.freedesktop.DBus.ObjectManager GetManagedObjects > /dev/null || break; echo -n .; done)
- real 0m26.843s
+ real 0m25.281s
- Regarding RSS size, just looking at the processes in similar
conditions, doesn't give a large difference. On my system they
consume about 19MB RSS. It seems that the new version has a
slightly smaller RSS size.
- 19356 RSS
+ 18660 RSS
2018-02-26 13:51:52 +01:00
|
|
|
nm_dbus_object_get_path (NM_DBUS_OBJECT (req)),
|
2013-11-06 10:35:59 -05:00
|
|
|
device,
|
|
|
|
|
nm_active_connection_get_subject (NM_ACTIVE_CONNECTION (req)),
|
2017-03-07 11:04:36 +01:00
|
|
|
NM_ACTIVATION_TYPE_MANAGED,
|
2018-03-28 17:18:04 +02:00
|
|
|
nm_active_connection_get_activation_reason (NM_ACTIVE_CONNECTION (req)),
|
core: improve and fix keeping connection active based on "connection.permissions"
By setting "connection.permissions", a profile is restricted to a
particular user.
That means for example, that another user cannot see, modify, delete,
activate or deactivate the profile. It also means, that the profile
will only autoconnect when the user is logged in (has a session).
Note that root is always able to activate the profile. Likewise, the
user is also allowed to manually activate the own profile, even if no
session currently exists (which can easily happen with `sudo`).
When the user logs out (the session goes away), we want do disconnect
the profile, however there are conflicting goals here:
1) if the profile was activate by root user, then logging out the user
should not disconnect the profile. The patch fixes that by not
binding the activation to the connection, if the activation is done
by the root user.
2) if the profile was activated by the owner when it had no session,
then it should stay alive until the user logs in (once) and logs
out again. This is already handled by the previous commit.
Yes, this point is odd. If you first do
$ sudo -u $OTHER_USER nmcli connection up $PROFILE
the profile activates despite not having a session. If you then
$ ssh guest@localhost nmcli device
you'll still see the profile active. However, the moment the SSH session
ends, a session closes and the profile disconnects. It's unclear, how to
solve that any better. I think, a user who cares about this, should not
activate the profile without having a session in the first place.
There are quite some special cases, in particular with internal
activations. In those cases we need to decide whether to bind the
activation to the profile's visibility.
Also, expose the "bind" setting in the D-Bus API. Note, that in the future
this flag may be modified via D-Bus API. Like we may also add related API
that allows to tweak the lifetime of the activation.
Also, I think we broke handling of connection visiblity with 37e8c53eeed
"core: Introduce helper class to track connection keep alive". This
should be fixed now too, with improved behavior.
Fixes: 37e8c53eeed579fe34a68819cd12f3295d581394
https://bugzilla.redhat.com/show_bug.cgi?id=1530977
2018-11-21 13:30:16 +01:00
|
|
|
initial_state_flags,
|
2013-11-06 10:35:59 -05:00
|
|
|
&error);
|
|
|
|
|
if (ac)
|
|
|
|
|
secondary_ac_list = g_slist_append (secondary_ac_list, g_object_ref (ac));
|
|
|
|
|
else {
|
2016-03-03 09:20:18 +01:00
|
|
|
_LOGW (LOGD_DEVICE, "secondary connection '%s (%s)' auto-activation failed: (%d) %s",
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
nm_settings_connection_get_id (sett_conn), sec_uuid,
|
2016-03-03 09:20:18 +01:00
|
|
|
error->code,
|
|
|
|
|
error->message);
|
2013-11-06 10:35:59 -05:00
|
|
|
g_clear_error (&error);
|
|
|
|
|
success = FALSE;
|
|
|
|
|
break;
|
|
|
|
|
}
|
2012-08-21 17:49:41 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (success && secondary_ac_list != NULL) {
|
|
|
|
|
secondary_data = pending_secondary_data_new (device, secondary_ac_list);
|
2013-08-22 10:10:17 -04:00
|
|
|
priv->pending_secondaries = g_slist_append (priv->pending_secondaries, secondary_data);
|
2012-08-21 17:49:41 +02:00
|
|
|
} else
|
2013-10-07 17:35:03 -05:00
|
|
|
g_slist_free_full (secondary_ac_list, g_object_unref);
|
2012-08-21 17:49:41 +02:00
|
|
|
|
|
|
|
|
return success;
|
|
|
|
|
}
|
|
|
|
|
|
2007-02-08 15:34:26 +00:00
|
|
|
static void
|
2008-07-11 10:28:53 +00:00
|
|
|
device_state_changed (NMDevice *device,
|
|
|
|
|
NMDeviceState new_state,
|
|
|
|
|
NMDeviceState old_state,
|
|
|
|
|
NMDeviceStateReason reason,
|
|
|
|
|
gpointer user_data)
|
2007-02-08 15:34:26 +00:00
|
|
|
{
|
2016-04-04 14:36:34 +02:00
|
|
|
NMPolicyPrivate *priv = user_data;
|
2016-09-29 13:49:01 +02:00
|
|
|
NMPolicy *self = _PRIV_TO_SELF (priv);
|
2017-09-25 22:20:47 +02:00
|
|
|
NMActiveConnection *ac;
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
NMSettingsConnection *sett_conn = nm_device_get_settings_connection (device);
|
2012-05-29 22:17:26 -05:00
|
|
|
NMIP4Config *ip4_config;
|
|
|
|
|
NMIP6Config *ip6_config;
|
2014-03-04 16:51:01 +01:00
|
|
|
NMSettingConnection *s_con = NULL;
|
2007-02-08 15:34:26 +00:00
|
|
|
|
2017-10-25 21:29:45 +02:00
|
|
|
switch (nm_device_state_reason_check (reason)) {
|
|
|
|
|
case NM_DEVICE_STATE_REASON_GSM_SIM_PIN_REQUIRED:
|
|
|
|
|
case NM_DEVICE_STATE_REASON_GSM_SIM_PUK_REQUIRED:
|
|
|
|
|
case NM_DEVICE_STATE_REASON_SIM_PIN_INCORRECT:
|
|
|
|
|
case NM_DEVICE_STATE_REASON_GSM_APN_FAILED:
|
2018-12-11 15:50:41 +01:00
|
|
|
/* Block autoconnection at settings level if there is any settings-specific
|
|
|
|
|
* error reported by the modem (e.g. wrong SIM-PIN or wrong APN). Do not block
|
|
|
|
|
* autoconnection at settings level for errors in the device domain (e.g.
|
|
|
|
|
* a missing SIM or wrong modem initialization).
|
2017-10-25 21:29:45 +02:00
|
|
|
*/
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
if (sett_conn) {
|
|
|
|
|
nm_settings_connection_autoconnect_blocked_reason_set (sett_conn,
|
2017-11-24 10:53:39 +01:00
|
|
|
NM_SETTINGS_AUTO_CONNECT_BLOCKED_REASON_FAILED,
|
|
|
|
|
TRUE);
|
2017-10-25 21:29:45 +02:00
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
2008-07-11 10:28:53 +00:00
|
|
|
switch (new_state) {
|
2008-04-07 Dan Williams <dcbw@redhat.com>
* include/NetworkManager.h
- Remove the DOWN and CANCELLED device states
- Add UNMANAGED and UNAVAILABLE device states
- Document the device states
* introspection/nm-device.xml
src/nm-device-interface.c
src/nm-device-interface.h
- Add the 'managed' property
* test/nm-tool.c
- (detail_device): print out device state
* src/NetworkManagerSystem.h
src/backends/NetworkManagerArch.c
src/backends/NetworkManagerDebian.c
src/backends/NetworkManagerFrugalware.c
src/backends/NetworkManagerGentoo.c
src/backends/NetworkManagerMandriva.c
src/backends/NetworkManagerPaldo.c
src/backends/NetworkManagerRedHat.c
src/backends/NetworkManagerSlackware.c
src/backends/NetworkManagerSuSE.c
- (nm_system_device_get_system_config, nm_system_device_get_disabled
nm_system_device_free_system_config): remove; they were unused and
their functionality should be re-implemented in each distro's
system settings service plugin
* src/nm-gsm-device.c
src/nm-gsm-device.h
src/nm-cdma-device.c
src/nm-cdma-device.h
- (*_new): take the 'managed' argument
* src/nm-device.c
- (nm_device_set_address): remove, fold into nm_device_bring_up()
- (nm_device_init): start in unmanaged state, not disconnected
- (constructor): don't start device until the system settings service
has had a chance to figure out if the device is managed or not
- (nm_device_deactivate, nm_device_bring_up, nm_device_bring_down):
don't set device state here, let callers handle that as appropriate
- (nm_device_dispose): don't touch the device if it's not managed
- (set_property, get_property, nm_device_class_init): implement the
'managed' property
- (nm_device_state_changed): bring the device up if its now managed,
and deactivate it if it used to be active
- (nm_device_get_managed, nm_device_set_managed): do the right thing
with the managed state
* src/nm-hal-manager.c
- (wired_device_creator, wireless_device_creator, modem_device_creator):
take initial managed state and pass it along to device constructors
- (create_device_and_add_to_list): get managed state and pass to
type creators
* src/nm-device-802-11-wireless.c
- (real_can_activate): fold in most of
nm_device_802_11_wireless_can_activate()
- (can_scan): can't scan in UNAVAILABLE or UNMANAGED
- (link_timeout_cb): instead of deactivating, change device state and
let the device state handler to it
- (real_update_hw_address): clean up
- (state_changed_cb): when entering UNAVAILABLE state, schedule an idle
handler to transition to DISCONNECTED if the device isn't rfkilled
* src/nm-device-802-3-ethernet.c
- (set_carrier): move above callers and get rid of prototype
- (device_state_changed): when entering UNAVAILABLE state, schedule an
idle handler to transition to DISCONNECTED if the device has a
carrier
- (real_update_hw_address): clean up
- (link_timeout_cb, ppp_state_changed): change state instead of calling
deactivation directly as deactivation doesn't change state anymore
* src/NetworkManagerPolicy.c
- (schedule_activate_check): yay, remove wireless_enabled hack since
the NMManager and wireless devices work that out themselves now
- (device_state_changed): change to a switch and update for new device
states
- (device_carrier_changed): remove; device handles this now through
state changes
- (device_added): don't care about carrier any more; the initial
activation check will happen when the device transitions to
DISCONNECTED
* src/nm-manager.c
- (dispose): clear unmanaged devices
- (handle_unmanaged_devices): update unmanaged device list and toggle
the managed property on each device when needed
- (system_settings_properties_changed_cb): handle signals from the
system settings service
- (system_settings_get_unmanaged_devices_cb): handle callback from
getting the unmanaged device list method call
- (query_unmanaged_devices): ask the system settings service for its
list of unmanaged devices
- (nm_manager_name_owner_changed, initial_get_connections): get unmanaged
devices
- (manager_set_wireless_enabled): push rfkill state down to wireless
devices directly and let them handle the necessary state transitions
- (manager_device_state_changed): update for new device states
- (nm_manager_add_device): set initial rfkill state on wireless devices
- (nm_manager_remove_device): don't touch the device if it's unmanaged
- (nm_manager_activate_connection): return error if the device is
unmanaged
- (nm_manager_sleep): handle new device states correctly; don't change
the state of unavailable/unmanaged devices
* libnm-glib/nm-device-802-11-wireless.c
- (state_changed_cb): update for new device states
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@3540 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-04-08 02:58:02 +00:00
|
|
|
case NM_DEVICE_STATE_FAILED:
|
2009-09-09 12:21:55 -07:00
|
|
|
/* Mark the connection invalid if it failed during activation so that
|
|
|
|
|
* it doesn't get automatically chosen over and over and over again.
|
|
|
|
|
*/
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
if ( sett_conn
|
2011-11-17 22:56:51 -06:00
|
|
|
&& old_state >= NM_DEVICE_STATE_PREPARE
|
|
|
|
|
&& old_state <= NM_DEVICE_STATE_ACTIVATED) {
|
2020-07-16 11:58:20 +02:00
|
|
|
gboolean blocked = FALSE;
|
2017-10-04 13:57:15 +02:00
|
|
|
int tries;
|
core: fix race of blocking autoconnect for no-secrets when a new secret-agent registers
When activation of the connection fails with no-secrets, we block
autoconnect due to that. However, NMPolicy also unblocks such
autoconnect, whenever a new secret-agent registers. The reason
is obviously, that the new secret-agent might be able to provide
the previously missing secrets.
However, there is a race between
- making the secret request, failing activation and blocking autoconnect
- new secret-agent registers
If the secret-agent registers after making the request, but before we
block autoconnect, then autoconnect stays blocked.
[1511468634.5759] device (wlp4s0): state change: config -> need-auth (reason 'none', sys-iface-state: 'managed')
[1511468634.5772] device (wlp4s0): No agents were available for this request.
[1511468638.4082] agent-manager: req[0x55ea7e58a5d0, :1.32/org.kde.plasma.networkmanagement/1000]: agent registered
[1511468638.4082] policy: re-enabling autoconnect for all connections with failed secrets
[1511468664.6280] device (wlp4s0): state change: need-auth -> failed (reason 'no-secrets', sys-iface-state: 'managed')
[1511468664.6287] policy: connection 'tuxmobil' now blocked from autoconnect due to no secrets
Note the long timing between making the secret request and the
activation failure. This race already existed before, but now with
WPS push-button method enabled by default, the duraction of the
activation is much longer and the race is easy to hit.
https://bugzilla.gnome.org/show_bug.cgi?id=790571
2017-11-27 09:33:32 +01:00
|
|
|
guint64 con_v;
|
2011-04-01 13:30:37 -05:00
|
|
|
|
2017-02-23 15:19:03 +01:00
|
|
|
if (nm_device_state_reason_check (reason) == NM_DEVICE_STATE_REASON_NO_SECRETS) {
|
core: fix race of blocking autoconnect for no-secrets when a new secret-agent registers
When activation of the connection fails with no-secrets, we block
autoconnect due to that. However, NMPolicy also unblocks such
autoconnect, whenever a new secret-agent registers. The reason
is obviously, that the new secret-agent might be able to provide
the previously missing secrets.
However, there is a race between
- making the secret request, failing activation and blocking autoconnect
- new secret-agent registers
If the secret-agent registers after making the request, but before we
block autoconnect, then autoconnect stays blocked.
[1511468634.5759] device (wlp4s0): state change: config -> need-auth (reason 'none', sys-iface-state: 'managed')
[1511468634.5772] device (wlp4s0): No agents were available for this request.
[1511468638.4082] agent-manager: req[0x55ea7e58a5d0, :1.32/org.kde.plasma.networkmanagement/1000]: agent registered
[1511468638.4082] policy: re-enabling autoconnect for all connections with failed secrets
[1511468664.6280] device (wlp4s0): state change: need-auth -> failed (reason 'no-secrets', sys-iface-state: 'managed')
[1511468664.6287] policy: connection 'tuxmobil' now blocked from autoconnect due to no secrets
Note the long timing between making the secret request and the
activation failure. This race already existed before, but now with
WPS push-button method enabled by default, the duraction of the
activation is much longer and the race is easy to hit.
https://bugzilla.gnome.org/show_bug.cgi?id=790571
2017-11-27 09:33:32 +01:00
|
|
|
/* we want to block the connection from auto-connect if it failed due to no-secrets.
|
|
|
|
|
* However, if a secret-agent registered, since the connection made the last
|
|
|
|
|
* secret-request, we do not block it. The new secret-agent might not yet
|
|
|
|
|
* been consulted, and it may be able to provide the secrets.
|
|
|
|
|
*
|
|
|
|
|
* We detect this by using a version-id of the agent-manager, which increments
|
2018-03-06 13:04:00 +00:00
|
|
|
* whenever new agents register. Note that the agent-manager's version-id is
|
|
|
|
|
* never zero and strictly increasing.
|
|
|
|
|
*
|
|
|
|
|
* A connection's version-id of zero means that the connection never tried to request secrets.
|
|
|
|
|
* That can happen when nm_settings_connection_get_secrets() fails early without actually
|
|
|
|
|
* consulting any agents.
|
|
|
|
|
*/
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
con_v = nm_settings_connection_get_last_secret_agent_version_id (sett_conn);
|
core: fix race of blocking autoconnect for no-secrets when a new secret-agent registers
When activation of the connection fails with no-secrets, we block
autoconnect due to that. However, NMPolicy also unblocks such
autoconnect, whenever a new secret-agent registers. The reason
is obviously, that the new secret-agent might be able to provide
the previously missing secrets.
However, there is a race between
- making the secret request, failing activation and blocking autoconnect
- new secret-agent registers
If the secret-agent registers after making the request, but before we
block autoconnect, then autoconnect stays blocked.
[1511468634.5759] device (wlp4s0): state change: config -> need-auth (reason 'none', sys-iface-state: 'managed')
[1511468634.5772] device (wlp4s0): No agents were available for this request.
[1511468638.4082] agent-manager: req[0x55ea7e58a5d0, :1.32/org.kde.plasma.networkmanagement/1000]: agent registered
[1511468638.4082] policy: re-enabling autoconnect for all connections with failed secrets
[1511468664.6280] device (wlp4s0): state change: need-auth -> failed (reason 'no-secrets', sys-iface-state: 'managed')
[1511468664.6287] policy: connection 'tuxmobil' now blocked from autoconnect due to no secrets
Note the long timing between making the secret request and the
activation failure. This race already existed before, but now with
WPS push-button method enabled by default, the duraction of the
activation is much longer and the race is easy to hit.
https://bugzilla.gnome.org/show_bug.cgi?id=790571
2017-11-27 09:33:32 +01:00
|
|
|
if ( con_v == 0
|
2020-07-16 11:58:20 +02:00
|
|
|
|| con_v == nm_agent_manager_get_agent_version_id (priv->agent_mgr)) {
|
|
|
|
|
_LOGD (LOGD_DEVICE, "connection '%s' now blocked from autoconnect due to no secrets",
|
|
|
|
|
nm_settings_connection_get_id (sett_conn));
|
|
|
|
|
nm_settings_connection_autoconnect_blocked_reason_set (sett_conn,
|
|
|
|
|
NM_SETTINGS_AUTO_CONNECT_BLOCKED_REASON_NO_SECRETS,
|
|
|
|
|
TRUE);
|
|
|
|
|
blocked = TRUE;
|
|
|
|
|
}
|
|
|
|
|
} else if (nm_device_state_reason_check (reason) == NM_DEVICE_STATE_REASON_DEPENDENCY_FAILED) {
|
|
|
|
|
/* A connection that fails due to dependency-failed is not
|
|
|
|
|
* able to reconnect until the master connection activates
|
|
|
|
|
* again; when this happens, the master clears the blocked
|
|
|
|
|
* reason for all its slaves in activate_slave_connections()
|
|
|
|
|
* and tries to reconnect them. For this to work, the slave
|
|
|
|
|
* should be marked as blocked when it fails with
|
|
|
|
|
* dependency-failed.
|
|
|
|
|
*/
|
|
|
|
|
_LOGD (LOGD_DEVICE, "connection '%s' now blocked from autoconnect due to failed dependency",
|
|
|
|
|
nm_settings_connection_get_id (sett_conn));
|
|
|
|
|
nm_settings_connection_autoconnect_blocked_reason_set (sett_conn,
|
|
|
|
|
NM_SETTINGS_AUTO_CONNECT_BLOCKED_REASON_FAILED,
|
|
|
|
|
TRUE);
|
|
|
|
|
blocked = TRUE;
|
core: fix race of blocking autoconnect for no-secrets when a new secret-agent registers
When activation of the connection fails with no-secrets, we block
autoconnect due to that. However, NMPolicy also unblocks such
autoconnect, whenever a new secret-agent registers. The reason
is obviously, that the new secret-agent might be able to provide
the previously missing secrets.
However, there is a race between
- making the secret request, failing activation and blocking autoconnect
- new secret-agent registers
If the secret-agent registers after making the request, but before we
block autoconnect, then autoconnect stays blocked.
[1511468634.5759] device (wlp4s0): state change: config -> need-auth (reason 'none', sys-iface-state: 'managed')
[1511468634.5772] device (wlp4s0): No agents were available for this request.
[1511468638.4082] agent-manager: req[0x55ea7e58a5d0, :1.32/org.kde.plasma.networkmanagement/1000]: agent registered
[1511468638.4082] policy: re-enabling autoconnect for all connections with failed secrets
[1511468664.6280] device (wlp4s0): state change: need-auth -> failed (reason 'no-secrets', sys-iface-state: 'managed')
[1511468664.6287] policy: connection 'tuxmobil' now blocked from autoconnect due to no secrets
Note the long timing between making the secret request and the
activation failure. This race already existed before, but now with
WPS push-button method enabled by default, the duraction of the
activation is much longer and the race is easy to hit.
https://bugzilla.gnome.org/show_bug.cgi?id=790571
2017-11-27 09:33:32 +01:00
|
|
|
}
|
|
|
|
|
|
2020-07-16 11:58:20 +02:00
|
|
|
if (!blocked) {
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
tries = nm_settings_connection_autoconnect_retries_get (sett_conn);
|
2017-10-04 13:57:15 +02:00
|
|
|
if (tries > 0) {
|
|
|
|
|
_LOGD (LOGD_DEVICE, "connection '%s' failed to autoconnect; %d tries left",
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
nm_settings_connection_get_id (sett_conn), tries - 1);
|
|
|
|
|
_connection_autoconnect_retries_set (self, sett_conn, tries - 1);
|
core: fix race of blocking autoconnect for no-secrets when a new secret-agent registers
When activation of the connection fails with no-secrets, we block
autoconnect due to that. However, NMPolicy also unblocks such
autoconnect, whenever a new secret-agent registers. The reason
is obviously, that the new secret-agent might be able to provide
the previously missing secrets.
However, there is a race between
- making the secret request, failing activation and blocking autoconnect
- new secret-agent registers
If the secret-agent registers after making the request, but before we
block autoconnect, then autoconnect stays blocked.
[1511468634.5759] device (wlp4s0): state change: config -> need-auth (reason 'none', sys-iface-state: 'managed')
[1511468634.5772] device (wlp4s0): No agents were available for this request.
[1511468638.4082] agent-manager: req[0x55ea7e58a5d0, :1.32/org.kde.plasma.networkmanagement/1000]: agent registered
[1511468638.4082] policy: re-enabling autoconnect for all connections with failed secrets
[1511468664.6280] device (wlp4s0): state change: need-auth -> failed (reason 'no-secrets', sys-iface-state: 'managed')
[1511468664.6287] policy: connection 'tuxmobil' now blocked from autoconnect due to no secrets
Note the long timing between making the secret request and the
activation failure. This race already existed before, but now with
WPS push-button method enabled by default, the duraction of the
activation is much longer and the race is easy to hit.
https://bugzilla.gnome.org/show_bug.cgi?id=790571
2017-11-27 09:33:32 +01:00
|
|
|
} else if (tries != 0) {
|
2017-10-04 13:57:15 +02:00
|
|
|
_LOGD (LOGD_DEVICE, "connection '%s' failed to autoconnect; infinite tries left",
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
nm_settings_connection_get_id (sett_conn));
|
2017-10-04 13:57:15 +02:00
|
|
|
}
|
2011-04-01 18:50:58 -05:00
|
|
|
}
|
2011-04-01 13:30:37 -05:00
|
|
|
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
/* FIXME(copy-on-write-connection): avoid modifying NMConnection instances and share them via copy-on-write. */
|
|
|
|
|
nm_connection_clear_secrets (nm_settings_connection_get_connection (sett_conn));
|
2007-10-08 18:07:36 +00:00
|
|
|
}
|
2008-04-07 Dan Williams <dcbw@redhat.com>
* include/NetworkManager.h
- Remove the DOWN and CANCELLED device states
- Add UNMANAGED and UNAVAILABLE device states
- Document the device states
* introspection/nm-device.xml
src/nm-device-interface.c
src/nm-device-interface.h
- Add the 'managed' property
* test/nm-tool.c
- (detail_device): print out device state
* src/NetworkManagerSystem.h
src/backends/NetworkManagerArch.c
src/backends/NetworkManagerDebian.c
src/backends/NetworkManagerFrugalware.c
src/backends/NetworkManagerGentoo.c
src/backends/NetworkManagerMandriva.c
src/backends/NetworkManagerPaldo.c
src/backends/NetworkManagerRedHat.c
src/backends/NetworkManagerSlackware.c
src/backends/NetworkManagerSuSE.c
- (nm_system_device_get_system_config, nm_system_device_get_disabled
nm_system_device_free_system_config): remove; they were unused and
their functionality should be re-implemented in each distro's
system settings service plugin
* src/nm-gsm-device.c
src/nm-gsm-device.h
src/nm-cdma-device.c
src/nm-cdma-device.h
- (*_new): take the 'managed' argument
* src/nm-device.c
- (nm_device_set_address): remove, fold into nm_device_bring_up()
- (nm_device_init): start in unmanaged state, not disconnected
- (constructor): don't start device until the system settings service
has had a chance to figure out if the device is managed or not
- (nm_device_deactivate, nm_device_bring_up, nm_device_bring_down):
don't set device state here, let callers handle that as appropriate
- (nm_device_dispose): don't touch the device if it's not managed
- (set_property, get_property, nm_device_class_init): implement the
'managed' property
- (nm_device_state_changed): bring the device up if its now managed,
and deactivate it if it used to be active
- (nm_device_get_managed, nm_device_set_managed): do the right thing
with the managed state
* src/nm-hal-manager.c
- (wired_device_creator, wireless_device_creator, modem_device_creator):
take initial managed state and pass it along to device constructors
- (create_device_and_add_to_list): get managed state and pass to
type creators
* src/nm-device-802-11-wireless.c
- (real_can_activate): fold in most of
nm_device_802_11_wireless_can_activate()
- (can_scan): can't scan in UNAVAILABLE or UNMANAGED
- (link_timeout_cb): instead of deactivating, change device state and
let the device state handler to it
- (real_update_hw_address): clean up
- (state_changed_cb): when entering UNAVAILABLE state, schedule an idle
handler to transition to DISCONNECTED if the device isn't rfkilled
* src/nm-device-802-3-ethernet.c
- (set_carrier): move above callers and get rid of prototype
- (device_state_changed): when entering UNAVAILABLE state, schedule an
idle handler to transition to DISCONNECTED if the device has a
carrier
- (real_update_hw_address): clean up
- (link_timeout_cb, ppp_state_changed): change state instead of calling
deactivation directly as deactivation doesn't change state anymore
* src/NetworkManagerPolicy.c
- (schedule_activate_check): yay, remove wireless_enabled hack since
the NMManager and wireless devices work that out themselves now
- (device_state_changed): change to a switch and update for new device
states
- (device_carrier_changed): remove; device handles this now through
state changes
- (device_added): don't care about carrier any more; the initial
activation check will happen when the device transitions to
DISCONNECTED
* src/nm-manager.c
- (dispose): clear unmanaged devices
- (handle_unmanaged_devices): update unmanaged device list and toggle
the managed property on each device when needed
- (system_settings_properties_changed_cb): handle signals from the
system settings service
- (system_settings_get_unmanaged_devices_cb): handle callback from
getting the unmanaged device list method call
- (query_unmanaged_devices): ask the system settings service for its
list of unmanaged devices
- (nm_manager_name_owner_changed, initial_get_connections): get unmanaged
devices
- (manager_set_wireless_enabled): push rfkill state down to wireless
devices directly and let them handle the necessary state transitions
- (manager_device_state_changed): update for new device states
- (nm_manager_add_device): set initial rfkill state on wireless devices
- (nm_manager_remove_device): don't touch the device if it's unmanaged
- (nm_manager_activate_connection): return error if the device is
unmanaged
- (nm_manager_sleep): handle new device states correctly; don't change
the state of unavailable/unmanaged devices
* libnm-glib/nm-device-802-11-wireless.c
- (state_changed_cb): update for new device states
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@3540 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-04-08 02:58:02 +00:00
|
|
|
break;
|
|
|
|
|
case NM_DEVICE_STATE_ACTIVATED:
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
if (sett_conn) {
|
2011-04-01 18:50:58 -05:00
|
|
|
/* Reset auto retries back to default since connection was successful */
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
nm_settings_connection_autoconnect_retries_reset (sett_conn);
|
2010-10-19 19:32:01 +02:00
|
|
|
|
2010-04-06 12:48:53 -07:00
|
|
|
/* And clear secrets so they will always be requested from the
|
|
|
|
|
* settings service when the next connection is made.
|
|
|
|
|
*/
|
2015-07-14 16:53:24 +02:00
|
|
|
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
/* FIXME(copy-on-write-connection): avoid modifying NMConnection instances and share them via copy-on-write. */
|
|
|
|
|
nm_connection_clear_secrets (nm_settings_connection_get_connection (sett_conn));
|
2010-04-06 12:48:53 -07:00
|
|
|
}
|
|
|
|
|
|
2012-05-29 22:17:26 -05:00
|
|
|
/* Add device's new IPv4 and IPv6 configs to DNS */
|
|
|
|
|
|
2013-08-22 10:10:17 -04:00
|
|
|
nm_dns_manager_begin_updates (priv->dns_manager, __func__);
|
2012-05-29 22:17:26 -05:00
|
|
|
|
|
|
|
|
ip4_config = nm_device_get_ip4_config (device);
|
2012-11-14 14:48:07 +01:00
|
|
|
if (ip4_config)
|
2019-02-09 11:46:47 +01:00
|
|
|
_dns_manager_set_ip_config (priv->dns_manager, NM_IP_CONFIG_CAST (ip4_config), NM_DNS_IP_CONFIG_TYPE_DEFAULT, device);
|
2012-05-29 22:17:26 -05:00
|
|
|
ip6_config = nm_device_get_ip6_config (device);
|
2012-11-14 14:48:07 +01:00
|
|
|
if (ip6_config)
|
2019-02-09 11:46:47 +01:00
|
|
|
_dns_manager_set_ip_config (priv->dns_manager, NM_IP_CONFIG_CAST (ip6_config), NM_DNS_IP_CONFIG_TYPE_DEFAULT, device);
|
2012-11-14 14:48:07 +01:00
|
|
|
|
2016-03-29 10:46:07 +02:00
|
|
|
update_routing_and_dns (self, FALSE);
|
2012-05-29 22:17:26 -05:00
|
|
|
|
2013-08-22 10:10:17 -04:00
|
|
|
nm_dns_manager_end_updates (priv->dns_manager, __func__);
|
2008-04-07 Dan Williams <dcbw@redhat.com>
* include/NetworkManager.h
- Remove the DOWN and CANCELLED device states
- Add UNMANAGED and UNAVAILABLE device states
- Document the device states
* introspection/nm-device.xml
src/nm-device-interface.c
src/nm-device-interface.h
- Add the 'managed' property
* test/nm-tool.c
- (detail_device): print out device state
* src/NetworkManagerSystem.h
src/backends/NetworkManagerArch.c
src/backends/NetworkManagerDebian.c
src/backends/NetworkManagerFrugalware.c
src/backends/NetworkManagerGentoo.c
src/backends/NetworkManagerMandriva.c
src/backends/NetworkManagerPaldo.c
src/backends/NetworkManagerRedHat.c
src/backends/NetworkManagerSlackware.c
src/backends/NetworkManagerSuSE.c
- (nm_system_device_get_system_config, nm_system_device_get_disabled
nm_system_device_free_system_config): remove; they were unused and
their functionality should be re-implemented in each distro's
system settings service plugin
* src/nm-gsm-device.c
src/nm-gsm-device.h
src/nm-cdma-device.c
src/nm-cdma-device.h
- (*_new): take the 'managed' argument
* src/nm-device.c
- (nm_device_set_address): remove, fold into nm_device_bring_up()
- (nm_device_init): start in unmanaged state, not disconnected
- (constructor): don't start device until the system settings service
has had a chance to figure out if the device is managed or not
- (nm_device_deactivate, nm_device_bring_up, nm_device_bring_down):
don't set device state here, let callers handle that as appropriate
- (nm_device_dispose): don't touch the device if it's not managed
- (set_property, get_property, nm_device_class_init): implement the
'managed' property
- (nm_device_state_changed): bring the device up if its now managed,
and deactivate it if it used to be active
- (nm_device_get_managed, nm_device_set_managed): do the right thing
with the managed state
* src/nm-hal-manager.c
- (wired_device_creator, wireless_device_creator, modem_device_creator):
take initial managed state and pass it along to device constructors
- (create_device_and_add_to_list): get managed state and pass to
type creators
* src/nm-device-802-11-wireless.c
- (real_can_activate): fold in most of
nm_device_802_11_wireless_can_activate()
- (can_scan): can't scan in UNAVAILABLE or UNMANAGED
- (link_timeout_cb): instead of deactivating, change device state and
let the device state handler to it
- (real_update_hw_address): clean up
- (state_changed_cb): when entering UNAVAILABLE state, schedule an idle
handler to transition to DISCONNECTED if the device isn't rfkilled
* src/nm-device-802-3-ethernet.c
- (set_carrier): move above callers and get rid of prototype
- (device_state_changed): when entering UNAVAILABLE state, schedule an
idle handler to transition to DISCONNECTED if the device has a
carrier
- (real_update_hw_address): clean up
- (link_timeout_cb, ppp_state_changed): change state instead of calling
deactivation directly as deactivation doesn't change state anymore
* src/NetworkManagerPolicy.c
- (schedule_activate_check): yay, remove wireless_enabled hack since
the NMManager and wireless devices work that out themselves now
- (device_state_changed): change to a switch and update for new device
states
- (device_carrier_changed): remove; device handles this now through
state changes
- (device_added): don't care about carrier any more; the initial
activation check will happen when the device transitions to
DISCONNECTED
* src/nm-manager.c
- (dispose): clear unmanaged devices
- (handle_unmanaged_devices): update unmanaged device list and toggle
the managed property on each device when needed
- (system_settings_properties_changed_cb): handle signals from the
system settings service
- (system_settings_get_unmanaged_devices_cb): handle callback from
getting the unmanaged device list method call
- (query_unmanaged_devices): ask the system settings service for its
list of unmanaged devices
- (nm_manager_name_owner_changed, initial_get_connections): get unmanaged
devices
- (manager_set_wireless_enabled): push rfkill state down to wireless
devices directly and let them handle the necessary state transitions
- (manager_device_state_changed): update for new device states
- (nm_manager_add_device): set initial rfkill state on wireless devices
- (nm_manager_remove_device): don't touch the device if it's unmanaged
- (nm_manager_activate_connection): return error if the device is
unmanaged
- (nm_manager_sleep): handle new device states correctly; don't change
the state of unavailable/unmanaged devices
* libnm-glib/nm-device-802-11-wireless.c
- (state_changed_cb): update for new device states
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@3540 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-04-08 02:58:02 +00:00
|
|
|
break;
|
2008-04-22 02:38:05 +00:00
|
|
|
case NM_DEVICE_STATE_UNMANAGED:
|
|
|
|
|
case NM_DEVICE_STATE_UNAVAILABLE:
|
2012-05-29 22:17:26 -05:00
|
|
|
if (old_state > NM_DEVICE_STATE_DISCONNECTED)
|
2016-03-29 10:46:07 +02:00
|
|
|
update_routing_and_dns (self, FALSE);
|
2011-04-19 00:31:53 -05:00
|
|
|
break;
|
2014-01-03 14:04:54 -05:00
|
|
|
case NM_DEVICE_STATE_DEACTIVATING:
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
if (sett_conn) {
|
2017-11-24 10:33:23 +01:00
|
|
|
NMSettingsAutoconnectBlockedReason blocked_reason = NM_SETTINGS_AUTO_CONNECT_BLOCKED_REASON_NONE;
|
2017-11-09 16:16:53 +01:00
|
|
|
|
2017-11-24 10:33:23 +01:00
|
|
|
switch (nm_device_state_reason_check (reason)) {
|
|
|
|
|
case NM_DEVICE_STATE_REASON_USER_REQUESTED:
|
|
|
|
|
blocked_reason = NM_SETTINGS_AUTO_CONNECT_BLOCKED_REASON_USER_REQUEST;
|
|
|
|
|
break;
|
|
|
|
|
case NM_DEVICE_STATE_REASON_DEPENDENCY_FAILED:
|
|
|
|
|
blocked_reason = NM_SETTINGS_AUTO_CONNECT_BLOCKED_REASON_FAILED;
|
|
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
if (blocked_reason != NM_SETTINGS_AUTO_CONNECT_BLOCKED_REASON_NONE) {
|
2017-11-09 16:16:53 +01:00
|
|
|
_LOGD (LOGD_DEVICE, "blocking autoconnect of connection '%s': %s",
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
nm_settings_connection_get_id (sett_conn),
|
2019-01-10 09:17:28 +01:00
|
|
|
NM_UTILS_LOOKUP_STR_A (nm_device_state_reason_to_str,
|
|
|
|
|
nm_device_state_reason_check (reason)));
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
nm_settings_connection_autoconnect_blocked_reason_set (sett_conn, blocked_reason, TRUE);
|
2013-12-17 10:09:47 -05:00
|
|
|
}
|
|
|
|
|
}
|
2016-10-31 23:31:14 +01:00
|
|
|
ip6_remove_device_prefix_delegations (self, device);
|
2014-01-03 14:04:54 -05:00
|
|
|
break;
|
2008-04-07 Dan Williams <dcbw@redhat.com>
* include/NetworkManager.h
- Remove the DOWN and CANCELLED device states
- Add UNMANAGED and UNAVAILABLE device states
- Document the device states
* introspection/nm-device.xml
src/nm-device-interface.c
src/nm-device-interface.h
- Add the 'managed' property
* test/nm-tool.c
- (detail_device): print out device state
* src/NetworkManagerSystem.h
src/backends/NetworkManagerArch.c
src/backends/NetworkManagerDebian.c
src/backends/NetworkManagerFrugalware.c
src/backends/NetworkManagerGentoo.c
src/backends/NetworkManagerMandriva.c
src/backends/NetworkManagerPaldo.c
src/backends/NetworkManagerRedHat.c
src/backends/NetworkManagerSlackware.c
src/backends/NetworkManagerSuSE.c
- (nm_system_device_get_system_config, nm_system_device_get_disabled
nm_system_device_free_system_config): remove; they were unused and
their functionality should be re-implemented in each distro's
system settings service plugin
* src/nm-gsm-device.c
src/nm-gsm-device.h
src/nm-cdma-device.c
src/nm-cdma-device.h
- (*_new): take the 'managed' argument
* src/nm-device.c
- (nm_device_set_address): remove, fold into nm_device_bring_up()
- (nm_device_init): start in unmanaged state, not disconnected
- (constructor): don't start device until the system settings service
has had a chance to figure out if the device is managed or not
- (nm_device_deactivate, nm_device_bring_up, nm_device_bring_down):
don't set device state here, let callers handle that as appropriate
- (nm_device_dispose): don't touch the device if it's not managed
- (set_property, get_property, nm_device_class_init): implement the
'managed' property
- (nm_device_state_changed): bring the device up if its now managed,
and deactivate it if it used to be active
- (nm_device_get_managed, nm_device_set_managed): do the right thing
with the managed state
* src/nm-hal-manager.c
- (wired_device_creator, wireless_device_creator, modem_device_creator):
take initial managed state and pass it along to device constructors
- (create_device_and_add_to_list): get managed state and pass to
type creators
* src/nm-device-802-11-wireless.c
- (real_can_activate): fold in most of
nm_device_802_11_wireless_can_activate()
- (can_scan): can't scan in UNAVAILABLE or UNMANAGED
- (link_timeout_cb): instead of deactivating, change device state and
let the device state handler to it
- (real_update_hw_address): clean up
- (state_changed_cb): when entering UNAVAILABLE state, schedule an idle
handler to transition to DISCONNECTED if the device isn't rfkilled
* src/nm-device-802-3-ethernet.c
- (set_carrier): move above callers and get rid of prototype
- (device_state_changed): when entering UNAVAILABLE state, schedule an
idle handler to transition to DISCONNECTED if the device has a
carrier
- (real_update_hw_address): clean up
- (link_timeout_cb, ppp_state_changed): change state instead of calling
deactivation directly as deactivation doesn't change state anymore
* src/NetworkManagerPolicy.c
- (schedule_activate_check): yay, remove wireless_enabled hack since
the NMManager and wireless devices work that out themselves now
- (device_state_changed): change to a switch and update for new device
states
- (device_carrier_changed): remove; device handles this now through
state changes
- (device_added): don't care about carrier any more; the initial
activation check will happen when the device transitions to
DISCONNECTED
* src/nm-manager.c
- (dispose): clear unmanaged devices
- (handle_unmanaged_devices): update unmanaged device list and toggle
the managed property on each device when needed
- (system_settings_properties_changed_cb): handle signals from the
system settings service
- (system_settings_get_unmanaged_devices_cb): handle callback from
getting the unmanaged device list method call
- (query_unmanaged_devices): ask the system settings service for its
list of unmanaged devices
- (nm_manager_name_owner_changed, initial_get_connections): get unmanaged
devices
- (manager_set_wireless_enabled): push rfkill state down to wireless
devices directly and let them handle the necessary state transitions
- (manager_device_state_changed): update for new device states
- (nm_manager_add_device): set initial rfkill state on wireless devices
- (nm_manager_remove_device): don't touch the device if it's unmanaged
- (nm_manager_activate_connection): return error if the device is
unmanaged
- (nm_manager_sleep): handle new device states correctly; don't change
the state of unavailable/unmanaged devices
* libnm-glib/nm-device-802-11-wireless.c
- (state_changed_cb): update for new device states
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@3540 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-04-08 02:58:02 +00:00
|
|
|
case NM_DEVICE_STATE_DISCONNECTED:
|
2013-12-05 04:57:01 -05:00
|
|
|
/* Reset retry counts for a device's connections when carrier on; if cable
|
|
|
|
|
* was unplugged and plugged in again, we should try to reconnect.
|
|
|
|
|
*/
|
2017-02-23 15:19:03 +01:00
|
|
|
if ( nm_device_state_reason_check (reason) == NM_DEVICE_STATE_REASON_CARRIER
|
|
|
|
|
&& old_state == NM_DEVICE_STATE_UNAVAILABLE)
|
2017-11-22 14:52:54 +01:00
|
|
|
reset_autoconnect_all (self, device, FALSE);
|
2011-05-30 12:07:41 +01:00
|
|
|
|
2017-12-16 15:11:16 +01:00
|
|
|
if (old_state > NM_DEVICE_STATE_DISCONNECTED)
|
2016-03-29 10:46:07 +02:00
|
|
|
update_routing_and_dns (self, FALSE);
|
2012-05-29 22:17:26 -05:00
|
|
|
|
2011-04-19 00:31:53 -05:00
|
|
|
/* Device is now available for auto-activation */
|
2016-03-29 10:46:07 +02:00
|
|
|
schedule_activate_check (self, device);
|
2008-04-07 Dan Williams <dcbw@redhat.com>
* include/NetworkManager.h
- Remove the DOWN and CANCELLED device states
- Add UNMANAGED and UNAVAILABLE device states
- Document the device states
* introspection/nm-device.xml
src/nm-device-interface.c
src/nm-device-interface.h
- Add the 'managed' property
* test/nm-tool.c
- (detail_device): print out device state
* src/NetworkManagerSystem.h
src/backends/NetworkManagerArch.c
src/backends/NetworkManagerDebian.c
src/backends/NetworkManagerFrugalware.c
src/backends/NetworkManagerGentoo.c
src/backends/NetworkManagerMandriva.c
src/backends/NetworkManagerPaldo.c
src/backends/NetworkManagerRedHat.c
src/backends/NetworkManagerSlackware.c
src/backends/NetworkManagerSuSE.c
- (nm_system_device_get_system_config, nm_system_device_get_disabled
nm_system_device_free_system_config): remove; they were unused and
their functionality should be re-implemented in each distro's
system settings service plugin
* src/nm-gsm-device.c
src/nm-gsm-device.h
src/nm-cdma-device.c
src/nm-cdma-device.h
- (*_new): take the 'managed' argument
* src/nm-device.c
- (nm_device_set_address): remove, fold into nm_device_bring_up()
- (nm_device_init): start in unmanaged state, not disconnected
- (constructor): don't start device until the system settings service
has had a chance to figure out if the device is managed or not
- (nm_device_deactivate, nm_device_bring_up, nm_device_bring_down):
don't set device state here, let callers handle that as appropriate
- (nm_device_dispose): don't touch the device if it's not managed
- (set_property, get_property, nm_device_class_init): implement the
'managed' property
- (nm_device_state_changed): bring the device up if its now managed,
and deactivate it if it used to be active
- (nm_device_get_managed, nm_device_set_managed): do the right thing
with the managed state
* src/nm-hal-manager.c
- (wired_device_creator, wireless_device_creator, modem_device_creator):
take initial managed state and pass it along to device constructors
- (create_device_and_add_to_list): get managed state and pass to
type creators
* src/nm-device-802-11-wireless.c
- (real_can_activate): fold in most of
nm_device_802_11_wireless_can_activate()
- (can_scan): can't scan in UNAVAILABLE or UNMANAGED
- (link_timeout_cb): instead of deactivating, change device state and
let the device state handler to it
- (real_update_hw_address): clean up
- (state_changed_cb): when entering UNAVAILABLE state, schedule an idle
handler to transition to DISCONNECTED if the device isn't rfkilled
* src/nm-device-802-3-ethernet.c
- (set_carrier): move above callers and get rid of prototype
- (device_state_changed): when entering UNAVAILABLE state, schedule an
idle handler to transition to DISCONNECTED if the device has a
carrier
- (real_update_hw_address): clean up
- (link_timeout_cb, ppp_state_changed): change state instead of calling
deactivation directly as deactivation doesn't change state anymore
* src/NetworkManagerPolicy.c
- (schedule_activate_check): yay, remove wireless_enabled hack since
the NMManager and wireless devices work that out themselves now
- (device_state_changed): change to a switch and update for new device
states
- (device_carrier_changed): remove; device handles this now through
state changes
- (device_added): don't care about carrier any more; the initial
activation check will happen when the device transitions to
DISCONNECTED
* src/nm-manager.c
- (dispose): clear unmanaged devices
- (handle_unmanaged_devices): update unmanaged device list and toggle
the managed property on each device when needed
- (system_settings_properties_changed_cb): handle signals from the
system settings service
- (system_settings_get_unmanaged_devices_cb): handle callback from
getting the unmanaged device list method call
- (query_unmanaged_devices): ask the system settings service for its
list of unmanaged devices
- (nm_manager_name_owner_changed, initial_get_connections): get unmanaged
devices
- (manager_set_wireless_enabled): push rfkill state down to wireless
devices directly and let them handle the necessary state transitions
- (manager_device_state_changed): update for new device states
- (nm_manager_add_device): set initial rfkill state on wireless devices
- (nm_manager_remove_device): don't touch the device if it's unmanaged
- (nm_manager_activate_connection): return error if the device is
unmanaged
- (nm_manager_sleep): handle new device states correctly; don't change
the state of unavailable/unmanaged devices
* libnm-glib/nm-device-802-11-wireless.c
- (state_changed_cb): update for new device states
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@3540 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-04-08 02:58:02 +00:00
|
|
|
break;
|
2011-12-07 13:15:24 +01:00
|
|
|
|
|
|
|
|
case NM_DEVICE_STATE_PREPARE:
|
|
|
|
|
/* Reset auto-connect retries of all slaves and schedule them for
|
|
|
|
|
* activation. */
|
2016-03-29 10:46:07 +02:00
|
|
|
activate_slave_connections (self, device);
|
2017-09-25 22:20:47 +02:00
|
|
|
|
|
|
|
|
/* Now that the device state is progressing, we don't care
|
|
|
|
|
* anymore for the AC state. */
|
|
|
|
|
ac = (NMActiveConnection *) nm_device_get_act_request (device);
|
|
|
|
|
if (ac && g_hash_table_remove (priv->pending_active_connections, ac)) {
|
|
|
|
|
g_signal_handlers_disconnect_by_func (ac, pending_ac_state_changed, self);
|
|
|
|
|
g_object_weak_unref (G_OBJECT (ac), pending_ac_gone, self);
|
|
|
|
|
g_object_unref (self);
|
|
|
|
|
}
|
2011-12-07 13:15:24 +01:00
|
|
|
break;
|
2014-01-23 10:30:50 -05:00
|
|
|
case NM_DEVICE_STATE_IP_CONFIG:
|
|
|
|
|
/* We must have secrets if we got here. */
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
if (sett_conn)
|
|
|
|
|
nm_settings_connection_autoconnect_blocked_reason_set (sett_conn, NM_SETTINGS_AUTO_CONNECT_BLOCKED_REASON_ALL, FALSE);
|
2014-01-23 10:30:50 -05:00
|
|
|
break;
|
2012-08-21 17:49:41 +02:00
|
|
|
case NM_DEVICE_STATE_SECONDARIES:
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
if (sett_conn)
|
|
|
|
|
s_con = nm_connection_get_setting_connection (nm_settings_connection_get_connection (sett_conn));
|
|
|
|
|
if ( s_con
|
|
|
|
|
&& nm_setting_connection_get_num_secondaries (s_con) > 0) {
|
2012-08-21 17:49:41 +02:00
|
|
|
/* Make routes and DNS up-to-date before activating dependent connections */
|
2016-03-29 10:46:07 +02:00
|
|
|
update_routing_and_dns (self, FALSE);
|
2012-08-21 17:49:41 +02:00
|
|
|
|
|
|
|
|
/* Activate secondary (VPN) connections */
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
if (!activate_secondary_connections (self,
|
|
|
|
|
nm_settings_connection_get_connection (sett_conn),
|
|
|
|
|
device)) {
|
|
|
|
|
nm_device_queue_state (device,
|
|
|
|
|
NM_DEVICE_STATE_FAILED,
|
2012-10-05 16:41:02 -04:00
|
|
|
NM_DEVICE_STATE_REASON_SECONDARY_CONNECTION_FAILED);
|
settings: use delegation instead of inheritance for NMSettingsConnection and NMConnection
NMConnection is an interface, which is implemented by the types
NMSimpleConnection (libnm-core), NMSettingsConnection (src) and
NMRemoteConnection (libnm).
NMSettingsConnection does a lot of things already:
1) it "is-a" NMDBusObject and exports the API of a connection profile
on D-Bus
2) it interacts with NMSettings and contains functionality
for tracking the profiles.
3) it is the base-class of types like NMSKeyfileConnection and
NMIfcfgConnection. These handle how the profile is persisted
on disk.
4) it implements NMConnection interface, to itself track the
settings of the profile.
3) and 4) would be better implemented via delegation than inheritance.
Address 4) and don't let NMSettingsConnection implemente the NMConnection
interface. Instead, a settings-connection references now a NMSimpleConnection
instance, to which it delegates for keeping the actual profiles.
Advantages:
- by delegating, there is a clearer separation of what
NMSettingsConnection does. For example, in C we often required
casts from NMSettingsConnection to NMConnection. NMConnection
is a very trivial object with very little logic. When we have
a NMConnection instance at hand, it's good to know that it is
*only* that simple instead of also being an entire
NMSettingsConnection instance.
The main purpose of this patch is to simplify the code by separating
the NMConnection from the NMSettingsConnection. We should generally
be aware whether we handle a NMSettingsConnection or a trivial
NMConnection instance. Now, because NMSettingsConnection no longer
"is-a" NMConnection, this distinction is apparent.
- NMConnection is implemented as an interface and we create
NMSimpleConnection instances whenever we need a real instance.
In GLib, interfaces have a performance overhead, that we needlessly
pay all the time. With this change, we no longer require
NMConnection to be an interface. Thus, in the future we could compile
a version of libnm-core for the daemon, where NMConnection is not an
interface but a GObject implementation akin to NMSimpleConnection.
- In the previous implementation, we cannot treat NMConnection immutable
and copy-on-write.
For example, when NMDevice needs a snapshot of the activated
profile as applied-connection, all it can do is clone the entire
NMSettingsConnection as a NMSimpleConnection.
Likewise, when we get a NMConnection instance and want to keep
a reference to it, we cannot do that, because we never know
who also references and modifies the instance.
By separating NMSettingsConnection we could in the future have
NMConnection immutable and copy-on-write, to avoid all unnecessary
clones.
2018-08-11 11:08:17 +02:00
|
|
|
}
|
2012-08-21 17:49:41 +02:00
|
|
|
} else
|
2012-10-05 16:41:02 -04:00
|
|
|
nm_device_queue_state (device, NM_DEVICE_STATE_ACTIVATED,
|
|
|
|
|
NM_DEVICE_STATE_REASON_NONE);
|
2012-08-21 17:49:41 +02:00
|
|
|
break;
|
2011-12-07 13:15:24 +01:00
|
|
|
|
2008-04-07 Dan Williams <dcbw@redhat.com>
* include/NetworkManager.h
- Remove the DOWN and CANCELLED device states
- Add UNMANAGED and UNAVAILABLE device states
- Document the device states
* introspection/nm-device.xml
src/nm-device-interface.c
src/nm-device-interface.h
- Add the 'managed' property
* test/nm-tool.c
- (detail_device): print out device state
* src/NetworkManagerSystem.h
src/backends/NetworkManagerArch.c
src/backends/NetworkManagerDebian.c
src/backends/NetworkManagerFrugalware.c
src/backends/NetworkManagerGentoo.c
src/backends/NetworkManagerMandriva.c
src/backends/NetworkManagerPaldo.c
src/backends/NetworkManagerRedHat.c
src/backends/NetworkManagerSlackware.c
src/backends/NetworkManagerSuSE.c
- (nm_system_device_get_system_config, nm_system_device_get_disabled
nm_system_device_free_system_config): remove; they were unused and
their functionality should be re-implemented in each distro's
system settings service plugin
* src/nm-gsm-device.c
src/nm-gsm-device.h
src/nm-cdma-device.c
src/nm-cdma-device.h
- (*_new): take the 'managed' argument
* src/nm-device.c
- (nm_device_set_address): remove, fold into nm_device_bring_up()
- (nm_device_init): start in unmanaged state, not disconnected
- (constructor): don't start device until the system settings service
has had a chance to figure out if the device is managed or not
- (nm_device_deactivate, nm_device_bring_up, nm_device_bring_down):
don't set device state here, let callers handle that as appropriate
- (nm_device_dispose): don't touch the device if it's not managed
- (set_property, get_property, nm_device_class_init): implement the
'managed' property
- (nm_device_state_changed): bring the device up if its now managed,
and deactivate it if it used to be active
- (nm_device_get_managed, nm_device_set_managed): do the right thing
with the managed state
* src/nm-hal-manager.c
- (wired_device_creator, wireless_device_creator, modem_device_creator):
take initial managed state and pass it along to device constructors
- (create_device_and_add_to_list): get managed state and pass to
type creators
* src/nm-device-802-11-wireless.c
- (real_can_activate): fold in most of
nm_device_802_11_wireless_can_activate()
- (can_scan): can't scan in UNAVAILABLE or UNMANAGED
- (link_timeout_cb): instead of deactivating, change device state and
let the device state handler to it
- (real_update_hw_address): clean up
- (state_changed_cb): when entering UNAVAILABLE state, schedule an idle
handler to transition to DISCONNECTED if the device isn't rfkilled
* src/nm-device-802-3-ethernet.c
- (set_carrier): move above callers and get rid of prototype
- (device_state_changed): when entering UNAVAILABLE state, schedule an
idle handler to transition to DISCONNECTED if the device has a
carrier
- (real_update_hw_address): clean up
- (link_timeout_cb, ppp_state_changed): change state instead of calling
deactivation directly as deactivation doesn't change state anymore
* src/NetworkManagerPolicy.c
- (schedule_activate_check): yay, remove wireless_enabled hack since
the NMManager and wireless devices work that out themselves now
- (device_state_changed): change to a switch and update for new device
states
- (device_carrier_changed): remove; device handles this now through
state changes
- (device_added): don't care about carrier any more; the initial
activation check will happen when the device transitions to
DISCONNECTED
* src/nm-manager.c
- (dispose): clear unmanaged devices
- (handle_unmanaged_devices): update unmanaged device list and toggle
the managed property on each device when needed
- (system_settings_properties_changed_cb): handle signals from the
system settings service
- (system_settings_get_unmanaged_devices_cb): handle callback from
getting the unmanaged device list method call
- (query_unmanaged_devices): ask the system settings service for its
list of unmanaged devices
- (nm_manager_name_owner_changed, initial_get_connections): get unmanaged
devices
- (manager_set_wireless_enabled): push rfkill state down to wireless
devices directly and let them handle the necessary state transitions
- (manager_device_state_changed): update for new device states
- (nm_manager_add_device): set initial rfkill state on wireless devices
- (nm_manager_remove_device): don't touch the device if it's unmanaged
- (nm_manager_activate_connection): return error if the device is
unmanaged
- (nm_manager_sleep): handle new device states correctly; don't change
the state of unavailable/unmanaged devices
* libnm-glib/nm-device-802-11-wireless.c
- (state_changed_cb): update for new device states
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@3540 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-04-08 02:58:02 +00:00
|
|
|
default:
|
|
|
|
|
break;
|
2007-10-08 18:07:36 +00:00
|
|
|
}
|
2013-08-22 13:06:51 -04:00
|
|
|
|
2018-06-28 18:05:05 +02:00
|
|
|
check_activating_active_connections (self);
|
2007-02-08 15:34:26 +00:00
|
|
|
}
|
|
|
|
|
|
2008-03-11 22:26:46 +00:00
|
|
|
static void
|
2018-01-08 15:44:05 +01:00
|
|
|
device_ip_config_changed (NMDevice *device,
|
|
|
|
|
NMIPConfig *new_config,
|
|
|
|
|
NMIPConfig *old_config,
|
|
|
|
|
gpointer user_data)
|
2008-03-11 22:26:46 +00:00
|
|
|
{
|
2016-04-04 14:36:34 +02:00
|
|
|
NMPolicyPrivate *priv = user_data;
|
2016-09-29 13:49:01 +02:00
|
|
|
NMPolicy *self = _PRIV_TO_SELF (priv);
|
2018-01-08 15:44:05 +01:00
|
|
|
int addr_family;
|
2012-05-29 22:17:26 -05:00
|
|
|
|
2018-01-08 15:44:05 +01:00
|
|
|
nm_assert (new_config || old_config);
|
2018-05-25 11:10:13 +02:00
|
|
|
nm_assert (!new_config || NM_IS_IP_CONFIG (new_config, AF_UNSPEC));
|
|
|
|
|
nm_assert (!old_config || NM_IS_IP_CONFIG (old_config, AF_UNSPEC));
|
2012-05-29 22:17:26 -05:00
|
|
|
|
2018-01-08 15:44:05 +01:00
|
|
|
if (new_config) {
|
|
|
|
|
addr_family = nm_ip_config_get_addr_family (new_config);
|
|
|
|
|
nm_assert (!old_config || addr_family == nm_ip_config_get_addr_family (old_config));
|
|
|
|
|
} else
|
|
|
|
|
addr_family = nm_ip_config_get_addr_family (old_config);
|
2012-05-29 22:17:26 -05:00
|
|
|
|
2013-08-22 10:10:17 -04:00
|
|
|
nm_dns_manager_begin_updates (priv->dns_manager, __func__);
|
2012-05-29 22:17:26 -05:00
|
|
|
|
2017-03-03 18:03:47 +01:00
|
|
|
/* We catch already all the IP events registering on the device state changes but
|
|
|
|
|
* the ones where the IP changes but the device state keep stable (i.e., activated):
|
|
|
|
|
* ignore IP config changes but when the device is in activated state.
|
2018-09-15 07:20:54 -04:00
|
|
|
* Prevents unnecessary changes to DNS information.
|
2012-05-29 22:17:26 -05:00
|
|
|
*/
|
2017-03-03 18:03:47 +01:00
|
|
|
if (nm_device_get_state (device) == NM_DEVICE_STATE_ACTIVATED) {
|
2013-09-06 11:56:41 +02:00
|
|
|
if (old_config != new_config) {
|
|
|
|
|
if (new_config)
|
2019-02-09 11:46:47 +01:00
|
|
|
_dns_manager_set_ip_config (priv->dns_manager, new_config, NM_DNS_IP_CONFIG_TYPE_DEFAULT, device);
|
2017-12-16 15:11:16 +01:00
|
|
|
if (old_config)
|
2018-01-08 15:44:05 +01:00
|
|
|
nm_dns_manager_set_ip_config (priv->dns_manager, old_config, NM_DNS_IP_CONFIG_TYPE_REMOVED);
|
2013-09-06 11:56:41 +02:00
|
|
|
}
|
2018-01-08 15:44:05 +01:00
|
|
|
update_ip_dns (self, addr_family);
|
|
|
|
|
if (addr_family == AF_INET)
|
|
|
|
|
update_ip4_routing (self, TRUE);
|
|
|
|
|
else
|
|
|
|
|
update_ip6_routing (self, TRUE);
|
|
|
|
|
update_system_hostname (self,
|
|
|
|
|
addr_family == AF_INET
|
|
|
|
|
? "ip4 conf"
|
|
|
|
|
: "ip6 conf");
|
2013-09-06 11:56:41 +02:00
|
|
|
} else {
|
|
|
|
|
/* Old configs get removed immediately */
|
|
|
|
|
if (old_config)
|
2018-01-08 15:44:05 +01:00
|
|
|
nm_dns_manager_set_ip_config (priv->dns_manager, old_config, NM_DNS_IP_CONFIG_TYPE_REMOVED);
|
2012-05-29 22:17:26 -05:00
|
|
|
}
|
|
|
|
|
|
2013-08-22 10:10:17 -04:00
|
|
|
nm_dns_manager_end_updates (priv->dns_manager, __func__);
|
2008-03-11 22:26:46 +00:00
|
|
|
}
|
|
|
|
|
|
2016-10-31 23:31:14 +01:00
|
|
|
/*****************************************************************************/
|
|
|
|
|
|
2012-05-14 15:32:54 +02:00
|
|
|
static void
|
|
|
|
|
device_autoconnect_changed (NMDevice *device,
|
|
|
|
|
GParamSpec *pspec,
|
|
|
|
|
gpointer user_data)
|
|
|
|
|
{
|
2016-04-04 14:36:34 +02:00
|
|
|
NMPolicyPrivate *priv = user_data;
|
2016-09-29 13:49:01 +02:00
|
|
|
NMPolicy *self = _PRIV_TO_SELF (priv);
|
2016-04-04 14:36:34 +02:00
|
|
|
|
2017-11-02 16:39:42 +01:00
|
|
|
schedule_activate_check (self, device);
|
2012-05-14 15:32:54 +02:00
|
|
|
}
|
|
|
|
|
|
2007-02-09 08:50:35 +00:00
|
|
|
static void
|
2014-02-24 18:10:18 -06:00
|
|
|
device_recheck_auto_activate (NMDevice *device, gpointer user_data)
|
2007-02-09 08:50:35 +00:00
|
|
|
{
|
2016-04-04 14:36:34 +02:00
|
|
|
NMPolicyPrivate *priv = user_data;
|
2016-09-29 13:49:01 +02:00
|
|
|
NMPolicy *self = _PRIV_TO_SELF (priv);
|
2016-04-04 14:36:34 +02:00
|
|
|
|
|
|
|
|
schedule_activate_check (self, device);
|
2011-08-02 15:48:24 -04:00
|
|
|
}
|
|
|
|
|
|
2010-10-27 16:57:31 -05:00
|
|
|
static void
|
2016-04-04 13:44:30 +02:00
|
|
|
devices_list_unregister (NMPolicy *self, NMDevice *device)
|
2008-03-07 Dan Williams <dcbw@redhat.com>
First pass of multiple active device support. Expect bugs.
* src/nm-ip4-config.c
src/nm-ip4-config.h
- (nm_ip4_config_get_secondary, nm_ip4_config_set_secondary): remove;
there are better ways to do this in the named manager
* src/nm-device.c
src/nm-device.h
- (nm_device_can_activate): return whether the device can activate a
connection right now; taking into account things like carrier state
and rfkill state
- (nm_device_get_best_auto_connection): renamed from
nm_device_get_best_connection
- (real_act_stage4_get_ip4_config): MTU stuff is now handled in the
device subclasses themselves, so that each device can override the
MTU from it's NMSetting subclass if needed
- (nm_device_set_ip4_config): set MTU when setting up routes and stuff
in NetworkManagerSystem.c, not here
* src/named-manager/nm-named-manager.c
src/named-manager/nm-named-manager.h
- (nm_named_manager_name_owner_changed,
nm_named_manager_dbus_connection_changed): fix for changes to
rewrite_resolv_conf()
- (compute_nameservers): don't need the NMNamedManager at all, remove
from parameter list
- (merge_one_ip4_config): new function; merge ip4 configs together
- (rewrite_resolv_conf): write out resolv.conf from all the stored
ip4 configs; the VPN config takes precedence, then the best
device config, then the rest of the configs
- (get_domain_for_config): take the NMNamedManager as an argument
to check whether the config is the VPN config
- (add_ip4_config_to_named): fixups for removal of the 'secondary'
attribute from ip4 configs
- (add_all_ip4_configs_to_named): add all the configs in priority order
- (remove_ip4_config_from_named): fix for changes to
get_domain_for_config()
- (nm_named_manager_add_ip4_config): assign the config to the right slot
based on its type; callers must pass in the type now
- (get_last_default_domain): remove, unused
- (nm_named_manager_remove_ip4_config): handle config slots correctly
* src/nm-device-802-11-wireless.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): handle MTU override
* src/nm-device-802-3-ethernet.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): new function; handle MTU override
* src/vpn-manager/nm-vpn-connection.c
- (nm_vpn_connection_ip4_config_get): don't need to set the 'secondary'
attribute on the ip4 config
* src/NetworkManagerPolicy.c
- (nm_policy_auto_get_best_device): remove
- (nm_policy_device_change_check): remove
- (update_default_route): new function; set the default route via
the specified device
- (get_device_priority): new function; return the priority number of
a device type WRT which one should have the default route. Order is
(highest to lowest) wired, wireless, GSM, CDMA.
- (update_routing_and_dns): new function; determine which device should
have the default route, then update the routing table and DNS
- (maybe_auto_activate_device): new function; if a device is now
available for activation, find out what connection it would like to
activate and do it
- (schedule_activate_check): new function; if a device can be activated
now, schedule the activation. Each device may have only one
pending activation at a given time.
- (device_state_changed): if activation was canceled, try again,
possibly with another connection; if the device was activated,
update routing and DNS; if the device was deactivated, try again
with another connection
- (device_carrier_changed): if there is no carrier, deactivate the
device; otherwise schedule an activation check for the device
- (wireless_networks_changed): schedule an activation check for the
device
- (device_added): keep track of the signal handler IDs so they can
be removed when the device goes away
- (device_removed): remove any signal handlers that might be attached
to the device; update routing and DNS
- (schedule_activate_all): new function
- (connections_added, connection_added, connection_updated): when
connections change, schedule all devices for an activation check
- (connection_removed): when a device is deactivated because its
connection was removed, schedule another activation check for it
- (nm_policy_destroy): destroy pending activations and disconnect
all device signal handlers
* src/nm-manager.c
- (nm_manager_activate_device): if the device was already actived,
deactivate it
- (deactivate_old_device): remove
- (connection_added_default_handler, impl_manager_activate_device):
don't deactivate other devices when activating this one
* src/backends/NetworkManagerGentoo.c
src/backends/NetworkManagerFrugalware.c
src/backends/NetworkManagerPaldo.c
src/backends/NetworkManagerRedHat.c
src/backends/NetworkManagerSlackware.c
src/backends/NetworkManagerArch.c
src/backends/NetworkManagerSuSE.c
src/backends/NetworkManagerDebian.c
- (nm_system_get_mtu): remove; MTU should be provided through the
distro's system settings service plugin instead
- (nm_system_device_add_default_route_via_device): remove
- (nm_system_device_add_default_route_via_device_with_iface): remove
- (nm_system_device_replace_default_route): new function; call
generic implementation
* src/backends/NetworkManagerGeneric.c
src/backends/NetworkManagerGeneric.h
- (nm_generic_device_add_default_route_via_device,
nm_generic_device_add_default_route_via_device_with_iface): remove
- (nm_generic_device_replace_default_route): replace the default route
with the given route via some gateway
* src/NetworkManagerSystem.c
src/NetworkManagerSystem.h
- (nm_system_device_set_from_ip4_config): let the policy handle updates
to routing and DNS; but set the MTU here
- (nm_system_vpn_device_set_from_ip4_config): set the route with the
ip_iface of the active device; use the standard MTU setting function
- (nm_system_set_mtu): remove
- (nm_system_device_set_mtu): consolidate MTU setting code in one place
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@3391 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-03-07 19:41:32 +00:00
|
|
|
{
|
2016-04-04 14:36:34 +02:00
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
|
|
|
|
|
|
|
|
|
g_signal_handlers_disconnect_by_data ((GObject *) device, priv);
|
2016-04-04 13:44:30 +02:00
|
|
|
}
|
2008-03-07 Dan Williams <dcbw@redhat.com>
First pass of multiple active device support. Expect bugs.
* src/nm-ip4-config.c
src/nm-ip4-config.h
- (nm_ip4_config_get_secondary, nm_ip4_config_set_secondary): remove;
there are better ways to do this in the named manager
* src/nm-device.c
src/nm-device.h
- (nm_device_can_activate): return whether the device can activate a
connection right now; taking into account things like carrier state
and rfkill state
- (nm_device_get_best_auto_connection): renamed from
nm_device_get_best_connection
- (real_act_stage4_get_ip4_config): MTU stuff is now handled in the
device subclasses themselves, so that each device can override the
MTU from it's NMSetting subclass if needed
- (nm_device_set_ip4_config): set MTU when setting up routes and stuff
in NetworkManagerSystem.c, not here
* src/named-manager/nm-named-manager.c
src/named-manager/nm-named-manager.h
- (nm_named_manager_name_owner_changed,
nm_named_manager_dbus_connection_changed): fix for changes to
rewrite_resolv_conf()
- (compute_nameservers): don't need the NMNamedManager at all, remove
from parameter list
- (merge_one_ip4_config): new function; merge ip4 configs together
- (rewrite_resolv_conf): write out resolv.conf from all the stored
ip4 configs; the VPN config takes precedence, then the best
device config, then the rest of the configs
- (get_domain_for_config): take the NMNamedManager as an argument
to check whether the config is the VPN config
- (add_ip4_config_to_named): fixups for removal of the 'secondary'
attribute from ip4 configs
- (add_all_ip4_configs_to_named): add all the configs in priority order
- (remove_ip4_config_from_named): fix for changes to
get_domain_for_config()
- (nm_named_manager_add_ip4_config): assign the config to the right slot
based on its type; callers must pass in the type now
- (get_last_default_domain): remove, unused
- (nm_named_manager_remove_ip4_config): handle config slots correctly
* src/nm-device-802-11-wireless.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): handle MTU override
* src/nm-device-802-3-ethernet.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): new function; handle MTU override
* src/vpn-manager/nm-vpn-connection.c
- (nm_vpn_connection_ip4_config_get): don't need to set the 'secondary'
attribute on the ip4 config
* src/NetworkManagerPolicy.c
- (nm_policy_auto_get_best_device): remove
- (nm_policy_device_change_check): remove
- (update_default_route): new function; set the default route via
the specified device
- (get_device_priority): new function; return the priority number of
a device type WRT which one should have the default route. Order is
(highest to lowest) wired, wireless, GSM, CDMA.
- (update_routing_and_dns): new function; determine which device should
have the default route, then update the routing table and DNS
- (maybe_auto_activate_device): new function; if a device is now
available for activation, find out what connection it would like to
activate and do it
- (schedule_activate_check): new function; if a device can be activated
now, schedule the activation. Each device may have only one
pending activation at a given time.
- (device_state_changed): if activation was canceled, try again,
possibly with another connection; if the device was activated,
update routing and DNS; if the device was deactivated, try again
with another connection
- (device_carrier_changed): if there is no carrier, deactivate the
device; otherwise schedule an activation check for the device
- (wireless_networks_changed): schedule an activation check for the
device
- (device_added): keep track of the signal handler IDs so they can
be removed when the device goes away
- (device_removed): remove any signal handlers that might be attached
to the device; update routing and DNS
- (schedule_activate_all): new function
- (connections_added, connection_added, connection_updated): when
connections change, schedule all devices for an activation check
- (connection_removed): when a device is deactivated because its
connection was removed, schedule another activation check for it
- (nm_policy_destroy): destroy pending activations and disconnect
all device signal handlers
* src/nm-manager.c
- (nm_manager_activate_device): if the device was already actived,
deactivate it
- (deactivate_old_device): remove
- (connection_added_default_handler, impl_manager_activate_device):
don't deactivate other devices when activating this one
* src/backends/NetworkManagerGentoo.c
src/backends/NetworkManagerFrugalware.c
src/backends/NetworkManagerPaldo.c
src/backends/NetworkManagerRedHat.c
src/backends/NetworkManagerSlackware.c
src/backends/NetworkManagerArch.c
src/backends/NetworkManagerSuSE.c
src/backends/NetworkManagerDebian.c
- (nm_system_get_mtu): remove; MTU should be provided through the
distro's system settings service plugin instead
- (nm_system_device_add_default_route_via_device): remove
- (nm_system_device_add_default_route_via_device_with_iface): remove
- (nm_system_device_replace_default_route): new function; call
generic implementation
* src/backends/NetworkManagerGeneric.c
src/backends/NetworkManagerGeneric.h
- (nm_generic_device_add_default_route_via_device,
nm_generic_device_add_default_route_via_device_with_iface): remove
- (nm_generic_device_replace_default_route): replace the default route
with the given route via some gateway
* src/NetworkManagerSystem.c
src/NetworkManagerSystem.h
- (nm_system_device_set_from_ip4_config): let the policy handle updates
to routing and DNS; but set the MTU here
- (nm_system_vpn_device_set_from_ip4_config): set the route with the
ip_iface of the active device; use the standard MTU setting function
- (nm_system_set_mtu): remove
- (nm_system_device_set_mtu): consolidate MTU setting code in one place
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@3391 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-03-07 19:41:32 +00:00
|
|
|
|
2016-04-04 13:44:30 +02:00
|
|
|
static void
|
|
|
|
|
devices_list_register (NMPolicy *self, NMDevice *device)
|
|
|
|
|
{
|
2016-04-04 14:36:34 +02:00
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
|
|
|
|
|
2016-04-04 13:44:30 +02:00
|
|
|
/* Connect state-changed with _after, so that the handler is invoked after other handlers. */
|
2016-04-04 14:36:34 +02:00
|
|
|
g_signal_connect_after (device, NM_DEVICE_STATE_CHANGED, (GCallback) device_state_changed, priv);
|
2018-01-08 15:44:05 +01:00
|
|
|
g_signal_connect (device, NM_DEVICE_IP4_CONFIG_CHANGED, (GCallback) device_ip_config_changed, priv);
|
|
|
|
|
g_signal_connect (device, NM_DEVICE_IP6_CONFIG_CHANGED, (GCallback) device_ip_config_changed, priv);
|
2016-10-31 23:31:14 +01:00
|
|
|
g_signal_connect (device, NM_DEVICE_IP6_PREFIX_DELEGATED, (GCallback) device_ip6_prefix_delegated, priv);
|
|
|
|
|
g_signal_connect (device, NM_DEVICE_IP6_SUBNET_NEEDED, (GCallback) device_ip6_subnet_needed, priv);
|
2016-04-04 14:36:34 +02:00
|
|
|
g_signal_connect (device, "notify::" NM_DEVICE_AUTOCONNECT, (GCallback) device_autoconnect_changed, priv);
|
|
|
|
|
g_signal_connect (device, NM_DEVICE_RECHECK_AUTO_ACTIVATE, (GCallback) device_recheck_auto_activate, priv);
|
2007-02-09 08:50:35 +00:00
|
|
|
}
|
|
|
|
|
|
2007-02-08 15:34:26 +00:00
|
|
|
static void
|
|
|
|
|
device_added (NMManager *manager, NMDevice *device, gpointer user_data)
|
|
|
|
|
{
|
2016-04-04 14:36:34 +02:00
|
|
|
NMPolicyPrivate *priv = user_data;
|
2016-09-29 13:49:01 +02:00
|
|
|
NMPolicy *self = _PRIV_TO_SELF (priv);
|
2007-02-08 15:34:26 +00:00
|
|
|
|
2016-04-04 13:44:30 +02:00
|
|
|
g_return_if_fail (NM_IS_POLICY (self));
|
|
|
|
|
|
|
|
|
|
priv = NM_POLICY_GET_PRIVATE (self);
|
|
|
|
|
|
2018-01-02 15:47:37 +01:00
|
|
|
if (!g_hash_table_add (priv->devices, device))
|
2016-04-04 13:44:30 +02:00
|
|
|
g_return_if_reached ();
|
|
|
|
|
|
|
|
|
|
devices_list_register (self, device);
|
2007-02-08 15:34:26 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void
|
|
|
|
|
device_removed (NMManager *manager, NMDevice *device, gpointer user_data)
|
|
|
|
|
{
|
2016-04-04 14:36:34 +02:00
|
|
|
NMPolicyPrivate *priv = user_data;
|
2016-09-29 13:49:01 +02:00
|
|
|
NMPolicy *self = _PRIV_TO_SELF (priv);
|
2017-11-22 12:15:24 +01:00
|
|
|
ActivateData *data;
|
2009-02-06 15:38:09 -05:00
|
|
|
|
2016-10-31 23:31:14 +01:00
|
|
|
/* XXX is this needed? The delegations are cleaned up
|
|
|
|
|
* on transition to deactivated too. */
|
|
|
|
|
ip6_remove_device_prefix_delegations (self, device);
|
|
|
|
|
|
2009-02-06 15:38:09 -05:00
|
|
|
/* Clear any idle callbacks for this device */
|
2017-11-22 12:15:24 +01:00
|
|
|
data = find_pending_activation (self, device);
|
|
|
|
|
if (data && data->autoactivate_id)
|
|
|
|
|
activate_data_free (data);
|
2008-03-07 Dan Williams <dcbw@redhat.com>
First pass of multiple active device support. Expect bugs.
* src/nm-ip4-config.c
src/nm-ip4-config.h
- (nm_ip4_config_get_secondary, nm_ip4_config_set_secondary): remove;
there are better ways to do this in the named manager
* src/nm-device.c
src/nm-device.h
- (nm_device_can_activate): return whether the device can activate a
connection right now; taking into account things like carrier state
and rfkill state
- (nm_device_get_best_auto_connection): renamed from
nm_device_get_best_connection
- (real_act_stage4_get_ip4_config): MTU stuff is now handled in the
device subclasses themselves, so that each device can override the
MTU from it's NMSetting subclass if needed
- (nm_device_set_ip4_config): set MTU when setting up routes and stuff
in NetworkManagerSystem.c, not here
* src/named-manager/nm-named-manager.c
src/named-manager/nm-named-manager.h
- (nm_named_manager_name_owner_changed,
nm_named_manager_dbus_connection_changed): fix for changes to
rewrite_resolv_conf()
- (compute_nameservers): don't need the NMNamedManager at all, remove
from parameter list
- (merge_one_ip4_config): new function; merge ip4 configs together
- (rewrite_resolv_conf): write out resolv.conf from all the stored
ip4 configs; the VPN config takes precedence, then the best
device config, then the rest of the configs
- (get_domain_for_config): take the NMNamedManager as an argument
to check whether the config is the VPN config
- (add_ip4_config_to_named): fixups for removal of the 'secondary'
attribute from ip4 configs
- (add_all_ip4_configs_to_named): add all the configs in priority order
- (remove_ip4_config_from_named): fix for changes to
get_domain_for_config()
- (nm_named_manager_add_ip4_config): assign the config to the right slot
based on its type; callers must pass in the type now
- (get_last_default_domain): remove, unused
- (nm_named_manager_remove_ip4_config): handle config slots correctly
* src/nm-device-802-11-wireless.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): handle MTU override
* src/nm-device-802-3-ethernet.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): new function; handle MTU override
* src/vpn-manager/nm-vpn-connection.c
- (nm_vpn_connection_ip4_config_get): don't need to set the 'secondary'
attribute on the ip4 config
* src/NetworkManagerPolicy.c
- (nm_policy_auto_get_best_device): remove
- (nm_policy_device_change_check): remove
- (update_default_route): new function; set the default route via
the specified device
- (get_device_priority): new function; return the priority number of
a device type WRT which one should have the default route. Order is
(highest to lowest) wired, wireless, GSM, CDMA.
- (update_routing_and_dns): new function; determine which device should
have the default route, then update the routing table and DNS
- (maybe_auto_activate_device): new function; if a device is now
available for activation, find out what connection it would like to
activate and do it
- (schedule_activate_check): new function; if a device can be activated
now, schedule the activation. Each device may have only one
pending activation at a given time.
- (device_state_changed): if activation was canceled, try again,
possibly with another connection; if the device was activated,
update routing and DNS; if the device was deactivated, try again
with another connection
- (device_carrier_changed): if there is no carrier, deactivate the
device; otherwise schedule an activation check for the device
- (wireless_networks_changed): schedule an activation check for the
device
- (device_added): keep track of the signal handler IDs so they can
be removed when the device goes away
- (device_removed): remove any signal handlers that might be attached
to the device; update routing and DNS
- (schedule_activate_all): new function
- (connections_added, connection_added, connection_updated): when
connections change, schedule all devices for an activation check
- (connection_removed): when a device is deactivated because its
connection was removed, schedule another activation check for it
- (nm_policy_destroy): destroy pending activations and disconnect
all device signal handlers
* src/nm-manager.c
- (nm_manager_activate_device): if the device was already actived,
deactivate it
- (deactivate_old_device): remove
- (connection_added_default_handler, impl_manager_activate_device):
don't deactivate other devices when activating this one
* src/backends/NetworkManagerGentoo.c
src/backends/NetworkManagerFrugalware.c
src/backends/NetworkManagerPaldo.c
src/backends/NetworkManagerRedHat.c
src/backends/NetworkManagerSlackware.c
src/backends/NetworkManagerArch.c
src/backends/NetworkManagerSuSE.c
src/backends/NetworkManagerDebian.c
- (nm_system_get_mtu): remove; MTU should be provided through the
distro's system settings service plugin instead
- (nm_system_device_add_default_route_via_device): remove
- (nm_system_device_add_default_route_via_device_with_iface): remove
- (nm_system_device_replace_default_route): new function; call
generic implementation
* src/backends/NetworkManagerGeneric.c
src/backends/NetworkManagerGeneric.h
- (nm_generic_device_add_default_route_via_device,
nm_generic_device_add_default_route_via_device_with_iface): remove
- (nm_generic_device_replace_default_route): replace the default route
with the given route via some gateway
* src/NetworkManagerSystem.c
src/NetworkManagerSystem.h
- (nm_system_device_set_from_ip4_config): let the policy handle updates
to routing and DNS; but set the MTU here
- (nm_system_vpn_device_set_from_ip4_config): set the route with the
ip_iface of the active device; use the standard MTU setting function
- (nm_system_set_mtu): remove
- (nm_system_device_set_mtu): consolidate MTU setting code in one place
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@3391 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-03-07 19:41:32 +00:00
|
|
|
|
2016-04-04 13:44:30 +02:00
|
|
|
if (g_hash_table_remove (priv->devices, device))
|
|
|
|
|
devices_list_unregister (self, device);
|
2008-03-07 Dan Williams <dcbw@redhat.com>
First pass of multiple active device support. Expect bugs.
* src/nm-ip4-config.c
src/nm-ip4-config.h
- (nm_ip4_config_get_secondary, nm_ip4_config_set_secondary): remove;
there are better ways to do this in the named manager
* src/nm-device.c
src/nm-device.h
- (nm_device_can_activate): return whether the device can activate a
connection right now; taking into account things like carrier state
and rfkill state
- (nm_device_get_best_auto_connection): renamed from
nm_device_get_best_connection
- (real_act_stage4_get_ip4_config): MTU stuff is now handled in the
device subclasses themselves, so that each device can override the
MTU from it's NMSetting subclass if needed
- (nm_device_set_ip4_config): set MTU when setting up routes and stuff
in NetworkManagerSystem.c, not here
* src/named-manager/nm-named-manager.c
src/named-manager/nm-named-manager.h
- (nm_named_manager_name_owner_changed,
nm_named_manager_dbus_connection_changed): fix for changes to
rewrite_resolv_conf()
- (compute_nameservers): don't need the NMNamedManager at all, remove
from parameter list
- (merge_one_ip4_config): new function; merge ip4 configs together
- (rewrite_resolv_conf): write out resolv.conf from all the stored
ip4 configs; the VPN config takes precedence, then the best
device config, then the rest of the configs
- (get_domain_for_config): take the NMNamedManager as an argument
to check whether the config is the VPN config
- (add_ip4_config_to_named): fixups for removal of the 'secondary'
attribute from ip4 configs
- (add_all_ip4_configs_to_named): add all the configs in priority order
- (remove_ip4_config_from_named): fix for changes to
get_domain_for_config()
- (nm_named_manager_add_ip4_config): assign the config to the right slot
based on its type; callers must pass in the type now
- (get_last_default_domain): remove, unused
- (nm_named_manager_remove_ip4_config): handle config slots correctly
* src/nm-device-802-11-wireless.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): handle MTU override
* src/nm-device-802-3-ethernet.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): new function; handle MTU override
* src/vpn-manager/nm-vpn-connection.c
- (nm_vpn_connection_ip4_config_get): don't need to set the 'secondary'
attribute on the ip4 config
* src/NetworkManagerPolicy.c
- (nm_policy_auto_get_best_device): remove
- (nm_policy_device_change_check): remove
- (update_default_route): new function; set the default route via
the specified device
- (get_device_priority): new function; return the priority number of
a device type WRT which one should have the default route. Order is
(highest to lowest) wired, wireless, GSM, CDMA.
- (update_routing_and_dns): new function; determine which device should
have the default route, then update the routing table and DNS
- (maybe_auto_activate_device): new function; if a device is now
available for activation, find out what connection it would like to
activate and do it
- (schedule_activate_check): new function; if a device can be activated
now, schedule the activation. Each device may have only one
pending activation at a given time.
- (device_state_changed): if activation was canceled, try again,
possibly with another connection; if the device was activated,
update routing and DNS; if the device was deactivated, try again
with another connection
- (device_carrier_changed): if there is no carrier, deactivate the
device; otherwise schedule an activation check for the device
- (wireless_networks_changed): schedule an activation check for the
device
- (device_added): keep track of the signal handler IDs so they can
be removed when the device goes away
- (device_removed): remove any signal handlers that might be attached
to the device; update routing and DNS
- (schedule_activate_all): new function
- (connections_added, connection_added, connection_updated): when
connections change, schedule all devices for an activation check
- (connection_removed): when a device is deactivated because its
connection was removed, schedule another activation check for it
- (nm_policy_destroy): destroy pending activations and disconnect
all device signal handlers
* src/nm-manager.c
- (nm_manager_activate_device): if the device was already actived,
deactivate it
- (deactivate_old_device): remove
- (connection_added_default_handler, impl_manager_activate_device):
don't deactivate other devices when activating this one
* src/backends/NetworkManagerGentoo.c
src/backends/NetworkManagerFrugalware.c
src/backends/NetworkManagerPaldo.c
src/backends/NetworkManagerRedHat.c
src/backends/NetworkManagerSlackware.c
src/backends/NetworkManagerArch.c
src/backends/NetworkManagerSuSE.c
src/backends/NetworkManagerDebian.c
- (nm_system_get_mtu): remove; MTU should be provided through the
distro's system settings service plugin instead
- (nm_system_device_add_default_route_via_device): remove
- (nm_system_device_add_default_route_via_device_with_iface): remove
- (nm_system_device_replace_default_route): new function; call
generic implementation
* src/backends/NetworkManagerGeneric.c
src/backends/NetworkManagerGeneric.h
- (nm_generic_device_add_default_route_via_device,
nm_generic_device_add_default_route_via_device_with_iface): remove
- (nm_generic_device_replace_default_route): replace the default route
with the given route via some gateway
* src/NetworkManagerSystem.c
src/NetworkManagerSystem.h
- (nm_system_device_set_from_ip4_config): let the policy handle updates
to routing and DNS; but set the MTU here
- (nm_system_vpn_device_set_from_ip4_config): set the route with the
ip_iface of the active device; use the standard MTU setting function
- (nm_system_set_mtu): remove
- (nm_system_device_set_mtu): consolidate MTU setting code in one place
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@3391 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-03-07 19:41:32 +00:00
|
|
|
|
2012-05-29 22:17:26 -05:00
|
|
|
/* Don't update routing and DNS here as we've already handled that
|
|
|
|
|
* for devices that need it when the device's state changed to UNMANAGED.
|
|
|
|
|
*/
|
2008-03-07 Dan Williams <dcbw@redhat.com>
First pass of multiple active device support. Expect bugs.
* src/nm-ip4-config.c
src/nm-ip4-config.h
- (nm_ip4_config_get_secondary, nm_ip4_config_set_secondary): remove;
there are better ways to do this in the named manager
* src/nm-device.c
src/nm-device.h
- (nm_device_can_activate): return whether the device can activate a
connection right now; taking into account things like carrier state
and rfkill state
- (nm_device_get_best_auto_connection): renamed from
nm_device_get_best_connection
- (real_act_stage4_get_ip4_config): MTU stuff is now handled in the
device subclasses themselves, so that each device can override the
MTU from it's NMSetting subclass if needed
- (nm_device_set_ip4_config): set MTU when setting up routes and stuff
in NetworkManagerSystem.c, not here
* src/named-manager/nm-named-manager.c
src/named-manager/nm-named-manager.h
- (nm_named_manager_name_owner_changed,
nm_named_manager_dbus_connection_changed): fix for changes to
rewrite_resolv_conf()
- (compute_nameservers): don't need the NMNamedManager at all, remove
from parameter list
- (merge_one_ip4_config): new function; merge ip4 configs together
- (rewrite_resolv_conf): write out resolv.conf from all the stored
ip4 configs; the VPN config takes precedence, then the best
device config, then the rest of the configs
- (get_domain_for_config): take the NMNamedManager as an argument
to check whether the config is the VPN config
- (add_ip4_config_to_named): fixups for removal of the 'secondary'
attribute from ip4 configs
- (add_all_ip4_configs_to_named): add all the configs in priority order
- (remove_ip4_config_from_named): fix for changes to
get_domain_for_config()
- (nm_named_manager_add_ip4_config): assign the config to the right slot
based on its type; callers must pass in the type now
- (get_last_default_domain): remove, unused
- (nm_named_manager_remove_ip4_config): handle config slots correctly
* src/nm-device-802-11-wireless.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): handle MTU override
* src/nm-device-802-3-ethernet.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): new function; handle MTU override
* src/vpn-manager/nm-vpn-connection.c
- (nm_vpn_connection_ip4_config_get): don't need to set the 'secondary'
attribute on the ip4 config
* src/NetworkManagerPolicy.c
- (nm_policy_auto_get_best_device): remove
- (nm_policy_device_change_check): remove
- (update_default_route): new function; set the default route via
the specified device
- (get_device_priority): new function; return the priority number of
a device type WRT which one should have the default route. Order is
(highest to lowest) wired, wireless, GSM, CDMA.
- (update_routing_and_dns): new function; determine which device should
have the default route, then update the routing table and DNS
- (maybe_auto_activate_device): new function; if a device is now
available for activation, find out what connection it would like to
activate and do it
- (schedule_activate_check): new function; if a device can be activated
now, schedule the activation. Each device may have only one
pending activation at a given time.
- (device_state_changed): if activation was canceled, try again,
possibly with another connection; if the device was activated,
update routing and DNS; if the device was deactivated, try again
with another connection
- (device_carrier_changed): if there is no carrier, deactivate the
device; otherwise schedule an activation check for the device
- (wireless_networks_changed): schedule an activation check for the
device
- (device_added): keep track of the signal handler IDs so they can
be removed when the device goes away
- (device_removed): remove any signal handlers that might be attached
to the device; update routing and DNS
- (schedule_activate_all): new function
- (connections_added, connection_added, connection_updated): when
connections change, schedule all devices for an activation check
- (connection_removed): when a device is deactivated because its
connection was removed, schedule another activation check for it
- (nm_policy_destroy): destroy pending activations and disconnect
all device signal handlers
* src/nm-manager.c
- (nm_manager_activate_device): if the device was already actived,
deactivate it
- (deactivate_old_device): remove
- (connection_added_default_handler, impl_manager_activate_device):
don't deactivate other devices when activating this one
* src/backends/NetworkManagerGentoo.c
src/backends/NetworkManagerFrugalware.c
src/backends/NetworkManagerPaldo.c
src/backends/NetworkManagerRedHat.c
src/backends/NetworkManagerSlackware.c
src/backends/NetworkManagerArch.c
src/backends/NetworkManagerSuSE.c
src/backends/NetworkManagerDebian.c
- (nm_system_get_mtu): remove; MTU should be provided through the
distro's system settings service plugin instead
- (nm_system_device_add_default_route_via_device): remove
- (nm_system_device_add_default_route_via_device_with_iface): remove
- (nm_system_device_replace_default_route): new function; call
generic implementation
* src/backends/NetworkManagerGeneric.c
src/backends/NetworkManagerGeneric.h
- (nm_generic_device_add_default_route_via_device,
nm_generic_device_add_default_route_via_device_with_iface): remove
- (nm_generic_device_replace_default_route): replace the default route
with the given route via some gateway
* src/NetworkManagerSystem.c
src/NetworkManagerSystem.h
- (nm_system_device_set_from_ip4_config): let the policy handle updates
to routing and DNS; but set the MTU here
- (nm_system_vpn_device_set_from_ip4_config): set the route with the
ip_iface of the active device; use the standard MTU setting function
- (nm_system_set_mtu): remove
- (nm_system_device_set_mtu): consolidate MTU setting code in one place
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@3391 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-03-07 19:41:32 +00:00
|
|
|
}
|
|
|
|
|
|
2016-10-02 18:22:50 +02:00
|
|
|
/*****************************************************************************/
|
2012-08-22 18:34:45 -05:00
|
|
|
|
|
|
|
|
static void
|
2016-03-29 10:46:07 +02:00
|
|
|
vpn_connection_activated (NMPolicy *self, NMVpnConnection *vpn)
|
2012-08-22 18:34:45 -05:00
|
|
|
{
|
2016-03-29 10:46:07 +02:00
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2012-08-22 18:34:45 -05:00
|
|
|
NMIP4Config *ip4_config;
|
|
|
|
|
NMIP6Config *ip6_config;
|
|
|
|
|
|
2013-12-19 11:05:43 -05:00
|
|
|
nm_dns_manager_begin_updates (priv->dns_manager, __func__);
|
2012-08-22 18:34:45 -05:00
|
|
|
|
|
|
|
|
ip4_config = nm_vpn_connection_get_ip4_config (vpn);
|
|
|
|
|
if (ip4_config)
|
2017-12-16 15:11:16 +01:00
|
|
|
nm_dns_manager_set_ip_config (priv->dns_manager, NM_IP_CONFIG_CAST (ip4_config), NM_DNS_IP_CONFIG_TYPE_VPN);
|
2012-08-22 18:34:45 -05:00
|
|
|
|
|
|
|
|
ip6_config = nm_vpn_connection_get_ip6_config (vpn);
|
|
|
|
|
if (ip6_config)
|
2017-12-16 15:11:16 +01:00
|
|
|
nm_dns_manager_set_ip_config (priv->dns_manager, NM_IP_CONFIG_CAST (ip6_config), NM_DNS_IP_CONFIG_TYPE_VPN);
|
2012-08-22 18:34:45 -05:00
|
|
|
|
2016-03-29 10:46:07 +02:00
|
|
|
update_routing_and_dns (self, TRUE);
|
2012-08-22 18:34:45 -05:00
|
|
|
|
2013-12-19 11:05:43 -05:00
|
|
|
nm_dns_manager_end_updates (priv->dns_manager, __func__);
|
2012-08-22 18:34:45 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void
|
2016-03-29 10:46:07 +02:00
|
|
|
vpn_connection_deactivated (NMPolicy *self, NMVpnConnection *vpn)
|
2012-08-22 18:34:45 -05:00
|
|
|
{
|
2016-03-29 10:46:07 +02:00
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2013-08-01 10:59:42 -05:00
|
|
|
NMIP4Config *ip4_config;
|
|
|
|
|
NMIP6Config *ip6_config;
|
2012-08-22 18:34:45 -05:00
|
|
|
|
2013-12-19 11:05:43 -05:00
|
|
|
nm_dns_manager_begin_updates (priv->dns_manager, __func__);
|
2012-08-22 18:34:45 -05:00
|
|
|
|
|
|
|
|
ip4_config = nm_vpn_connection_get_ip4_config (vpn);
|
2017-12-16 15:11:16 +01:00
|
|
|
if (ip4_config)
|
|
|
|
|
nm_dns_manager_set_ip_config (priv->dns_manager, NM_IP_CONFIG_CAST (ip4_config), NM_DNS_IP_CONFIG_TYPE_REMOVED);
|
2012-08-22 18:34:45 -05:00
|
|
|
|
|
|
|
|
ip6_config = nm_vpn_connection_get_ip6_config (vpn);
|
2017-12-16 15:11:16 +01:00
|
|
|
if (ip6_config)
|
|
|
|
|
nm_dns_manager_set_ip_config (priv->dns_manager, NM_IP_CONFIG_CAST (ip6_config), NM_DNS_IP_CONFIG_TYPE_REMOVED);
|
2012-08-22 18:34:45 -05:00
|
|
|
|
2016-03-29 10:46:07 +02:00
|
|
|
update_routing_and_dns (self, TRUE);
|
2012-08-22 18:34:45 -05:00
|
|
|
|
2013-12-19 11:05:43 -05:00
|
|
|
nm_dns_manager_end_updates (priv->dns_manager, __func__);
|
2013-09-27 10:12:08 -05:00
|
|
|
}
|
2012-08-22 18:34:45 -05:00
|
|
|
|
2013-09-27 10:12:08 -05:00
|
|
|
static void
|
libnm, core, cli, tui: fix the capitalization of various types
GLib/Gtk have mostly settled on the convention that two-letter
acronyms in type names remain all-caps (eg, "IO"), but longer acronyms
become initial-caps-only (eg, "Tcp").
NM was inconsistent, with most long acronyms using initial caps only
(Adsl, Cdma, Dcb, Gsm, Olpc, Vlan), but others using all caps (DHCP,
PPP, PPPOE, VPN). Fix libnm and src/ to use initial-caps only for all
three-or-more-letter-long acronyms (and update nmcli and nmtui for the
libnm changes).
2014-06-26 13:44:36 -04:00
|
|
|
vpn_connection_state_changed (NMVpnConnection *vpn,
|
|
|
|
|
NMVpnConnectionState new_state,
|
|
|
|
|
NMVpnConnectionState old_state,
|
2017-03-11 14:39:07 +01:00
|
|
|
NMActiveConnectionStateReason reason,
|
2016-03-29 10:46:07 +02:00
|
|
|
NMPolicy *self)
|
2013-09-27 10:12:08 -05:00
|
|
|
{
|
|
|
|
|
if (new_state == NM_VPN_CONNECTION_STATE_ACTIVATED)
|
2016-03-29 10:46:07 +02:00
|
|
|
vpn_connection_activated (self, vpn);
|
2013-09-27 10:12:08 -05:00
|
|
|
else if (new_state >= NM_VPN_CONNECTION_STATE_FAILED) {
|
|
|
|
|
/* Only clean up IP/DNS if the connection ever got past IP_CONFIG */
|
|
|
|
|
if (old_state >= NM_VPN_CONNECTION_STATE_IP_CONFIG_GET &&
|
|
|
|
|
old_state <= NM_VPN_CONNECTION_STATE_ACTIVATED)
|
2016-03-29 10:46:07 +02:00
|
|
|
vpn_connection_deactivated (self, vpn);
|
2013-09-27 10:12:08 -05:00
|
|
|
}
|
2012-08-22 18:34:45 -05:00
|
|
|
}
|
|
|
|
|
|
2014-09-26 16:03:06 -05:00
|
|
|
static void
|
2016-03-29 10:46:07 +02:00
|
|
|
vpn_connection_retry_after_failure (NMVpnConnection *vpn, NMPolicy *self)
|
2014-09-26 16:03:06 -05:00
|
|
|
{
|
2016-03-29 10:46:07 +02:00
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2014-09-26 16:03:06 -05:00
|
|
|
NMActiveConnection *ac = NM_ACTIVE_CONNECTION (vpn);
|
2015-07-14 16:53:24 +02:00
|
|
|
NMSettingsConnection *connection = nm_active_connection_get_settings_connection (ac);
|
2014-09-26 16:03:06 -05:00
|
|
|
GError *error = NULL;
|
|
|
|
|
|
|
|
|
|
/* Attempt to reconnect VPN connections that failed after being connected */
|
|
|
|
|
if (!nm_manager_activate_connection (priv->manager,
|
|
|
|
|
connection,
|
|
|
|
|
NULL,
|
|
|
|
|
NULL,
|
2016-09-07 17:47:26 +02:00
|
|
|
NULL,
|
2014-09-26 16:03:06 -05:00
|
|
|
nm_active_connection_get_subject (ac),
|
2017-03-07 11:04:36 +01:00
|
|
|
NM_ACTIVATION_TYPE_MANAGED,
|
2018-03-28 17:18:04 +02:00
|
|
|
nm_active_connection_get_activation_reason (ac),
|
core: improve and fix keeping connection active based on "connection.permissions"
By setting "connection.permissions", a profile is restricted to a
particular user.
That means for example, that another user cannot see, modify, delete,
activate or deactivate the profile. It also means, that the profile
will only autoconnect when the user is logged in (has a session).
Note that root is always able to activate the profile. Likewise, the
user is also allowed to manually activate the own profile, even if no
session currently exists (which can easily happen with `sudo`).
When the user logs out (the session goes away), we want do disconnect
the profile, however there are conflicting goals here:
1) if the profile was activate by root user, then logging out the user
should not disconnect the profile. The patch fixes that by not
binding the activation to the connection, if the activation is done
by the root user.
2) if the profile was activated by the owner when it had no session,
then it should stay alive until the user logs in (once) and logs
out again. This is already handled by the previous commit.
Yes, this point is odd. If you first do
$ sudo -u $OTHER_USER nmcli connection up $PROFILE
the profile activates despite not having a session. If you then
$ ssh guest@localhost nmcli device
you'll still see the profile active. However, the moment the SSH session
ends, a session closes and the profile disconnects. It's unclear, how to
solve that any better. I think, a user who cares about this, should not
activate the profile without having a session in the first place.
There are quite some special cases, in particular with internal
activations. In those cases we need to decide whether to bind the
activation to the profile's visibility.
Also, expose the "bind" setting in the D-Bus API. Note, that in the future
this flag may be modified via D-Bus API. Like we may also add related API
that allows to tweak the lifetime of the activation.
Also, I think we broke handling of connection visiblity with 37e8c53eeed
"core: Introduce helper class to track connection keep alive". This
should be fixed now too, with improved behavior.
Fixes: 37e8c53eeed579fe34a68819cd12f3295d581394
https://bugzilla.redhat.com/show_bug.cgi?id=1530977
2018-11-21 13:30:16 +01:00
|
|
|
( nm_active_connection_get_state_flags (ac)
|
|
|
|
|
& NM_ACTIVATION_STATE_FLAG_LIFETIME_BOUND_TO_PROFILE_VISIBILITY),
|
2014-09-26 16:03:06 -05:00
|
|
|
&error)) {
|
2016-03-03 09:20:18 +01:00
|
|
|
_LOGW (LOGD_DEVICE, "VPN '%s' reconnect failed: %s",
|
|
|
|
|
nm_settings_connection_get_id (connection),
|
2018-04-24 11:20:03 +02:00
|
|
|
error->message ?: "unknown");
|
2014-09-26 16:03:06 -05:00
|
|
|
g_clear_error (&error);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-22 18:34:45 -05:00
|
|
|
static void
|
|
|
|
|
active_connection_state_changed (NMActiveConnection *active,
|
|
|
|
|
GParamSpec *pspec,
|
2016-03-29 10:46:07 +02:00
|
|
|
NMPolicy *self)
|
2012-08-22 18:34:45 -05:00
|
|
|
{
|
2013-09-27 10:12:08 -05:00
|
|
|
NMActiveConnectionState state = nm_active_connection_get_state (active);
|
|
|
|
|
|
|
|
|
|
if (state == NM_ACTIVE_CONNECTION_STATE_ACTIVATED)
|
2016-03-29 10:46:07 +02:00
|
|
|
process_secondaries (self, active, TRUE);
|
2013-09-27 10:12:08 -05:00
|
|
|
else if (state == NM_ACTIVE_CONNECTION_STATE_DEACTIVATED)
|
2016-03-29 10:46:07 +02:00
|
|
|
process_secondaries (self, active, FALSE);
|
2012-08-22 18:34:45 -05:00
|
|
|
}
|
|
|
|
|
|
2018-10-10 17:10:01 +02:00
|
|
|
static void
|
2018-11-20 14:51:22 +01:00
|
|
|
active_connection_keep_alive_changed (NMKeepAlive *keep_alive,
|
2018-10-10 17:10:01 +02:00
|
|
|
GParamSpec *pspec,
|
|
|
|
|
NMPolicy *self)
|
|
|
|
|
{
|
2018-11-20 14:51:22 +01:00
|
|
|
NMPolicyPrivate *priv;
|
|
|
|
|
NMActiveConnection *ac;
|
2018-10-10 17:10:01 +02:00
|
|
|
GError *error = NULL;
|
|
|
|
|
|
2018-11-20 14:51:22 +01:00
|
|
|
nm_assert (NM_IS_POLICY (self));
|
|
|
|
|
nm_assert (NM_IS_KEEP_ALIVE (keep_alive));
|
|
|
|
|
nm_assert (NM_IS_ACTIVE_CONNECTION (nm_keep_alive_get_owner (keep_alive)));
|
|
|
|
|
|
|
|
|
|
if (nm_keep_alive_is_alive (keep_alive))
|
2018-11-17 13:35:02 +01:00
|
|
|
return;
|
|
|
|
|
|
2018-11-20 14:51:22 +01:00
|
|
|
ac = nm_keep_alive_get_owner (keep_alive);
|
|
|
|
|
|
|
|
|
|
if (nm_active_connection_get_state (ac) > NM_ACTIVE_CONNECTION_STATE_ACTIVATED)
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
priv = NM_POLICY_GET_PRIVATE (self);
|
|
|
|
|
|
|
|
|
|
if (!nm_manager_deactivate_connection (priv->manager,
|
|
|
|
|
ac,
|
|
|
|
|
NM_DEVICE_STATE_REASON_CONNECTION_REMOVED,
|
|
|
|
|
&error)) {
|
|
|
|
|
_LOGW (LOGD_DEVICE, "connection '%s' is no longer kept alive, but error deactivating it: %s",
|
|
|
|
|
nm_active_connection_get_settings_connection_id (ac),
|
|
|
|
|
error->message);
|
|
|
|
|
g_clear_error (&error);
|
2018-10-10 17:10:01 +02:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-22 18:34:45 -05:00
|
|
|
static void
|
|
|
|
|
active_connection_added (NMManager *manager,
|
|
|
|
|
NMActiveConnection *active,
|
|
|
|
|
gpointer user_data)
|
|
|
|
|
{
|
2016-04-04 14:36:34 +02:00
|
|
|
NMPolicyPrivate *priv = user_data;
|
2016-09-29 13:49:01 +02:00
|
|
|
NMPolicy *self = _PRIV_TO_SELF (priv);
|
2018-11-20 14:51:22 +01:00
|
|
|
NMKeepAlive *keep_alive;
|
2013-09-27 10:12:08 -05:00
|
|
|
|
|
|
|
|
if (NM_IS_VPN_CONNECTION (active)) {
|
|
|
|
|
g_signal_connect (active, NM_VPN_CONNECTION_INTERNAL_STATE_CHANGED,
|
|
|
|
|
G_CALLBACK (vpn_connection_state_changed),
|
2016-03-29 10:46:07 +02:00
|
|
|
self);
|
2014-09-26 16:03:06 -05:00
|
|
|
g_signal_connect (active, NM_VPN_CONNECTION_INTERNAL_RETRY_AFTER_FAILURE,
|
|
|
|
|
G_CALLBACK (vpn_connection_retry_after_failure),
|
2016-03-29 10:46:07 +02:00
|
|
|
self);
|
2013-09-27 10:12:08 -05:00
|
|
|
}
|
2012-08-22 18:34:45 -05:00
|
|
|
|
2018-11-20 14:51:22 +01:00
|
|
|
keep_alive = nm_active_connection_get_keep_alive (active);
|
|
|
|
|
|
device: arm keep-alive instance when queuing active-connection for activation
Now that the keep-alive instance defaults to ALIVE by default, we can
always arm it when starting to activate the active-connection.
The keep-alive instance may have been armed earlier already:
for example, when binding its lifetime to a D-Bus name or
when watching the connection's visible state.
However, at the moment when we queue the active-connection for
activation, we also want to make sure that the keep-alive instance is
armed. It is nicer for consistancy reasons.
Note, that nm_keep_alive_arm() has no effect if nm_keep_alive_disarm()
was called earlier already. Also note, that NMActiveConnection will
disarm the keep-alive instance, when changing to a state greater than
ACTIVATED. So, all works together nicely.
Also, no longer arm the keep-alive instance in the constructor of
NMActiveConnection. It would essentially mean, that the instances
is aremd very early.
Also, as alternative point of interest, arm the keep-alive instance
when registering the signal handler in "nm-policy.c".
2018-11-20 16:01:32 +01:00
|
|
|
nm_keep_alive_arm (keep_alive);
|
|
|
|
|
|
2012-08-22 18:34:45 -05:00
|
|
|
g_signal_connect (active, "notify::" NM_ACTIVE_CONNECTION_STATE,
|
|
|
|
|
G_CALLBACK (active_connection_state_changed),
|
2016-03-29 10:46:07 +02:00
|
|
|
self);
|
2018-11-20 14:51:22 +01:00
|
|
|
g_signal_connect (keep_alive,
|
|
|
|
|
"notify::" NM_KEEP_ALIVE_ALIVE,
|
2018-10-10 17:10:01 +02:00
|
|
|
G_CALLBACK (active_connection_keep_alive_changed),
|
|
|
|
|
self);
|
2018-11-20 14:51:22 +01:00
|
|
|
active_connection_keep_alive_changed (keep_alive, NULL, self);
|
2012-08-22 18:34:45 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void
|
|
|
|
|
active_connection_removed (NMManager *manager,
|
|
|
|
|
NMActiveConnection *active,
|
|
|
|
|
gpointer user_data)
|
|
|
|
|
{
|
2016-04-04 14:36:34 +02:00
|
|
|
NMPolicyPrivate *priv = user_data;
|
2016-09-29 13:49:01 +02:00
|
|
|
NMPolicy *self = _PRIV_TO_SELF (priv);
|
2013-09-27 10:12:08 -05:00
|
|
|
|
|
|
|
|
g_signal_handlers_disconnect_by_func (active,
|
|
|
|
|
vpn_connection_state_changed,
|
2016-03-29 10:46:07 +02:00
|
|
|
self);
|
2014-09-26 16:03:06 -05:00
|
|
|
g_signal_handlers_disconnect_by_func (active,
|
|
|
|
|
vpn_connection_retry_after_failure,
|
2016-03-29 10:46:07 +02:00
|
|
|
self);
|
2012-08-22 18:34:45 -05:00
|
|
|
g_signal_handlers_disconnect_by_func (active,
|
|
|
|
|
active_connection_state_changed,
|
2016-03-29 10:46:07 +02:00
|
|
|
self);
|
2018-11-20 14:51:22 +01:00
|
|
|
g_signal_handlers_disconnect_by_func (nm_active_connection_get_keep_alive (active),
|
2018-11-20 15:47:24 +01:00
|
|
|
active_connection_keep_alive_changed,
|
|
|
|
|
self);
|
2012-08-22 18:34:45 -05:00
|
|
|
}
|
|
|
|
|
|
2016-10-02 18:22:50 +02:00
|
|
|
/*****************************************************************************/
|
2012-08-22 18:34:45 -05:00
|
|
|
|
2016-04-13 14:14:36 +02:00
|
|
|
static gboolean
|
|
|
|
|
schedule_activate_all_cb (gpointer user_data)
|
2008-03-07 Dan Williams <dcbw@redhat.com>
First pass of multiple active device support. Expect bugs.
* src/nm-ip4-config.c
src/nm-ip4-config.h
- (nm_ip4_config_get_secondary, nm_ip4_config_set_secondary): remove;
there are better ways to do this in the named manager
* src/nm-device.c
src/nm-device.h
- (nm_device_can_activate): return whether the device can activate a
connection right now; taking into account things like carrier state
and rfkill state
- (nm_device_get_best_auto_connection): renamed from
nm_device_get_best_connection
- (real_act_stage4_get_ip4_config): MTU stuff is now handled in the
device subclasses themselves, so that each device can override the
MTU from it's NMSetting subclass if needed
- (nm_device_set_ip4_config): set MTU when setting up routes and stuff
in NetworkManagerSystem.c, not here
* src/named-manager/nm-named-manager.c
src/named-manager/nm-named-manager.h
- (nm_named_manager_name_owner_changed,
nm_named_manager_dbus_connection_changed): fix for changes to
rewrite_resolv_conf()
- (compute_nameservers): don't need the NMNamedManager at all, remove
from parameter list
- (merge_one_ip4_config): new function; merge ip4 configs together
- (rewrite_resolv_conf): write out resolv.conf from all the stored
ip4 configs; the VPN config takes precedence, then the best
device config, then the rest of the configs
- (get_domain_for_config): take the NMNamedManager as an argument
to check whether the config is the VPN config
- (add_ip4_config_to_named): fixups for removal of the 'secondary'
attribute from ip4 configs
- (add_all_ip4_configs_to_named): add all the configs in priority order
- (remove_ip4_config_from_named): fix for changes to
get_domain_for_config()
- (nm_named_manager_add_ip4_config): assign the config to the right slot
based on its type; callers must pass in the type now
- (get_last_default_domain): remove, unused
- (nm_named_manager_remove_ip4_config): handle config slots correctly
* src/nm-device-802-11-wireless.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): handle MTU override
* src/nm-device-802-3-ethernet.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): new function; handle MTU override
* src/vpn-manager/nm-vpn-connection.c
- (nm_vpn_connection_ip4_config_get): don't need to set the 'secondary'
attribute on the ip4 config
* src/NetworkManagerPolicy.c
- (nm_policy_auto_get_best_device): remove
- (nm_policy_device_change_check): remove
- (update_default_route): new function; set the default route via
the specified device
- (get_device_priority): new function; return the priority number of
a device type WRT which one should have the default route. Order is
(highest to lowest) wired, wireless, GSM, CDMA.
- (update_routing_and_dns): new function; determine which device should
have the default route, then update the routing table and DNS
- (maybe_auto_activate_device): new function; if a device is now
available for activation, find out what connection it would like to
activate and do it
- (schedule_activate_check): new function; if a device can be activated
now, schedule the activation. Each device may have only one
pending activation at a given time.
- (device_state_changed): if activation was canceled, try again,
possibly with another connection; if the device was activated,
update routing and DNS; if the device was deactivated, try again
with another connection
- (device_carrier_changed): if there is no carrier, deactivate the
device; otherwise schedule an activation check for the device
- (wireless_networks_changed): schedule an activation check for the
device
- (device_added): keep track of the signal handler IDs so they can
be removed when the device goes away
- (device_removed): remove any signal handlers that might be attached
to the device; update routing and DNS
- (schedule_activate_all): new function
- (connections_added, connection_added, connection_updated): when
connections change, schedule all devices for an activation check
- (connection_removed): when a device is deactivated because its
connection was removed, schedule another activation check for it
- (nm_policy_destroy): destroy pending activations and disconnect
all device signal handlers
* src/nm-manager.c
- (nm_manager_activate_device): if the device was already actived,
deactivate it
- (deactivate_old_device): remove
- (connection_added_default_handler, impl_manager_activate_device):
don't deactivate other devices when activating this one
* src/backends/NetworkManagerGentoo.c
src/backends/NetworkManagerFrugalware.c
src/backends/NetworkManagerPaldo.c
src/backends/NetworkManagerRedHat.c
src/backends/NetworkManagerSlackware.c
src/backends/NetworkManagerArch.c
src/backends/NetworkManagerSuSE.c
src/backends/NetworkManagerDebian.c
- (nm_system_get_mtu): remove; MTU should be provided through the
distro's system settings service plugin instead
- (nm_system_device_add_default_route_via_device): remove
- (nm_system_device_add_default_route_via_device_with_iface): remove
- (nm_system_device_replace_default_route): new function; call
generic implementation
* src/backends/NetworkManagerGeneric.c
src/backends/NetworkManagerGeneric.h
- (nm_generic_device_add_default_route_via_device,
nm_generic_device_add_default_route_via_device_with_iface): remove
- (nm_generic_device_replace_default_route): replace the default route
with the given route via some gateway
* src/NetworkManagerSystem.c
src/NetworkManagerSystem.h
- (nm_system_device_set_from_ip4_config): let the policy handle updates
to routing and DNS; but set the MTU here
- (nm_system_vpn_device_set_from_ip4_config): set the route with the
ip_iface of the active device; use the standard MTU setting function
- (nm_system_set_mtu): remove
- (nm_system_device_set_mtu): consolidate MTU setting code in one place
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@3391 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-03-07 19:41:32 +00:00
|
|
|
{
|
2016-04-13 14:14:36 +02:00
|
|
|
NMPolicy *self = user_data;
|
2016-03-29 10:46:07 +02:00
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2018-03-29 08:52:45 +02:00
|
|
|
const CList *tmp_lst;
|
core: track devices in manager via embedded CList
Instead of using a GSList for tracking the devices, use a CList.
I think a CList is in most cases the more suitable data structure
then GSList:
- you can find out in O(1) whether the object is linked. That
is nice, for example to assert in NMDevice's destructor that
the object was unlinked, and we will use that later in
nm_manager_get_device_by_path().
- you can unlink the element in O(1) and you can unlink the
element without having access to the link's head
- Contrary to GSList, this does not require an extra slice
allocation for the link node. It quite possibliy consumes
slightly less memory because the CList structure is embedded
in a struct that we already allocate. Even if slice allocation
would be perfect to only consume 2*sizeof(gpointer) for the link
note, it would at most be as-good as CList. Quite possibly,
there is an overhead though.
- CList possibly has better memory locality, because the link
structure and the data are close to each other.
Something which could be seen as disavantage, is that with CList
one device can only be tracked in one NMManager instance at a time.
But that is fine. There exists only one NMManager instance for now,
and even if we would ever introduce multiple managers, we probably
would not associate one NMDevice instance with multiple managers.
The advantages are arguably not huge, but CList is IMHO clearly the
more suited data structure. No need to stick to a suboptimal data
structure for the job. Refactor it.
2018-03-23 21:51:07 +01:00
|
|
|
NMDevice *device;
|
2007-02-08 15:34:26 +00:00
|
|
|
|
2016-04-13 14:14:36 +02:00
|
|
|
priv->schedule_activate_all_id = 0;
|
|
|
|
|
|
2018-03-29 08:52:45 +02:00
|
|
|
nm_manager_for_each_device (priv->manager, device, tmp_lst)
|
core: track devices in manager via embedded CList
Instead of using a GSList for tracking the devices, use a CList.
I think a CList is in most cases the more suitable data structure
then GSList:
- you can find out in O(1) whether the object is linked. That
is nice, for example to assert in NMDevice's destructor that
the object was unlinked, and we will use that later in
nm_manager_get_device_by_path().
- you can unlink the element in O(1) and you can unlink the
element without having access to the link's head
- Contrary to GSList, this does not require an extra slice
allocation for the link node. It quite possibliy consumes
slightly less memory because the CList structure is embedded
in a struct that we already allocate. Even if slice allocation
would be perfect to only consume 2*sizeof(gpointer) for the link
note, it would at most be as-good as CList. Quite possibly,
there is an overhead though.
- CList possibly has better memory locality, because the link
structure and the data are close to each other.
Something which could be seen as disavantage, is that with CList
one device can only be tracked in one NMManager instance at a time.
But that is fine. There exists only one NMManager instance for now,
and even if we would ever introduce multiple managers, we probably
would not associate one NMDevice instance with multiple managers.
The advantages are arguably not huge, but CList is IMHO clearly the
more suited data structure. No need to stick to a suboptimal data
structure for the job. Refactor it.
2018-03-23 21:51:07 +01:00
|
|
|
schedule_activate_check (self, device);
|
2016-04-13 14:14:36 +02:00
|
|
|
|
|
|
|
|
return G_SOURCE_REMOVE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void
|
|
|
|
|
schedule_activate_all (NMPolicy *self)
|
|
|
|
|
{
|
|
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
|
|
|
|
|
|
|
|
|
/* always restart the idle handler. That way, we settle
|
|
|
|
|
* all other events before restarting to activate them. */
|
|
|
|
|
nm_clear_g_source (&priv->schedule_activate_all_id);
|
|
|
|
|
priv->schedule_activate_all_id = g_idle_add (schedule_activate_all_cb, self);
|
2007-02-08 15:34:26 +00:00
|
|
|
}
|
|
|
|
|
|
2007-10-07 23:33:28 +00:00
|
|
|
static void
|
2010-10-27 20:05:23 -05:00
|
|
|
connection_added (NMSettings *settings,
|
2013-12-05 04:57:01 -05:00
|
|
|
NMSettingsConnection *connection,
|
2007-08-13 01:45:47 +00:00
|
|
|
gpointer user_data)
|
2007-05-07 15:17:45 +00:00
|
|
|
{
|
2016-04-04 14:36:34 +02:00
|
|
|
NMPolicyPrivate *priv = user_data;
|
2016-09-29 13:49:01 +02:00
|
|
|
NMPolicy *self = _PRIV_TO_SELF (priv);
|
2013-12-05 04:57:01 -05:00
|
|
|
|
2016-03-29 10:46:07 +02:00
|
|
|
schedule_activate_all (self);
|
2007-08-13 01:45:47 +00:00
|
|
|
}
|
|
|
|
|
|
2012-03-01 18:54:22 +01:00
|
|
|
static void
|
2017-04-20 18:30:01 +02:00
|
|
|
firewall_state_changed (NMFirewallManager *manager,
|
2017-04-20 18:52:06 +02:00
|
|
|
gboolean initialized_now,
|
2017-04-20 18:30:01 +02:00
|
|
|
gpointer user_data)
|
2012-03-01 18:54:22 +01:00
|
|
|
{
|
2016-03-29 10:46:07 +02:00
|
|
|
NMPolicy *self = (NMPolicy *) user_data;
|
|
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2018-03-29 08:52:45 +02:00
|
|
|
const CList *tmp_lst;
|
core: track devices in manager via embedded CList
Instead of using a GSList for tracking the devices, use a CList.
I think a CList is in most cases the more suitable data structure
then GSList:
- you can find out in O(1) whether the object is linked. That
is nice, for example to assert in NMDevice's destructor that
the object was unlinked, and we will use that later in
nm_manager_get_device_by_path().
- you can unlink the element in O(1) and you can unlink the
element without having access to the link's head
- Contrary to GSList, this does not require an extra slice
allocation for the link node. It quite possibliy consumes
slightly less memory because the CList structure is embedded
in a struct that we already allocate. Even if slice allocation
would be perfect to only consume 2*sizeof(gpointer) for the link
note, it would at most be as-good as CList. Quite possibly,
there is an overhead though.
- CList possibly has better memory locality, because the link
structure and the data are close to each other.
Something which could be seen as disavantage, is that with CList
one device can only be tracked in one NMManager instance at a time.
But that is fine. There exists only one NMManager instance for now,
and even if we would ever introduce multiple managers, we probably
would not associate one NMDevice instance with multiple managers.
The advantages are arguably not huge, but CList is IMHO clearly the
more suited data structure. No need to stick to a suboptimal data
structure for the job. Refactor it.
2018-03-23 21:51:07 +01:00
|
|
|
NMDevice *device;
|
2012-03-01 18:54:22 +01:00
|
|
|
|
2017-04-20 18:52:06 +02:00
|
|
|
if (initialized_now) {
|
|
|
|
|
/* the firewall manager was initializing, but all requests
|
|
|
|
|
* so fare were queued and are already sent. No need to
|
|
|
|
|
* re-update the firewall zone of the devices. */
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
2017-04-20 18:30:01 +02:00
|
|
|
if (!nm_firewall_manager_get_running (manager))
|
|
|
|
|
return;
|
|
|
|
|
|
2012-03-01 18:54:23 +01:00
|
|
|
/* add interface of each device to correct zone */
|
2018-03-29 08:52:45 +02:00
|
|
|
nm_manager_for_each_device (priv->manager, device, tmp_lst)
|
core: track devices in manager via embedded CList
Instead of using a GSList for tracking the devices, use a CList.
I think a CList is in most cases the more suitable data structure
then GSList:
- you can find out in O(1) whether the object is linked. That
is nice, for example to assert in NMDevice's destructor that
the object was unlinked, and we will use that later in
nm_manager_get_device_by_path().
- you can unlink the element in O(1) and you can unlink the
element without having access to the link's head
- Contrary to GSList, this does not require an extra slice
allocation for the link node. It quite possibliy consumes
slightly less memory because the CList structure is embedded
in a struct that we already allocate. Even if slice allocation
would be perfect to only consume 2*sizeof(gpointer) for the link
note, it would at most be as-good as CList. Quite possibly,
there is an overhead though.
- CList possibly has better memory locality, because the link
structure and the data are close to each other.
Something which could be seen as disavantage, is that with CList
one device can only be tracked in one NMManager instance at a time.
But that is fine. There exists only one NMManager instance for now,
and even if we would ever introduce multiple managers, we probably
would not associate one NMDevice instance with multiple managers.
The advantages are arguably not huge, but CList is IMHO clearly the
more suited data structure. No need to stick to a suboptimal data
structure for the job. Refactor it.
2018-03-23 21:51:07 +01:00
|
|
|
nm_device_update_firewall_zone (device);
|
2012-03-01 18:54:22 +01:00
|
|
|
}
|
|
|
|
|
|
2013-02-05 17:29:15 +01:00
|
|
|
static void
|
|
|
|
|
dns_config_changed (NMDnsManager *dns_manager, gpointer user_data)
|
|
|
|
|
{
|
2016-03-29 10:46:07 +02:00
|
|
|
NMPolicy *self = (NMPolicy *) user_data;
|
|
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2013-02-05 17:29:15 +01:00
|
|
|
|
|
|
|
|
/* Restart a thread for reverse-DNS lookup after we are signalled that
|
|
|
|
|
* DNS changed. Because the result from a previous run may not be right
|
|
|
|
|
* (race in updating DNS and doing the reverse lookup).
|
|
|
|
|
*/
|
|
|
|
|
|
2017-04-23 19:30:07 +02:00
|
|
|
nm_clear_g_cancellable (&priv->lookup.cancellable);
|
2013-02-05 17:29:15 +01:00
|
|
|
|
|
|
|
|
/* Re-start the hostname lookup thread if we don't have hostname yet. */
|
2017-04-23 19:30:07 +02:00
|
|
|
if (priv->lookup.addr) {
|
2014-08-29 10:44:00 +02:00
|
|
|
char *str = NULL;
|
2017-02-17 16:22:30 +01:00
|
|
|
gs_free char *hostname = NULL;
|
|
|
|
|
|
|
|
|
|
/* Check if the hostname was externally set */
|
2017-04-23 19:54:39 +02:00
|
|
|
if ( (hostname = _get_hostname (self))
|
2017-02-17 16:22:30 +01:00
|
|
|
&& nm_utils_is_specific_hostname (hostname)
|
|
|
|
|
&& !nm_streq0 (hostname, priv->last_hostname)) {
|
2017-04-23 19:30:07 +02:00
|
|
|
g_clear_object (&priv->lookup.addr);
|
2017-02-17 16:22:30 +01:00
|
|
|
return;
|
|
|
|
|
}
|
2013-11-25 15:18:02 -06:00
|
|
|
|
2016-03-03 09:20:18 +01:00
|
|
|
_LOGD (LOGD_DNS, "restarting reverse-lookup thread for address %s",
|
2017-04-23 19:30:07 +02:00
|
|
|
(str = g_inet_address_to_string (priv->lookup.addr)));
|
2013-11-25 15:18:02 -06:00
|
|
|
g_free (str);
|
2013-02-24 13:51:56 +01:00
|
|
|
|
2017-04-23 19:24:21 +02:00
|
|
|
lookup_by_address (self);
|
2013-02-05 17:29:15 +01:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2007-10-06 04:31:13 +00:00
|
|
|
static void
|
2010-10-27 20:05:23 -05:00
|
|
|
connection_updated (NMSettings *settings,
|
2016-04-13 16:03:06 +02:00
|
|
|
NMSettingsConnection *connection,
|
|
|
|
|
gboolean by_user,
|
2007-10-06 04:31:13 +00:00
|
|
|
gpointer user_data)
|
2015-09-07 14:17:39 +02:00
|
|
|
{
|
2016-04-04 14:36:34 +02:00
|
|
|
NMPolicyPrivate *priv = user_data;
|
2016-09-29 13:49:01 +02:00
|
|
|
NMPolicy *self = _PRIV_TO_SELF (priv);
|
2018-03-29 08:52:45 +02:00
|
|
|
const CList *tmp_lst;
|
2015-09-07 12:02:54 +02:00
|
|
|
NMDevice *device = NULL;
|
core: track devices in manager via embedded CList
Instead of using a GSList for tracking the devices, use a CList.
I think a CList is in most cases the more suitable data structure
then GSList:
- you can find out in O(1) whether the object is linked. That
is nice, for example to assert in NMDevice's destructor that
the object was unlinked, and we will use that later in
nm_manager_get_device_by_path().
- you can unlink the element in O(1) and you can unlink the
element without having access to the link's head
- Contrary to GSList, this does not require an extra slice
allocation for the link node. It quite possibliy consumes
slightly less memory because the CList structure is embedded
in a struct that we already allocate. Even if slice allocation
would be perfect to only consume 2*sizeof(gpointer) for the link
note, it would at most be as-good as CList. Quite possibly,
there is an overhead though.
- CList possibly has better memory locality, because the link
structure and the data are close to each other.
Something which could be seen as disavantage, is that with CList
one device can only be tracked in one NMManager instance at a time.
But that is fine. There exists only one NMManager instance for now,
and even if we would ever introduce multiple managers, we probably
would not associate one NMDevice instance with multiple managers.
The advantages are arguably not huge, but CList is IMHO clearly the
more suited data structure. No need to stick to a suboptimal data
structure for the job. Refactor it.
2018-03-23 21:51:07 +01:00
|
|
|
NMDevice *dev;
|
2015-09-07 12:02:54 +02:00
|
|
|
|
2016-04-13 16:03:06 +02:00
|
|
|
if (by_user) {
|
|
|
|
|
/* find device with given connection */
|
2018-03-29 08:52:45 +02:00
|
|
|
nm_manager_for_each_device (priv->manager, dev, tmp_lst) {
|
2016-04-13 16:03:06 +02:00
|
|
|
if (nm_device_get_settings_connection (dev) == connection) {
|
|
|
|
|
device = dev;
|
|
|
|
|
break;
|
|
|
|
|
}
|
2015-09-07 12:02:54 +02:00
|
|
|
}
|
2011-10-07 15:58:09 +02:00
|
|
|
|
2016-04-13 16:03:06 +02:00
|
|
|
if (device)
|
|
|
|
|
nm_device_reapply_settings_immediately (device);
|
|
|
|
|
|
|
|
|
|
/* Reset auto retries back to default since connection was updated */
|
2017-10-24 08:26:25 +02:00
|
|
|
nm_settings_connection_autoconnect_retries_reset (connection);
|
2016-04-13 16:03:06 +02:00
|
|
|
}
|
2015-07-14 16:53:24 +02:00
|
|
|
|
2016-04-13 16:03:06 +02:00
|
|
|
schedule_activate_all (self);
|
2013-12-12 20:00:38 +01:00
|
|
|
}
|
|
|
|
|
|
2007-08-13 01:45:47 +00:00
|
|
|
static void
|
2017-11-23 21:30:09 +01:00
|
|
|
_deactivate_if_active (NMPolicy *self, NMSettingsConnection *connection)
|
2007-08-13 01:45:47 +00:00
|
|
|
{
|
2017-11-23 21:30:09 +01:00
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
|
|
|
|
NMActiveConnection *ac;
|
2018-11-21 10:51:21 +01:00
|
|
|
const CList *tmp_list, *tmp_safe;
|
2017-11-23 21:30:09 +01:00
|
|
|
GError *error = NULL;
|
|
|
|
|
|
|
|
|
|
nm_assert (NM_IS_SETTINGS_CONNECTION (connection));
|
2007-10-06 04:26:33 +00:00
|
|
|
|
2018-11-21 10:51:21 +01:00
|
|
|
nm_manager_for_each_active_connection_safe (priv->manager, ac, tmp_list, tmp_safe) {
|
2007-10-06 04:26:33 +00:00
|
|
|
|
2017-11-23 21:30:09 +01:00
|
|
|
if ( nm_active_connection_get_settings_connection (ac) == connection
|
2018-03-29 16:47:05 +02:00
|
|
|
&& (nm_active_connection_get_state (ac) <= NM_ACTIVE_CONNECTION_STATE_ACTIVATED)) {
|
2017-11-23 21:30:09 +01:00
|
|
|
if (!nm_manager_deactivate_connection (priv->manager,
|
2017-01-26 14:20:01 +01:00
|
|
|
ac,
|
2012-08-22 17:21:56 -05:00
|
|
|
NM_DEVICE_STATE_REASON_CONNECTION_REMOVED,
|
|
|
|
|
&error)) {
|
2016-03-03 09:20:18 +01:00
|
|
|
_LOGW (LOGD_DEVICE, "connection '%s' disappeared, but error deactivating it: (%d) %s",
|
|
|
|
|
nm_settings_connection_get_id (connection),
|
|
|
|
|
error ? error->code : -1,
|
|
|
|
|
error ? error->message : "(unknown)");
|
2012-08-22 17:21:56 -05:00
|
|
|
g_clear_error (&error);
|
|
|
|
|
}
|
2008-03-07 Dan Williams <dcbw@redhat.com>
First pass of multiple active device support. Expect bugs.
* src/nm-ip4-config.c
src/nm-ip4-config.h
- (nm_ip4_config_get_secondary, nm_ip4_config_set_secondary): remove;
there are better ways to do this in the named manager
* src/nm-device.c
src/nm-device.h
- (nm_device_can_activate): return whether the device can activate a
connection right now; taking into account things like carrier state
and rfkill state
- (nm_device_get_best_auto_connection): renamed from
nm_device_get_best_connection
- (real_act_stage4_get_ip4_config): MTU stuff is now handled in the
device subclasses themselves, so that each device can override the
MTU from it's NMSetting subclass if needed
- (nm_device_set_ip4_config): set MTU when setting up routes and stuff
in NetworkManagerSystem.c, not here
* src/named-manager/nm-named-manager.c
src/named-manager/nm-named-manager.h
- (nm_named_manager_name_owner_changed,
nm_named_manager_dbus_connection_changed): fix for changes to
rewrite_resolv_conf()
- (compute_nameservers): don't need the NMNamedManager at all, remove
from parameter list
- (merge_one_ip4_config): new function; merge ip4 configs together
- (rewrite_resolv_conf): write out resolv.conf from all the stored
ip4 configs; the VPN config takes precedence, then the best
device config, then the rest of the configs
- (get_domain_for_config): take the NMNamedManager as an argument
to check whether the config is the VPN config
- (add_ip4_config_to_named): fixups for removal of the 'secondary'
attribute from ip4 configs
- (add_all_ip4_configs_to_named): add all the configs in priority order
- (remove_ip4_config_from_named): fix for changes to
get_domain_for_config()
- (nm_named_manager_add_ip4_config): assign the config to the right slot
based on its type; callers must pass in the type now
- (get_last_default_domain): remove, unused
- (nm_named_manager_remove_ip4_config): handle config slots correctly
* src/nm-device-802-11-wireless.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): handle MTU override
* src/nm-device-802-3-ethernet.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): new function; handle MTU override
* src/vpn-manager/nm-vpn-connection.c
- (nm_vpn_connection_ip4_config_get): don't need to set the 'secondary'
attribute on the ip4 config
* src/NetworkManagerPolicy.c
- (nm_policy_auto_get_best_device): remove
- (nm_policy_device_change_check): remove
- (update_default_route): new function; set the default route via
the specified device
- (get_device_priority): new function; return the priority number of
a device type WRT which one should have the default route. Order is
(highest to lowest) wired, wireless, GSM, CDMA.
- (update_routing_and_dns): new function; determine which device should
have the default route, then update the routing table and DNS
- (maybe_auto_activate_device): new function; if a device is now
available for activation, find out what connection it would like to
activate and do it
- (schedule_activate_check): new function; if a device can be activated
now, schedule the activation. Each device may have only one
pending activation at a given time.
- (device_state_changed): if activation was canceled, try again,
possibly with another connection; if the device was activated,
update routing and DNS; if the device was deactivated, try again
with another connection
- (device_carrier_changed): if there is no carrier, deactivate the
device; otherwise schedule an activation check for the device
- (wireless_networks_changed): schedule an activation check for the
device
- (device_added): keep track of the signal handler IDs so they can
be removed when the device goes away
- (device_removed): remove any signal handlers that might be attached
to the device; update routing and DNS
- (schedule_activate_all): new function
- (connections_added, connection_added, connection_updated): when
connections change, schedule all devices for an activation check
- (connection_removed): when a device is deactivated because its
connection was removed, schedule another activation check for it
- (nm_policy_destroy): destroy pending activations and disconnect
all device signal handlers
* src/nm-manager.c
- (nm_manager_activate_device): if the device was already actived,
deactivate it
- (deactivate_old_device): remove
- (connection_added_default_handler, impl_manager_activate_device):
don't deactivate other devices when activating this one
* src/backends/NetworkManagerGentoo.c
src/backends/NetworkManagerFrugalware.c
src/backends/NetworkManagerPaldo.c
src/backends/NetworkManagerRedHat.c
src/backends/NetworkManagerSlackware.c
src/backends/NetworkManagerArch.c
src/backends/NetworkManagerSuSE.c
src/backends/NetworkManagerDebian.c
- (nm_system_get_mtu): remove; MTU should be provided through the
distro's system settings service plugin instead
- (nm_system_device_add_default_route_via_device): remove
- (nm_system_device_add_default_route_via_device_with_iface): remove
- (nm_system_device_replace_default_route): new function; call
generic implementation
* src/backends/NetworkManagerGeneric.c
src/backends/NetworkManagerGeneric.h
- (nm_generic_device_add_default_route_via_device,
nm_generic_device_add_default_route_via_device_with_iface): remove
- (nm_generic_device_replace_default_route): replace the default route
with the given route via some gateway
* src/NetworkManagerSystem.c
src/NetworkManagerSystem.h
- (nm_system_device_set_from_ip4_config): let the policy handle updates
to routing and DNS; but set the MTU here
- (nm_system_vpn_device_set_from_ip4_config): set the route with the
ip_iface of the active device; use the standard MTU setting function
- (nm_system_set_mtu): remove
- (nm_system_device_set_mtu): consolidate MTU setting code in one place
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@3391 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-03-07 19:41:32 +00:00
|
|
|
}
|
2007-10-06 04:26:33 +00:00
|
|
|
}
|
2007-05-07 15:17:45 +00:00
|
|
|
}
|
|
|
|
|
|
2010-06-04 13:05:27 -07:00
|
|
|
static void
|
2010-10-27 20:05:23 -05:00
|
|
|
connection_removed (NMSettings *settings,
|
2015-07-14 16:53:24 +02:00
|
|
|
NMSettingsConnection *connection,
|
2010-10-27 15:47:10 -05:00
|
|
|
gpointer user_data)
|
|
|
|
|
{
|
2017-11-28 13:30:22 +01:00
|
|
|
NMPolicyPrivate *priv = user_data;
|
|
|
|
|
NMPolicy *self = _PRIV_TO_SELF (priv);
|
|
|
|
|
|
|
|
|
|
_deactivate_if_active (self, connection);
|
2010-10-27 15:47:10 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void
|
2017-12-04 16:05:26 +01:00
|
|
|
connection_flags_changed (NMSettings *settings,
|
|
|
|
|
NMSettingsConnection *connection,
|
|
|
|
|
gpointer user_data)
|
2010-06-04 13:05:27 -07:00
|
|
|
{
|
2016-04-04 14:36:34 +02:00
|
|
|
NMPolicyPrivate *priv = user_data;
|
2016-09-29 13:49:01 +02:00
|
|
|
NMPolicy *self = _PRIV_TO_SELF (priv);
|
2010-10-27 15:47:10 -05:00
|
|
|
|
2017-12-05 12:10:16 +01:00
|
|
|
if (NM_FLAGS_HAS (nm_settings_connection_get_flags (connection),
|
2018-04-05 20:02:45 +02:00
|
|
|
NM_SETTINGS_CONNECTION_INT_FLAGS_VISIBLE)) {
|
2017-12-05 12:20:41 +01:00
|
|
|
if (!nm_settings_connection_autoconnect_is_blocked (connection))
|
|
|
|
|
schedule_activate_all (self);
|
2018-10-10 17:10:01 +02:00
|
|
|
}
|
2010-06-04 13:05:27 -07:00
|
|
|
}
|
|
|
|
|
|
2011-06-28 15:48:12 +02:00
|
|
|
static void
|
|
|
|
|
secret_agent_registered (NMSettings *settings,
|
|
|
|
|
NMSecretAgent *agent,
|
|
|
|
|
gpointer user_data)
|
|
|
|
|
{
|
2017-11-27 09:07:28 +01:00
|
|
|
NMPolicy *self = NM_POLICY (user_data);
|
2013-12-05 04:57:01 -05:00
|
|
|
|
2011-06-28 15:48:12 +02:00
|
|
|
/* The registered secret agent may provide some missing secrets. Thus we
|
|
|
|
|
* reset retries count here and schedule activation, so that the
|
|
|
|
|
* connections failed due to missing secrets may re-try auto-connection.
|
|
|
|
|
*/
|
2017-11-22 14:52:54 +01:00
|
|
|
if (reset_autoconnect_all (self, NULL, TRUE))
|
|
|
|
|
schedule_activate_all (self);
|
2011-06-28 15:48:12 +02:00
|
|
|
}
|
|
|
|
|
|
2018-06-28 18:05:05 +02:00
|
|
|
NMActiveConnection *
|
|
|
|
|
nm_policy_get_default_ip4_ac (NMPolicy *self)
|
2016-03-28 16:21:07 +02:00
|
|
|
{
|
2018-06-28 18:05:05 +02:00
|
|
|
return NM_POLICY_GET_PRIVATE (self)->default_ac4;
|
2016-03-28 16:21:07 +02:00
|
|
|
}
|
|
|
|
|
|
2018-06-28 18:05:05 +02:00
|
|
|
NMActiveConnection *
|
|
|
|
|
nm_policy_get_default_ip6_ac (NMPolicy *self)
|
2016-03-28 16:21:07 +02:00
|
|
|
{
|
2018-06-28 18:05:05 +02:00
|
|
|
return NM_POLICY_GET_PRIVATE (self)->default_ac6;
|
2016-03-28 16:21:07 +02:00
|
|
|
}
|
|
|
|
|
|
2018-06-28 18:05:05 +02:00
|
|
|
NMActiveConnection *
|
|
|
|
|
nm_policy_get_activating_ip4_ac (NMPolicy *self)
|
2016-03-28 16:21:07 +02:00
|
|
|
{
|
2018-06-28 18:05:05 +02:00
|
|
|
return NM_POLICY_GET_PRIVATE (self)->activating_ac4;
|
2016-03-28 16:21:07 +02:00
|
|
|
}
|
|
|
|
|
|
2018-06-28 18:05:05 +02:00
|
|
|
NMActiveConnection *
|
|
|
|
|
nm_policy_get_activating_ip6_ac (NMPolicy *self)
|
2016-03-28 16:21:07 +02:00
|
|
|
{
|
2018-06-28 18:05:05 +02:00
|
|
|
return NM_POLICY_GET_PRIVATE (self)->activating_ac6;
|
2016-03-28 16:21:07 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*****************************************************************************/
|
|
|
|
|
|
2017-04-23 18:35:24 +02:00
|
|
|
NM_UTILS_LOOKUP_STR_DEFINE_STATIC (_hostname_mode_to_string, NMPolicyHostnameMode,
|
|
|
|
|
NM_UTILS_LOOKUP_DEFAULT_NM_ASSERT ("unknown"),
|
|
|
|
|
NM_UTILS_LOOKUP_STR_ITEM (NM_POLICY_HOSTNAME_MODE_NONE, "none"),
|
|
|
|
|
NM_UTILS_LOOKUP_STR_ITEM (NM_POLICY_HOSTNAME_MODE_DHCP, "dhcp"),
|
|
|
|
|
NM_UTILS_LOOKUP_STR_ITEM (NM_POLICY_HOSTNAME_MODE_FULL, "full"),
|
|
|
|
|
);
|
|
|
|
|
|
|
|
|
|
/*****************************************************************************/
|
|
|
|
|
|
2016-03-28 16:21:07 +02:00
|
|
|
static void
|
|
|
|
|
get_property (GObject *object, guint prop_id,
|
|
|
|
|
GValue *value, GParamSpec *pspec)
|
|
|
|
|
{
|
2016-03-29 10:46:07 +02:00
|
|
|
NMPolicy *self = NM_POLICY (object);
|
|
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2016-03-28 16:21:07 +02:00
|
|
|
|
|
|
|
|
switch (prop_id) {
|
2018-06-28 18:05:05 +02:00
|
|
|
case PROP_DEFAULT_IP4_AC:
|
|
|
|
|
g_value_set_object (value, priv->default_ac4);
|
2016-03-28 16:21:07 +02:00
|
|
|
break;
|
2018-06-28 18:05:05 +02:00
|
|
|
case PROP_DEFAULT_IP6_AC:
|
|
|
|
|
g_value_set_object (value, priv->default_ac6);
|
2016-03-28 16:21:07 +02:00
|
|
|
break;
|
2018-06-28 18:05:05 +02:00
|
|
|
case PROP_ACTIVATING_IP4_AC:
|
|
|
|
|
g_value_set_object (value, priv->activating_ac4);
|
2016-03-28 16:21:07 +02:00
|
|
|
break;
|
2018-06-28 18:05:05 +02:00
|
|
|
case PROP_ACTIVATING_IP6_AC:
|
|
|
|
|
g_value_set_object (value, priv->activating_ac6);
|
2016-03-28 16:21:07 +02:00
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2016-03-29 10:36:51 +02:00
|
|
|
static void
|
|
|
|
|
set_property (GObject *object, guint prop_id,
|
|
|
|
|
const GValue *value, GParamSpec *pspec)
|
|
|
|
|
{
|
2016-03-29 10:46:07 +02:00
|
|
|
NMPolicy *self = NM_POLICY (object);
|
|
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2016-03-29 10:36:51 +02:00
|
|
|
|
|
|
|
|
switch (prop_id) {
|
|
|
|
|
case PROP_MANAGER:
|
|
|
|
|
/* construct-only */
|
|
|
|
|
priv->manager = g_value_get_object (value);
|
|
|
|
|
g_return_if_fail (NM_IS_MANAGER (priv->manager));
|
|
|
|
|
break;
|
|
|
|
|
case PROP_SETTINGS:
|
|
|
|
|
/* construct-only */
|
|
|
|
|
priv->settings = g_value_dup_object (value);
|
|
|
|
|
g_return_if_fail (NM_IS_SETTINGS (priv->settings));
|
|
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2016-03-28 16:21:07 +02:00
|
|
|
/*****************************************************************************/
|
|
|
|
|
|
|
|
|
|
static void
|
2016-03-29 10:46:07 +02:00
|
|
|
nm_policy_init (NMPolicy *self)
|
2016-03-28 16:21:07 +02:00
|
|
|
{
|
2016-09-29 13:49:01 +02:00
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2019-04-11 18:21:40 +02:00
|
|
|
gs_free char *hostname_mode = NULL;
|
2017-03-02 19:18:49 +01:00
|
|
|
|
2017-11-22 12:15:24 +01:00
|
|
|
c_list_init (&priv->pending_activation_checks);
|
|
|
|
|
|
2017-04-17 20:17:45 +02:00
|
|
|
priv->netns = g_object_ref (nm_netns_get ());
|
|
|
|
|
|
2017-04-23 00:40:46 +02:00
|
|
|
priv->hostname_manager = g_object_ref (nm_hostname_manager_get ());
|
|
|
|
|
|
2017-03-02 19:18:49 +01:00
|
|
|
hostname_mode = nm_config_data_get_value (NM_CONFIG_GET_DATA_ORIG,
|
|
|
|
|
NM_CONFIG_KEYFILE_GROUP_MAIN,
|
|
|
|
|
NM_CONFIG_KEYFILE_KEY_MAIN_HOSTNAME_MODE,
|
|
|
|
|
NM_CONFIG_GET_VALUE_STRIP | NM_CONFIG_GET_VALUE_NO_EMPTY);
|
|
|
|
|
if (nm_streq0 (hostname_mode, "none"))
|
|
|
|
|
priv->hostname_mode = NM_POLICY_HOSTNAME_MODE_NONE;
|
|
|
|
|
else if (nm_streq0 (hostname_mode, "dhcp"))
|
|
|
|
|
priv->hostname_mode = NM_POLICY_HOSTNAME_MODE_DHCP;
|
|
|
|
|
else /* default - full mode */
|
|
|
|
|
priv->hostname_mode = NM_POLICY_HOSTNAME_MODE_FULL;
|
2016-04-04 13:44:30 +02:00
|
|
|
|
2017-11-15 16:06:43 +01:00
|
|
|
priv->devices = g_hash_table_new (nm_direct_hash, NULL);
|
|
|
|
|
priv->pending_active_connections = g_hash_table_new (nm_direct_hash, NULL);
|
2016-10-31 23:31:14 +01:00
|
|
|
priv->ip6_prefix_delegations = g_array_new (FALSE, FALSE, sizeof (IP6PrefixDelegation));
|
|
|
|
|
g_array_set_clear_func (priv->ip6_prefix_delegations, clear_ip6_prefix_delegation);
|
2016-03-28 16:21:07 +02:00
|
|
|
}
|
|
|
|
|
|
2016-03-29 10:36:51 +02:00
|
|
|
static void
|
|
|
|
|
constructed (GObject *object)
|
2007-02-08 15:34:26 +00:00
|
|
|
{
|
2016-03-29 10:46:07 +02:00
|
|
|
NMPolicy *self = NM_POLICY (object);
|
|
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2017-02-17 15:33:27 +01:00
|
|
|
char *hostname = NULL;
|
2007-02-08 15:34:26 +00:00
|
|
|
|
2010-01-27 17:13:35 -08:00
|
|
|
/* Grab hostname on startup and use that if nothing provides one */
|
2017-04-23 19:54:39 +02:00
|
|
|
if ((hostname = _get_hostname (self))) {
|
2017-02-17 16:22:30 +01:00
|
|
|
/* init last_hostname */
|
|
|
|
|
priv->last_hostname = hostname;
|
|
|
|
|
|
2010-01-27 17:13:35 -08:00
|
|
|
/* only cache it if it's a valid hostname */
|
2017-02-17 15:33:27 +01:00
|
|
|
if (nm_utils_is_specific_hostname (hostname))
|
2017-02-17 16:22:30 +01:00
|
|
|
priv->orig_hostname = g_strdup (hostname);
|
2010-01-27 17:13:35 -08:00
|
|
|
}
|
2017-04-23 20:31:31 +02:00
|
|
|
_LOGT (LOGD_DNS, "hostname-original: set to %s%s%s",
|
|
|
|
|
NM_PRINT_FMT_QUOTE_STRING (priv->orig_hostname));
|
2010-01-27 17:13:35 -08:00
|
|
|
|
2017-11-27 09:07:28 +01:00
|
|
|
priv->agent_mgr = g_object_ref (nm_agent_manager_get ());
|
|
|
|
|
|
2015-08-10 17:07:40 +02:00
|
|
|
priv->firewall_manager = g_object_ref (nm_firewall_manager_get ());
|
2017-04-20 18:30:01 +02:00
|
|
|
g_signal_connect (priv->firewall_manager, NM_FIREWALL_MANAGER_STATE_CHANGED,
|
|
|
|
|
G_CALLBACK (firewall_state_changed), self);
|
2011-10-07 15:58:09 +02:00
|
|
|
|
2015-01-05 17:20:23 +01:00
|
|
|
priv->dns_manager = g_object_ref (nm_dns_manager_get ());
|
2013-12-17 09:24:00 -05:00
|
|
|
nm_dns_manager_set_initial_hostname (priv->dns_manager, priv->orig_hostname);
|
2016-04-04 16:03:13 +02:00
|
|
|
priv->config_changed_id = g_signal_connect (priv->dns_manager, NM_DNS_MANAGER_CONFIG_CHANGED,
|
2016-03-29 10:46:07 +02:00
|
|
|
G_CALLBACK (dns_config_changed), self);
|
2013-02-05 17:29:15 +01:00
|
|
|
|
2017-04-23 19:30:07 +02:00
|
|
|
priv->lookup.resolver = g_resolver_get_default ();
|
2013-02-24 13:51:56 +01:00
|
|
|
|
2017-04-23 14:20:37 +02:00
|
|
|
g_signal_connect (priv->hostname_manager, "notify::" NM_HOSTNAME_MANAGER_HOSTNAME, (GCallback) hostname_changed, priv);
|
|
|
|
|
|
2016-04-04 14:36:34 +02:00
|
|
|
g_signal_connect (priv->manager, "notify::" NM_MANAGER_SLEEPING, (GCallback) sleeping_changed, priv);
|
|
|
|
|
g_signal_connect (priv->manager, "notify::" NM_MANAGER_NETWORKING_ENABLED, (GCallback) sleeping_changed, priv);
|
|
|
|
|
g_signal_connect (priv->manager, NM_MANAGER_INTERNAL_DEVICE_ADDED, (GCallback) device_added, priv);
|
|
|
|
|
g_signal_connect (priv->manager, NM_MANAGER_INTERNAL_DEVICE_REMOVED, (GCallback) device_removed, priv);
|
|
|
|
|
g_signal_connect (priv->manager, NM_MANAGER_ACTIVE_CONNECTION_ADDED, (GCallback) active_connection_added, priv);
|
|
|
|
|
g_signal_connect (priv->manager, NM_MANAGER_ACTIVE_CONNECTION_REMOVED, (GCallback) active_connection_removed, priv);
|
2016-03-29 10:46:07 +02:00
|
|
|
|
2017-12-04 16:05:26 +01:00
|
|
|
g_signal_connect (priv->settings, NM_SETTINGS_SIGNAL_CONNECTION_ADDED, (GCallback) connection_added, priv);
|
|
|
|
|
g_signal_connect (priv->settings, NM_SETTINGS_SIGNAL_CONNECTION_UPDATED, (GCallback) connection_updated, priv);
|
|
|
|
|
g_signal_connect (priv->settings, NM_SETTINGS_SIGNAL_CONNECTION_REMOVED, (GCallback) connection_removed, priv);
|
|
|
|
|
g_signal_connect (priv->settings, NM_SETTINGS_SIGNAL_CONNECTION_FLAGS_CHANGED, (GCallback) connection_flags_changed, priv);
|
2017-11-27 09:07:28 +01:00
|
|
|
|
|
|
|
|
g_signal_connect (priv->agent_mgr, NM_AGENT_MANAGER_AGENT_REGISTERED, G_CALLBACK (secret_agent_registered), self);
|
2011-01-21 14:20:38 -06:00
|
|
|
|
2016-03-29 10:36:51 +02:00
|
|
|
G_OBJECT_CLASS (nm_policy_parent_class)->constructed (object);
|
2017-04-23 18:35:24 +02:00
|
|
|
|
|
|
|
|
_LOGD (LOGD_DNS, "hostname-mode: %s", _hostname_mode_to_string (priv->hostname_mode));
|
2016-03-29 10:36:51 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
NMPolicy *
|
|
|
|
|
nm_policy_new (NMManager *manager, NMSettings *settings)
|
|
|
|
|
{
|
|
|
|
|
g_return_val_if_fail (NM_IS_MANAGER (manager), NULL);
|
|
|
|
|
g_return_val_if_fail (NM_IS_SETTINGS (settings), NULL);
|
2016-03-29 10:39:47 +02:00
|
|
|
|
|
|
|
|
return g_object_new (NM_TYPE_POLICY,
|
|
|
|
|
NM_POLICY_MANAGER, manager,
|
|
|
|
|
NM_POLICY_SETTINGS, settings,
|
|
|
|
|
NULL);
|
2007-02-08 15:34:26 +00:00
|
|
|
}
|
|
|
|
|
|
2013-08-22 10:10:17 -04:00
|
|
|
static void
|
|
|
|
|
dispose (GObject *object)
|
|
|
|
|
{
|
2016-03-29 10:46:07 +02:00
|
|
|
NMPolicy *self = NM_POLICY (object);
|
|
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
2016-04-04 13:44:30 +02:00
|
|
|
GHashTableIter h_iter;
|
|
|
|
|
NMDevice *device;
|
2017-11-22 12:15:24 +01:00
|
|
|
ActivateData *data, *data_safe;
|
2008-02-07 18:41:42 +00:00
|
|
|
|
2017-04-23 19:30:07 +02:00
|
|
|
nm_clear_g_cancellable (&priv->lookup.cancellable);
|
|
|
|
|
g_clear_object (&priv->lookup.addr);
|
|
|
|
|
g_clear_object (&priv->lookup.resolver);
|
2008-09-18 Dan Williams <dcbw@redhat.com>
Implement support for honoring configured and automatic hostnames, and for
setting the configured hostname.
* introspection/nm-ip4-config.xml
src/nm-ip4-config.c
src/nm-ip4-config.h
src/dhcp-manager/nm-dhcp-manager.c
- Remove useless hostname property; it's not really part of the IPv4
config
* introspection/nm-settings-system.xml
libnm-glib/nm-dbus-settings-system.c
libnm-glib/nm-dbus-settings-system.h
- Add SetHostname() call to system settings D-Bus interface
- Add Hostname property to system settings D-Bus interface
- (nm_dbus_settings_system_save_hostname,
nm_dbus_settings_system_get_hostname): implement
* src/nm-device.c
src/nm-device.h
- (nm_device_get_dhcp4_config): implement
* src/nm-manager.c
src/nm-manager.h
- Fetch and track system settings service hostname changes, and proxy
the changes via a GObject property of the manager
* system-settings/src/nm-system-config-interface.c
system-settings/src/nm-system-config-interface.h
- Replace nm_system_config_interface_supports_add() with a capabilities
bitfield
* system-settings/src/nm-system-config-error.c
system-settings/src/nm-system-config-error.h
- Add additional errors
* system-settings/src/dbus-settings.c
system-settings/src/dbus-settings.h
- (get_property, nm_sysconfig_settings_class_init): add hostname
property; first plugin returning a hostname wins
- (impl_settings_add_connection): use plugin capabilities instead of
nm_system_config_interface_supports_add()
- (impl_settings_save_hostname): implement hostname saving
* src/NetworkManagerPolicy.c
- (lookup_thread_run_cb, lookup_thread_worker, lookup_thread_new,
lookup_thread_die): implement an asynchronous hostname lookup thread
which given an IPv4 address tries to look up the hostname for that
address with reverse DNS
- (get_best_device): split out best device code from
update_routing_and_dns()
- (update_etc_hosts): update /etc/hosts with the machine's new hostname
to preserve the 127.0.0.1 reverse mapping that so many things require
- (set_system_hostname): set a given hostname
- (update_system_hostname): implement hostname policy; a configured
hostname (from the system settings service) is used if available,
otherwise an automatically determined hostname from DHCP, VPN, etc.
If there was no automatically determined hostname, reverse DNS of
the best device's IP address will be used, and as a last resort the
hostname 'localhost.localdomain' is set.
- (update_routing_and_dns): use get_best_device(); update the system
hostname when the network config changes
- (hostname_changed): update system hostname if the system settings
service signals a hostname change
- (nm_policy_new): list for system settings service hostname changes
- (nm_policy_destroy): ensure that an in-progress hostname lookup thread
gets told to die
* system-settings/plugins/keyfile/plugin.c
system-settings/plugins/ifcfg-suse/plugin.c
- (get_property, sc_plugin_ifcfg_class_init): implement hostname and
capabilities properties
* system-settings/plugins/ifcfg-fedora/shvar.c
- (svOpenFile): re-enable R/W access of ifcfg files since the plugin
writes out /etc/sysconfig/network now
* system-settings/plugins/ifcfg-fedora/plugin.c
- (plugin_get_hostname): get hostname from /etc/sysconfig/network
- (plugin_set_hostname): save hostname to /etc/sysconfig/network
- (sc_network_changed_cb): handle changes to /etc/sysconfig/network
- (sc_plugin_ifcfg_init): monitor /etc/sysconfig/network for changes
- (get_property, set_property, sc_plugin_ifcfg_class_init): implement
hostname get/set and capabilities get
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@4077 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-09-18 15:16:44 +00:00
|
|
|
|
2018-06-28 18:05:05 +02:00
|
|
|
nm_clear_g_object (&priv->default_ac4);
|
|
|
|
|
nm_clear_g_object (&priv->default_ac6);
|
|
|
|
|
nm_clear_g_object (&priv->activating_ac4);
|
|
|
|
|
nm_clear_g_object (&priv->activating_ac6);
|
2017-09-25 22:20:47 +02:00
|
|
|
g_clear_pointer (&priv->pending_active_connections, g_hash_table_unref);
|
2017-09-06 15:41:12 +02:00
|
|
|
|
2017-11-22 12:15:24 +01:00
|
|
|
c_list_for_each_entry_safe (data, data_safe, &priv->pending_activation_checks, pending_lst)
|
|
|
|
|
activate_data_free (data);
|
2008-02-07 18:41:42 +00:00
|
|
|
|
2013-08-22 10:10:17 -04:00
|
|
|
g_slist_free_full (priv->pending_secondaries, (GDestroyNotify) pending_secondary_data_free);
|
|
|
|
|
priv->pending_secondaries = NULL;
|
2012-08-21 17:49:41 +02:00
|
|
|
|
2015-08-10 17:07:40 +02:00
|
|
|
if (priv->firewall_manager) {
|
2017-04-20 18:30:01 +02:00
|
|
|
g_signal_handlers_disconnect_by_func (priv->firewall_manager, firewall_state_changed, self);
|
2015-08-10 17:07:40 +02:00
|
|
|
g_clear_object (&priv->firewall_manager);
|
2013-08-22 10:10:17 -04:00
|
|
|
}
|
2011-10-07 15:58:09 +02:00
|
|
|
|
2017-11-27 09:07:28 +01:00
|
|
|
if (priv->agent_mgr) {
|
|
|
|
|
g_signal_handlers_disconnect_by_func (priv->agent_mgr, secret_agent_registered, self);
|
|
|
|
|
g_clear_object (&priv->agent_mgr);
|
|
|
|
|
}
|
|
|
|
|
|
2013-08-22 10:10:17 -04:00
|
|
|
if (priv->dns_manager) {
|
2016-03-29 10:36:51 +02:00
|
|
|
nm_clear_g_signal_handler (priv->dns_manager, &priv->config_changed_id);
|
2015-01-05 17:20:23 +01:00
|
|
|
g_clear_object (&priv->dns_manager);
|
2013-08-22 10:10:17 -04:00
|
|
|
}
|
2013-02-05 17:29:15 +01:00
|
|
|
|
2016-04-04 13:44:30 +02:00
|
|
|
g_hash_table_iter_init (&h_iter, priv->devices);
|
|
|
|
|
if (g_hash_table_iter_next (&h_iter, (gpointer *) &device, NULL)) {
|
|
|
|
|
g_hash_table_iter_remove (&h_iter);
|
|
|
|
|
devices_list_unregister (self, device);
|
2008-03-07 Dan Williams <dcbw@redhat.com>
First pass of multiple active device support. Expect bugs.
* src/nm-ip4-config.c
src/nm-ip4-config.h
- (nm_ip4_config_get_secondary, nm_ip4_config_set_secondary): remove;
there are better ways to do this in the named manager
* src/nm-device.c
src/nm-device.h
- (nm_device_can_activate): return whether the device can activate a
connection right now; taking into account things like carrier state
and rfkill state
- (nm_device_get_best_auto_connection): renamed from
nm_device_get_best_connection
- (real_act_stage4_get_ip4_config): MTU stuff is now handled in the
device subclasses themselves, so that each device can override the
MTU from it's NMSetting subclass if needed
- (nm_device_set_ip4_config): set MTU when setting up routes and stuff
in NetworkManagerSystem.c, not here
* src/named-manager/nm-named-manager.c
src/named-manager/nm-named-manager.h
- (nm_named_manager_name_owner_changed,
nm_named_manager_dbus_connection_changed): fix for changes to
rewrite_resolv_conf()
- (compute_nameservers): don't need the NMNamedManager at all, remove
from parameter list
- (merge_one_ip4_config): new function; merge ip4 configs together
- (rewrite_resolv_conf): write out resolv.conf from all the stored
ip4 configs; the VPN config takes precedence, then the best
device config, then the rest of the configs
- (get_domain_for_config): take the NMNamedManager as an argument
to check whether the config is the VPN config
- (add_ip4_config_to_named): fixups for removal of the 'secondary'
attribute from ip4 configs
- (add_all_ip4_configs_to_named): add all the configs in priority order
- (remove_ip4_config_from_named): fix for changes to
get_domain_for_config()
- (nm_named_manager_add_ip4_config): assign the config to the right slot
based on its type; callers must pass in the type now
- (get_last_default_domain): remove, unused
- (nm_named_manager_remove_ip4_config): handle config slots correctly
* src/nm-device-802-11-wireless.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): handle MTU override
* src/nm-device-802-3-ethernet.c
- (real_can_activate): new function
- (real_get_best_auto_connection): renamed from real_get_best_connection
- (real_act_stage4_get_ip4_config): new function; handle MTU override
* src/vpn-manager/nm-vpn-connection.c
- (nm_vpn_connection_ip4_config_get): don't need to set the 'secondary'
attribute on the ip4 config
* src/NetworkManagerPolicy.c
- (nm_policy_auto_get_best_device): remove
- (nm_policy_device_change_check): remove
- (update_default_route): new function; set the default route via
the specified device
- (get_device_priority): new function; return the priority number of
a device type WRT which one should have the default route. Order is
(highest to lowest) wired, wireless, GSM, CDMA.
- (update_routing_and_dns): new function; determine which device should
have the default route, then update the routing table and DNS
- (maybe_auto_activate_device): new function; if a device is now
available for activation, find out what connection it would like to
activate and do it
- (schedule_activate_check): new function; if a device can be activated
now, schedule the activation. Each device may have only one
pending activation at a given time.
- (device_state_changed): if activation was canceled, try again,
possibly with another connection; if the device was activated,
update routing and DNS; if the device was deactivated, try again
with another connection
- (device_carrier_changed): if there is no carrier, deactivate the
device; otherwise schedule an activation check for the device
- (wireless_networks_changed): schedule an activation check for the
device
- (device_added): keep track of the signal handler IDs so they can
be removed when the device goes away
- (device_removed): remove any signal handlers that might be attached
to the device; update routing and DNS
- (schedule_activate_all): new function
- (connections_added, connection_added, connection_updated): when
connections change, schedule all devices for an activation check
- (connection_removed): when a device is deactivated because its
connection was removed, schedule another activation check for it
- (nm_policy_destroy): destroy pending activations and disconnect
all device signal handlers
* src/nm-manager.c
- (nm_manager_activate_device): if the device was already actived,
deactivate it
- (deactivate_old_device): remove
- (connection_added_default_handler, impl_manager_activate_device):
don't deactivate other devices when activating this one
* src/backends/NetworkManagerGentoo.c
src/backends/NetworkManagerFrugalware.c
src/backends/NetworkManagerPaldo.c
src/backends/NetworkManagerRedHat.c
src/backends/NetworkManagerSlackware.c
src/backends/NetworkManagerArch.c
src/backends/NetworkManagerSuSE.c
src/backends/NetworkManagerDebian.c
- (nm_system_get_mtu): remove; MTU should be provided through the
distro's system settings service plugin instead
- (nm_system_device_add_default_route_via_device): remove
- (nm_system_device_add_default_route_via_device_with_iface): remove
- (nm_system_device_replace_default_route): new function; call
generic implementation
* src/backends/NetworkManagerGeneric.c
src/backends/NetworkManagerGeneric.h
- (nm_generic_device_add_default_route_via_device,
nm_generic_device_add_default_route_via_device_with_iface): remove
- (nm_generic_device_replace_default_route): replace the default route
with the given route via some gateway
* src/NetworkManagerSystem.c
src/NetworkManagerSystem.h
- (nm_system_device_set_from_ip4_config): let the policy handle updates
to routing and DNS; but set the MTU here
- (nm_system_vpn_device_set_from_ip4_config): set the route with the
ip_iface of the active device; use the standard MTU setting function
- (nm_system_set_mtu): remove
- (nm_system_device_set_mtu): consolidate MTU setting code in one place
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@3391 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2008-03-07 19:41:32 +00:00
|
|
|
}
|
|
|
|
|
|
2013-08-30 17:57:56 -05:00
|
|
|
/* The manager should have disposed of ActiveConnections already, which
|
|
|
|
|
* will have called active_connection_removed() and thus we don't need
|
|
|
|
|
* to clean anything up. Assert that this is TRUE.
|
|
|
|
|
*/
|
2017-11-23 21:30:09 +01:00
|
|
|
nm_assert (c_list_is_empty (nm_manager_get_active_connections (priv->manager)));
|
2013-08-22 10:10:17 -04:00
|
|
|
|
2016-01-04 09:46:22 +01:00
|
|
|
nm_clear_g_source (&priv->reset_retries_id);
|
2016-04-13 14:14:36 +02:00
|
|
|
nm_clear_g_source (&priv->schedule_activate_all_id);
|
2012-08-22 18:34:45 -05:00
|
|
|
|
2013-08-22 10:10:17 -04:00
|
|
|
g_clear_pointer (&priv->orig_hostname, g_free);
|
|
|
|
|
g_clear_pointer (&priv->cur_hostname, g_free);
|
2017-02-17 16:22:30 +01:00
|
|
|
g_clear_pointer (&priv->last_hostname, g_free);
|
2011-06-17 12:43:28 +02:00
|
|
|
|
2017-04-23 14:20:37 +02:00
|
|
|
if (priv->hostname_manager) {
|
|
|
|
|
g_signal_handlers_disconnect_by_data (priv->hostname_manager, priv);
|
|
|
|
|
g_clear_object (&priv->hostname_manager);
|
|
|
|
|
}
|
|
|
|
|
|
2016-04-04 14:07:46 +02:00
|
|
|
if (priv->settings) {
|
2016-04-04 14:36:34 +02:00
|
|
|
g_signal_handlers_disconnect_by_data (priv->settings, priv);
|
2016-04-04 14:07:46 +02:00
|
|
|
g_clear_object (&priv->settings);
|
2016-04-04 14:07:46 +02:00
|
|
|
|
|
|
|
|
/* we don't clear priv->manager as we don't own a reference to it,
|
|
|
|
|
* that is, NMManager must outlive NMPolicy anyway.
|
|
|
|
|
*
|
|
|
|
|
* Hence, we unsubscribe the signals here together with the signals
|
|
|
|
|
* for settings. */
|
2016-04-04 14:36:34 +02:00
|
|
|
g_signal_handlers_disconnect_by_data (priv->manager, priv);
|
2016-04-04 14:07:46 +02:00
|
|
|
}
|
2010-01-27 17:13:35 -08:00
|
|
|
|
2016-11-09 17:58:11 +01:00
|
|
|
if (priv->ip6_prefix_delegations) {
|
|
|
|
|
g_array_free (priv->ip6_prefix_delegations, TRUE);
|
|
|
|
|
priv->ip6_prefix_delegations = NULL;
|
|
|
|
|
}
|
2016-10-31 23:31:14 +01:00
|
|
|
|
2016-03-29 10:36:51 +02:00
|
|
|
nm_assert (NM_IS_MANAGER (priv->manager));
|
|
|
|
|
|
2013-08-22 10:10:17 -04:00
|
|
|
G_OBJECT_CLASS (nm_policy_parent_class)->dispose (object);
|
2007-02-08 15:34:26 +00:00
|
|
|
}
|
2007-08-15 01:59:19 +00:00
|
|
|
|
2016-04-04 13:44:30 +02:00
|
|
|
static void
|
|
|
|
|
finalize (GObject *object)
|
|
|
|
|
{
|
|
|
|
|
NMPolicy *self = NM_POLICY (object);
|
|
|
|
|
NMPolicyPrivate *priv = NM_POLICY_GET_PRIVATE (self);
|
|
|
|
|
|
|
|
|
|
g_hash_table_unref (priv->devices);
|
|
|
|
|
|
|
|
|
|
G_OBJECT_CLASS (nm_policy_parent_class)->finalize (object);
|
2017-04-17 20:17:45 +02:00
|
|
|
|
|
|
|
|
g_object_unref (priv->netns);
|
2016-04-04 13:44:30 +02:00
|
|
|
}
|
|
|
|
|
|
2013-08-22 10:10:17 -04:00
|
|
|
static void
|
|
|
|
|
nm_policy_class_init (NMPolicyClass *policy_class)
|
|
|
|
|
{
|
|
|
|
|
GObjectClass *object_class = G_OBJECT_CLASS (policy_class);
|
|
|
|
|
|
2013-08-22 13:06:51 -04:00
|
|
|
object_class->get_property = get_property;
|
2016-03-29 10:36:51 +02:00
|
|
|
object_class->set_property = set_property;
|
|
|
|
|
object_class->constructed = constructed;
|
2013-08-22 10:10:17 -04:00
|
|
|
object_class->dispose = dispose;
|
2016-04-04 13:44:30 +02:00
|
|
|
object_class->finalize = finalize;
|
2013-08-22 13:06:51 -04:00
|
|
|
|
2016-03-29 10:36:51 +02:00
|
|
|
obj_properties[PROP_MANAGER] =
|
|
|
|
|
g_param_spec_object (NM_POLICY_MANAGER, "", "",
|
|
|
|
|
NM_TYPE_MANAGER,
|
|
|
|
|
G_PARAM_WRITABLE |
|
|
|
|
|
G_PARAM_CONSTRUCT_ONLY |
|
|
|
|
|
G_PARAM_STATIC_STRINGS);
|
|
|
|
|
obj_properties[PROP_SETTINGS] =
|
|
|
|
|
g_param_spec_object (NM_POLICY_SETTINGS, "", "",
|
|
|
|
|
NM_TYPE_SETTINGS,
|
|
|
|
|
G_PARAM_WRITABLE |
|
|
|
|
|
G_PARAM_CONSTRUCT_ONLY |
|
|
|
|
|
G_PARAM_STATIC_STRINGS);
|
2018-06-28 18:05:05 +02:00
|
|
|
obj_properties[PROP_DEFAULT_IP4_AC] =
|
|
|
|
|
g_param_spec_object (NM_POLICY_DEFAULT_IP4_AC, "", "",
|
|
|
|
|
NM_TYPE_ACTIVE_CONNECTION,
|
2016-03-28 20:49:18 +02:00
|
|
|
G_PARAM_READABLE |
|
|
|
|
|
G_PARAM_STATIC_STRINGS);
|
2018-06-28 18:05:05 +02:00
|
|
|
obj_properties[PROP_DEFAULT_IP6_AC] =
|
|
|
|
|
g_param_spec_object (NM_POLICY_DEFAULT_IP6_AC, "", "",
|
2016-03-28 20:49:18 +02:00
|
|
|
NM_TYPE_DEVICE,
|
|
|
|
|
G_PARAM_READABLE |
|
|
|
|
|
G_PARAM_STATIC_STRINGS);
|
2018-06-28 18:05:05 +02:00
|
|
|
obj_properties[PROP_ACTIVATING_IP4_AC] =
|
|
|
|
|
g_param_spec_object (NM_POLICY_ACTIVATING_IP4_AC, "", "",
|
2016-03-28 20:49:18 +02:00
|
|
|
NM_TYPE_DEVICE,
|
|
|
|
|
G_PARAM_READABLE |
|
|
|
|
|
G_PARAM_STATIC_STRINGS);
|
2018-06-28 18:05:05 +02:00
|
|
|
obj_properties[PROP_ACTIVATING_IP6_AC] =
|
|
|
|
|
g_param_spec_object (NM_POLICY_ACTIVATING_IP6_AC, "", "",
|
2016-03-28 20:49:18 +02:00
|
|
|
NM_TYPE_DEVICE,
|
|
|
|
|
G_PARAM_READABLE |
|
|
|
|
|
G_PARAM_STATIC_STRINGS);
|
2016-09-29 13:49:01 +02:00
|
|
|
|
2016-03-28 20:49:18 +02:00
|
|
|
g_object_class_install_properties (object_class, _PROPERTY_ENUMS_LAST, obj_properties);
|
2013-08-22 10:10:17 -04:00
|
|
|
}
|